[Kirjoita yrityksen nimi]

Essays on cultural economics

[Tiedoston alaotsikko]

seppo[Valitse pvm.]

Contents

0 Introduction.....................................................................................................6

0.1 Essay 1: Critics review or preceding week’s admissions explaining movie admissions.........................10

0.2 Essay 2: Demand for ice hockey, the factors explaining attendance of ice hockey games in Finland...11

0.3 Essay 3: Fan loyalty in Finnish ice hockey.............................................................................................15

0.4 Essay 4: Spectators of performing arts – who is sitting in the auditorium?..........................................16

0.5 Essay 5: Are the spectators of performing arts and the spectators of the movies the same?.............20

0.6 Conclusions...........................................................................................................................................23

1 Critics´ reviews or preceding week’s admissions explaining movie admissions........................................................................................................29

1.1 Introduction..........................................................................................................................................29

1.2 Literature review..................................................................................................................................30

1.3 Empirical model and variables..............................................................................................................34

1.4 Estimation and results..........................................................................................................................38

1.5 Conclusions and suggestions................................................................................................................45

2 Demand for ice hockey, the factors explaining attendance of ice hockey games in Finland...............................................................................................56

2.1 Introduction..........................................................................................................................................57

2.2 Literature..............................................................................................................................................60

2.3 A model explaining attendance............................................................................................................64

2.4 Estimation.............................................................................................................................................68

2.5 Conclusions and suggestions................................................................................................................71

3 Fan loyalty in Finnish Ice Hockey..................................................................87

3.1 Introduction..........................................................................................................................................87

3.2 Fan loyalty or brand loyalty – stochastic frontier analysis -..................................................................89

3.3 Stylized facts: Finnish Ice Hockey..........................................................................................................91

3.4 Estimation and results..........................................................................................................................96

3.5 Conclusions.........................................................................................................................................101

4. Spectators of performing arts – who is sitting in the auditorium?.............108

4.1 Introduction........................................................................................................................................108

4.2 Method and sample............................................................................................................................112

4.3 Conclusions and evaluation................................................................................................................120

1

5. Are the spectators of performing arts and the spectators of movies the same?..............................................................................................................141

5.1 Introduction........................................................................................................................................141

5.2 Literature review................................................................................................................................146

5.3 The method and sample.....................................................................................................................149

5.4 Results................................................................................................................................................154

5.5 Conclusions.........................................................................................................................................161

2

Tables

Table 1: Sports consumption in Finland 2007.................................................11Table 2 Attendance popularity and correlation among adult population in Finland. 2005-2006...........................................................................................12Table 3 (Table1.1) Overview of top 10 films in 2003 in Finland, source: Finnish Film Foundation................................................................................................35Table 4: (Table 1.2) Descriptive statistics and sources of variables, * weekly, ** non-zero observations...................................................................................38Table 5: (Table 1.3) Fixed Effect and Random Effect Models (Park 2008)......39Table 6: (Table 1.4) Estimation results, full sample, n = 1060........................40Table 7: (Table 1.5) Estimation results, all movies with previous admission in Helsinki, n = 515..............................................................................................42Table 8: (Table 1.6) Estimation results, all movies critically reviewed and with previous week’s Helsinki admission, n = 205..................................................43Table 9: (Table 1.7) Estimation results, all movies critically reviewed and with previous week’s Helsinki admission, n = 205..................................................44Table 10: (Table 1.8) Distributors’ premieres in 2001 – 2003..........................51Table 11: (Table 1.9) Descriptive statistics for critical review rank (scale 1 – “top” to 10 – ”lowest”)......................................................................................52Table 12: (Table 1.10) Duration of movie run, quantiles..................................53Table 13: (Table 1.11) Estimation results, n = 345..........................................54Table 14: (Table 1.12) Estimation results, all movies critically reviewed and with previous week’s Helsinki admission, n = 205..........................................55Table 15: (Table 2.1) Regular season 2007 – 2008 average attendance and capacity statictics, source: Jääkiekkokirja 2007-2008 ja Jääkiekkokirja 2008-2009..................................................................................................................80Table 16: (Table 2.2) Variables, measurement, source and expected sign......81Table 17: (Table2.3) Variables, means, standard deviations and correlation matrix. ATT = attendance, Price (€), Dist = distance between home team’s and visitor’s stadiums along road (km), Temp = max tempature Unempl = monthly regional unemployment rate(%), HomePop = home town population, VisiPop = vistor’s town population, HPoint = points per game, home team, before the game VPoint = visitor’s points per game, before the game, HomeG = number of games, home team, before the game, VisiG = number of games,

3

visitor, before the game , Last3H = points from 3 last games, home team, Last3V = points from 3 last games, visitor. The number of observations = 392...........................................................................................................................82Table 18: (Table 2.4a) Some estimation results...............................................83Table 19: (Table 2.4b) Some estimation results...............................................84Table 20: (Table 2.4c) Some estimation results...............................................85Table 21: (Table 3.1) Average attendance statistics........................................92Table 22: (Table 3.2) Variables, means, standard deviations and correlation matrix................................................................................................................94Table 23: (Table 3.3) OLS results, dependent variable is log(Attendance), n = 406....................................................................................................................96Table 24: (Table 3.4) Estimation results, dependent variable is log(Attendance), n = 406..................................................................................98Table 25: (Table 2.5) Inefficiency scores of teams.........................................100Table 26: (Table 2.6) Correlation matrix of selected variables......................101Table 27: (Table 2.7) Average attendance, home games, regular seasons....105Table 28: (Table 4.1) Culture and physical education hobbies 1981, 1991 and 1999................................................................................................................126Table 29: (Table 4.2) Kulttuuripuntari (culture barometer) 1999:.................127Table 30: (Table 4.3) Suomen Teatterit (Taloustutkimus), survey on visits to theatre, opera or ballet during the past 12 months, years 1985, 1998, 2001, 2004 and 2007................................................................................................128Table 31: (Table 4.4) Eurobarometer 56.0: August-September 2001, n = 1024.........................................................................................................................130Table 32: (Table 4.5) ISSP 2007, ”How often in your leisure do you go to concerts, exhibitions, theatre etc.?”...............................................................131Table 33: (Table 4.6) Visitor density, ANOVA (significance in parenthesis). .132Table 34: (Table 4.7) Visitor density, Anova and Manova, Women and Men separately.......................................................................................................134Table 35: (Table 4.8) Multivariate logit analysis............................................135Table 36: (Table 4.9) Multivariate logit analysis............................................136Table 37: (Table 4.10) ISSP 2007, ”How often on your leisure do you go to see sport events on the location (ice hockey, football, athletics, motor racing etc.)? n = 1355.........................................................................................................137Table 38: (Table 4.11) Multivariate logit analysis..........................................138Table 39: (Table 4.12) Bivariate probit analysis,............................................139

4

Table 40: (Table 4.13) Bivariate probit analysis,............................................140Table 41: (Table 5.1) Spectators of movies at the cinema and performing atrs (concert, theatre, art exhibition) in Finland, recent surveys.........................166Table 42: (Table 5.2) descriptive statistics of age-group and education variables..........................................................................................................167Table 43: (Table 5.3) Average monthly household and personal gross incomes........................................................................................................................168Table 44: (Table 5.4) Descriptive statistics of some explanatory variables.. .169Table 45: (Table 5.5) Bivariate probit analysis...............................................170Table 46: (Table 5.6) Bivariate probit analysis ,.............................................173Table 47: (Table 5.7) Multinomial logit (MNL) analysis.................................176Table 48: (Table 5.8) Multinomial logit (MNL) analysis,................................180

5

Figures

Figure 1: Value added of culture in 2007, EUR/capita in NUTS3 and Capital regions in Finland.............................................................................................18Figure 2: (Figure 2.1) Actual attendance for HIFK (Series1) and Model 6 (Series2)............................................................................................................86Figure 3: (Figure 2.2) Actual attendance for HIFK (ADM), Model 6 and simulations 1 and 2...........................................................................................86Figure 4: Nuts areas.......................................................................................154Figure 5: (Figure 5.1) Direct and indirect marginal effect of age-cohorts on highbrow art consumption..............................................................................158Figure 6: (Figure 5.2) Direct and indirect marginal effects of education on highbrow art consumption..............................................................................158

6

0 Introduction

Stigler and Becker (1977) introduced the concept of cultural capital developed by the aggregation of past consumption of cultural goods. The accumulation function can be considered to be similar to that of human capital, i.e. formal education. Part of culture consumed can be considered as investment in the future cultural capital. In other words, not all cultural spending is consumed within a year, whereas can be consumed and accumulated over a longer period. Cultural behaviour is determined by the consumer’s budget, time, social, physical constraints and formal education (Frey 2000). A central feature is also the variety in cultural consumptions and its accumulation. People with higher education have on average less leisure than those with lower education. At the same time the higher educated have a bigger variety at leisure and therefore also in cultural consumption (Ruuskanen 2004, 136).

A large amount of economic and sociological research has been done to classify different cultural consumption patterns. The economics of cultural consumption has traditionally focused on explaining attendance figures and studying the socioeconomic characteristics of the audience. Audience and participation surveys often argue that performing arts audiences consist of relatively wealthy citizens while the audiences of sport events and cinemas are different. However, cultural consumption is not just about going to see art exhibitions, opera or theatrical performances. Some of the consumers prefer sport events and films. Sport events, especially football and ice hockey matches, are favoured by middle-class males and cinema lovers are young students.

Cultural consumption is thus connected with leisure activities of consumers. They might choose to go to the cinema, go to an ice hockey match, go to the opera or theatrical performances, etc. depending on their preferences and the amount of leisure time and incomes. Time constraints are related to (i) the place of residence, (ii) to the leisure time. In Finland art institutions, like

7

opera and theatre houses, are located in bigger cities, but there are some theatre groups making tours in the countryside. Still the place of residence is a very important factor to explain consumers’ cultural participation decisions. Several studies have shown that there is a substantial seasonal variation in leisure time use. During the winter, leisure is more sports oriented while during summer more socially oriented (Niemi and Pääkkönen 1992). Men’s ice hockey is the most popular sport in terms of total attendance. The regular season in the highest league in Finland begins in September and ends in March. After that there are some play-off matches in March and April. Also the movie attendance statistics in the first essay show that the summer is the weakest time by attendance. Most citizens have their holidays in summer, but they do not seem to go to cultural or sport events even if there was plenty of leisure.

The socio-economic status of the consumer clearly has an effect on leisure activities. The unemployed have more leisure but less income. Ruuskanen (2004) has shown that both net wage and the yearly income of the spouse have a negative effect on the joint time spent together, the number of children reduces joint leisure time of spouses, and university education increases the time spent together in leisure. Both the age and health situation of consumers have an impact on the leisure time and how active the leisure is (Piekkola and Ruuskanen 2006). Both unemployed and employed older men are more active in leisure if they consider themselves healthy, but the relation is not so obvious for younger men. Older women are more active during their leisure than younger women. Taking care of small children does not restrict any more. Also teenaged girls are more active in participation in cultural activities except the movies (Pääkkönen 2010, 234) although the amount of leisure is lower for girls than for boys.

Limited leisure time restricts and these cultural events are substitutes to some extent. However, are art exhibitions, opera or theatrical performances substitutes or rather complements? Budget constraints can limit the participation so that consumers can only choose one cultural event and therefore different events are substitutes. But, on the other hand, some

8

culturally oriented consumers can accumulate positive experiences and these can induce further consumption. In this case cultural events are complements. Complementarity is also likely to vary depending on the socio-economic class. Ruuskanen (2004) has shown that skilled workers are more engaged in several types of activities while the time use of low-educated is more monotone. The sociology of cultural participation has shown that consumers can be classified into three groups: omnivore, paucivore and inactive (Alderson, Junisbai and Heacock 2007). The omnivores are active in all cultural consumption, from cinema to classical music. The paucivores go to see all kinds of cultural activities but less than the omnivores. To the omnivores cultural events are complements.

Irrespective of the cultural events being complements or substitutes, the quality of the event is important from the viewpoint of enjoying. Advertising provides direct information about the characteristics of products with search qualities, their main attributes can be determined by visual or tactile inspection (e.g. clothes) or by a test drive or trial (car). Advertising may convey hard facts, vague claims or a favourable impression of a product. The informational content of advertising depends on whether consumers can determine the quality of that product before buying. If the consumer can value a product’s quality by inspection before buying it, the product has search qualities or the product is a search good. However, if the consumer must consume the product to determine its quality, the product has experience qualities or the product is an experience good (Nelson 1970). Experience goods must be consumed before their quality can be determined (e.g. processed foods, software programs, and gymnastic exercises). The early writers in the 1950’s considered advertising as being manipulative (Kaldor 1950) and therefore it reduces competition and welfare since advertising persuades consumers to purchase more heavily advertised products even though there is no quality difference between otherwise equal or comparable products. The price of the highly advertised products rises and therefore the advertising serves as an entry-deterring mechanism. If advertising is predatory, the incumbent firm is capable of creating an entry-deterring strategy (Cubbin 1981). More recent authors propose that advertising serves as a tool for transmitting information from producers to consumers about

9

differentiated brands and therefore reducing the search costs and increasing welfare (Benham 1972).

Nelson (1974) argues that producers of high-quality experience goods can spend more money on advertising because first-time consumers are more likely to be satisfied with the quality and will make repeat purchases, than with low-quality experience goods. Consumers are not dependent on the information received through producers’ advertising when they buy search goods since they receive that information by inspection or trial. So the effects of advertising vary between search goods and experience goods and there is more intensive advertising with experience goods. On the other hand, Schmalensee (1978) argues that low-quality brands are more frequently purchased and low-quality producers advertise more intensively. The recent rise of social media has substantially changed the media usage of advertising campaigns. Consumers trust more on recommendations from other consumers, e.g. word-of-mouth or blogs than on paid advertising (Viljakainen, Bäck and Lindqvist 2008 or Karjaluoto 2010).

Producers (distributors, importers) can use other means to signal about the quality of their products, not just advertising but also product labelling or branding, reputation, guarantees or expert ratings. Some fruits and vegetables are sold without a brand name. Consumers might assume that a banana is a banana and there is a little variation between producers or countries of origin. However, there are strong and universal brands also, like Chiquita and consumers attach this brand and trustworthy quality. If a large proportion of sales is generated by customers who do not repeat their purchases, like tourists, the reputation of a shop matters less since few customers are familiar with the shop’s reputation (Carlton & Perloff 1990, 530). To the contrary consumers who repeat purchases are willing to repurchase cultural and other goods if their past experiences are positive and producers’ signals have less importance. One essay shows evidence that especially committed ice hockey fans repeat their purchases and regularly go to the ice hockey stadium. The impact of advertising on the match attendance is minor.

10

Different terminologies have been used to rank tastes, like: highbrow – middlebrow – lowbrow, or high – popular, or legitimate – vulgar. This division has been used frequently in the sociology of cultural consumption. With Swedish data Bihagen and Katz-Gerro (2000) show that women are more active in highbrow consumption (opera, dance or theatrical performances) and men in low, like watching television (entertainment, sport). Highbrow television (documentary, culture, news) and lowbrow culture (films) are less connected to gender and formal education, but Warde and Gayo-Gal (2009) show that these are strongly related to age. The omnivore group is associated with legitimate taste that is aesthetically the most valuable. Omnivorousness increases with age up to around 50 and strongly diminishes among those over 70.

In Finland, the economics of culture has been less studied. There are a few surveys on the cinema spectators (Suomalaisen elokuvan markkinat ja kilpailukyky 1999, Kotimaisen elokuvan yleisöt –tutkimus 2010), theatre and opera audiences (Kivekäs 1991, Suomalaisten teatterissa käynti 2007, Mikkonen and Pasanen 2009), audiences of sport events (e.g. Kansallinen liikuntatutkimus 2010) and a substantial amount of sociological studies on the cultural consumption (e.g. recently Virtanen 2007 or Purhonen, Gronow and Rahkonen 2010). Most surveys present descriptive statistics of the audience, but there are virtually no studies that use more advanced econometric methods. Using frequency and contingency tables, the analysis of variance and logistic regression methods Virtanen (2007) showed that education, age and socio-economic status have important explanation power in highbrow cultural consumption in the whole European Union area. However, these variables can explain only 10 – 15 per cent of the variation in consumption choices. Purhonen, Gronow and Rahkonen (2010) showed using logistic and Poisson regression analysis that regardless of how omnivorousness is operationalized, different socio-economic variables are better to explain literature taste than musical taste. The socio-economic variables are gender, age group, education and the place of residence. Income level is not significant.

11

The purpose of this study is to use advanced econometric methods to explain cultural consumption choices in Finland and to get more information on this topic. The study is a combination of five separate papers in cultural economics. The cultural capital related to movies and attendance to cinema audience is the topic in the first study. Conventionally it is argued that the biggest group in the cinema audience consists of young people of age 15 – 24 (Suomalaisen elokuvan markkinat ja kilpailukyky, 1999, 89). What is the role of public information on the decision to go to the cinema? Will they read critical reviews from the newspapers before they make the decision between different movies in the repertoire? The second and the third essay are studying the audiences of the ice hockey matches in the men’s champion league (Sm-liiga). What is the role of the winning percentage of the home team and of the visitor team on the attendance? Since typically the audience is male dominant who read carefully the sport pages in the newspapers where the series situation is published, that information might have an important impact on the attendance figures. The fourth and fifth essays are studying what is the composition of the audiences of highbrow arts in relation to cinema and sports. Are these audiences different and how? The essays draw a picture of omnivore, paucivore and inactive consumers and especially on how sensitive the omnivore consumption patterns are to the various background variables such as age, education and gender. In the concluding chapter we also draw some tentative conclusions on how inactive people can be encouraged to consume or invest more in cultural capital.

0.1 Essay 1: Critics review or preceding week’s admissions explaining movie

admissions

The first essay considers movie attendance in Finland in 2003 explaining the number of spectators of the 20 most popular films in each week. The total number of films on distribution was 225 but with only 177 premieres. Many films had the first evening during preceding December in 2002 since it is widely known that the Christmas season is top time. Seasonal variation is

12

high, during summer the attendance figures are the lowest. Variation across films is also large. The aggregate attendance in 2003 was roughly 7.7 million and the top 10 got roughly 42 % of the spectators. Since both the weekly (time series) and movie specific (cross-section) variation is substantial, conventional regression methods are unsuitable. A panel data analysis enables regression analysis with both time-series and cross-sectional dimension.

In 2003 the average duration of movie runs in Finland was four months for the top 10 films and roughly one month for the median film in respect of the spectator number. Hence, spectators have had enough time to reveal the necessary information on the quality of the film from various sources. The essay studies the role of word-of-mouth and critical reviews in explaining movie attendance. Critical reviews are published in the weekly magazine supplement ‘Nyt’ for the newspaper that has the largest circulation in Finland, Helsingin Sanomat. World-of-mouth is measured by the previous week’s attendance figure at the cinemas in Helsinki. Since more than a fourth of young audience (age-group: 15-24) are heavy users and since they read less newspapers than older citizens in Finland, the role of critics review is probably lower than the role of word-of-mouth. Consumers in general rely more on the word-of-mouth than on other forms of information (Viljakainen, Bäck and Lindqvist 2008, 25). The first essay verifies that when the world-of-mouth is taken into account, the critics review is not a significant variable to explain movie attendance. Since admission figures are typically the highest during the first weeks, a variable “weeks since released” is used to control this peak. The analysis shows that it is significant, as well as the price variable. The price elasticity of weekly movie admission is roughly -1 which shows some monopoly pricing potential. Panel data analysis also indicates that the fixed effects model is the most suitable for explaining weekly movie admissions in Finland in 2003.

0.2 Essay 2: Demand for ice hockey, the factors explaining attendance of ice

hockey games in Finland

13

The second essay studies the ice hockey match attendance during the regular season 2007 – 2008 in Finland. The ice hockey league, ‘SM-liiga’ is the most important live cultural event or series of cultural events in terms of attendance per event. Yearly movies attendance is three times bigger compared to the ice hockey matches but these and opera or theatrical performances are live events. At the opera and theatre there is a manuscript that they follow and thus the possibility of surprises is smaller, but a match has more uncertainty. The home team might win or lose the match depending on the quality of the team and the visitor among others. The regular season usually begins in September and ends in the following March. The number of regular season matches was 392. The total attendance was 1,964,626 i.e. 5,012 per match. Besides these matches there were play-off matches in March and April, but these matches were left out from the examination due to the different nature of these events. The Finnish data on the recent International Social Survey Programme (ISSP 2007) reveals that almost 40 % of the population never goes to see a sports activity (ice hockey, football, athletics, motor racing etc.), less than 8 % attends several times a month and the rest (i.e. more than 50 %) occasionally. The same survey also shows that physical exercise (active sport consumption) is more common than passive sport consumption (table 1: Sports consumption in Finland 2007).

Table 1: Sports consumption in Finland 2007

Daily Several times

a week

Several times

a month

Occasionally Never Total, n

How often do you

attend a sports

activity?

4 (0.3%) 17 (1.3%) 82 (6.2%) 691 (52.3%) 526 (39.8%) 1320

How often do you

attend a sports

activity? (Female)

1 (0.1%) 5 (0.7%) 38 (5.1%) 327 (44.2%) 369 (49.9%) 740

How often do you

attend a sports

activity? (Male)

3 (0.5%) 12 (2.1%) 43 (7.6%) 358 (63.3%) 150 (26.5%) 566

How often do you

exercise sports?

301 (22.6%) 546 (41.0%) 272 (20.5%) 183 (13.8%) 28 (2.1%) 1330

14

How often do you

exercise sports?

(Female)

189 (25.5%) 308 (41.5%) 144 (19.4%) 92 (12.4%) 9 (1.2%) 742

How often do you

exercise sports? (Male)

106 (18.5%) 230 (40.1%) 127 (22.2%) 91 (15.9%) 19 (3.3%) 573

Source: ISSP 2007. Own calculations.

There is a significant difference between genders so that males are more active in passive sport consumption (attendance, Mann-Whitley U-test, z= -8,430, sig. = 0,000), while females are more active exercisers (Mann-Whitney U-test, z=-3,858, sig=0,000). Active and passive (attendance) sports consumption are only slightly positively correlated (Kendall’s τ = 0,054, n = 1314, sig. = 0,028). There is also a negative relationship between age and passive sports consumption (Spearman’s ρ = -0,182, n = 1265, sig. = 0,000). For female the negative relationship is somewhat stronger (Spearman’s ρ = -0,193, n = 724, sig. = 0,000) than for male (Spearman’s ρ = -0,179, n = 540, sig. = 0,000). Another survey (Liikuntatutkimus 2005-2006, Sport Survey: Adult Population) on adult population sport consumption – both active and passive – in Finland was carried out a few years ago1. The sample size was 5510. In this survey 44% responded that they had not attended any sports event between February 2005 and January 2006. The most popular sports in terms of attendance were ice hockey (25.5%), football (16.9%), athletics (10.6%), skiing (6.5%) and Finnish rule baseball (5%). The largest positive correlation is between ice hockey and football attendance. Attendance and income level (8 categories from the lowest to the highest) are not correlated (not reported here).

Table 2 Attendance popularity and correlation among adult population in Finland. 2005-2006.

Popularity Ice Hockey Football Athletics Skiing F Rule

Baseball

Ice Hockey 25.5%

F: 14.6%

M: 36.4%

1

Football 16.9% 0.323 (0.000) 1

1 Recent (February – March 2007) Eurobarometer 67.1 reports that almost 56 % in the sample (n = 1054 in Finland)

had not attended any sport event during the last 12 month period. The figure was lower for male (44%) than for

female (65%). 15

F: 11.0%

M: 22.8%

F: 0.353 (0.000)

M:0.193(0.000)

Athletics 10.6%

F: 9.9%

M: 11.3%

0.093 (0.000)

F: 0.133 (0.000)

M: 0.031

(0.108)

0.123 (0.000)

F: 0.156 (0.000)

M:0.074 (0.000)

1

Skiing 6.5%

F: 6.3%

M: 6.6%

0.009 (0.517)

F: 0.002 (0.909)

M: 0.015

(0.431)

0.022 (0.110)

F:0.019 (0.315)

M: 0.024 (0.216)

0.147 (0.000)

F:0.150 (0.000)

M: 0.143

(0.000)

1

F Rule Baseball 5.0%

F: 3.9%

M: 6.1%

0.098 (0.000)

F: 0.085 (0.000)

M: 0.096

(0.000)

0.056 (0.000)

F: 0.050 (0.009)

M: 0.049 (0.010)

0.056 (0.000)

F: 0.053 (0.006)

M: 0.058

(0.002)

0.014 (0.295)

F: 0.024 (0.212)

M: 0.001 (0.942)

1

Observations 5510. Significance in parenthesis. Legend: F = female n = 2754, M = male n = 2756

Tables 1 and 2 indicate that more than half of the adult population (age between 15 and 74 in ISSP 2007) have attended a sport activity and for half of these an ice hockey match has been that event.

The sociology of sport consumption has revealed that there are substantial motive differences between genders. A well-known classification is Sport Fan Motivation Scale (SFMS) by Wann (1995). There are eight motives: eustress (i.e. the need for positive stress), self-esteem (i.e. the desire to maintain a positive self-concept through team success), escape (i.e. sport as diversion from bored everyday life), entertainment, economic (i.e. gamble on the events), aesthetic (i.e. sport as an art), group affiliation (i.e. belongingness need), and family (i.e. opportunities to spend time with family). Wann conducted a quantitative examination with a 23-item Likert scale questionnaire. Using confirmatory factor analysis the above mentioned eight internally consistent, reliable and criterion valid motives were found. The original sample consisted of primarily of university college students. Several studies, however, confirmed the results (e.g. Wann, Schrader & Wilson 1999, Wann, Royalty & Rochelle 2002, Wann, Robinson, Dick & Gillentine 2003, Ridinger & Funk 2006, Wann, Grieve, Zapalac & Pease 2008 or Koo & Hardin 2008). Eustress, self-esteem and group affiliation motives were more associated with team and aggressive sport type (e.g. football, ice hockey) rather than individual and nonaggressive sport type. On the other hand, aesthetic motive was associated with individual and nonaggressive sport type

16

(e.g. figure skating, tennis). Wann, Schrader & Wilson (1999) also classify sport spectators as intrinsically or extrinsically oriented. Fans that enjoy sport because of its aesthetics and artistic movement (intrinsic) may not bother of their favorite team’s or individual’s poor performance since the aesthetic performance of the event is present regardless of the outcome. On the other hand extrinsic fans (self-esteem, economic motives) could find it unpleasant to watch their favorite team’s games unless the team is victorious. Self-respect and self-fulfillment are more associated with women’s team spectators (Kahle, Duncan, Dalakas & Aiken 2001) while self-indulgence is more men’s team spectators’ attribute. The opportunity to spend time with family or sense of belonging or socialization is attributes associated with women’s sport spectators (Kahle, Duncan, Dalakas & Aiken 2001 or Ridinger & Funk 2006). Females seem to be more sport fans for social reasons (Dietz-Uhler, Harrick, End & Jacquemotte 2000), while males are more likely to be fans because they play sports and want to acquire sport information (e.g. read sport pages in newspapers).

The second essay (”Demand for ice hockey, the factors explaining attendance of ice hockey games in Finland”) particularly studies among others the effects of public information on the ice hockey attendance figures. Since men typically read the sport pages of newspapers, the home team’s performance is well known. The performance is operationalized as the points per game measure (success). There are four alternatives: a win within the normal playing time (60 minutes) produces 3 points, an win within extension time (> 60 min) or a penalty shot win produces 2 points, a lost within extension time or after penalty shots produces 1 point, and a lost within normal playtime gives 0.

The results indicate that both the population of the town of the home team and of the visitor have a statistically significant effect on attendance. The distance between home team’s town and visiting team’s town is also significant, i.e. local games have a bigger attendance than other games. The demand is not elastic with respect to the ticket price. Loyal supporters have a season ticket, but the share of season ticket holders in the audience is not

17

known. Success of both the home team and the visitor has an effect: home team’s success with a positive and visitor’s with a negative coefficient. The number of plays already played has a negative effect. Weekday effect is important: the attendance is bigger during Saturdays. Also the day temperature has an effect: the colder, the bigger attendance. That effect is small but still statistically significant. The local unemployment rate has no effect, and the success factor of the last three games (the form guide) does not seem to explain as well as the success factor of all games played.

With caution it can be argued that ticket price has a slight effect on attendance, since demand seems to be inelastic. However, the price variable is not the actual average price since this data was not available. The price variable used in the estimations is the ticket price to the best seats. As the season goes on and more games have been played, attendance seems to diminish but the estimated coefficient is low even though significant. Team’s success seems to attract a bigger attendance, while visitor’s success has the opposite effect. Spectators are willing to see live game in the stadium if they expect that home team will win the game. The local unemployment rate has no effect on attendance, while weather conditions measured by the outside temperature are a significant variable. Colder weather attracts more spectators. However, the estimated coefficient is minor but significant. The estimation results reveal that the models can explain about two thirds of actual attendance based on the coefficient of determination.

0.3 Essay 3: Fan loyalty in Finnish ice hockey

Fan loyalty is the topic of the third essay. Sport has become more professional over the years. Sport managers view their teams or leagues as brands to be managed. A product or service is considered as a brand if the name, logo, sign or slogan increases the value of that product or service. The psychological aspect in the consumer’s mind, the brand image, consists of all information and associations with a product or service. The third essay studies fan loyalty in Finnish men’s ice hockey during the regular season 2008-2009

18

using stochastic frontier analysis. Most teams in the highest ice hockey league are local monopolies, but there are two teams in Helsinki which might be substitutes since the distance between their stadiums is less than 3 km. Moreover, there is one team in the neighbouring city, Espoo, whose stadium is at a distance of about 13 km from the previous. In addition there are two teams in Tampere with a shared stadium. However, some teams are local monopolies and some teams meet higher competition. Therefore brand loyalty or fan loyalty might differ according to the competitive position, and the aim of this research is to study the relationship between the fan loyalty and the competitive position of teams. The competitive position is defined here as the geographical distance between teams’ stadiums.

The teams in the champion league generally raise funds from not just gate revenues but also from merchandise sales, the sales of broadcast rights and commercial sponsorships. Loyal fans use various fan products such as fan shits and scarves. Broadcast rights are usually sold by the league association and the broadcast revenue is shared among the teams. Sponsorship revenue depends on the popularity of the team, which in turn is associated with larger market base, i.e. larger home town population. Depken (2000, 2001) measures fan loyalty by efficiency score in stochastic frontier analysis, Winfree, McCluskey, Mitterhammer and Fort (2004) by the permanency of successive years’ attendance and Brandes, Franck and Theiler (2010) by mean match tickets per market size. Also direct surveys to get self-revealed levels of fan loyalty have been used. Wakefield and Sloan (1995) show that fan loyalty increases home game attendance. The third essay follows Depken and uses a panel data of Finnish men’s champion league ice hockey attendance during the regular season 2008-2009 with stochastic frontier analysis. There were 406 games played during that season beginning in September 2008 and ending in March 2009.

The explanatory variables used in this study are conventional and consistent with other studies (for a review, see Borland and MacDonald 2003 or Simmons 2006): home town population, visitor’s town population, distance between teams’ home stadiums, the winning percentage of the home team and

19

of the visitor team, the game round, and the local unemployment rate. The time specific variables are weekday dummies and the outside temperature. The error term has two components ε*i = εi – ln(λi) in which εi is the random error term that captures noise as well as team and time-specific unobserved heterogeneity (Greene 2005).The inefficiency term λi in the stochastic frontier is time invariant and team specific. Two possible distributions have frequently been used (see Greene 2008, 538): the absolute value of a normally distributed variable (“half-normal”) and an exponentially distributed variable. The distributions are asymmetric. However, the problem with stochastic frontier analysis is that the error term distribution assumption has its effects on the magnitude of the measure of the fan loyalty. If the team specific term is fixed, one of the teams is considered strong (as 100 % strong) in the sense of fan loyalty. Fans are committed. The fan loyalty of the other teams is relative to the best-practice team(s) in the sample (cf. Last and Wetzel 2010).

The fixed effects model is here more plausible since it captures both the relevant explanatory variables for attendance and the inefficiency scores. The estimated coefficients of the explanatory variables are in line with those reported in the previous literature. Since the team loyalty scores are correlated with the distance measure, the fans are more committed to ice hockey and not to a particular team. The brand of ice hockey is stronger than the brand of an individual team. This is consistent with the results of Bauer, Sauer and Exler (2005) that show that non-product-related attributes (e.g. stadium and regional provenance) are more important for fan loyalty than product-related attributes like players, success, and general team performance. It is clear that fan loyalty offers opportunities for monopoly pricing.

0.4 Essay 4: Spectators of performing arts – who is sitting in the auditorium?

The fourth essay examines the performing arts audiences using a bivariate probit and multivariate logit analysis. According to the statistics, around 5 per cent of the Finns go to see performing arts (art exhibition, opera or

20

theatrical performances) diligently and roughly 80 per cent occasionally (ISSP 2007). Audience and participation surveys argue that participation is segmented. Highbrow consumption is related to gender, age and formal education. Women are more active in highbrow art consumption, while men favour sports. The purpose of the fourth essay is to analyse differences in the visitor density in more detail. Can differences be observed between the regions when, for example, the effect of the educational background is taken into account? A bivariate probit model is useful because it estimates simultaneously two equations in cultural participation decisions. It also allows to study whether there is significant correlation between the equations’ random disturbances. With this method, the principal characteristics of the performing arts and the sport events audiences can be identified. Using Finnish data a study like this has not been conducted earlier.

The ISSP 2007 survey was carried out between 18th September and 11th

December 2007 through a mail questionnaire in Finland. The ISSP is a continuous programme of cross-national collaboration on social science surveys. The surveys are internationally integrated. In Finland the ISSP surveys are carried out together by three institutions: Finnish Social Science Data Archive, The Department of Social Research at the University of Tampere and the Interview and Survey Services of Statistics Finland. The cultural participation questions in the ISSP survey were: “How many times in the past twelve months have you seen an art exhibition, opera or theatrical performance?” and “How many times in the past twelve months have you been attending a sport event (ice hockey, football, athletics, motor race, etc.)?” Five alternatives were given: ‘Every day’, ‘Several times a week’, ‘Several times a month’, ‘Less often’ or ‘Never in the last twelve months’. However, it is widely known that the categories “every day” or “several times a week” or “several times a month” get a small number of respondents and it is reasonable to combine these categories with “less often” (e.g. Vander Stichele and Laermans 2006). One step further is to assume that the error terms of two explanatory models are correlated. One model is estimated for highbrow (ballet, dance performance, opera) and another for sports (lowbrow). The first step in the essay is to use the multivariate analysis of variance (MANOVA) to

21

simply compare the variance between the sample means explained by explanatory variables.

There are differences in the highbrow visitor density of across separate groups: “often”, “less often” and “never”. The results of the multivariate analysis of variance show that gender, age, education and the place of residence are significantly different across separate groups. However, the multivariate analysis of variance only reveals that there are differences, but it does not show the direction of the effect, i.e. it does not show whether for example women are more active than men in highbrow performing art consumption. The multinomial logit model (MNL) is the second step in the essay to find out what is the direction of the explanatory variables on art consumption. The explanatory variables in MNL are the following: gender, classified age, education and the classified place of province. The classification is needed since there are good reasons to assume that the effect of age is not linear. On the contrary previous studies have shown that middle-aged people are most active performing art consumers. According to the Statistics Finland, the economy of culture (value added per capital in 2007) is highly concentrated (Figure 1).

Figure 1: Value added of culture in 2007, EUR/capita in NUTS3 and Capital regions in Finland

22

Source: Statistics Finland

The metropolitan area has clearly the highest value added of culture per capita and 52) % of the culture labour is located there. In the MNL analysis one region must be considered as the reference value and the effects of region variables are relative to this reference region. The reference values in the MNL analysis are the following: Rural area exemplified by Northern Finland (FI1A according to the NUTS-2 classification), pupil or student (education 1), young person (age: 15-24). The region variables are mainly compatible with the NUTS-2 classification except for the provinces of Uusimaa and Ahvenanmaa.

23

The results of the MNL show that the ones that “often” go to performing art performances or exhibitions have graduated from the upper secondary school (edu5) or have a bachelor’s degree (university of applied sciences, edu7 or university, edu8) or have a master’s degree (edu9). Middle-aged people (age between 45 and 54 or between 55 and 64) go most diligently. Gender is important: women are more active than men. The above mentioned socio-economic variables, college level education (edu6) and somewhat younger age (age between 35 and 44) are significant to classify “less often” group from other visitor density groups. Regional differences are significant. The citizens of the province of Uusimaa or the region of Eastern Finland are the most active. A conclusion from the MNL models is that a crucial feature to classify into not attending and attending groups is at least upper secondary school. Furthermore, the separating feature between less often and often groups is at least a bachelor’s degree and 45 year age. Women on average are more active in highbrow art consumption. Furthermore, the essay studies what the roles of gender and other socio-economic variables are in sport events’ attendance.

The visitor density of sport events attendance is also investigated using a MNL model. Following the participating arts model, the sports events model has three groups: “often”, “less often” and “never”. Gender separates, but men are significantly more active than women. This result is in line with the participation motive models (Wann 1995) and with the statistics of the most popular sport events. Ice hockey and football are the most popular sports in terms of attendance and both could be classified as aggressive. A low education level (elementary school, edu2 or comprehensive school, edu3) is typical for those that are the most active and age less than 45. The results are mainly contrary to the performing arts participation results. However, the performing arts visitor density is added as an explanatory variable; it has a positive coefficient meaning that these two cultural segments have a common feature. Those that are active in highbrow art consumption are also active in sport event consumption. This is especially true for those that are “less often” goers. High education seems to be the common feature. There are no regional differences in sport consumption. The findings are consistent with the time-use survey evidence that highly-educated perform more activities and these include the consumption of cultural capital (Ruuskanen 2004).

24

Since there is a common factor in both participating arts (art exhibition, opera and theatrical performances) and sport events consumption, the bivariate profit model must be used to study the participation equations simultaneously. The fundamental difference between the multinomial logit models and bivariate probit models is to assume that the error terms of the two explanatory models are correlated. One model is estimated for highbrow (ballet, dance performance, opera) and the other for sports (lowbrow). The multinomial logit model estimates only one equation to explain cultural consumption, but it allows more than two categories (‘often’, ‘less often’ and ‘never’), while the bivariate probit model assumes that there is a binary variable to be explained. If the disturbances of the bivariate equations are correlated, both the direct marginal effects and the indirect marginal effects can be evaluated. The general specification for a two-equation model assuming the binary choice is (Greene 2008, 817):

y1¿=x1

' β1+ε 1 , then y1=1 if y1¿>0 ,∧ y1=0otherwise

y2¿=x2

' β2+ε2 , then y2=1 if y2¿>0 ,∧ y2=0otherwise

E [ε 1|x1 , x2 ]=E [ ε2|x1 , x2 ]=0

Var [ε1|x1 , x2 ]=Var [ε2|x1 , x2 ]=1Cov [ ε1, ε2|x1 , x2 ]=ρ

The marginal effects of each explanatory variable are more reasonable since both the direct marginal effect and the indirect marginal effect can be estimated. Since education for example has an effect on both cultural segments (arts and sports), the indirect effect reveals whether these cultural segments are substitutes or complements. If the direct marginal effect of (say) master’s degree education (edu9) is positive for arts and indirect marginal effect is negative, the arts and sports consumption are substitutes for this socio-economic group. The results of the bivariate probit model confirm the effects of gender, education and age. Women are active in highbrow consumption and men in sport events consumption. Direct marginal

25

effects of the education are significant if the education level is equal to or higher than upper secondary (college, a bachelor’s or master’s degree). The threshold age is 35. People older than 35 prefer arts and they diminish sport events consumption. The indirect marginal effects of education levels 6, 7 and 9 (a college diploma, and a bachelor’s degree from a university of applied sciences or a master’s degree) reveal that these citizens consider arts and sport events as substitutes. The correlation coefficient ρ of the error terms of the equations is 0.382 showing that the audiences of arts and sports have a common feature.

0.5 Essay 5: Are the spectators of performing arts and the spectators of the

movies the same?

The fifth essay uses a similar framework than the fourth essay, but the comparison is made between performing arts and cinema, and the effects of household incomes and family background have been added as explanatory variables. The marginal effects of the socio-economic variables on the performing art consumption in the multivariate logit model are examined using the ISSP 2007 survey data.

The descriptive statistics of the explanatory variables reveal that age (age-group) and education are related. Most of the youngest in the sample were pupils or students (at a comprehensive, an upper secondary, a vocational school or at a college) and correspondingly the oldest had a rather low education (elementary or comprehensive school). A college level education was mainly replaced by bachelor’s degree education in the early 1990’s and therefore people having a bachelor’s degree from a polytechnic (university of applied sciences) are somewhat younger than persons having a college diploma. People less than 50 –years old on average have a (better and) longer education than people older than 50. Age and education are related with the household or personal incomes. Middle-aged and high-educated seem to have the highest incomes (including all social security contributions, e.g. child benefit that may explain why the age-group 30-34 has the highest incomes).

26

There are some differences in education between genders. Men are somewhat less educated than women. Since the income variable in the sample includes all social security contributions (e.g. child benefit) the number of children is used as an explanatory variable. There are two different variables: the number of less than 6-year-old children and the number of 7-17-year-old children.

The results of the bivariate probit analysis when the age-cohort 50-54 and elementary school (edu2) are considered as reference value (i.e. the constant in the equation) show that the two spectator groups are not independent since the correlation coefficient of the error terms ρ = 0.625. Hence the hypothesis that the spectators of movies and arts belong to independent groups can be rejected. There are common characteristics, a common background which could be called an intrinsic culture orientation. If a person likes art exhibitions, opera and theatrical performances, she also likes to see movies at the cinema and vice versa, given that the institutions in the region offer these events. Those that are inactive and culture orientated do not go to exhibitions or performances and to the cinema. However, there are some particular effects that are related with exhibitions and performances or with movies. The importance of gender is very strong: females are more active in both arts (highbrow) and movies. The direct marginal effect of gender (female) is positive but the indirect marginal effect is negative. Both the direct and indirect marginal effects have been reported only for the highbrow art (art exhibition, opera and theatrical performances). The negative indirect effect describes the preference of seeing a film on the cinema. These leisure time activities are to some extent substitutes. Marital status matters: married or common-law married citizens go more often to see highbrow art than single people.

If the effect of age on cultural consumption is relative to the age cohort 50-54, all younger cohorts prefer more movies and only the oldest (70-74) seem to go less often to the cinema than the reference group. The indirect marginal effect of age on highbrow art is negative for each younger age cohort. The direct marginal effects of cohorts are not significant. The results indicate that age is not a relevant variable to classify highbrow art consumption into active and

27

inactive groups. Education seems to be very important to classify culture consumption structures. When the reference level is elementary school (edu2), citizens with any other education level are significantly more active in culture consumption, in both directions: highbrow art and movies. Omnivores have a higher level of education. Household’s size matters only indirectly to highbrow art consumption since bigger families seem to favour movies. The number of small children (less than 7) or older children (7-17) significantly reduces both culture consumption segments. The household incomes (or personal incomes, not reported here) are not significant.

The age cohorts 30-34 and 35-39 are most omnivore, but this indication is unreliable to some extent. The results of the MNL analysis confirm the importance of gender. Females are more active to go to an arts exhibition, opera and/or theatrical performances. Both the marginal effects of the gender variable or over individuals show that females most often belong to the group ‘less often’ (occasionally). The only marital status variable to classify into three groups is ‘married’. There are no differences if the person is single or living in common-law marriage. Married people most often belong to the group ‘less often’. The age cohort 25-29 is most passive in going to see performing arts. Surprisingly the older age-cohorts (55-59, 65-69 and 70-74) are most active. The oldest seem to strongly classify into totally not-going and actively going groups, but the probability of belonging into ‘less often’ –group is the lowest. Education is very important to classify performing arts consumption.

28

0.6 Conclusions

The essays have added information on the economics of culture in Finland. The sophisticated methods have been used. The role of public, non-advertising information on the cultural consumption decisions has been underlined. Essay 1 gives evidence that critical reviews published in the newspapers do not have an impact on movie admissions when the world-of-mouth is taken into account. In the ice hockey case, considered in essays 2 and 3, however, the public information in the form of series situation or the winning percentage of the team has an impact on the attendance figure. Furthermore, there is some evidence in essay 3 that ice hockey fans are more committed to the sport (species) than to the team. The brand of ice hockey is stronger than the brand of an individual team.

It is reasonable to assume that the marginal costs of most cultural events are almost zero, and the producers or distributors should maximize revenues. The cinema ownership in Finland is very concentrated and this leads to strategic behaviour. Essay 1 shows that movie attendance has price elasticity minus one, which as such follows the optimal pricing rule of monopolies. Movies are the most homothetic product in cultural capital, at least in respect of the most popular movies, though the number of annual attendees varies to significantly greater degree than in ice hockey or in highbrow culture. In the latter case, the policy is usually to satisfy all supply by means of special discounts e.g. in last minute reservations.

Essay 1 also shows that a wide release with extensive advertising should be used with mainstream films. Since the word-of-mouth is important the bad experiences of low quality films has a smaller effect on attendance. A hit-and-run strategy should be used with lower quality films while a platform release with a small number of initial screens should be favoured with high quality films. This is compatible with the results of Schmalensee (1978) who argue that low quality products should be advertised more intensively. The live

29

opera performances from the New York Metropolitan Opera that could be seen in the biggest towns’ cinemas in Finland were launched using a platform release. During the first year only live performances were offered but during the second season also encore performances were given some days later. Since the performing art audiences are highly educated the ticket prices for these cinema opera performances are substantially higher than the normal cinema tickets.

The audience composition can be studied using bivariate probit analysis. This analysis is an important method to classify audiences of different cultural events and simultaneously to classify consumers into omnivore and other groups. Essay 4 indicates that the time constraints on leisure activities are connected with the number of children in the family. Formal education is an important factor to classify consumers into different groups. Highly educated are more active in highbrow consumption, but according to essay 4 they go less often to sport events. However, the omnivorousness increases with formal education. Overall, essays 1 through 5 reveal that gender differences are important in both the performing art consumption and the sport consumption.

Essay 4 shows that sport consumption has similar types of characteristics although it is less elitist and typically favoured by men. 25% of total population attend at least once a year in the ice hockey match and the preferences are fairly price inelastic. Fans are loyal to the local ice hockey team, also to a large extent irrespective of its success or failure. The latter results only in limited substitution of a less successful team with the one with better recent performance. Fans are still loyal to ice hockey and not to the team. All this offers opportunities for monopoly pricing and we have observed in recent years substantial increase in the ticket prices.

Intellectual assets including cultural capital are highly agglomerated in the greater Helsinki area (Helsinki, Espoo and Vantaa): 52% of cultural capital and according to a recent study 48% of intangible capital is located in the capital region with a population share of 30% (Piekkola, 2011). These areas

30

also have on average more educated inhabitants with human capital. Cultural capital has been shown to be sensitive to the human capital of the attendees. In essays 4 and 5 highly skilled are typically more active than low educated and are engaged in several types of activities including cultural events.

One can also categorize the cultural capital in terms of the degree of experience consumption. Movies stand out again as the cultural capital with the least surprise content and hence most close to experience consumption. Highbrow cultural capital is less frequent and includes the biggest unknown element, although in certain dimensions sports are the most unpredictable.

All these findings are of big importance in the evaluation of cultural policy and subsidies for cultural consumption. The inelastic part is quite insensitive to subsidies and it is expected that rather high share of subsidies in the form of cultural spending voucher (kulttuuriseteli) benefit the price elastic cultural activities such as theatre. Both the state and local authorities subsidy production of highbrow cultural events but they should reallocate the policy towards consumption in terms of cultural spending voucher subsidies and diminish direct production related subsidies. The institutions should also offer last minute places to low-income students with reasonable discounts. It is also noteworthy that price elasticity is likely to be lower for high educated that can better afford allocate their time to several activities. On the other hand, highly educated people are more time constrained and do more voluntary work.

The study also shows clear substitutability between highbrow cultural capital and movies. Cultural capital is highly concentrated in the greater Helsinki area and there is every reason to believe that a fairer regional distribution will lead to much greater demand for highbrow cultural capital. Cultural capital policy is important regionally also because it has been shown that older people consume it less and hence the demand for cultural capital can be subject to dramatic changes in the rapidly ageing areas, many of which are located in rural areas in eastern and northern Finland. The substitutability between highbrow cultural consumption and sport events attendance is less

31

low; this explains why sport events demand falls into inelastic part while the movie attendance has price elasticity minus one.

Is cultural capital enjoyment or long-term investment? This study shows in many respects low price elasticity and hence the importance of non-monetary reasons for cultural capital consumption. Clearly price mechanism has only limited role or can lead easily to monopoly pricing rules, where our price elasticities show some evidence in movie attendance. This is supported by the surprisingly limited role that critics have on the consumer decisions. It is also noteworthy that the price elasticity in Finnish movie consumption has been found to be lower than that observed in the Great Britain or the United States (Davis 2002, 2006).

References

Alderson, Arthur S., Azamat Junisbai and Isaac Heacock (2007): Social status and cultural consumption in the United States. Poetics 35, 191-212

Bauer. Hans H.. Nicola E. Sauer & Stefanie Exler (2005): The loyalty of German soccer fans: does a team’s brand image matter? International Journal of Sports Marketing & Sponsorship. vol 7. issue 1. 14-22

Benham, L. (1972.: The Effects of Advertising on the Price of Eye-Glasses. Journal of Law and Economics 15: 337-352

Bihagen, Erik and Tally Katz-Gerro (2000): Culture consumption in Sweden: The stability of gender differences. Poetics 27, 327-349

Borland. Jeffery & Robert Macdonald (2003): Demand for sport. Oxford Review of Economic Policy. 19. 478-502

Brandes. Leif. Egon Franck & Philipp Theiler (2010): The Group Size and Loyalty of Football Fans: A Two-Stage Estimation Procedure to Compare Customer Potential Across Teams. University of Zurich. Institute for Strategy and Business Economics (ISU). No 126

Carlton, D. and J. Perloff (1990): Modern industrial organization. Harper Collins

Cubbin, John S. (1981): Advertising and the theory of entry barriers. Economica 48: 289-298

Davis, Peter (2002): Estimating multi-way error components models with unbalanced data structures. Journal of Econometrics 106: 67-95

32

Davis, Peter (2006): Measuring the Business Stealing, Cannibalization and Market Expansion Effects of Entry in the U.S. Motion Picture Exhibition Market. Journal of Industrial Economics 54: 293-321

Depken. Craig A. II (2000): Fan Loyalty and Stadium Funding in Professional Basketball. Journal of Sports Economics. vol. 1. 124-138

Depken. Craig A. II (2001): Research Notes: Fan Loyalty in Professional Sports: An Extention of the National Football League. Journal of Sports Economics. vol. 2. 275-284

Dietz-Uhler, Beth, Elizabeth A. Harrick, Christian End & Lindy Jacquemotte (2000): Sex differences in sport fan behaviour and reasons for being a sport fan. Journal of Sport Behavior, 12, 219-231

Eurobarometer 67.1, available at http://www.fsd.uta.fi/aineistot/kvdata/eb.html

Frey, Bruno S. (2000): Arts & economics. Analysis & Cultural Policy. Springer Verlag

Greene. William (2005): Fixed and Random Effects in Stochastic Frontier Models. Journal of Productivity Analysis. 23. 7-32

Greene. William H. (2008): Econometric analysis. 6th Edition. Pearson International Edition

ISSP 2007: vapaa-aika ja urheilu: Suomen aineisto [elektroninen aineisto]. FSD2330, versio 1.0 (2008-03-18). Blom, Raimo (Tampereen yliopisto. Sosiologian ja sosiaalipsykologian laitos) & Melin, Harri (Turun yliopisto. Sosiologian laitos) & Tanskanen, Eero (Tilastokeskus. Haastattelu- ja tutkimuspalvelut) [tekijä(t)]. Tampere: Yhteiskuntatieteellinen tietoaineisto [jakaja], 2008

Kahle, Lynn, Marc Duncan, Vassilis Dalakas & Damon Aiken (2001): The social values of fans for men’s versus women’s university basketball. Sport Marketing Quarterly, 10, 156-162

Kaldor, N. (1950): The Economic Aspects of Advertising. Review of Economic Studies 18: 1-27

Kansallinen Liikuntatutkimus 2005-2006: aikuisväestö [elektroninen aineisto]. FSD2233, versio 1.0 (2007-01-29).Espoo: Suomen Gallup [aineistonkeruu], 2006. Helsingin kaupunki, Nuori Suomi, Opetusministeriö, Suomen Kuntoliikuntaliitto, Suomen Liikunta ja Urheilu, Suomen Olympiakomitea [tuottajat].Tampere: Yhteiskuntatieteellinen tietoarkisto [jakaja], 2007.

Kansallinen liikuntatutkimus 2010, available at http://www.slu.fi/liikuntapolitiikka/liikuntatutkimus2/

Karjaluoto, Heikki (2010): Digitaalinen markkinointiviestintä. Esimerkkejä parhaista käytännöistä yritys- ja kuluttajamarkkinointiin. WSOY

Kivekäs, Kyllikki (1991) Alueoopperat ja yleisöt. Taiteen keskustoimikunta, Tilastotietoa taiteesta 5A

33

Koo, Gi-Yong & Robin Hardin (2008): Difference in interrelationship between spectators’ motives and behavioral intentions based on emotional attachment. Sport Marketing Quarterly, 17, 30-43

Kotimaisen elokuvan yleisöt –tutkimus 2010, available at www.ses.fi

Mikkonen, Jenni ja Katja Pasanen (2009) Selvitys Savonlinnan oopperajuhlien asiakasprofiileista ja aluellisesta vaikuttavuudesta. East side story – Puhtia itäsuomalaiseen tapahtumamatkailuun – hankkeen tutkimusraportti. Itä-Suomen yliopisto

Nelson, Philip (1970): Information and Consumer Behavior, Journal of Political Economy 78: 311-329

Nelson, Philip (1974): Advertising as information. Journal of Political Economy, 81: 729-754

Niemi, Iiris and Hannu Pääkkönen (1992): Vuotuinen ajankäyttö. Statistics Finland, Studies 183

Official Statistics of Finland (OSF): Cultural statistics [e-publication]. Cultural labour force 2007. Helsinki: Statistics Finland [referred: 21.2.2011]. Access method: http://www.stat.fi/til/klt/2007/01/klt_2007_01_2010-02-10_tie_001_en.html.

Piekkola (2011) Intangible capital agglomeration and economic growth: An Analysis of Regions in Finland. Innodrive working paper No. xx

Piekkola H. and O.-P. Ruuskanen (2006): Työtä eri elämänvaiheissa ja ajankäyttö. Äidit ja ikääntyvät. Sosiaali- ja terveysministeriön selvityksiä 2006: 73

Purhonen, Semi, Jukka Gronow and Keijo Rahkonen (2010): Nordic democracy of taste? Cultural omnivorousness in musical and literary taste preferences in Finland. Poetics 38, 266-298

Pääkkönen, Hannu (2010): Perheiden aika ja ajankäyttö. Tutkimuksia kokonaistyöajasta, vapaaehtoistyöstä, lapsista ja kiireestä. Tilastokeskus, tutkimuksia

Ridinger, Lynn L. & Daniel C. Funk (2006): Looking at gender differences through the lens of sport spectators. Sport Marketing Quarterly, 15, 155-166

Ruuskanen, Olli-Pekka (2004): An econometric analysis of time use in Finnish households. Helsinki School of Economics. Acta Universitatis Oeconomicae Helsingiensis A-246

Schmalensee, R. (1978): A Model of Advertising and Product Quality. Journal of Political Economy 86: 485-503

Simmons. Rob (2006): The demand for spectator sports. In Handbook on the Economics of Sport. edited by Wladimir Andreff & Stefan Szymanski. 77-89. Edward Elgar

34

Stigler, George and Gary S. Becker (1977): De Gustibus Non est Disputandum. American Economic Review 67, 76-90

Suomalaisen elokuvan markkinat ja kilpailukyky 1999, F&L Research

Suomen Teatterit ry (2007) Suomalaisten teaterissa käynti. Taloustutkimus Oy

Wakefield. Kirk L. & Hugh J. Sloan (1995): The Effects of Team Loyalty and Selected Stadium Factors on Spectator Attendance. Journal of Sport Management. vol. 9. 153-172

Vander Stichele, Alexander and Rudi Laermans (2006): Cultural participation in Flanders: Testing the cultural omnivore thesis with population data. Poetics 34, 45-64

Wann, Daniel L. (1995): Preliminary validation of the sport fan motivation scale. Journal of Sport & Social Issues, 19, 29-42

Wann, Daniel L., Frederick G. Grieve, Ryan K. Zapalac & Dale G. Pease (2008): Motivational Profiles of Sport Fans of Different Sports. Sport Marketing Quarterly, 17, 6-19

Wann, Daniel L., Matthwe J. Robinson, Ronald J. Dick & Andrew J. Gillentine (2003): Motives and Points of Attachment: Fans Versus Spectators in Intercollegiate Athletics. Sport Marketing Quarterly, 12, 217-227

Wann, Daniel L., Joel L. Royalty & Al R. Rochelle (2002): Using motivation and team identification to predict sport fans’ emotional responses to team performance. Journal of Sport Behavior, 25, 207-216

Wann, Daniel L., Michael P. Schrader & Anthony M. Wilson (1999): Sport fan motivation: Questionnaire validation, comparisons by sport, and relationship to athletic motivation. Journal of Sport Behavior, 22, 114-139

Warde, Alan and Modesto Gayo-Cal (2009): The anatomy of cultural omnivorousness: The case of the United Kingdom. Poetics 37, 119-145

Viljakainen, Anna, Asta Bäck & Ulf Lindqvist (2008): Media ja mainonta vuoteen 2013. VTT tiedotteita – research notes 2450

Winfree. Jason A.. Jill J. McCluskey. Ron C. Mittelhammer & Rodney Fort (2004): Location and attendance in major league baseball. Applied Economics 36.2117-2124.

Virtanen, Taru (2007) Across and beyond the bounds of taste on cultural consumption patterns in the European Union. Turun kauppakorkeakoulu Sarja A-11: 2007

35

1 Critics´ reviews or preceding week’s admissions explaining movie admissions

1.1 Introduction

Critics and their opinions or critical reviews have been shown to have an impact on movie admissions. Critics or reviewers are typically invited to an early screening of the film and write critics before the film opens to the public. This information is important for many experience goods, like restaurants, theaters, books, movies. Other information is also available after the first night. Word-of-mouth has been recognized as one of the prime resources of information transmission. It is natural that critical reviews have an impact on premiere weekend’s movie admissions, while word-of-mouth is more important to explain overall (long run) admissions (Basuroy, Desai & Talukdar 2006).

The star power of actors, director power and awards or nominations for awards are movie-related characteristics that have been shown to have an impact on movie admissions (Hennig-Thurau, Houston and Walsh 2007). Production budget seems to correlate with opening weekend screens and post-filming actions like advertising have been found to bring about success to theatrical box office revenue (Hennig-Thurau, Houston and Walsh 2006). Conventional economics postulates that price affects demand, however, in the movie admission or movie box office literature the effect of price has been neglected. One possible explanation for this shortage is that theaters seem to use uniform pricing (Orbach and Einav 2007). There are several possible explanations for using uniform pricing. Demand uncertainty might result in uniform pricing. Consumers might assume that prices reflect quality; low prices for low quality movies and high prices for high quality movies. To avoid this signaling, distributors choose uniform prices. Another explanation is that selling packages of several tickets would require monitoring mechanisms to prevent using low-price tickets and watching high-priced movies. Therefore, uniform pricing is often used. But in Finland movie theater tickets are not totally uniform. Typically the ticket price for children and conscripts is lower and also during weekdays prices might be lower than weekends. A single

36

ticket is cheaper within a package of several tickets; therefore, the actual average price of a display is not uniform. However, a large portion of distributors’ profits come from adjunct sales (e.g. popcorn, confectionery), and the price setting has minor importance (Chen 2009). It is also true during the last decade that the role of video (DVD, Blu-Ray) rental has increased the share in film producers’ profits.

The objective is to study what the relevance of critical reviews and word-of-mouth is in explaining movie admissions. The contribution of this study is that Finnish weekly panel data is used to evaluate the role of critical reviews in weekly movie admissions with having other control variables, like the number of screens, the ticket price and so on. With the panel data, conventional regression analysis cannot be used since the results might be biased. The benefits of using the panel data are that (1) individual heterogeneity can be controlled, (2) estimated parameters are more efficient, and (3) with the panel data the dynamics of adjustment can be studied better (Baltagi 2008, 6-7). The panel data suggests that individuals or movies are heterogeneous. Time-series and cross-section studies that do not control heterogeneity might yield biased results.

The variables in the panel data of this study are partially conventional and partially new. Critical reviews published in newspapers and word-of-mouth measured as previous week’s admission in Helsinki theaters are among the conventional variables. Weekly ticket price measured as the ratio of box office revenue and admissions is the new candidate to explain weekly admissions. The results of this study indicate that when word-of-mouth is taken into account, critical reviews do not seem to significantly explain weekly movie admissions. Since admission figures are typically highest during the first weeks, a variable “weeks since released” is used to control for this peak. The analysis shows that it is significant, as well as the price variable. The price elasticity of weekly movie admission is roughly -1. The panel data analysis also indicates that the fixed effects model is the most suitable for explaining weekly movie admissions in Finland in 2003.

The article continues with a literature review, the presentation of the empirical model and variables. This is followed by an analysis of why panel

37

data models have been used. Section 4 presents estimation results, and section 5 concludes.

1.2 Literature review

The correlation analysis of Eliashberg and Shugan (1997) has been very influential in explaining critical reviews on movie admissions. Critics could act as opinion leaders (influencers) who are considered more experienced and having more knowledge of the quality of movies. On the other hand, critics could act merely as predictors without any impact on early box office revenue. Influencers have an impact on early box office revenues, while predictors have an impact on overall box office revenues. The impact of critic reviews has been found positive in many studies. Basuroy, Desai and Talukdar (2006) also consider the impact of consensus on critics’ opinion (from Variety magazine) on movie admissions.

Eliashberg and Shugan (1997) assess only the aggregate impact of critics, while Boatwright, Basuroy and Kamakura (2007) also consider individual critics reviews published in newspapers and magazines with large circulation, like Entertainment Weekly, LA Times, Chicago Tribune or New York Times. TV shows have an even bigger audience, but Reinstein and Snyder (2005) do not find a significant impact of critics on movie admissions when such reviews are revealed in television talk shows (national, USA). A big majority of studies on movie admissions have been carried out with US data but there are interesting examples with other countries. Elberse and Eliashberg (2003) estimate demand (box office revenue) and supply (number of screens) equations for several countries: the USA, France, Germany, Spain and the UK. Critical reviews have a significant and positive impact on premiere week’s demand in the USA and UK but a negative impact on first week’s supply. The impact is not significant for France, Germany or Spain.

With Dutch data Gemser, Van Oostrum and Leenders (2007) show that the number and size of film reviews have an impact on art house premier week’s revenue and also overall box office revenue, while the same impact for mainstream movies is valid only for overall or long run box office revenue. On

38

the other hand, Hennig-Thurau, Houston and Walsh (2006) present results that indicate a positive impact of critical reviews on short- term theatrical box office but not on long- term box office nor video rental revenues of movies released during 1999-2001 in the USA. Recently Elliott and Simmons (2008) showed that higher average critic ratings are associated with greater box office revenues and increased advertising in the UK. They also remark that advertising is greater for films with higher US opening revenues and higher budgets.

Hence the influence of film reviews is supported in many studies, but d’Astous, Carú, Koll and Sigué (2005) argue that the influence depends strongly on cultural dimension. Using Hofstede’s (1984) theoretical framework in predicting consumers’ movie attendance they show that based on differences in power distance between Austria, Canada, Colombia and Italy, Austrian and Canadian moviegoers are more susceptible to value-expressive social influence than Colombian or Italian audience. The impact of consultation on film reviews is stronger among Austrian and Canadian moviegoers. Consistent with the level of uncertainty avoidance, Canadian audience appreciates more movie genres than Austrian, Colombian or Italian audience. Consumers with higher uncertainty avoidance are more brand loyal. In addition, d’Astous, Colbert and Nobert (2007) propose that moviegoers may be influenced by the movie’s country of origin when they search for information about new movies. It is also true that critical reviews might be biased. Ravid, Wald and Basuroy (2006) propose that several critics are significantly affected by the film distributor’s identity. High budget films seem to get more reviews, but these reviews are worse than average, also films with star decoration tend to get more reviews with positive assessment.

A hypothesis can be set as a summary: positive critic reviews have a positive effect on the spectator number.

Word-of-mouth (WOM) has also a powerful effect on movie admissions. Basuroy, Desai and Talukdar (2006) measure WOM as the cumulative number of screens since its release and they find a positive effect. WOM incorporates three effects: valence (positive, neutral, negative), volume of mouth-to-mouth discussion and persuasiveness of WOM generated. Neelamegham and Chingagunta (1999) on the contrary find no significant results between weekly

39

revenue and WOM measured as cumulative viewership and they argue that cumulative viewership is not a good proxy for WOM. Elberse and Eliashberg (2003) used previous week’s average revenue per screen as a proxy for WOM and they find significant positive results. Liu (2006) proposes that the volume of WOM (from Yahoo Movies Web site) offers significant explanatory power for both weekly and overall box office revenue, but the valence of WOM (measured as percentages of positive and negative messages) is not significant. WOM is more trustworthy than advertising or critical reviews since it comes from other moviegoers. Recently Duan, Gu and Whinston (2008) show that box office revenue of a movie and online WOM valence (measured on a daily basis from three web site sources: Variety.com, Yahoo!Movies and BoxOfficeMojo.com) have a significant impact on WOM volume which in turn leads to higher box office revenues. Moul (2007) proposes that WOM accounts for 10 % of the variation in the consumer expectations of movies, while distribution related effects, like the number of screens, release time and movie fixed effects, like star power, production budget comprise the great majority of observed variation in movie admissions.

The second hypothesis is therefore: Word-of-Mouth has a powerful effect on admissions.

The information flow through WOM affects also supply. The number of screens must adapt as demand develops dynamically. The prior screen decisions made before the actual release must be adjusted as the attendance number is known during the first weeks after premiere. The demand – supply dynamics in the movie industry, however, is subject to high uncertainty (De Vany and Walls 1996) but DeVany and Lee (2001) show that WOM can be a credible means to share information about good and bad movies.

Movie related elements like the star power of actors (Bagella and Becchetti 1999, Neelamegham and Chintagunta 1999, Walls 2005, Elberse 2007 or Meiseberg, Erhmann and Dormann 2008), director power (Bagella and Becchetti 1999 or Jansen 2005) or awards/nominations (Deuchert, Adjamah & Pauly 2005) seem to correlate with higher box office revenue or movie admissions but the evidence is, however, mixed (see e.g. Elberse and Eliashberg (2003), Hennig-Thurau, Houston and Walsh (2006) or McKenzie 2009). Bagella and Becchetti (1999) show that the star power of actors and

40

directors have a positive impact on admission but, on the contrary, McKenzie (2009) reports the insignificance. Deuchert, Adjamah and Pauly (2005) prove that nominations generate extra income, while awards do not have this effect. On the other hand, Lee (2009) has recently proved that there is a negative relationship between drama awards and box office revenues as the cultural distance between the USA and the country where the movie is shown grows. There is also strong evidence for a relationship between weekly revenues, opening week revenues or cumulative revenues and the number of screens (Elberse and Eliashberg 2003). Sequels also seem to collect a greater admission figure than contemporaneous non-sequels (Basuroy and Chatterjee 2008).

Movie distributors seem to release more hits (blockbusters) during high season, like the beginning of summer and during the Christmas holiday season. Collins, Hand and Snell (2002) show that action, adventure, horror or romantic comedy movies are more often blockbusters than other genres of movies. Einav (2007) proposes that roughly two-thirds of the seasonal variation can be explained by underlying demand. The rest i.e. a third is associated with the number and quality of movies. Einav also shows that wide release is often associated with heavy advertising, while word-of-mouth is more important in platform release. Wide release begins with a large number of screens with extensive national advertising. But only few widely released films are successful so that they are running many weeks (DeVany and Walls 1997). Platform release begins with a small number of initial screens and expands to additional screens and also to rural areas. Typically the movies with platform release cannot be classified as mainstream movies, or actors are not well known stars. The production budget of a movie or prior advertising also seem to correlate with the number of premiere week screens (Elberse and Eliashberg 2003).

Only few studies have considered the role of the price of the ticket. Davis (2002) estimates that the theater price elasticities of demand are about -3. The six theaters in the sample displayed different number of movies ranging from 2 to 9 during a six-week period. Davis (2006) presents also similar consumer price sensitivity results. In Dewenter and Westermann (2005) the price elasticity is about -2½ with German long-term (1950-2002) annual data.

41

Several different methods have been used to study the relationship between box office revenue or movie admission and the explanatory variables, like correlation analysis (Eliashberg and Shugan (1997)), OLS, or partial least squares (Hennig-Thurau, Houston and Walsh (2006)).

Elberse and Eliashberg (2003) used OLS, 2SLS and 3SLS to explain the supply of movies (screens as dependent variable) or demand for movies (revenues as dependent variable) with various predictors (budget, stars, director, advertising expenditure, reviews, etc.). Both 2SLS and 3SLS take into account the endogeneity and simultaneity of screens and revenues. OLS is inconsistent since the endogeneous variable screens used as explaining variable in the revenues equation is correlated with the error term of the same equation. Such correlation may occur when the dependent variable causes at least one of the regressors ("reverse" causation), when there are relevant explanatory variables which are omitted from the model, or when the covariates are subject to measurement error. Since the error terms across equations may be correlated, a 3SLS method is more efficient than 2SLS. Elliott and Simmons (2008) also use 3SLS method to estimate simultaneously supply (opening screens), advertising and demand (total revenue).

Recently Einav (2007) estimated a nested logit demand model for weekly market shares for movies. Nested logit is a suitable method to assort two or more choice problems. With this model Einav distinguishes seasonality (first level: to go to a movie) and the quality of a movie (second level: to choose between different movies). Therefore, the second level endogeneous variable is the market share of each movie. Also Ainslie, Drèze and Zufryden (2005) have estimated the market share of a movie using a random effects logit model with a gamma diffusion pattern. As consumers make the decision to see a movie, the time to decide and the time to act is derived from gamma distribution. They show that the impact of screens on movie sales may be lower than previously thought because screens act as a proxy for seasonality. Another interesting model is presented by Neelamegham and Chintagunta (1999). They use a Poisson count data model with the number of screens, distribution strategy, genre of a movie and stars explaining movie admissions. They find that the number of screens is the most important factor on admissions. An interesting model to predict box office success with neural networks is presented by Sharda and Delen (2006). Their neural network

42

approach is suitable to classify movies into nine different categories ranging from flop (box office revenue less than USD 1 million ) to blockbuster (revenue more than USD 200 million).

Davis (2002) uses the error components model (ECM) with unbalanced panel data. The data consists of sales, price and theater characteristics for six movie theaters and for a six-week period. A multinomial logit of demand for theaters is estimated and both own and cross price elasticities are reported. Theater demand is rather price sensitive, cross price elasticity between theaters not in the same group is practically zero but within the group cross price elasticities are positive and rather large. Recently Davis (2006) showed using the generalized method of moments (GMM) and a multinomial logit (MNL) demand model that low cross price elasticities between theaters is associated with (high) travel costs.

As a summary of the theoretical and previous empirical literature, the following equation is reasonable:

admission = f(critics, WOM, Z)

in which Z includes the other explanatory variables. The main focus is the role of critical reviews (critics) and Word-of-Mouth (WOM). What is the importance of critical reviews as the WOM is taken into account?

1.3 Empirical model and variables

The empirical study focuses on assessing the effects of various factors on weekly movie admissions in Finland in 2003. The Finnish Film Foundation (FFF) collects data from various distributors and importers. In 2003 the total number of films in distribution was 225 with only 177 premieres. Only 14 premieres were domestic, but the share of domestic movies in total admission was about 22%. Domestic film “Bad Boys – A True Story” got the biggest admission number: 614097 with roughly € 4.4 M total box office revenue. The ultimate week was the last week (53rd. i.e. Friday 26th December 2003 to Thursday 1st January 2004) when the top 20 movies collected 296495 admissions. The lowest figure was 48135 at the end of June. During the

43

ultimate week “Lord of The Rings: Return of The King” had 165502 admissions in 68 screens and “Underworld” was the last on the top 20 list with an audience of 606 in 2 screens. The median weekly admission was 138361 and the median screen number was 368 in 2003. Table 1 presents an overview of top 10 films in 2003 in Finland. The sample in this study has 1060 observations, there were 53 weeks with 20 biggest admission movies.

Table 3 (Table1.1) Overview of top 10 films in 2003 in Finland, source: Finnish Film Foundation

Original title of the film Release

date

Screens Total

gross

box office

Admission

s

Country

of Origin

Distributor

Bad Boys – A True Story (local) 17.1.2003 55 4413507 614097 Finland BVI

Lord of The Rings: The Two Towers 18.12.2002 58 3610000 467644 USA SF/FS

Lord of The Rings: Return of the

King

17.12.2003 68 3060269 355739 USA SF

The Matrix – Reloaded 21.5.2003 55 2364215 334206 USA SMD

Bruce Almighty 25.7.2003 32 2103080 279485 USA SF

Johnny English 11.4.2003 45 1912100 260643 UK UIP

Sibelius 12.9.2003 50 1885625 257031 Finland BVI

Pirates of The Caribbean 29.8.2003 44 1865774 245252 USA BVI

Piglet’s BIG Movie 29.8.2003 48 1398415 228421 USA BVI

Helmiä ja sikoja (local) 29.8.2003 40 1586939 213385 Finland Nordisk

Film

Previous empirical evidence (good surveys: Hennig-Thurau, Walsh and Wruck 2001 and Eliashberg, Elberse and Leenders 2006) has shown that the demand for movies is determined by several factors. On the supply side, the number of screens is probably the most important factor. Once the movie production has been completed it is ready for distribution. The launch stage includes both the physical distribution of the prints to the theaters and the marketing activities. Einav (2007) points out that a wide release is associated with heavy advertising, while platform or narrower release is more often associated with information diffusion through word-of-mouth (WOM).

It is here merely assumed that the number of screens is positively associated with movie admissions. Weekly movie admissions and the number of screens (“prints this week”) were collected by FFF which is the source of the data. Prints this week can include several showings during that week, typically there are some showings during the weekends, e.g. one at 3 p.m., the second

44

at 6 p.m. and the last at 9 p.m. Hence the number of screens underestimates the actual showings.

Expert reviews or ratings (critical reviews) and previous week’s movie admission (WOM) can convey some information about the quality of a movie. Critical reviews can influence consumers in their selection process. This is the influence effect. On the other hand, reviews can forecast whether the film becomes a success or not. This is the prediction effect of critical reviews (Eliashberg and Shugan 1997). Different proxies have been used to measure WOM in the literature. In this study critics’ reviews have been published weekly on Fridays in “Nyt”, which is a supplement to Helsingin Sanomat that has the largest newspaper circulation in Finland. In 2003 the subscription number was about 420,000, i.e. almost every twelfth Finnish citizen receives this newspaper home delivered. There are five reviewers that independently judge films in other newspapers than Nyt which simply collects and republishes these reviews. Three are Finnish and their critics are published in different newspapers and magazines: Helena Ylänen (Helsingin Sanomat), Antti Lindqvist (TV-maailma), and Tapani Maskula (Turun Sanomat). Helena Lindblad publishes her critics in Sweden (Dagens Nyheter) and Derek Malcolm in the UK (Guardian). Their judgement is published as stars ranging from 5 (superior) to 1 (loss of time). The average number of stars is published weekly and films are in descending order. The most liked film is on the top of the table and the least liked film is on the bottom. Each week 10 movies are valued. For 43 movies the stars indicator is shown only once but there are movies for which the stars indicator is published in more than ten succeeding Nyt2. 133 movies were critically reviewed in Nyt. Ylänen reviewed 65, Lindqvist 118, Maskula 105, Lindblad 77 and Malcolm 75. But in the panel sample (20 top movies, 53 weeks, i.e. 1060 observations) there are e.g. 211 non-zero observations of Ylänen’s critical reviews. The average value of critical reviews is used as an explaining variable in the estimations.

Word-of-mouth is also based on tables printed in Nyt. The previous week’s top 10 admission figures at theatres in Helsinki are listed on the same page as critical reviews. Typically the share of theatres in Helsinki in total admissions

2 Descriptive statistics for critical reviews is given in the appendix (table 2). It reveals that the critics of many “lower quality” is published only once or twice since the mean of critical review rank is decreasing in time (weeks).

45

is about 35-40 %.3 Both the actual number of admissions and ranking from 1 to 10 is printed. The film with the biggest admission in Helsinki theatres is ranked as number 1, and so on. Since that information is on the same page as critical reviews, both of these variables are used to explain next week’s movie admissions in whole Finland.

The proxy for word-of-mouth in this study (previous week’s attendance in Helsinki) has a connection to what have been used elsewhere: cumulative number of screens since its release (Basuroy, Desai and Talukdar 2006), cumulative viewership (Neelamegham and Chingagunta 1999), and previous week’s average revenue per screen (Elberse and Eliashberg 2003). Herr, Kardes & Kim (1991) or Grewal, Cline & Davies (2003) show that anecdotal information presented in a face-to-face manner (vivid WOM) has a greater impact on product judgments than the same information presented in printed form (e.g. advertising, critical reviews)4. In this study it is assumed that previous week’s attendance in Helsinki theaters is a suitable measure for vivid WOM.

Seasonal variation is very important since many blockbusters are released during the high season. The highest movie admission month in Finland has been January during a five-year period from 2003 to 2007 and June has been the lowest.

The weekly admission number is shown in appendix in figure 1. It reveals that the Christmas season and the end of May (the school year end) and late July/early August (the summer holiday end) are the peaks in movie admission. A proxy variable for seasonal variation is the number of all screens for all movies that week. Admission is highest typically during the first weeks for blockbusters (e.g. Ainslie, Drèze and Zufryden 2006). The life cycle of sleeper movies is different since demand peaks later; weeks 4 and 5 from the release demand is highest. The mean duration of a movie run is typically 7 to 10 weeks in Western countries (Neelamegham and Chintagunta 1999, table 1). A control variable to take the life cycle effect into account is needed: weeks

3 In 2005 three important cities, Helsinki, Tampere and Turku had a 56% share in total admissions and a 57 % share in gross box revenue. Source: European Cinema Yearbook 2006

4 On the importance of WOM vs. public information, see Hidalgo, Castro & Rodriguez-Sickert (2006)

46

since released. The median duration run of films with the biggest admission number in Finland is 17 weeks for the ultimate top 10 (1st to 10th) and roughly 10 weeks for the following 3 quantiles (from 11th to 40th)5.

Descriptive statistics and the hypothesis (expected signs) are summarized in table 2. The sample consists of 53 weeks with 20 top movies each week. The price variable is simply box office revenue/admission which takes into account both the difference between the price of using packages of several tickets and normal tickets as well as children/conscripts’ lower prices compared with normal prices.6 For some cases, especially among the lowest box office films, revenue data was not available and some approximation was needed. Either previous week’s revenue was used or revenue was set lower than the lowest reported revenue. Only less than 10 films the revenue data were missing and therefore price variables are approximated. Since all the films in the sample have not been critically evaluated or listed on Helsinki top 10, there are lots of zero observations. For the entire sample a dummy variable “not critically reviewed” (NOTCR) or “not top10” (NOTHK) is used. Otherwise the logarithmic values of the variables are used and therefore the estimated parameters are elasticities.

5 See appendix 3.

6 The percentiles (min, 10th, 20th, .., med, 60th, 70th, … max) in the price variable are: 1 – 5,95 – 6,52 – 6,83 – 7,07 – 7,27 (med) – 7,42 – 7,56 – 7,66 – 7,79 – 10,47 (max).

47

Table 4: (Table 1.2) Descriptive statistics and sources of variables, * weekly, ** non-zero observations

Variable Mean Median sd min max valid

observations

source expected

sign

Weekly Admission 6783,97 2240 14003,

4

65 165502 1060 FFF

Screens (SCR) 17,10 10 15,33 1 70 1060 FFF +

All Screens (ALLSCR) 341,94 368 72,92 176 471 1060 FFF +

Box office revenue (BOR) 50005 15825 109700 390 116581

4

1060 FFF

Price = BOR/Admission (PRICE) 7,04 7,27 0,88 1,00 10,47 1060 -

Critical reviews, average (CA) 2,83 3 0,90 1 5 133* Nyt +

Critical reviews, average (CA) 0,96 0 1,48 0 4,7 1060 Nyt +

Critical reviews, average (CA) 2,98 3 0,87 1 4,7 340** Nyt +

WOM (previous week’s admission in

Helsinki) (HKIADM)

2391,12 1500 2606,4

0

239 21271 520** Nyt +

WOM (previous week’s admission in

Helsinki) (HKIADM)

1173 0 2181,6

3

0 21271 1060 Nyt +

WOM (previous week’s admission in

Helsinki, rank) (TOP10)

5,44 5 2,86 1 10 520** Nyt -

WOM (previous week’s admission in

Helsinki, rank) (TOP10)

2,67 0 3,38 0 10 1060 Nyt -

Weeks since released (WEEKSREL) 8,25 5 8,73 0 56 1060 FFF -

1.4 Estimation and results

Since the data has both time-series (weekly) and cross-sectional (different movies) dimension, conventional regression analysis cannot be used. Panel data analysis enables regression analysis with both time-series and cross-sectional dimension. Panel data can have group effects (movies), time effects or both. Panel data models estimate fixed and/or random effects models using dummy variables. The core difference between the fixed and random effect models lies in the role of dummies. If dummies are considered as a part of the intercept, it is a fixed effect model. In a random effect model, the dummies act as an error term7. The fixed effect model examines movie differences in intercepts, assuming the same slopes and constant variance across the movies. Fixed effect models use least square dummy variables (LSDV), within effect, and between effect estimation methods. Thus, ordinary least squares 7 Hun Myoung Park: Linear Regression Models for Panel Data Using SAS, STATA, LIMDEP, and SPSS. http://www.indiana.edu/~statmath/stat/all/panel/panel.pdf accessed 5th February 2008

48

(OLS) regressions with dummies, in fact, are fixed effect models. The random effect model, by contrast, estimates variance components for groups and error, assuming the same intercept and slopes. The difference among groups (or time periods) lies in the variance of the error term. This model is estimated by generalized least squares (GLS) when the variance structure among genres is known. The feasible generalized least squares (FGLS) method is used to estimate the variance structure when the variance structure among genres is not known. Fixed effects are tested by the F test, while random effects are examined by the Lagrange multiplier (LM) test (Breusch and Pagan 1980). If the null hypothesis is not rejected, the pooled OLS regression is favoured. The Hausman specification test (Hausman 1978) compares fixed effect and random effect models. Table 3 (Park 2008) compares the fixed effect and random effect models. Group effect models create dummies using grouping variables (movie). If one grouping variable is considered, it is called a one-way fixed or random group effects model. Two-way group effect models have two sets of dummy variables, one for a grouping variable and the other for a time variable.

Table 5: (Table 1.3) Fixed Effect and Random Effect Models (Park 2008)

Fixed Effect Model Random Effect Model

Functional form

assuming νit ~

IID(0,σν2)

yit = ( +α μi)+Xit’ + β νit yit = + Xα it

’ +( β μi+ νit)

Intercepts Varying across groups (movies)

and/or times (weeks)

Constant

Error variances Constant Varying across groups

and/or times

Slopes Constant Constant

Estimation LSDV, within effect, between effect GLS, FGLS

Hypothesis test Incremental F test Breusch-Pagan LM test

The least square dummy variable (LSDV) model, however, becomes problematic when there are many groups or subjects in the panel data. If the total number of periods is fixed and the total number of observations is vast, only the coefficients of regressors are consistent. The coefficients of dummy variables are not consistent since the number of these parameters increases as N increases (Greene 2008, 197). This is the so called incidental parameter problem. Too many dummy variables may weaken the model for adequately

49

powerful statistical tests. Under this circumstance LSDV is useless and another method might be used: the within effect model which does not use dummy variables but uses deviations from group means.

The estimation results for the full sample with three different models are presented in table 4: conventional regression (OLS) analysis, fixed effects model and random effects model with all explanatory variables.

50

Table 6: (Table 1.4) Estimation results, full sample, n = 1060

Model OLS without

group dummy

variables

LSDV, Fixed effects

model (FEM)

Random effects model

(REM)

Log SCR 0,780

(0,021)***

0,899

(0,030)***

0,862

(0,240)***

Log ALLSCR 0,314

(0,076)***

0,139

(0,089)

0,238

(0,079)**

Log PRICE 0,042

(0,125)

-0,088

(0,112)

0,001

(0,108)

Log WEEKSREL -0,378

(0,022)***

-0,711

(0,027)***

-0,619

(0,023)***

Log HKIADM 0,177

(0,025)***

0,081

(0,022)***

0,091

(0,021)***

Log TOP10 -0,253

(0,049)***

-0,184

(0,041)***

-0,218

(0,040)***

Log CA 0,250

(0,099)**

-0,025

(0,100)

0,096

(0,093)

NotCR -0,095

(0,139)

-0,345

(0,129)**

-0,249

(0,123)*

NotHK 0,551

(0,234)**

0,096

(0,205)

0,085

(0,197)

Constant 4,10

(0,523)***

5,42

(0,531)***

Depending variable is log of weekly admissions, n = 1060

Standard deviations in parenthesis

Adjusted R-sq 0,811 0,897 0,784

F-test 507,55*** 61,39***

Diagnostic LL 1777,81*** 2576,50***

Test statistics for the Classical Model

Constant term

only (1)

Log Likelihood

= -1720,44

LM test vs. Model (3)

395,01***

Group effects

only (2)

LL = -1461,90 Hausman test (FEM vs.

REM): 122,39***

X– variables

only (3)

LL = -831,54

X-and group

effects (4)

LL = -432,19

Hypothesis

tests

(2) vs. (1) LR test

517,09***

F test

3.99***

(3) vs. (1) 1777,80**

*

507,55***

(4) vs. (1) 2576.50**

*

61,38***

(4) vs. (2) 2059,41**

*

601,84***

51

(4) vs. (3) 798,69*** 7,07***

The test statistics indicate that fixed effects model is favoured. The number of screens, weeks since released, last week’s admission in Helsinki theatres, movies shown in Helsinki top 10 listing and a dummy variable “not critically reviewed” are significant and correctly signed variables to explain weekly movie admissions. In Helsinki top 10 listing the movie with the biggest previous week’s admissions is numbered as 1, the movie with the second biggest admission is numbered as 2, and so on up to 10. Hence TOP10 variable should get a negative coefficient. Since the model is log-linear, other than dummy parameters are elasticises. Each movie has a different intercept (not shown). Since the other dummy variable “not top 10” (NOTHK) is not significant, another estimation is carried out beginning with the second week since released (i.e. NOTHK = 0). The sample size is now significantly lower, there are 515 observations.

52

Table 7: (Table 1.5) Estimation results, all movies with previous admission in Helsinki, n = 515

Model OLS without

group dummy

variables

LSDV, Fixed effects

model (FEM)

Random effects model

(REM)

Log SCR 0,613

(0,031)***

0,665

(0,043)***

0,710

(0,032)***

Log ALLSCR 0,177

(0,112)

-0,034

(0,111)

0,167

(0,098)*

Log PRICE -0,773

(0,295)**

-0,408

(0,245)

-0,511

(0,232)*

Log WEEKSREL -0,169

(0,043)***

-1,006

(0,050)***

-0,713

(0,041)***

Log HKIADM 0,487

(0,057)***

0,077

(0,042)

0,148

(0,041)***

Log TOP10 -0,160

(0,082)

-0,235

(0,061)***

-0,268

(0,058)***

Log CA 0,444

(0,104)***

0,056

(0,088)

0,242

(0,082)**

NotCR 0,422

(0,143)**

0,030

(0,108)

0,162

(0,103)

Constant 3,73

(0,712)***

6,757

(0,652)***

Depending variable is log of weekly admissions, n = 515

Standard deviations in parenthesis

Adjusted R-sq 0,837 0,946 0,777

F-test 332,20*** 81,03***

Diagnostic LL 943,96*** 1638,67***

Test statistics for the Classical Model

Constant term

only (1)

Log Likelihood

= -781,75

LM test vs. Model (3)

167,74***

Group effects

only (2)

LL = -554,44 Hausman test (FEM vs.

REM): 151,36***

X– variables

only (3)

LL = -309,78

X-and group

effects (4)

LL = 37,57

Hypothesis

tests

(2) vs. (1) LR test

454,62***

F test

5,45***

(3) vs. (1) 943,95*** 332,19***

(4) vs. (1) 1638,67**

*

81,02***

(4) vs. (2) 1184,05**

*

448,27***

(4) vs. (3) 694,71*** 10,76***

53

Fixed effects model is favoured, and screens, weeks since released and movie shown in Helsinki top 10 lists are significant, but the only dummy variable “not critically reviewed” is not significant. Therefore a new estimation is carried out without dummy variables.

Table 8: (Table 1.6) Estimation results, all movies critically reviewed and with previous week’s Helsinki admission, n = 205

Model OLS without

group dummy

variables

LSDV, Fixed effects

model (FEM)

Random effects model

(REM)

Log SCR 0,647

(0,053)***

0,531

(0,095)***

0,775

(0,051)***

Log ALLSCR 0,028

(0,172)

0,072

(0,174)

0,171

(0,146)

Log PRICE -0,361

(0,522)

-1,123

(0,373)**

-1,133

(0,339)**

Log WEEKSREL -0,275

(0,075)***

-1,134

(0,068)***

-0,851

(0,060)***

Log HKIADM 0,621

(0,085)***

0,163

(0,054)**

0,248

(0,051)***

Log TOP10 0,091

(0,133)

-0,004

(0,856)

-0,000

(0,080)

Log CA 0,380

(0,113)**

0,133

(0,114)

0,196

(0,097)*

Constant 2,519

(1,214)*

6,994

(0,950)***

Depending variable is log of weekly admissions, n = 205

Standard deviations in parenthesis

Adjusted R-sq 0,852 0,966

F-test 169,39 89,70***

Diagnostic LL 399,47 779,77***

Test statistics for the Classical Model

Constant term

only (1)

Log Likelihood

= -325,59

LM test vs. Model (3)

60,44***

Group effects

only (2)

LL = -151,72 Hausman test (FEM vs.

REM): 78,36***

X– variables

only (3)

LL = -125,86

X-and group

effects (4)

LL = 64,28

Hypothesis

tests

(2) vs. (1) LR test

347,74***

F test

10,68***

(3) vs. (1) 399,47*** 169,39**

54

*

(4) vs. (1) 779,76*** 89,70***

(4) vs. (2) 432,02*** 141,44**

*

(4) vs. (3) 380,29*** 12,31**

55

Table 9: (Table 1.7) Estimation results, all movies critically reviewed and with previous week’s Helsinki admission, n = 205

Model OLS without

group dummy

variables

LSDV, Fixed effects

model (FEM)

Random effects model

(REM)

Log SCR 0,631

(0,050)***

0,528

(0,092)***

0,775

(0,049)***

Log PRICE -0,284

(0,505)

-1,058

(0,329)***

-0,988

(0,313)**

Log WEEKSREL -0,268

(0,073)***

-1,139

(0,066)**

-0,860

(0,058)***

Log HKIADM 0,586

(0,058)***

0,166

(0,038)***

0,252

(0,036)***

Log CA 0,381

(0,111)***

0,136

(0,112)

0,205

(0,096)*

Constant 2,977

(1,064)**

7,677

(0,715)***

Depending variable is log of weekly admissions, n = 205

Standard deviations in parenthesis

Adjusted R-sq 0,853 0,967 0,802

F-test 238,43*** 93,69***

Diagnostic LL 398,64*** 779,51**

Test statistics for the Classical Model

Constant term

only (1)

Log Likelihood

= -325,59

LM test vs. Model (3)

60,84***

Group effects

only (2)

LL = -151,72 Hausman test (FEM vs.

REM): 79,72***

X– variables

only (3)

LL = -126,27

X-and group

effects (4)

LL = 64,15

Hypothesis

tests

(2) vs. (1) LR test

347,74***

F test

10,68***

(3) vs. (1) 398,63*** 238,43**

*

(4) vs. (1) 779,50*** 93,69***

(4) vs. (2) 431,76*** 200,62**

*

(4) vs. (3) 380,86*** 12,53***

The estimation results in table 6 and 7 indicate that word-of-mount measured as last week’s admission in Helsinki theatres seem to explain movie admissions, but critical reviews published in newspaper “Nyt” is not significant in fixed effects model that is favoured (Hausman test). Movie

56

admission is price sensitive with approximately -1 price elasticity. Test statistics for the classical model indicate that conventional regression analysis (OLS) without group dummy variables is not suitable for explaining weekly movie admissions. The t-statistics for critical reviews variable that illustrates significance is misleading due to misspecified model.

With the Finnish data, movie admission is inelastic with respect to number of screens. The screen variable does not take into account the number of actual seats in the hall. Blockbusters with a vast admission are shown in larger auditoriums and with more daily showings than arts movies. Increasing the number of screens is not as flexible as increasing daily showings if the movie turns out be a blockbuster. If the number of screens is still increased, these are probably with lower number of actual seats and therefore the relative admission increase is lower, and that might explain the inelasticity.

Two important hypotheses were imposed. Positive critical reviews should have a positive impact on movie attendance, but the results indicate that this is not true. On the contrary, when the Word-of-Mouth (second hypothesis) is taken into account critics do not explain attendance.

1.5 Conclusions and suggestions

In the movie admission or movie box office literature the importance of word-of-mouth has been well documented. Word-of-mouth has a positive effect on movie admissions (Elberse and Eliashberg 2003, Basuroy, Desai and Talukdar 2006, Liu 2006, Moul 2007, Duan, Gu and Whinston 2008). The evidence on the impact of critical reviews on movie admissions is mixed. Eliashberg and Shugan (1997) argue that critics could act as influencers or predictors. Influencers can predict opening box office revenue, while predictors can classify films either to successful or not-successful films in terms of revenue in the longer term. Hence the impact of critical reviews is not uniform. Some predict well short- term revenue and some better long- term revenue. Not only the existence of reviews but also the variation or consensus of critics can have an impact on admission (Basuroy, Desai and Talukdar 2006). The impact is also different depending on genre (Gemser, van Oostrum and Leenders 2007),

57

country of origin (d’Astous, Colbert and Nobert 2007, King 2007) and cultural dimension (d’Astous, Carú, Koll and Sigué 2005). Critical reviews may be biased towards distributor’s identity (Ravid, Wald and Basuroy 2006). This study shows with weekly Finnish data and using panel data estimation methods that word-of-mouth has a significant impact on movie admissions, while critical reviews have not. The critical review variable is the average value of five independent critics published in newspaper Nyt. The impact of an individual critic’s reviews has not been tested in this study and it needs to be done in the future. Are there differences among different genres? Are action movie lovers (younger and) less liable to rely on critical reviews and more liable to rely on word-of-mouth than drama and/or romance audience? Collins and Hand (2005) show with the UK data that richer and younger people are most likely to go to the movies, also the residential neighborhood matters.

An important implication for movie distributors in Finland is that they should use a wide release strategy when the expected WOM is negative. In many cases, the release weekend is later than it is in larger and English spoken countries. Hence there is some knowledge about the WOM in other countries. With the wide release strategy, this negative WOM has less influence since the strategy puts more weight on the first week and the WOM has less circulation time. On the contrary, if the expected WOM is positive, movie distributors should use platform release with a small number of initial screens and expanding later.

The star power of actors, director power or awards or nominations for awards have not been tested with the Finnish data since the share of domestic films in 2003 was only 14 % in premieres or 22 % in total admissions. The biggest admission film in 2003 was domestic and several main actors had received Jussi Awards some years before. Jussi Award is the most important Finnish award. It remains an open question whether these awards or well-known actors have had any impact on admissions or box office revenue.

The role of theater ticket price has been missing in international movie admission literature. Although the variation in prices is rather small, this study shows that movie admission is price sensitive. Davis (2002) showed that the theater demand is elastic with respect to price (about -2,3 to -4,1). With

58

the Finnish data, movie demand is roughly unit elastic. Conventional regression (OLS) analysis does not bring about significant and reasonable price elasticity estimates. Only panel data methods, especially fixed effects models are suitable for producing proper estimates.

References

Ainslie, Andrew, Drèze, Xavier and Zufryden, Fred (2005): Modeling Movie Life Cycles and Market Share. Marketing Science 24, 3: 508-517

d’Astous, Alain, Carú, Antonella, Koll, Oliver and Sigué, Simon Pierre (2005): Moviegoers’ Consultation of Film Reviews in the Search for Information: A Multi-country Study. International Journal of Arts Management 7, 3: 32-45

d’Astous, Alain, Colbert, François and Nobert, Véronique (2007): Effects of Country-Genre Congruence on the Evaluation of Movies : The Moderating Role of Critical Reviews and Moviegoers’ Prior Knowledge. International Journal of Arts Management 10, 1: 45-51

Bagella, M and Becchetti, L. (1999): The Determinants of Motion Picture Box Office Performance: Evidence from Movies Produced in Italy. Journal of Cultural Economics 23, 4: 237-256

Baltagi, Badi H. (2008): Econometric Analysis of Panel Data. 4th Edition, John Wiley & Sons Ltd

Basuroy, Suman and Chatterjee, Subimal (2008): Fast and frequent: Investigating box office revenues of motion picture sequels. Journal of Business Research 61: 798-803

Basuroy, Suman, Desai, Kalpesh Kaushik and Talukdar, Debabrata (2006): An Empirical Investigation of Signaling in the Motion Picture Industry. Journal of Marketing Research 63: 287-295

Boatwright, Peter, Basuroy, Suman and Kamakura, Wagner (2007): Reviewing the reviewers: The impact of individual film critics on box office performance. Quantitative Marketing and Economics 5, 4: 401-425

59

Breusch, T. and Pagan, A. (1980): The LM Test and Its Applications to Model Specification in Econometrics. Review of Economic Studies 47: 239-254

Chen, Chien-Ping (2009): A Puzzle or a Choice: Uniform Pricing for Motion Pictures at the Box. Atlantic Economic Journal 37, 73-85

Collins Alan and Chris Hand (2005): Analyzing Moviegoing Demand: An Individual-level Cross-sectional Approach. Managerial and Decision Economics 26: 319-330

Collins Alan, Chris Hand and Martin C. Snell (2002): What Makes a Blockbuster? Economic Analysis of Film Success in the United Kingdom. Managerial and Decision Economics 23: 343-354

Davis, Peter (2002): Estimating multi-way error components models with unbalanced data structures. Journal of Econometrics 106: 67-95

Davis, Peter (2006): Spatial competition in retail markets: movie theaters. Rand Journal of Economics 37, 4: 964-982

Deuchert, Eva, Adjamah, Kossi and Pauly, Florian (2005): For Oscar Glory or Oscar Money? Journal of Cultural Economics 29: 159-176

DeVany, Arthur and Cassey Lee (2001): Quality signals in information cascades and the dynamics of the distribution of motion picture box office revenue. Journal of Economic Dynamics and Control 25: 593-614

DeVany, Arthur and W. David Walls (1996): Bose-Einstein dynamics and adaptive contracting in the motion picture industry. The Economic Journal 106: 1493-1514

DeVany, Arthur S. and W. David Walls (1997): The market for motion pictures: Rank, revenue, and survival. Economic Inquiry 35: 783-797

Dewenter, Ralf and Michael Westermann (2005): Cinema Demand in Germany. Journal of Cultural Economics 29: 213-231

60

Duan, Wenjing, Gu, Bin and Whinston, Andrew B. (2008): The dynamics of online word-of-mouth and product sales – An empirical investigation of the movie industry. Journal of Retailing 84, 2: 233-242

Einav, Liran (2007): Seasonality in the U.S. Motion Picture Industry. Rand Journal of Economics 38, 1: 127-145

Elberse, Anita (2007): The Power of Stars: Do Star Actors Drive the Success of Movies? Journal of Marketing 71, 4: 102-120

Elberse, Anita and Eliashberg, Jehoshua (2003): Demand and Supply Dynamics for Sequentially Released Products in International Markets: The Case of Motion Pictures. Marketing Science 22, 3: 329-354

Eliashberg, Jehoshua, Elberse, Anita and Leenders, Mark A. A.M. (2006): The Motion Picture Industry: Critical Issues in Practice, Current Research, and New Research Directions. Marketing Science 25, 6: 638-661

Eliashberg, Jehoshua and Shugan, Steven M. (1997): Film Critics: Influencers or Predictors. Journal of Marketing 61, 2: 68-78

Elliott, Caroline and Simmons, Rob (2008): Determinants of UK Box Office Success: The Impact of Quality Signals. Review of Industrial Organization 33, 2: 93-111

Gemser, Gerda, Van Oostrum, Martine and Leenders, Mark A.A.M. (2007): The impcat of film reviews on the box office performance of art house versus mainstream motion pictures. Journal of Cultural Economics 31: 43-63

Greene, William H. (2008): Econometric Analysis. 6th Edition, Pearson Prentice-Hall

Grewal, Rajdeep, Thomas W. Cline & Anthony Davies (2003): Early-Entrant Advantage, Word-of-Mouth Communication, Brand Similarity, and the Consumer Decision-Making Process. Journal of Consumer Psychology 13, 187-197

61

Hausman, J. (1978): Specification Test in Econometrics. Econometrica 46: 1251-1271

Hennig-Thurau, Thorsten, Houston, Mark B. & Walsh, Gianfranco (2006). The Differing Roles of Success Drivers Across Sequential Channels: An Appication to the Motion Picture Industry. Journal of the Academy of Marketing Science, 34 (4): 559-574

Hennig-Thurau, Thorsten, Houston, Mark B. and Walsh, Gianfranco (2007): Determinants of Motion Picture Box Office and Profitability: An Interrelationship Approach. Review of Managerial Science 1, 1: 65-92

Hennig-Thurau, Thorsten, Walsh, Gianfranco and Wruck, Oliver (2001): An Investigation into the Factors Determining the Success of Service Innovations: The Case of Motion Pictures. Academy of Marketing Science Review 6: 1-23

Herr, Paul, Frank Kardes & John Kim (1991): Effects of word-of-mouth and product-attribute information on persuasion: An accessibility-diagnosticity perspective. Journal of Consumer Research 17, 454-462

Hidalgo R, César A., Alejandra Castro & Carlos Rodriguez-Sickert (2006): The effect of social interactions in the primary consumption life cycle of motion pictures. New Journal of Physics 8, 52

Hofstede, G. (1984): Dimensions of National Cultures in Fifty Countries and Three Regions. In Explications in Cross-cultural Psychology, Deregowski, J.B., Dziurawiec, S. and Annis, R.C. (eds). Swets & Zetilinger, p. 144-155

Holbrook, Morris B. (1999): Popular Appeal Versus Expert Judgments of Motion Pictures. Journal of Consumer Research 26: 144-155

Jansen, Christian (2005): The Performance of German Motion Pictures, Profits and Subsidies: Some Empirical Evidence. Journal of Cultural Economics 29: 191-212

King, Timothy (2007): Does film criticism affect box office earnings? Evidence from movies released in the U.S. in 2003. Journal of Cultural Economics 31: 171-186

62

Lee, Francis L.F. (2009): Cultural discount of cinematic achievement: the academy awards and U.S. movies’ East Asian box office. Journal of Cultural Economics 33, 4, 239-263

Liu, Yong (2006): Word of Mouth for Movies: Its Dynamics and Impact on Box Office Revenue. Journal of Marketing 70, 3: 74-89

McKenzie, Jordi (2009): Revealed Word-of-Mouth Demand and Adaptive Supply: Survival of Motion Pictures at the Australian Box Office. Mimeo, Journal of Cultural Economics 33: 279-299

Meiseberg, Brinja, Erhmann, Thomas and Dormann, Julian (2008): We Don´t Need Another Hero – Implications from Network Structure and Resource Commitment for Movie Performance. Schmalenbach Business Review 60: 74-98

Moul, Charles C. (2007): Measuring Word of Mouth’s Impact on Theatrical Movie Admissions. Journal of Economics & Management Strategy 16: 859-892

Neelamegham, Ramya and Chingagunta, Pradeep (1999): A Bayesian Model to Forecast New Product Performance in Domestic and International Markets. Marketing Science 18: 115-136

Nelson, Philip (1970): Information and Consumer Behavior. Journal of Political Economy 81: 311-329

Orbach, Barak Y. and Einav, Liran (2007): Uniform prices for differentiated goods: The case of the movie-theater industry. International Review of Law and Economics 27: 129-153

Park , Hun Myoung: Linear Regression Models for Panel Data Using SAS, STATA, LIMDEP, and SPSS. http://www.indiana.edu/~statmath/stat/all/panel/panel.pdf accessed 5th February 2008

Ravid, S. Abraham, Wald, John K. and Basuroy, Suman (2006): Distributors and Film Critics: Does it take two to Tango? Journal of Cultural Economics 30: 201-218

63

Reinstein, David and Snyder, Christopher M. (2005): The Influence of Expert Reviews on Consumer Demand for Experience Goods: A Case Study of Movie Critics. The Journal of Industrial Economics 53: 27-52

Sharda, Ramesh and Delen, Dursun (2006): Predicting box-office success of motion pictures with neural networks. Expert System with Applications 30: 243-254

Walls, W. David (2005): Modeling Movie Success when ‘Nobody Knows Anything’: Conditional Stable-Distribution Analysis of Film Returns. Journal of Cultural Economics 29: 177-190

Data sources: Finnish Film Foundation (www.ses.fi); Helsingin Sanomat, Nyt – available at Päivälehden museo, Ludviginkatu 2-4, Helsinki, Finland (www.paivalehdenmuseo.fi)

Estimation method: LIMDEP - NLOGIT 4.0 (www.limdep.com)

AppendicesFigure 1.1: Weekly Total Admission, Years 2003 to 2007

1234567891011121314151617181920212223242526272829303132333435363738394041424344454647484950515253

0

50000

100000

150000

200000

250000

300000

350000

weekAdmis03Admis04Admis05Admis06Admis07

Table 10: (Table 1.8) Distributors’ premieres in 2001 – 2003

64

Distributor 2001 200

2

2003 examples in 2003 (or late 2002)

Columbia Tristar Egmo 27 27 28 Terminator 3, Charlie’s Angels, Bad Boys 2

FS Film 28 28 26 Lord of The Rings: The Two Towers, Lord of The Rings: Return of The King

Buena Vista 12 20 24 Bad Boys – A True Story, Sibelius, Pirates of The Caribbean

Scanbox 6 16 19 The Hours, The Human Stain, A la Folie

Sandrew Metronome 26 25 19 The Matrix Reloaded, The Matrix Revolutions, Harry Potter and The Chamber…

Cinema Mondo 19 17 16 The Pianist, Spirited Away, Stupeur & Treblements

Kamras Film Group 10 15 12 Good Bye Lenin, Nirgendwo in Africa, Cidade de Deus

UIP 20 17 12 Johnny English, Ring, Catch Me If You Can

Future Film 9 9 11 Swimming Pool, Evil Dead, Les vacances M Hulot

Senso Films 9 11 4 L’Ultimo bacio, Movern Callar, Last Orders

Rest; Kinoscreen, Rapid

Eye Movie, Finnkino

5 7 6 Bella Martha aka Mostly Martha, Lejontämjaren, Pure

All premieres 171 192 177

65

Table 11: (Table 1.9) Descriptive statistics for critical review rank (scale 1 – “top” to 10 – ”lowest”)

Variable Mean Median sd mi

n

max valid observationms source notes

Critical review, rank,

1st occurrence, display

6,92 8 2,57 1 10 133 Nyt 43 films are reviewed

only once

Critical review, rank,

2nd display

6,11 6,5 2,50 1 10 90 Nyt Critical reviews (index: 1 to 5)

is shown twice for 27 films

Critical review, rank,

3rd display

5,75 6 2,66 1 10 63 Nyt

Critical review, rank,

4th display

5,33 5 2,96 1 10 51 Nyt

Critical review, rank,

5th display

5,14 4 2,93 1 10 37 Nyt

Critical review, rank,

6th display

4,67 4 2,90 1 10 27 Nyt

Critical review, rank,

7th display

3,90 3 2,85 1 10 19 Nyt

Critical review, rank,

7th display

3,88 3 2,87 1 10 16 Nyt

Critical review, rank,

8th display

2,75 2 1,93 1 7 12 Nyt

Critical review, rank,

9th display

3,4 2 2,46 1 8 10 Nyt

Critical review, rank,

10th display

3 3 1,58 1 5 9 Nyt 11 weeks: 1 film, 12 weeks: 2 films

14 weeks: 3 films, 15 weeks: 1 film

18 weeks: 1 film, 20 weeks: 1 film

66

Table 12: (Table 1.10) Duration of movie run, quantiles

Variable Mean Median Screens, five first

weeks, mean

Screens, first

week, mean

Screens,

second

week, mean

Screens, third

week, mean

Top 10, duration

of movie run, weeks

17,3 17 44,5 29,8 46,1 49,3

Films 11-20, duration

of movie run, weeks

13,8 10,5 39,0 31,6 43,2 45,5

Films 21-30, duration

of movie run, weeks

13,9 10,5 30,1 28,7 34,1 33,2

Films 31-40, duration

of movie run, weeks

10,9 9 28,3 25,6 31,2 30,2

Films 41-50, duration

of movie run, weeks

7,8 7,5 21,8 17,4 24,2 27,7

Films 51-60, duration

of movie run, weeks

10 10,5 12,3 9,9 13,6 13,4

Films 61-70, duration

of movie run, weeks

6,6 6,5 8,2 8,9 9,3 7,9

Films 71-80, duration

of movie run, weeks

5,6 5 8,2 10,0 11,7 8,9

Films 81-90, duration

of movie run, weeks

5,3 5 3,6 4,7 4,8 3,4

Films 91-100, duration

of movie run, weeks

3,4 3,5 4,0 6,1 5,1 4,7

Films 101-110,

duration

of movie run, weeks

4 4,5 4,5 5,8 5,4 5,0

Films 111-120,

duration

of movie run, weeks

3 3,5 2,0 3,4 2,9 1,8

Films 121-130,

duration

of movie run, weeks

1,5 2 2,3 6,3 4,9 0,5

67

Table 13: (Table 1.11) Estimation results, n = 345

Model OLS without

group dummy

variables

LSDV, Fixed effects

model (FEM)

Random effects model

(REM)

Log SCR 0,955

(0,034)***

0,691

(0,068)***

0,929

(0,038)***

Log ALLSCR 0,131

(0,141)

0,119

(0,169)

0,093

(0,139)

Log PRICE 1,174

(0,454)*

-0,942

(0,381)*

-0,249

(0,340)

Log WEEKSREL -0,471

(0,043)***

-1,079

(0,048)***

-0,893

(0,041)***

Log HKIADM 0,263

(0,079)**

0,123

(0,057)*

0,116

(0,052)*

Log TOP10 0,018

(0,119)

0,050

(0,092)

-0,009

(0,083)

Log CA 0,374

(0,111)****

0,123

(0,127)

0,195

(0,104)

NotHK 2,009

(0,762)**

0,985

(0,579)

0,859

(0,522)

Constant 1,210

(1,161)

6,214

(1,020)***

Depending variable is log of weekly admissions, n = 345

Standard deviations in parenthesis

Adjusted R-sq 0,838

F-test 217,86***

Diagnostic LL 628,75***

Test statistics for the Classical Model

Constant term

only (1)

Log Likelihood

= -589,43

LM test vs Model (3)

118,15***

Group effects

only (2)

LL = -312,10 Hausman test (FEM vs

REM): 103,30***

X– variables

only (3)

LL = -275,05

X-and group

effects (4)

LL = 0,667

Hypothesis

tests

(2) vs (1) LR test

554,65***

F test

10,61***

(3) vs (1) 628,74*** 217,85***

(4) vs (1) 1180,21**

*

70,22***

(4) vs (2) 625,56*** 155,19***

(4) vs (3) 551,46*** 10,15***

68

Table 14: (Table 1.12) Estimation results, all movies critically reviewed and with previous week’s Helsinki admission, n = 205

Model OLS without

group dummy

variables

LSDV, Fixed effects

model (FEM)

Random effects model

(REM)

Log SCR 0,642

(0,050)***

0,740

(0,103)***

0,866

(0,052)***

Log PRICE 0,052

(0,529)

-0,976

(0,313)**

-0,880

(0,301)**

Log WEEKSREL -0,284

(0,076)***

-1,150

(0,059)***

-0,959

(0,054)***

Log HKIADM 0,545

(0,060)***

0,125

(0,034)***

0,184

(0,033)***

Constant 3,129

(1,117)*

8,167

(0,665)***

Depending variable is log of weekly admissions, n = 205

Standard deviations in parenthesis

Adjusted R-sq 0,841 0,971 0,777

F-test 268,51*** 74,02***

Diagnostic LL 376,61*** 792,65***

Test statistics for the Classical Model

Constant term

only (1)

Log Likelihood

= -322,30

LM test vs Model (3)

72,04***

Group effects

only (2)

LL = -141,72 Hausman test (FEM vs

REM): 70,56***

X– variables

only (3)

LL = -133,99

X-and group

effects (4)

LL = 74,02

Hypothesis

tests

(2) vs (1) LR test

361,16***

F test

11,69***

(3) vs (1) 376,60*** 268,51**

*

(4) vs (1) 792,65*** 106,18**

*

(4) vs (2) 431,48*** 255,71**

*

(4) vs (3) 416,04*** 15,62***

69

2 Demand for ice hockey, the factors explaining attendance of ice hockey games in Finland

The study focuses on season 2007 – 2008 ice hockey league games in Finland.

The aim is to explain factors affecting attendance. During the season there

were 392 games played excluding playoff games. The total attendance number

was 1,964,626. Both the population of the home team and the visiting team

has a statistically significant effect on attendance, as well as distance between

home town and visitor’s town. Local games have a bigger attendance than

other games. The demand is not elastic with respect to the ticket price.

Success of both the home team and the visitor has an effect: home team’s

success with a positive and visitor’s with a negative coefficient. The number

of plays already played has a negative effect. Weekday effect is important: the

attendance is bigger during Saturdays. Also the day temperature has an

effect: the colder, the bigger attendance. That effect is small but still

statistically significant.

The unemployment rate has no effect, and the success factor of the last three

games does not seem to explain, neither does the success factor of all games

played.

Keywords: Ice hockey, Attendance, Finland, Temperature

70

2.1 Introduction

This paper uses regular season 2007 – 2008 Finnish ice hockey attendance

figures to examine the explaining determinants. Simple economic theory

suggests that the demand for attendance should depend on the ticket price of

the game and travel costs, the incomes of spectators, the prices of substitute

goods, and market size (Simmons 2006). Usually market size is measured by

local population. There is a wide range of literature on attendance of sports

events but not any with Finnish ice hockey data. A recent sport attendance

survey8 – both active consumption (participation in sport competitions or

being a member in a sport or gymnastic club) and passive consumption

(attendance) reveals that the most popular sports by attendance were ice

hockey (25.5%), football (16.9%), athletics (10.6%), skiing (6.5%) and Finnish-

rules baseball (5%). In this survey 44 % responded that they had not attended

any sports event between February 2005 and January 2006. A key

contribution of this paper is to show that both the market size (town

population) of the home and the visiting teams have an impact on attendance.

Previous success measured as points per game from the beginning of the

season is better to explain attendance than points per game from three last

games (the form guide), and temperature also matters although the games are

not played outdoors.

There are a few studies that have compared the attendance of sport activities

between men’s and women’s games or between genders. Most studies show

that there are more male spectators than female (see Vuolle, Telama & Laakso

8 Liikuntatutkimus 2005-2006, Sport Survey: Adult Population

71

1986, Gantz & Wenner 1991, Zhang, Pease, Hui & Michaud 1995, White &

Wilson 1999 or Thrane 2001). Women seem to favour women’s games and

men favour men’s games. (Kahle, Duncan, Dalakas & Aiken 2001). The

sociology of sport consumption has revealed that the motives for attending

women’s games and men’s games differ. Typically, the aesthetics of the game

or competition is more important for women’s team spectators and for female

spectators (Ridinger & Funk 2006) ,while e.g. tracking statistics is more

important for men (Fink, Trail & Anderson 2002). The relationship between

gender and active sport consumption, i.e. participation in sport competitions

or being a member in a sport or gymnastic club, reveals only minor

differences in Finland. Both genders are as active, but women seem to favour

more clubs of commercial purposes (e.g. gym with aerobics), while men are

more often members in sports associations that play games (Kansallinen

liikuntatutkimus 2005-2006). Gymnastics at home and within a gymnastic

association have been typically female, while fishing and hunting have been

male sport activities (Marin 1988).

The relationship between gender and passive sport consumption, i.e.

attendance at games, has been less studied. If there are gender differences

across games and if women spectators have different motives to attend sport

activities, the factors explaining attendance should be different. Since men

use more time in tracking statistics and reading about sports in daily

newspapers (Dietz-Uhler, Harrick, End & Jacquemotte 2000), a team’s

winning percentage or other previous performance measure of the team

should be less important to explain women’s teams’ attendance figure. There

are also differences between the importance of ticket pricing, friend influence

and family involvement in women’s and men’s (basketball in the USA) games

72

(Fink, Trail & Anderson 2002); hence the price elasticity of demand should

differ. Women’s games should be more ticket price sensitive.

There are 14 teams playing at the men’s highest level in the Finnish ice

hockey league. The regular season 2007 – 2008 was a four-fold series, i.e. 52

games per team, and teams located in Helsinki (HIFK and Jokerit) played

extra four mutual games; two at home stadium and two at visitor’s stadium. In

addition to that, the remaining 12 teams played extra four games in the

subdivisions of three teams. The subdivisions were 1) Blues (home city:

Espoo), Pelicans (Lahti), SaiPa (Lappeenranta), 2) HPK (Hämeenlinna), Ilves

(Tampere), Tappara (Tampere), 3) JYP (Jyväskylä), KalPa (Kuopio), Kärpät

(Oulu) and 4) Lukko (Rauma), TPS (Turku), Ässät (Pori)9. Altogether each team

played 28 home games and 28 games as visitor (Jääkiekkokirja 2008 - 2009,

55). The first regular season games were played in September 2007 and the

last in March 2008. After that some teams continued their games in play-offs

and the champion (Kärpät) was known in mid- April.

Jokerit from Helsinki got the biggest average attendance (8591 per game),

while the lowest figure was for HPK (3281 per game). Jokerit has the biggest

stadium (Hartwall Areena) in terms of capacity. The number of seats was

13506, while in Hämeenlinna (HPK) the number of seats was only 3214 but

with 1786 standing places – so altogether 5000 places. Table 1 summarizes

some statistics for the average attendance of each team during the regular

season 2007 – 2008.

9 The distance between the cities in these subdivisions are 1) Espoo – 114 km – Lahti – 152,5 km – Lappeenranta – 235,5 km – Espoo, 2) Hämeenlinna – 79,8 km – Tampere, 3) Jyväskylä – 148,9 km – Kuopio – 289,2 km – Oulu – 341,2 km – Jyväskylä, 4) Rauma – 87,4 km – Turku – 135,9 km – Pori – 50 km – Rauma.

73

(Table 1 about here)

Naturally bigger cities like Helsinki, Espoo, Tampere or Turku have a bigger

attendance potential, but this does not explain enough the variation in

attendance. It is also true based on the coefficient of variation that attendance

variation is much higher for Jokerit and KalPa than for Kärpät, JYP, HPK or

Pelicans.

Altogether, in regular season the number of games was 14 x 28 = 392 and the

total attendance was 1964626, i.e. 5012 per game. Other cultural events, like

theater, gathered a bigger admission in 2007: about 2.7 million but with a

bigger number of total presentations (about 13000) which equals 207 per

presentation. The Finnish national opera sold 162555 tickets to 198

presentations (about 820 per presentation). In the highest football league,

Veikkausliiga games, the admission number was 541612 with 182 games

(2976 per game). Hence, in terms of cultural events attendance in live

performances, ice hockey was the most important. However, going to the

movies was even more important since the total admission number was about

6.5 million in Finland (population 5.3 M) but there were not live performances

(Statistical Year Book 2008 Finland and www.veikkausliiga.fi). According to

statistical surveys made by national sports associations (SLU, Suomen

Liikunta ja Urheilu, published in Statistical Year Book 2008, Finland),

exercising ice hockey is not as usual as football, hence one might assume that

attendance in football games should be higher than ice hockey. However, the

aim of this study is not to compare different sports but to explain ice hockey

74

games attendance, especially the role of tracing statistics measured by team

performance is the core of this study. Newspapers and sport news on

television reveal this information and since most teams have typically three

games per week, this team performance information is revealed three times a

week during the regular ice hockey season.

The results indicate that both the population of the home team and the visitor

has a statistically significant effect on attendance as well as the distance

between home town and visitor’s town. Local games have a bigger attendance

than other games. The demand is not elastic with respect to the ticket price.

The success of both the home team and the visitor has an effect: home team’s

success with a positive and visitor’s with a negative coefficient. The number

of plays already played has a negative effect. Weekday effect is important: the

attendance is bigger during Saturdays. Also the day temperature has an

effect: the colder, the bigger attendance. That effect is small but still

statistically significant. The unemployment rate has no effect, and the success

factor of the last three games does not seem to explain or the success factor of

all games played.

2.2 Literature

There is a wide range of literature on attendance in cultural events. An

important and influential study explaining movie attendance by Eliashberg

and Shugan (1997) showed that attendance and the number of screens are

highly correlated. A bigger number of screens is associated with movie

attendance. Critics’ reviews published in media, like newspapers or

magazines, have been shown to have an impact on movie attendance. Also

75

spontaneous dispersal, like “word-of-mouth” is an important factor to explain

movie attendance (Elberse and Eliashberg 2003).

The literature explaining attendance in sport events, especially in the USA, is

also wide starting with Demmert (1973) and Noll (1974). Conventional

economic theory assumes that demand base measured as the incomes of the

relevant market population and market size (population) should have an

impact on attendance. Teams from bigger cities should have bigger

attendance if the venue capacity allows it. Many teams are local monopolies

with almost zero marginal costs of attendance. Hence maximizing profits

equals maximizing revenues, and the outcome should be to set ticket prices

high enough to ensure unitary price elasticity. Most studies still reveal that

sporting events are priced in the inelastic range (Krautmann & Berri 2007).

Coates and Harrison (2005) studied baseball (MLB) attendance with a panel

data throughout the years 1969 – 1996. The team’s home town population as

well as winning percentage is a positively significant variable to explain

attendance, while the ticket price is negatively associated. Different,

alternative ticket prices, like “gate” measured as the ratio of total box-office

income to total attendance or “seat” measured as weighted average of

different category seat prices have been used. Regardless of which price

measure is used, attendance is price inelastic. Incomes in home town do not

seem to be statistically significant. Also Coates and Humphreys (2007) have

similar results. Elsewhere Depken (2000) shows that attendance at MLB

baseball is positively significant with the incomes of the home town and

team’s payroll. Kahane and Shmanske (1997) report similar results and show

also that changes in teams’ structure (scorecard) has a negative effect on 76

attendance. Relatively big scorecard changes between seasons diminish

attendance. The distance between home town and visitor’s town has a

significant negative effect on attendance in MLB baseball (Knowles, Sherony

and Haupert 1992). The further away a visitor comes, the fewer spectators the

game attracts. During weekends there are more spectators than during

weekdays. With MLB baseball McDonald and Rascher (2000) show that sales

promotion has a positive but diminishing effect on attendance. Sales

promotion results in larger attendance, but excessive promotion are probably

too expensive in terms of profitability. The competitive balance of the league

is important for attendance (Schmidt and Berri 2001). If some teams “always”

win and some “always” lose the games, spectators’ motivation to attend falls.

Occasionally there have been strikes or lock-outs in baseball, which has not

had any significant and long-term effect on attendance (Schmidt and Berri

2002). During the season 1994 – 1995 even a six-month strike did not have

any long-term effects on attendance. Fans returned after the pause. This

result is valid in baseball but also in football (NFL) and ice hockey (NHL) as

shown by Schmidt and Berri (2004).

Attendance is rather price inelastic in other sports – not just baseball, like

football (NFL, Depken 2001), basketball (NBA, Coates and Humphreys 2007),

Australian rules football (Borland and Lye 1992), and English rugby football

(Carmichael, Millington and Simmons 1999), Spanish football (Garcia and

Rodriguez 2002). Teams seem to lower ticket prices to the inelastic range of

demand to increase non-ticket revenues, like intermediate time refreshment

sales or broadcasting revenues (Fort 2004 or Krautmann & Berri 2007).

77

On the other hand there is some evidence of both price elastic and price

inelastic demand in British football (Simmons 1996). Simmons argues that

there is a remarkable difference between season ticket holders and occasional

spectators that buy the ticket at the gate. Season ticket holders have lower

price sensitivity than occasional spectators. Negative income elasticity or

positive effect of unemployment rate on attendance has been found in several

studies: Baimbridge, Cameron and Dawson (1996) with British football;

Borland and Lye (1992) with Australian football; Falter and Perignon (2000)

with football in France but also a positive coefficient with home town incomes

in some sports, like basketball in the USA (NBA, Coates and Humpreys 2007)

or baseball in the USA (MLB, Depken 2000 or Coates and Harrison 2005) or

football in the USA (NFL, Depken 2001).

Some games are played outdoors and some indoors and therefore temperature

or weather conditions might have different effects on attendance. Especially

weather conditions have been shown to have an effect on sports attendance

outdoors. Baimbridge, Cameron and Dawson (1995) and Jones, Schofield and

Giles (2000) have shown that high temperature and rainless conditions have a

positive effect on rugby football attendance in the UK. Spanish football

attracts a larger audience when the weather is favorable (Garcia and

Rodriguez 2002) which is valid with Australian football (Borland and Lye

1992) or Finnish football (Iho and Heikkilä 2008). During spring the

attendance is larger in football in France (Falter and Perignon 2000). When it

is raining or snowing the attendance is lower in college football (National

College Athletic Association, NCAA, DeSchriver and Jensen 2002). To the

contrary Carmichael, Millington and Simmons (1999) show that low

78

temperature is associated with higher attendance in rugby football. Elsewhere

Baimbridge, Cameron and Dawson (1996) find no statistical association

between weather conditions and British football. Wilson and Sim (1995) show

that at the beginning of the season the attendance is higher than later in

football in Malaysia.

Winning probability or team’s success has a positive impact on attendance

(Boyd and Boyd 1998, Burdekin and Idson 1991, Coates and Harrison 2005,

Coates and Humphreys 2007, Depken 2000, Depken 2001, Kahane and

Shmanske 1997, McDonald and Rascher 2000, Simmons 1996) but also the

inverse relationship has been found (Baimbridge, Cameron and Dawson 1995).

In most cases there are more males than females in the audience. The focus

on the literature survey above is also more masculine than feminine. Most

surveys carried out reveal that one of the most important reasons to attend a

live game rather than watching through some media like television or radio is

to be involved in the success, the play-off games and winning the

championship. The ticket price is not so relevant (Hansen and Gauhier 1989).

Most of the games are played during weekends or evenings and therefore

entertainment is a very important motive to attend, especially for men since

true fans are male (Hall and O´Mahony 2006). For women the motive to

attend is to share time with friends and family (Fink, Trail and Anderson 2002

or Dietz-Uhler, Harrick, End and Jacquemotte 2000). Parking space, the

quality of seats and transportation possibilities in general to the stadium are

more important for female than for male. Appealing side services, like the

quality of parking space, the cleanliness of the stadium, adequate entrance

hall space and eating possibilities during intermediate times seem to increase 79

attendance (Wakefield and Sloan 1995). Men seem to fan only one team, their

favorite, to which they are rather loyal. Peer group acceptance is important

and men seem to have an emotional attitude towards the team (Bauer, Sauer

and Exler 2005). Supporters use fan shirts and scarves. Admiration is

associated with strong and enjoyable feelings (Heinonen 2005).

Unreasonableness and superfluity are essential. The brand equity of a team

has a significant positive impact on attendance in German football (Bauer,

Sauer and Schmitt 2004). Mustonen, Arms and Russell (1996) show that the

possibilities to see proficient ice hockey and support the team are among the

most important reasons to attend, while getting together and especially game

violence are far less important motives in Finland. The two latter are more

important in Canada.

The survey on the impact of temperature, winning percentage and venue

quality indicate that temperature matters, after all ice hockey is played

indoors, and the relationship might not be equal to what has been found with

outdoor sports events. To the contrary: when the temperature is high

especially occasional spectators have other, substitute alternatives (outdoor

activities) and that might diminish attendance. Winning percentage and brand

equity are associated and they should have a positive impact on attendance.

Since most spectators are male, venue quality should have less importance

and it is not considered as an explanatory factor.

80

2.3 A model explaining attendance

Based on the literature survey, the following model explaining attendance can

be formulated where the utility function uijkt of individual i to attend a home

team’s j game against visitor k on time t

uijkt = α1(yit – Pjkt) + α2GSFjk + α3SSFjt + α4TSFjt + εijkt

In the model yit measures spectator’s income and Pjkt is the ticket price (α1 >

0). GSFjk (game specific factors) describes interest towards the game that can

be measured with home team’s j winning ratio as well as visitor’s k winning

ratio in previous games. Based on Bauer, Sauer and Schmitt (2004) and

Coates and Harrison (2005), it is plausible to assume that interest towards the

game is higher when the home team has won the previous games. Both points

per game from the beginning of the season and points from the last three

games are suitable empirical measures for the winning ratio. The regular

season games yield points according to the following scheme: a win within

normal playtime (60 min) gives 3 points, a win within extension time (60 min

+) or a penalty shot win gives 2 points, a lost within extension time or after

penalty shots gives 1 point, and a lost within normal playtime gives 0. Interest

towards a game is larger when home town population or visitor’s town

population is higher (Coates and Harrison 2005), while a bigger distance

between home town and visitor’s town should lower interest towards the

game (Knowles, Sherony and Haupert 1992). It is plausible that local games,

like HIFK – Jokerit (both from Helsinki) or Ilves – Tappara (both from

Tampere) have (almost) full house. High unemployment rate in the region on

the one hand might reduce attendance due to lower average incomes but on

the other hand especially in France attendance in football games and

unemployment rate are positively correlated.

81

SSFt is measuring time from the beginning of the season. In early autumn

when the season begins games have high interest since the team has new

players and the lines are new (Wilson and Sim 1995). As time goes on, this

interest might diminish and hence attendance also goes down. The number of

games played since the beginning of the season is the empirical measure in

this study.

Attendance is rather inelastic with respect to price pjkt . Stadia or halls have

different price categories. During the season 2007 – 2008 e.g. the ticket price

of Blues’s (Espoo) home games on club seats (201-206) was normally €27, on

the second long side (207-211) €24, standing places (terraces) (212) €10,

gable seats lower (101-102) €18, normal seat upper (401-406) €14, disabled

persons €14, conscripts and students €10 (normal seats upper, not Blues –

HIFK, nor Blues – Jokerit), boxes for box owners €14, and children under 7

years free if they were sitting on parent’s knees. Since Espoo and Helsinki are

neighboring towns, the ticket prices for games against HIFK or Jokerit (both

from Helsinki) were €2 higher. These prices were valid only when the ticket

was bought in advance. When bought on entrance, there was €1 increase. For

empirical purposes the variation is very challenging and since there was no

data concerning the true distribution of seats taken, the empirical equivalent

of the price is usually the ticket price of the best seat including local game

excess fees. For Blues, this price is €27 or €29 with HIFK or Jokerit as visiting

team. However, throughout the regular season, ticket prices do not vary: at

the beginning of the season and at the end of season prices remain

unchanged, and there is no weekend premium, hence pjkt = pjk.

82

The proxy for the time specific factor (TSF) is partially the weather conditions

and partially weekday. A good, sunny weather brings about a larger

attendance than rainy weather in Spanish football (Garcia and Rodriguez

2002). However, ice hockey is played indoors and weather - here: the

temperature outside – might have an opposite effect. The maximum day

temperature in the nearest meteorological observation site is used to measure

the temperature. For other teams than Blues, HIFK, HPK and Jokerit, the

observation site is usually the airport of the home town. The airport in Oulu

(team: Kärpät) is located in the neighboring town, Oulunsalo and the

temperature for Blues (Espoo), HIFK and Jokerit (Helsinki) is measured at

Helsinki-Vantaa Airport which is located in the neighboring town, Vantaa. The

temperature for the team of Hämeenlinna, HPK, is measured in Jokioinen

which is about 50 km away from Hämeenlinna.

The weekday effect takes into account the fact that during weekends there is

usually a larger attendance.

Individuals have the other option not to attend an ice hockey game and the

utility for the alternative is

ui0t = α1(yit ) + α4TSFt + εi0t

Letting Uijkt = uijkt - ui0t. Individuals will choose to attend a game if Uijkt > 0.

Therefore the model to be estimated explain ice hockey games attendance

(ATT) is:

(1)ATT = β1(yit – Pjkt) + β 2GSFjk + β 3SSFjt + β 4TSFjt + Φijkt

83

A complete listing of variables is given in table 2. Variables, except for the

temperature and the weekday, are in logarithmic form and thus the parameter

coefficients in estimation results are elasticities.

(Table 2 about here)

Descriptive statistics and correlation on variables (before taking logarithms) are shown in table 3.

(table 3 about here)

Roughly 31 % of the games were played on Saturdays, about 27 % on

Thursdays and about 25 % on Tuesdays. In addition to that, a few games were

played on Mondays (< 3 %), Wednesdays (< 5 %), Fridays (> 6 %) and

Sundays (> 3 %). Somewhat more often there were Monday games in bigger

towns (r = 0,109) against teams from far away (r = 0,106). There seems to

have been more games on Fridays in bigger towns (r = 0,121) and that seems

to have been reducing Saturday games (r = -0,128). Otherwise the weekday

variables do not seem to correlate with other variables.

The correlation matrix reveals the ticket price seems to have been higher in

larger towns and it seems to have a positive relation with attendance. The

number of home team games and the number of visitor’s games were

(naturally) highly positively correlated. Points per game from the beginning of

the season (HPoint) and points from the last three games (Last3H) were also

positively correlated. The regional unemployment rate seems to have been

higher in areas with smaller towns. The temperature seems to have been

lower when the number of games has increased. Probably the relation is like

84

inverse U or inverse J. According to long-term statistics (1900 – 2000) the

temperature in Helsinki (Kaisaniemi observation site) has been + 11,1 Celsius

in September, + 6,2 Celsius in October, + 1,5 Celsius in November, - 2,1

Celsius in December, -4,7 Celsius in January, -5,7 Celsius in February and -2,2

in March (Ilmatieteen laitos 2009).

Since points per game (HPoint or VPoint) and the corresponding points from

the last three games (Last3H or Last3V) are strongly positively correlated and

these partially measure the same for empirical purposes, these are used as

alternative measures.

85

2.4 Estimation

Conventional regression analysis is used here but the results might be biased

due to heterogeneity. An alternative for conventional regression analysis is

panel data methods. The benefits of using panel data are that (1) individual

heterogeneity can be controlled, (2) estimated parameters are more efficient

and (3) with panel data the dynamics of adjustment can be studied better

(Baltagi 2008, 6-7). Time-series and cross-section studies that do not control

heterogeneity might yield biased results. However, since in this study the

population variable is constant for each team, i.e. for Blues (Espoo) the

population is regardless of the game always 238047 (population 31st

December.2007), the panel data estimation method (NLogit 4.0) will not

estimate the coefficient, and conventional regression analysis is used.

The first model uses the price of the ticket (LogPrice), the population in home

and visitor’s town (LogHPop and LogVPop), the distance from home team’s

stadium to visitor’s stadium (LogDist), round (LogHGame), teams’ success or

winning ratio (LogHPoin and LogVPoin) and the region´s unemployment rate

(LogUnemp), maximum day temperature in home town (Temp, note: not

logarithm) and dummies for Tuesday (TU) and Thursday (TH) or Saturday

(SA). All parameter estimates except the unemployment rate are statistically

significant, have the right sign and are plausible. Model 2 is otherwise similar

to Model 1 except that the weekday dummy is Saturday. In Models 3 and 4 the

success variable is points from three last games (LogHLast or LogVLast). In

these models visitor’s last three games points (LogVLast) do not seem to

significantly explain the attendance of the game, while home team’s last three

games points seem to explain. The Saturday effect is substantial: the audience

is about 10 – 11 percent larger than on Tuesdays or Thursdays. Other

86

weekday dummies (Monday, Wednesday, Friday or Sunday) are not significant

(not reported here).

Models 5 - 8 do not have the unemployment rate variable but the results seem

to be similar. Attendance is fairly inelastic with respect to ticket price (price

elasticity is about -0,25 … - 0,32 depending on the model), population

variables get a positive coefficient so that home town population elasticity is

about 8 times as high as the visitor’s town population elasticity. Distance

between the home team and the visitor’s town seem to be significant: the

longer the distance, the less attendance. However, the effect has only minor

importance since the distance elasticity is absolutely rather small. Still as the

distance increases from 50 km to 100 km, attendance diminishes by 2½

percent. Game round has a diminishing effect on attendance which was also

found by Wilson and Sim (1995) with Malaysian football.

Home team’s success or winning ratio (LogHPoin or LogHLast) seems to

attract more spectators since these variables get a positive coefficient

regardless of the model. The visitor’s success on the contrary seems to lessen

attendance (LogVPoin), but the recent success in terms of last three games

(LogVLast) does not seem to be significant. Ice hockey audience seems to

favour an assured win and not even games. If the home team falls into a losing

circle, the attendance also falls. The temperature has interesting effects on

attendance since low temperature seems to attract a bigger audience. Ice

hockey is a game played indoors, thus the effect of temperature is different

than in football or other outdoors sports.

87

(tables 4a,4b and 4c about here)

88

Models 5 and 6 are the most credible to explain ice hockey attendance, still

the coefficient of determination (R2) is only about 0,67. In addition model 6

seems to overestimate attendance for some teams - Blues, HIFK, HPK, JYP,

KalPa, Pelicans and SaiPa – and underestimate the rest – Ilves, Jokerit, Kärpät,

Lukko, Tappara, TPS and Ässät.

Figure 1 contains the true attendance of HIFK (Series1) and the estimate

made through Model 6 (Series2). Based on figure 1 the model is not able to

explain peaks and bottoms, i.e. large variation. The actual average attendance

(and standard deviation) for HIFK in regular season 2007 – 2008 was 6573

(1023), while the model 6 estimates are correspondingly 7161 (585). It is

possible to make some simulation exercises using model 6. The first

simulation shown in Figure 2 as SIM1, the game days for HIFK were

postponed by one, e.g. actual Tuesday’s 11th September 2007 game HIFK vs.

HPK to the following Friday 21st September 2007 when actually there was a

game HIFK vs. Blues and so on, except that the last regular season game of 4th

March 2008 (HIFK vs. TPS) was moved to 11th September 2007. Since most of

the games are played on Tuesdays, Thursdays and Saturdays, the first

simulation is trying to answer the question of what might be the result of

having the “Saturday games” on Tuesdays, and having the “Tuesday games”

on Thursdays and the “Thursday games” on Saturdays. This simulation is not

accurate since there have been some games on Mondays and so on. However,

simulation 1 reveals that the average attendance (and standard variation)

would be 7165 (822). Correspondingly, the other simulation 2 is made so that

the “Saturday games” have been moved to Thursdays, and the “Tuesday

89

games” to Saturdays and the “Thursday games” to Tuesdays. In this case, the

attendance to HIFK’s home games would be on average (std): 7169 (820).

With these simulations, the home team characteristics – population

(LogHPop), success (LogHPoin), round (LogHGame), the temperature (temp) -

are assumed not to change, while the visitor’s characteristics – ticket price

(LogPrice), population (LogVPop), distance (LogDist), success (LogVPoin) are

assumed to change.

2.5 Conclusions and suggestions

There is a rather wide variation in ice hockey game attendance between

weekdays. On Saturdays the attendance is about 10 – 11 percent higher than

during other conventional playdays, i.e. Tuesdays and Thursdays. In addition

to that there is also big variation across teams: Jokerit from Helsinki got the

biggest attendance, while HPK from Hämeenlinna got the lowest average

attendance. It is natural that home town population partially explains this,

since the elasticity of attendance with respect to home town population is

positive (about 0,336) but also the elasticity with respect to visitor’s town

population is positive (about 0,048). The visitor’s fans will attend the team’s

away games but distance matters. The bigger distance, the lower attendance.

With caution it can be argued that ticket price has a negative effect on

attendance, since demand seems to be inelastic. However, the price variable

90

is not the actual average price since this data was not available. The price

variable used in the estimations is the ticket price to the best seats. As the

season goes on and more games have been played, the attendance seems to

diminish but the estimated coefficient is low even though significant. Team’s

success seems to attract a bigger attendance, while visitor’s success has the

opposite effect. Spectators are willing to see a live game in the stadium if they

expect that home team will win the game. The unemployment rate has no

effect on attendance, while weather condition measured by the outside

temperature is a significant variable. Colder weather attracts more spectators.

However, the estimated coefficient is minor but significant.

The effect of mass media, e.g. television has been neglected in this study.

Some of the games were seen through cable television (Pay-tv) and some

through open commercial channels. In Finland, all television owners must pay

a TV fee which was during these years about € 200 – 230 per year. There is

some evidence that broadcasting through tv has a negative effect on live

attendance (Baimbridge, Cameron and Dawson 1995 or Carmichael,

Millington and Simmons 1999). Moreover, other activities like the premieres

of blockbuster movies or concerts by famous orchestras or rock bands might

lower attendance. However, these have not been taken into account.

The estimation results reveal that the models can explain about 2/3’s of actual

attendance based on the coefficient of determination. The models do not

explain whether fans are loyal to their teams. Will they abandon the team if

success is not good enough? What is loyalty, what is the effect of that on

attendance? Are the results robust with international data? These topics are 91

among those that should be studied. Using panel data analysis and methods is

also worth considering.

92

References

Baimbridge, Mark, Samuel Cameron & Peter Dawson (1995) Satellite

broadcasting and match attendance: the case of rugby league. Applied

Economics Letters, vol 2, 343-346

Baimbridge, Mark, Samuel Cameron & Peter Dawson (1996) Satellite

television and the demand for football: a whole new ball game? Scottish

Journal of Political Economy, vol 43, No 2, August, 317-333

Baltagi, Badi H. (2008): Econometric Analysis of Panel Data. 4th Edition, John Wiley & Sons Ltd

Bauer, Hans H., Nicola E. Sauer & Philipp Schmitt (2004) Customer-based

brand equity in the team sport industry. Operationalization and impact on the

economic success of sport teams. European Journal of Marketing, vol. 39, No.

5/6, 496-513

Bauer, Hans H., Nicola E. Sauer & Stefanie Exler (2005) The loyalty of

German soccer fans: does a team’s brand image matter? International Journal

of Sports Marketing & Sponsorship, vol 7, issue 1, 14-22

Borland, Jeff & Jenny Lye (1992) Attendance at Australian Rules football: a

panel study. Applied Economics, vol. 24, 1053-1058

93

Boyd, David W. & Laura A. Boyd (1998) The Home Field Advantage:

Implications for the Pricing of Tickets to Professional Team Sporting Events.

Journal of Economics and Finance, vol. 22, Numbers 2-3, Summer/Fall, 169-

179

Burdekin, Richard C.K. & Todd L. Idson (1991) Customer preferences,

attendance and the racial structure of professional basketball teams. Applied

Economics, vol. 23, 179-186

Carmichael, Fiona, Janet Millington & Roberts Simmons (1999) Elasticity of

demand for Rugby League attendance and the impact of BskyB. Applied

Economics Letters, vol. 6, 797-800

Coates, Dennis & Thane Harrison (2005) Baseball strikes and the Demand of

Attendance. Journal of Sports Economics, vol. 6, 282-302

Coates, Dennis & Brad R. Humphreys (2007) Ticket Prices, Concessions and

Attendance at Professional Sporting Events. International Journal of Sport

Finance, vol. 2, 161-170

Demmert, Henry G. (1973): The Economics of Professional Team Sports.

Lexington Books

Depken, Craig A. II (2000) Fan Loyalty and Stadium Funding in Professional

Basketball. Journal of Sports Economics, vol. 1, 124-138

94

Depken, Craig A. II (2001) Research Notes: Fan Loyalty in Professional

Sports: An Extention of the National Football League. Journal of Sports

Economics, vol. 2, 275-284

DeSchriver, Timothy D. & Paul E. Jensen (2002) Determinants of Spectator

Attendance at NCAA Division II Football Contests. Journal of Sport

Management, vol. 16, 311-330

Dietz-Uhler, Beth, Elizabeth A. Harrick, Christian End & Lindy Jacquemotte

(2000) Sex Differences in Sport Fan Behavior and Reasons for Being a Sport

Fan. Journal of Sport Behavior, vol. 23, 219-231

Elberse, Anita & Jehoshua Eliashberg (2003) Demand and Supply Dynamics

for Sequentially Related Products in International Markets: The Case of

Motion Pictures. Marketing Science, vol. 25, 6, 638-661

Eliashberg, Jehoshua & Steven M. Shugan (1997) Film Critics: Influencers or

Predictors. Journal of Marketing, vol. 61, 68-78

Falter, Jean-Marc & Christophe Pérignon (2000) Demand for football and

intramatch winning probability: an essay on the glorious uncertainty of sports.

Applied Economics, vol. 32, 1757-1765

Fink, Janet S., Galen T. Trail & Dean F. Anderson (2002) Environmental

Factors Associated With Spectator Attendance and Sport Consumption

95

Behavior: Gender and Team Differences. Sport Marketing Quarterly, vol. 11,

No 1, 8-19

Fort, Rodney (2004): Inelastic Sports Pricing. Managerial and Decision

Economics, 25, 87-94

Gantz, W & L.A. Wenner (1991): Men, women and sports: Audience

experiences and effects. Journal of Broadcasting and Electronic Media, 35,

233-243

García, Jaume & Plácido Rodríguez (2002) The Determinants of Football

Match Attendance Revisited: Empirical Evidence From the Spanish Football

League. Journal of Sports Economics, vol. 3, 18-38

Hall, John & Barry O´Mahony (2006) An empirical analysis of gender

differences in sports attendance motives. International Journal of Sports

Marketing & Sponsorship, vol 7, issue 4, 334-346

Hansen, Hal & Roger Gauthier (1989) Factors Affecting Attendance at

Professional Sport Events. Journal of Sport Management, vol. 3, 15-32

Heinonen, Harri (2005) Jalkapallon lumo – tutkimus suomalaisesta Everton-

faniudesta. Atena kustannus Oy.

96

Iho, Antti & Jaakko Heikkilä (2008) Has selling tickets in advance increased

attendance in the Finnish football league? University of Helsinki. Department

of Economics and Management. Discussion Papers n:o 21.

Kahane, Leo & Stephen Shmanske (1997) Team roster turnover and

attendance in major league football. Applied Economics, vol. 29, 425-431

Kahle, Lynn, Marc Duncan, Vassilis Dalakas & Damon Aiken (2001): The social

values of fans for men’s versus women’s university basketball. Sport

Marketing Quarterly, 10, 156-162

Kansallinen liikuntatutkimus 2005-2006. SLU:n julkaisusarja 5/06. 2006 (in Finnish)

Knowles, Glenn, Keith Sherony 6 Mike Haupert (1992) The demand for major

league baseball: a test of the uncertainty of the outcome hypothesis. The

American Economist, vol.36, no 2, 72-80

Krautmann, Anthony C. & David J. Berri (2007): Can We Find It at the

Concessions? Understanding Price Elasticity in Professional Sports. Journal of

Sports Economics, 8, 183-191

Liikuntatutkimus 2005-2006 : aikuisväestö [elektroninen aineisto]. FSD2233,

versio 1.0 (2007-01-29).

Helsingin kaupunki, Nuori Suomi, Opetusministeriö, Suomen

Kuntoliikuntaliitto, Suomen Liikunta ja Urheilu, Suomen Olympiakomitea & 97

Pehkonen, Juhani (Suomen Gallup) [tekijät]. Tampere : Yhteiskuntatieteellinen

tietoarkisto [jakaja], 2007. In Finnish

Marin, Marjatta (1988): Gender-Differences in sport movement in Finland.

International Review for Sociology of Sport, 23, 345-359

McDonald, Mark & Daniel Rascher (2000) Does Bat Day Make Cents? The

Effect of Promotions on the Demand for Major League Baseball. Journal of

Sport Management, vol. 14, 8-27

Mustonen, Anu, Robert L. Arms & Gordon W. Russell (1996) Predictors of

sports specators’ proclivity for riotous behaviour in Finland and Canada.

Personality and Individual Differences, vol. 21, Issue 4, 519-525

Noll, Roger G. (Editor, 1974): Government and the Sports Business. The

Brookings Institution

Ridinger, Lynn L. & Daniel C. Funk (2006): Looking at gender differences

through the lens of sport spectators. Sport Marketing Quarterly, 15, 155-166

Schmidt, Martin B. & David J. Berri (2001) Competitive Balance and

Attendance: The Case of Major League Baseball. Journal of Sports Economics,

vol. 2, 145-167

98

Schmidt, Martin B. & David J. Berri (2002) The impact of the 1981 and 1994-

1995 strikes on Major League Baseball attendance: a time-series analysis.

Applied Economics, vol. 34, 471-478

Schmidt, Martin B. & David J. Berri (2004) The Impact of Labor Strikes on

Consumer Demand: An Application to Professional Sports. The American

Economic Review, vol. 94, 1, 344-357

Simmons, Robert (1996) The demand for English league football: a club-level

analysis. Applied Economics, vol. 28, 139-155

Simmons, Rob (2006): The demand for spectator sports. In Handbook on the

Economics of Sport, pp. 77-89, edited by Wladimir Andreff and Stefan

Szymanski. Edward Elgar, Celtenham, UK

Thrane, Christer (2001): Sport spectatorship in Scandinavia: A class

phenomenon? International Review for the Sociology of Sport, 36, 149-163

Wakefield, Kirk L. & Hugh J. Sloan (1995) The Effects of Team Loyalty and

Selected Stadium Factors on Spectator Attendance. Journal of Sport

Management, vol. 9, 153-172

White, Philip & Brian Wilson (1999): Distinctions in the stands: An

investigation of Bourdieu’s ‘Habitus’, socioeconomic status and sport

99

spectatorship in Canada. International Review for the Sociology of Sport, 34,

245-264

Wilson, Peter & Benson Sim (1995) The demand for Semi-Pro League football

ni Malaysia 1989-91: a panel data approach. Applied Economics, vol. 27, 131-

138

Vuolle, P, R. Telama & L. Laakso (1986): Näin suomalaiset liikkuvat.

Liikunnan ja kansanterveyden julkaisuja, 50. Helsinki

Zhang, J., D. Pease, S. Hui & T. Michaud (1995) : Variables affecting the

spectator decision to attend NBA games. Sport Marketing Quarterly, 4, 29-39

Statistics:

Jääkiekkokirja 2007-2008. Egmont Kustannus 2008

Official Statistics, Finland 2008

http://www.tutiempo.net/

http://www.tem.fi

Estimation: Nlogit 4.0 (www.limdep.com)

100

Table 15: (Table 2.1) Regular season 2007 – 2008 average attendance and capacity statictics, source: Jääkiekkokirja 2007-2008 ja Jääkiekkokirja 2008-2009

team Home game average

attendance (in relation to

capacity)

Variation of attendance in home games:

min – max (std)

Coefficient of variation

Stadium capacity, seats

(sitting, standing, others)

Blues 4837 (70%) 3706 – 6530 (693,5)

V = 0,143

6914 (5633/230/1051 boxes)

HIFK 6573 (80%) 5219 – 8200 (1023,4)

V = 0,155

8200

HPK 3281 (65%) 2780 – 4395 (384,8)

V = 0,117

5000 (3214/1786)

Ilves 5914 (76%) 4584 – 7800 (1026,0)

V = 0,173

7800 (6635/1165)

Jokerit 8591 (64%) 6203 – 13464 (1890,4)

V = 0,220

13506

JYP 4054 (90%) 3347 – 4500 (343,8)

V = 0,084

4500 (2352/2148)

KalPa 3388 (65%) 2512 – 4911 (722,3)

V = 0,213

5225 (2767/2458)

Kärpät 6054 (92%) 5062 – 6614 (485,2)

V = 0,080

6614 (4760/1854)

Lukko 3733 (69%) 2901 – 5400 (658,8)

V = 0,174

5400 (3386/2014)

Pelicans 4252 (87%) 3505 – 4910 (485,2)

V = 0,114

4910 (3410/1500)

SaiPa 3557 (73%) 2881 – 4847 (533,4)

V = 0,149

4847 (2810/2025/12 wheelchair)

Tappara 5712 (73%) 4193 – 7800 (1074,1)

V = 0,187

7800 (6635/1165)

TPS 5978 (51%) 3919 – 8394 (1155,0)

V = 0,193

11820 (9042/2778)

Ässät 4234 (65%) 3287 – 6472 (828,8)

V = 0,195

6472 (3972/2500)

101

Table 16: (Table 2.2) Variables, measurement, source and expected sign

variable measure source expected sign

Game specific factor home town population 31.12.2007 (logHPop) Statistics Finland +

visitor’s town population 31.12.2007

(logVPop)

Statistics Finland +

distance between home town and visitor’s

town (logDist)

Stadium address

http://www.sm-liiga.fi

distance:

http://kartat.eniro.fi

-

home team’s points per game (logHPoin): if

zero, then replaced by 0,01

own calclulations based on

Jääkiekkokirja 2007-2008

+

visitor’s points per game (logVPoin): if zero,

then replaced by 0,01

own calclulations based on

Jääkiekkokirja 2007-2008

?

home team’s points from 3 last games

(logHLast): if zero, then replaced by 0,01

own calclulations based on

Jääkiekkokirja 2007-2008

+

visitor’s points from 3 last games (logVLast):

if zero, then replaced by 0,01

own calclulations based on

Jääkiekkokirja 2007-2008

?

Incomes regional unemployment rate (Unempl) http://www.tem.fi ?

Season specific factor played games since the beginning of th

season (logHGame) if zero, then replaced by

0,01

own calclulations based on

Jääkiekkokirja 2007-2008

-

ticket price, pjk ticket price (logPrice) Jääkiekkokirja 2007 - 2008 -

Time specific factor,

temperature, tempit

temperature at nearist observation site

(temp)

http://www.tutiempo.net/ ?

Time specific factor,

weekday

weekday, three dummies TU (tuessay) TH

thursday), SA (saturday)

TU –

TH –

SA +

102

Table 17: (Table2.3) Variables, means, standard deviations and correlation matrix. ATT = attendance, Price (€), Dist = distance between home team’s and visitor’s stadiums along road (km), Temp = max tempature Unempl = monthly regional unemployment rate(%), HomePop = home town population, VisiPop = vistor’s town population, HPoint = points per game, home team, before the game VPoint = visitor’s points per game, before the game, HomeG = number of

games, home team, before the game, VisiG = number of games, visitor, before the game , Last3H = points from 3 last games, home team, Last3V = points from 3 last games, visitor. The number of observations = 392.

variable mean std ATT Price Dist Temp Unempl HomePop VisiPop HPoint VPoint HomeG VisiG Last3H Last3V

ATT 5014,4 1712,8 1 0,619 -0,074 -0,013 -0,601 0,730 0,138 0,313 0,022 0,051 0,066 0,292 0,034

Price 25,4 4,3 1 -0,212 0,045 -0,681 0,833 0,109 0,125 -0,001 0,022 0,036 0,200 0,063

Dist 245,8 154,2 1 -0,081 0,204 -0,155 -0,154 0,065 0,071 0,027 0,023 0,012 0,054

Temp 4,5 5,6 1 -0,175 0,047 -0,001 -0,113 -0,107 -0,714 -0,714 -0,089 -0,117

Unempl 8,3 2,0 1 -0,788 0,015 -0,164 0,040 0,117 0,103 -0,190 0,036

HomePop 185320 168248 1 -0,012 0,192 -0,010 0,009 0,024 0,227 0,012

VisiPop 185320 168248 1 0,009 0,172 0,003 -0,009 0,026 0,224

HPoint 1,45 0,51 1 0,224 0,120 0,128 0,644 0,086

VPoint 1,48 0,53 1 0,119 0,113 0,151 0,643

HomeG 27,5 16,2 1 0,998 0,106 0,131

VisiG 27,5 16,2 1 0,109 0,125

Last3H 4,24 2,51 1 0,099

Last3V 4,42 2,52 1

103

Table 18: (Table 2.4a) Some estimation results

Model 1 Model 2 Model 3

variable coefficient std t-value P [|T|>t] coefficient std t-value P [|T|>t] coefficient std t-value P [|T|>t]

constant 4,785 0,403 11,87 0,000 4,596 0,406 11,31 0,000 4,876 0,418 11,68 0,000

LogPrice -0,272 0,110 -2,28 0,013 -0,265 0,111 -2,39 0,017 -0,324 0,112 -2,91 0,004

LogHPop 0,355 0,028 12,84 0,000 0,360 0,028 12,89 0,000 0,365 0,028 12,89 0,000

LogVPop 0,047 0,012 3,93 0,000 0,048 0,013 3,96 0,000 0,044 0,012 3,50 0,000

LogDist -0,037 0,009 -4,17 0,000 -0,038 0,009 -4,22 0,000 -0,036 0,009 -3,98 0,000

LogHGame -0,032 0,013 -2,44 0,015 -0,032 0,013 -2,41 0,016 -0,026 0,011 -2,44 0,015

LogHPoin 0,083 0,017 4,79 0,000 0,0846 0,017 4,87 0,000

LogVPoin -0,058 0,016 -3,57 0,000 -0,058 0,016 -3,52 0,000

LogHLast 0,014 0,006 2,47 0,014

LogVLast -0,003 0,005 -0,62 0,537

LogUnempl 0,083 0,060 1,39 0,166 0,087 0,061 1,44 0,151 0,071 0,062 1,15 0,253

Temp -0,005 0,002 -2,00 0,046 -0,005 0,002 -2,02 0,044 -0,004 0,002 -1,83 0,068

TU -0,111 0,023 -4,88 0,000 -0,106 0,023 -4,58 0,000

TH -0,122 0,022 -5,50 0,000 -0,114 0,023 -5,04 0,000

SA 0,116 0,020 5,75 0,000

N = 392 R2 = 0,677 F = 75,34 χ2 = 453,62 DW = 2,00 R2 = 0,672 F = 80,97 χ2 = 446,68 DW = 2,03 R2 = 0,661 F = 70,53 χ2 = 436,08 DW = 1,96

104

Table 19: (Table 2.4b) Some estimation results

Model 4 Model 5 Model 6

variable coefficient std t-value P [|T|>t] coefficient std t-value P [|T|>t] coefficient std t-value P [|T|>t]

constant 4,706 0,420 11,20 0,000 5,210 0,263 19,78 0,000 5,039 0,265 19,02 0,000

LogPrice -0,318 0,112 -2,83 0,005 -0,265 0,110 -2,42 0,016 -0,258 0,111 -2,33 0,020

LogHPop 0,369 0,029 12,92 0,000 0,332 0,022 15,01 0,000 0,336 0,022 15,04 0,000

LogVPop 0,044 0,013 3,54 0,000 0,047 0,012 3,93 0,000 0,048 0,012 3,96 0,000

LogDist -0,037 0,009 -4,02 0,000 -0,036 0,009 -4,06 0,000 -0,037 0,009 -4,11 0,000

LogHGame -0,026 0,011 -2,39 0,017 -0,033 0,013 -2,51 0,012 -0,033 0,013 -2,49 0,013

LogHPoin 0,083 0,017 4,80 0,000 0,085 0,017 4,87 0,000

LogVPoin -0,058 0,016 -3,53 0,000 -0,058 0,016 -3,49 0,001

LogHLast 0,015 0,006 2,70 0,007

LogVLast -0,003 0,005 -0,60 0,549

LogUnempl 0,073 0,062 1,17 0,243

Temp -0,004 0,002 -1,85 0,065 -0,005 0,002 -2,31 0,021 -0,005 0,002 -2,33 0,020

TU -0,110 0,023 -4,85 0,000

TH -0,122 0,022 -5,50 0,000

SA 0,109 0,021 5,29 0,000 0,115 0,020 5,71 0,000

N = 392 R2 = 0, 657 F = 75,96 χ2 = 429,83 DW = 1,98 R2 = 0, 676 F = 82,48 χ2 = 451,64 DW = 2,00 R2 = 0,671 F = 89,49 χ2 = 444,56 DW = 2,03

105

106

Table 20: (Table 2.4c) Some estimation results

Model 7 Model 8

variable coefficient std t-value P [|T|>t] coefficient std t-value P [|T|>t]

constant 5,240 0,271 19,28 0,000 5,080 0,273 18,62 0,000

LogPrice -0,318 0,111 -2,85 0,004 -0,311 0,112 -2,78 0,006

LogHPop 0,345 0,022 15,44 0,000 0,348 0,023 15,46 0,000

LogVPop 0,044 0,012 3,50 0,000 0,044 0,013 3,54 0,001

LogDist -0,036 0,009 -3,90 0,000 -0,036 0,009 -3,93 0,000

LogHGame -0,027 0,011 -2,54 0,009 -0,027 0,011 -2,50 0,013

LogHPoin

LogVPoin

LogHLast 0,014 0,005 2,60 0,010 0,016 0,006 2,83 0,005

LogVLast -0,003 0,005 -0,57 0,570 -0,003 0,005 -0,550 0,583

LogUnempl

Temp -0,005 0,002 -2,12 0,035 -0,005 0,002 -2,14 0,033

TU -0,106 0,023 -4,55 0,000

TH -0,114 0,023 -5,04 0,000

SA 0,108 0,021 5,27 0,000

N = 392 R2 = 0,661 F = 77,39 χ2 = 434,73 DW = 1,96 R2 = 0,657 F = 84,17 χ2 = 428,43 DW = 1,98

107

Figure 2: (Figure 2.1) Actual attendance for HIFK (Series1) and Model 6 (Series2)

11/9/2007

19/9/2007

27/9/2007

5/10/2007

13/10/2007

21/10/2007

29/10/2007

6/11/2007

14/11/2007

22/11/2007

30/11/2007

8/12/2007

16/12/2007

24/12/2007

1/1/2008

9/1/2008

17/1/2008

25/1/2008

2/2/2008

10/2/2008

18/2/2008

26/2/20080

1000

2000

3000

4000

5000

6000

7000

8000

9000

Series1Series2

Figure 3: (Figure 2.2) Actual attendance for HIFK (ADM), Model 6 and simulations 1 and 2

11/9/2007

19/9/2007

27/9/2007

5/10/2007

13/10/2007

21/10/2007

29/10/2007

6/11/2007

14/11/2007

22/11/2007

30/11/2007

8/12/2007

16/12/2007

24/12/2007

1/1/2008

9/1/2008

17/1/2008

25/1/2008

2/2/2008

10/2/2008

18/2/2008

26/2/20080

1000

2000

3000

4000

5000

6000

7000

8000

9000

10000

ADMModel6SIM1SIM2

108

3 Fan loyalty in Finnish Ice Hockey

3.1 Introduction

Sport has become more professional over the years. Sport managers view

their teams, leagues as brands to be managed. A product or service is

considered as a brand if the name, logo, sign or slogan increases the value of

that product or service. The psychological aspect in the consumer’s mind, the

brand image consists of all information and associations with a product or

service. The quality of brand associations is determined by the favourability,

uniqueness and strength. High levels of brand awareness and a positive

brand image should increase among others greater consumer loyalty (Keller

1993).

Brand knowledge has two dimensions: brand awareness and brand image.

Furthermore brand awareness can be classified into active (brand recall) and

passive (brand recognition) awareness. (Keller 1993). Commitment is the

emotional or psychological attachment to a brand. A sports consumer is

committed if he/she feels a deep and persistent attachment to his/her

favourite team and resists conflicting information or experience and the future

welfare of the team is important (Bauer, Sauer and Exler 2005). Committed

fans are loyal (Bauer, Sauer and Schmitt 2005).

This study studies fan loyalty in Finnish men’s ice hockey during the regular

season 2008-2009 using stochastic frontier analysis. Most teams in the

highest ice hockey league are local monopolies but there are two teams in

Helsinki which might be substitutes since the distance between their stadiums 109

is less than 3 km. Anecdotal evidence suggests that the fans of these teams

come from different districts partly due to different public transport to their

stadiums. One is located near to a railway station that can be reached easily

from the eastern parts of Helsinki while the other is easily reached through

bus services from the western parts of Helsinki. Moreover, there is one team

in the neighbouring city, Espoo, whose stadium is at a distance of about 13 km

from the previous. In addition there are two teams in Tampere with a shared

stadium. However, some teams are local monopolies, and some teams meet

higher competition. Therefore brand loyalty or fan loyalty might differ

according to competitive position and the aim of this research is to study the

relationship between fan loyalty and competitive position of teams.

Competitive position is defined here as the geographical distance between

teams’ stadiums. Consumers or spectators can be either loyal to ice hockey or

to a particular team. Those living in Helsinki or Tampere region where there

are at least two teams and are only loyal to (highest league) ice hockey have

the possibility not to remain loyal to a particular team.

Teams in the highest leagues generally get revenues not just gate revenues

but also from merchandise sales, sales of broadcast rights and commercial

sponsorships. Loyal fans use fan shits and scarves. Broadcast rights are

usually sold by the league association and the broadcast revenue is shared

among the teams. Sponsorship revenue is associated with larger attendance

which in turn is associated with larger market base, i.e. larger home town

population. This in enables increased budgets to spend on playing and

coaching that facilitates improved team performance (Buraimo, Forrest and

Simmons 2007). However, Bauer, Sauer and Exler (2005) show that among

110

German soccer fans the success of the team is not the central driving force of

a fan’s utility opinion.

Fan loyalty can be measured as permanency of successive years’ attendance

(Winfree, McCluskey, Mitterhammer and Fort 2004), mean match tickets per

market size (Brandes, Franck and Theiler 2010) or as an efficiency score in

stochastic frontier analysis (Depken 2000, 2001). Also direct surveys to get

self-revealed levels of fan loyalty have been used. Wakefield and Sloan (1995)

show that fan loyalty increases home game attendance. The novelty of this

study is that a panel data of Finnish men’s highest league ice hockey

attendance during the regular season 2008-2009 will be analysed using

stochastic frontier analysis. There were 406 games played during that season

beginning in September 2008 and ending in March 2009. The Finnish men’s

highest league, labelled “SM-Liiga” was a closed league with 14 teams, so the

last did not drop. The best 10 teams continued in play-off games and the

champion (JYP) was known in Mid April. During the regular season all teams

had 29 home games with total 1997019 spectators, i.e. on average 4919 per

game ranging from 8456 (Jokerit from Helsinki) to 3437 (SaiPa from

Lappeenranta).

3.2 Fan loyalty or brand loyalty – stochastic frontier analysis -

Following Depken (2000, 2001), a panel of Finnish men’s highest league ice

hockey for the regular season 2008 – 2009 is used in the estimation. A

conventional Cobb-Douglas form to explain attendance is used:

111

(1) ATT i=C[∏j=1

k

X ijβj]exp¿)[ 1λ i ]TSV ¿

As (1) - in which TSV* denote for time specific variables, like weekday dummy

and climate conditions (temperature) - is transformed by taking logs of both

sides, we get

(2)ln ATT i=lnC+∑j=1

k

β j ln X ij+∑j=1

k

γ jTSV ij+εi¿

In which C a constant term identical to all teams, the β j and γj are parameters

to be estimated, and ε*I = εi – ln(λi) is the error term.

The explanatory variables Xi used in this study are conventional and consistent

with other studies (for a review, see Borland and MacDonald 2003 or

Simmons 2006): home town population, visitor’s town population, distance

between teams’ home stadiums, the winning percentage of the home team and

of the visitor team, the game round, the unemployment rate. The time specific

variables are weekday dummies and the outside temperature measured in

Celsius. Due to the nature of these variables, they are not transformed by

taking logs. The error term has two components ε*I = εi – ln(λi) in which εi is

the random error term that captures noise as well as team and time-specific

unobserved heterogeneity (Greene 2005).The inefficiency term λi in the

stochastic frontier is time invariant and team specific. Two possible

distributions have frequently been used (see Greene 2008, 538): the absolute

value of a normally distributed variable (“half-normal*) and an exponentially

distributed variable. The distributions are asymmetric. However, the problem

with stochastic frontier analysis is that the error term distribution assumption

has its effects on the size of the fan loyalty. If the team specific term is fixed,

112

one of the teams is considered strong (as 100 % strong) in the sense of fan

loyalty. Fans are committed. The fan loyalty of the other teams is relative to

the best-practise team(s) in the sample (cf. Last and Wetzel 2010). The fan

loyalty estimates are sensitive to sample selection criteria and outliers. The

fixed effects approach is distribution free and it allows for correlation between

effects and time-specific regressors. The random effects approach maintains

the original distributional assumption. With a half-normal model the least

squares is unbiased and consistent and efficient among linear unbiased

estimators while the maximum likelihood estimator is not linear but it is more

efficient (Greene 2008, 539). The shortcoming with a random effects model is

that it has stronger distribution assumptions that the effects are time

invariant and uncorrelated with the explanatory variables in the model

(Greene 2005). Furthermore, these models tend to overestimate the disloyalty.

The assumption of time invariance is more problematic if the time series is

long, however, in this study the sample consists of game attendance during

the regular season 2008 – 2009, i.e. from September 2008 to March 2009.

As the disloyalty measure approaches 0, fans are more loyal, and the other

explanatory variables, especially winning percentage and other time specific

variables matter less. Teams with low levels of fan loyalty lose more

spectators as the quality of games goes down. The climate conditions, i.e. the

temperature and a worsening winning record are more relevant and the less

enthusiast spectators do not attend. There is a wide sports economics

literature that use frontier models but most of these focus on cost efficiency

(for a good survey, see Barros and Garcia-del-Barrio 2008) or technical

113

efficiency (Kahane 2005). The output typically is related to team performance,

like winning percentage and the inputs are cost related, like wages or the

number of coaches. Most of these study professional baseball or football

(soccer) in USA or UK.

There are few studies with ice hockey data and even less using frontier

analysis (Kahahe 2005). The pioneering attendance study of Noll (1974) has

been very influential. Jones and Ferguson (1988) showed with NHL data that

home town population, winning percentage and team related attributes like

way to play, the number of stars in the team are important to explain

attendance. The effect of population incomes was negative. However, Cocco

and Jones (1997) show that the effect of incomes was positive as expected.

Using frontier analysis Kahane (2005) shows that teams owned by

corporations are more efficient than team owned by individuals.

3.3 Stylized facts: Finnish Ice Hockey

There were 14 teams playing in the highest men’s ice hockey league in

Finland. Three of the teams were located in the metropolitan area of Helsinki,

two from Tampere and the rest are local monopolies.

114

Table 21: (Table 3.1) Average attendance statistics

Team Home game average attendance, regular season 2008-2009, n

= 29

Variation of home game attendance: min – max (std), coefficient

of variation

Home town population

1st

September 2008

Distance to the

nearest team, km

Blues 4651 3922 – 5722 (476.2)0.102

240275(Espoo)

12.6

HIFK 6324 5005 – 8200 (933.9)0.148

571887(Helsinki)

2.6

HPK 3780 3205 – 5360 (525.9)0.139

65941(Hämeenlin

na)

75.5

Ilves 5672 4197 – 7800 (936.9)0.165

208657(Tampere)

0

Jokerit 8463 6283 – 13464 (1672.5)0.198

571887(Helsinki)

2.6

JYP 4016 3531 – 4180 (192.3)0.048

127186(Jyväskylä)

147.2

KalPa 4599 3703 – 5225 (429.1)0.093

91601(Kuopio)

148.9

Kärpät 5741 4909 – 6614 (472.4)0.082

132726(Oulu)

289.2

Lukko 3708 3019 – 5400 (567.8)0.153

39757(Rauma)

50.0

Pelicans 4081 3422 – 4910 (470.8)0.115

99816(Lahti)

75.5

SaiPa 3437 2843 – 4847 (457.0)0.133

70267(Lappeenra

nta)

152.5

Tappara 5138 3718 – 7800 (998.4)0.194

208657(Tampere)

0

TPS 5139 3831 – 6813 (901.1)0.175

175279(Turku)

87.4

Ässät 4110 3001 – 6472 (723.0)0.176

76355(Pori)

50.0

Note: Ilves and Tappara from Tampere are using a common stadium, whereas HIFK 115

and Jokerit from Helsinki have separate stadiums. Distance is measured from the team’s stadium to the nearest.

The Coefficient of Variation of the attendance variable and the distance to the

next nearest team are negatively correlated (ρ = -0.669) which may indicate

that loyalty is positively associated with the competitive position, i.e. local

monopolies have more loyal fans. However, since team performance among

others has been shown to have an impact on attendance, a stochastic frontier

analysis is needed to obtain more certain view about this proposition. Most of

the games have been played on Tuesdays (28.6 %), on Thursdays (26.1%) and

on Saturdays (33%), some games on Fridays (32 games, i.e. 7,9%) and the rest

on Mondays (9), Wednesdays (5) and Sundays (4). The correlations of weekday

dummies with the other variables are negligible except that during Saturdays

the attendance is bigger. The teams from Helsinki (HIFK and Jokerit) have

had the biggest attendance but it has been declining during the last years (see

Appendix 1). Only two teams were able to increase average attendance

during the regular season 2008 – 2009: HPK by 15 % and KalPa by 36 %. The

spectator number of TPS (-14 %) and Tappara (-10 %) decreased most.

116

Table 22: (Table 3.2) Variables, means, standard deviations and correlation matrix

Variable

Mean Std Att Cap Price Dist Temp Unempl

HomePop

VisPop

PPGH PPGV FGH FGV HomeG

Att 4918.6

1500.9

1 0.71 0.29 -0.09 -0.06 -0.47 0.75 0.09 0.08 -0.01 0.10 -0.08 0.06

Cap 7066.6

2595.0

1 0.23 -0.08 0.07 -0.65 0.70 -0.04 -0.04 0.00 0.02 -0.05 -0.00

Price 28.2 4.1 1 -0.29 0.11 -0.52 0.52 0.11 0.14 -0.05 0.02 -0.05 -0.01

Dist 253.2 155.6 1 -0.08 0.21 -0.14 -0.13 0.09 0.10 0.05 0.09 0.00

Temp 3.0 6.3 1 -0.44 0.07 -0.02 -0.09 -0.08 -0.10 -0.07 -0.83

Unempl 8.4 2.2 1 -0.71 0.05 -0.05 0.03 -0.00 0.07 0.41

HomePop

192187

166893

1 -0.07 0.12 -0.00 0.07 -0.05 -0.02

VisPop 192198

166861

1 -0.02 0.16 -0.03 0.06 0.03

PPGH 1.47 0.44 1 0.14 0.52 0.10 0.13

PPGV 1.48 0.43 1 0.07 0.44 0.07

FGH 4.39 2.44 1 0.04 0.14

FGV 4.31 2.50 1 0.08

HomeG 29.5 16.8 1

Note : Att = home game attendance, Cap = Capacity of Stadium, Dist = distance between home team’s stadium and visitor’s 117

stadium (km), Unempl = province monthly unemployment rate, HomePop = home town population in the beginning of the month, VisPop = visitor’s town population in the beginning of the month, PPGH = points per game, home team, PPGV = points per game, visitor, FGH = form guide (3 last games), home team, FGV = form guide, visitor, HomeG = leg. Price is the ticket price of the best plain seats, not box seats, it is overrated since most of the seats are cheaper. n = 406

118

The home town population is positively correlated with the attendance, the

capacity of the stadium and the ticket price and negatively with the province

unemployment rate. These variables are also associated in other respects. The

key variable or the cause is the town population. The two alternative team

performance variables (for the home team: points per game, PPGH and the

form guide, FGH, or the corresponding variables for the visitor: PPGV and

FGV) are correlated. The temperature is negatively correlated with the leg,

i.e. in September when the season begins the temperature is higher than in

March when the regular season ends. The unemployment rate was increasing

during the season and the variables are positively correlated.

119

3.4 Estimation and results

The model is estimated first with OLS because of comparability and then with

MLE assuming that the inefficiency term is distributed half-normal.

Table 23: (Table 3.3) OLS results, dependent variable is log(Attendance), n = 406

Variable OLS OLS OLS OLS OLS OLS OLS OLS

LnCap 0.360 ***

(0.035)

0.270***

(0.032)

0.345***

(0,036)

0.264***

(0.033)

0.349***

(0.035)

0.263***

(0.032)

0.330***

(0.035)

0.254***

(0.033)

LnPrice -0.018(0.062)

-0.173**(0.058)

0.003(0.064)

-0.138**(0.059)

-0.009(0.063)

-0.162*

*(0.058

)

0.008(0.064)

-0.129*

*(0.059)

LnDist -0.032*

**(0.006)

-0.032**

*(0.006)

-0.031*

**(0.006)

-0.031**

*(0.006)

-0.033*

**(0.006)

-0.034*

**(0.006

)

-0.032*

**(0.006)

-0.033*

**(0.007)

Temp -0.003(0.002)

-0.006**(0,002)

-0.005*

*(0.002)

-0.007**

*(0.002)

-0.003(0.002)

-0.006*

**(0.002

)

-0.004*(0.002)

-0.007*

**(0.002)

LnUnempl 0.274***

(0.049)

0.247***

(0.050)

0.273***

(0.049)

0.243***

(0.050)

LnHomePop

0.233***

(0.016)

0.201***

(0.010)

0.234***

(0.016)

0.205***

(0.016)

0.236***

(0.016)

0.202***

(0.015)

0.238***

(0.016)

0.207***

(0.015)

LnVisPop 0.032***

0.037***

0.034***

0.038***

0.030**

0.034***

0.031**

0.036***

120

(0.009) (0,010) (0.010) (0.010) (0.094) (0.010)

(0.010) (0.010)

LnPPGH 0.027**

(0.012)

0.018(0,012)

0.034**

(0.012)

0.025*(0.012)

LnPPGV -0.040*

*(0,013)

-0.042**(0,013)

-0.042*

*(0.013)

-0.043**(0.013)

LnFGH 0.004(0.004)

0.002(0.004

)

0.005(0.004)

0.004(0.004)

LnFGV -0.008*(0.004)

-0.008*(0.004

)

-0.008*(0.004)

-0.008*(0.004)

LnHomeG -0.031(0.016)

-0.018(0,017)

-0.042*

*(0.016)

-0.028*(0.017)

-0.028*(0.014)

-0.022(0.014

)

-0.034*

*(0.014)

-0.028*(0.014)

Tuesday -0.120*

**(0.017)

-0.111**

*(0.017)

-0.117*

**(0.017)

-0.109*

**(0.017

)

Thursday -0.121*

**(0.017)

-0.116**

*(0,018)

-0.118*

**(0.017)

-0.111*

**(0.018

)

Saturday 0.113***

(0,015)

0.111***

(0.016)

0.109***

(0.015)

0.106***

(0.016)

Constant 1.97***

(0.482)

4,15***(0.291)

1.99***

(0.490)

3.96***(0.293)

2.02***

(0.483)

4.21***

(0.293)

2.08***(0.492)

4.023***

(0.294)

121

R2 0.747 0.728 0.737 0.721 0.743 0.724 0.733 0.717

LogL (Χ2) 238.97 (570.1

4)

223.63 (539.46

)

230.68 (553.5

4)

218.52 (529.23

)

236.17 (564.4

7)

220.74 (533.6

8)

227.37 (546.9

4)

215.52 (523.2

3)

All the variables, except the day of the week and the temperature, are logarithmic. ***, **, * denote 1%,5%,10% significance

122

The OLS estimated price elasticity is negative only if the unemployment

variable is not enclosed. These variables are highly negatively correlated

and therefore the OLS estimates for the price variable without the

unemployment variable are more plausible. The home town population

coefficient is positive as expected and roughly 5 - 6 times higher than

the visitor’s town population coefficient. The distance between the

towns is significantly negative. The temperature matters even though

the coefficient is tiny. The team performance measured from the

beginning of the season is more suitable than the form guide which

measures the performance of the last three games. Spectators can easily

observe the points per game variable through tracking statistics that are

shown in newspapers. During the season the spectator number

diminishes since the leg variable (HomeG) is negative. The difference

between Tuesday and Thursday games is significant with the Saturday

games.

Table 24: (Table 3.4) Estimation results, dependent variable is log(Attendance), n = 406

Variable OLS SF,MLE, fixed

SF, MLE, random

LnCap 0.264***(0.033)

0.264***(0.032)

0.279(2.92)

LnPrice -0.138**(0.059)

-0.138**(0.058)

0.325(2.63)

LnDist -0.031***(0.006)

-0.031***(0.006)

-0.041(0.032)

Temp -0.007***(0.002)

-0.007***(0.002)

-0.003(0.024)

LnHomePop 0.205***(0.016)

0.205***(0.015)

0.143(1.004)

LnVisPop 0.038*** 0.038*** 0.018

123

(0.010) (0.010) (0.019)

LnPPGH 0.025*(0.012)

0.025*(0.012)

-0.025(0.057)

LnPPGV -0.043**(0.013)

-0.043**(0.013)

-0.013(0.073)

LnHomeG -0.028*(0.017)

-0.028*(0.016)

0.003(0.280)

Saturday 0.111***(0.016)

0.111***(0.015)

0.107**(0.048)

Constant 3.96***(0.293)

3.96(4.63)

3.47(21.2)

σ 0.141 0.310

R2 0.721

LogL 218.52 218.52 293.01

124

The stochastic frontier model (Table 4) without panel data assumption

(fixed effects) yields almost totally similar results as OLS. The similarity

of the OLS and MLE estimates is not surprising since both methods

generate consistent estimates. With fixed effects model the inefficiency

of the team is relative to the best. With random effects model the

inefficiency score is E[u|e]. The inefficiency scores of the teams are

listed in table 5 below.

Table 25: (Table 2.5) Inefficiency scores of teams

TeamInefficiency, Fixed

effects model

Inefficiency, random effects

model

Blues 0,295 0.392

HIFK 0.172 0.306

HPK 0.199 0.294

Ilves 0.156 0.203

Jokerit 0.069 0.134

JYP 0.174 0.286

KalPa 0.062 0.119

Kärpät 0 0.015

Lukko 0.089 0.286

Pelicans 0.161 0.291

SaiPa 0.283 0.334

Tappara 0.209 0.318

TPS 0.278 0.376

Ässät 0.205 0.308

The inefficiency scores are positively correlated (ρ = 0.907), even the

random effects model is unsatisfactory. These inefficiency scores and

125

the coefficient of variation presented in table 1 measuring the variation

of teams’ attendance figures are associated with the distance measure

also presented in table 1.

126

Table 26: (Table 2.6) Correlation matrix of selected variables

Inefficiency. Fixed effects model

Inefficiency. random effects model

coefficient of variation

distance

Inefficiency. Fixed effects model

1 0.907 0.178 -0.371

Inefficiency. random effects model

1 0.213 -0.509

coefficient of variation

1 -0.669

distance 1

The correlation coefficients reveal that fan loyalty measured as

inefficiency scores or attendance’s coefficient of variation is associated

with the distance, i.e. competitive position of the team: the bigger the

distance, the bigger fan loyalty.

3.5 Conclusions

The purpose of this paper was to consider the relationship between fan

loyalty and the competitive position of the men’s highest league ice

hockey teams in Finland using stochastic frontier approach. Fan loyalty

was measured conversely as inefficiency score of the stochastic frontier

model explaining attendance of games. This approach is useful because

127

it reveals that there are differences in fan loyalty and its relation with

the competitive position is plausible.

The random effects model is unsatisfactory since the coefficients of the

variables are not significant and therefore inefficient. The fixed effects

model is more plausible since it captures both the relevant explanatory

variables for attendance and the inefficiency scores. The estimated

coefficients of the explanatory variables are in line with those reported

in the previous literature. Since the team loyalty scores seem to be

correlated with the distance measure, the fans are more committed to

ice hockey and not to a particular team. The brand of ice hockey is

stronger than the brand of an individual team. This is consistent with

the results of Bauer, Sauer and Exler (2005) who show that non-product-

related attributes (e.g. logo and club colours, club culture and tradition,

stadium and regional provenance) are more important for fan loyalty

than product-related attributes like players, success, general team

performance.

The coefficient of variation of the teams’ attendance number and the

distance measure are more negatively correlated than the inefficiency

scores obtained through the stochastic frontier models but still the used

approach is suitable to explain fan loyalty. However, a larger data is

needed to confirm the validity of the results.

128

References

Barros. Carlos Pestana & Pedro Garcia-del-Barrio (2008): Efficiency measurement of the English football Premier League with a random frontier model. Economic Modelling. 25. 994-1002

Bauer. Hans H.. Nicola E. Sauer & Philipp Schmitt (2005): Customer-based brand equity in the team sport industry. Operationalization and impact on the economic success of sport teams. European Journal of Marketing. vol. 39. No. 5/6. 496-513

Bauer. Hans H.. Nicola E. Sauer & Stefanie Exler (2005): The loyalty of German soccer fans: does a team’s brand image matter? International Journal of Sports Marketing & Sponsorship. vol 7. issue 1. 14-22

Borland. Jeffery & Robert Macdonald (2003): Demand for sport. Oxford Review of Economic Policy. 19. 478-502

Brandes. Leif. Egon Franck & Philipp Theiler (2010): The Group Size and Loyalty of Football Fans: A Two-Stage Estimation Procedure to Compare Customer Potential Across Teams. University of Zurich. Institute for Strategy and Business Economics (ISU). No 126

Buraimo. Babatunde. David Forrest & Robert Simmons (2007): Freedom of Entry. Market Size. and Competitive Outcome: Evidence form English Soccer. Southern Economic Journal vol. 74. 204-213

Cocco. Angelo & J.C.H. Jones (1997): On going south: the economics of survival and relocation of small market NHL franchises in Canada. Applied Economics. 29. 1537-1552

Depken. Craig A. II (2000): Fan Loyalty and Stadium Funding in Professional Basketball. Journal of Sports Economics. vol. 1. 124-138

Depken. Craig A. II (2001): Research Notes: Fan Loyalty in Professional Sports: An Extention of the National Football League. Journal of Sports Economics. vol. 2. 275-284

129

Greene. William (2005): Fixed and Random Effects in Stochastic Frontier Models. Journal of Productivity Analysis. 23. 7-32

Greene. William H. (2008): Econometric analysis. 6th Edition. Pearson International Edition

Jones. J.C.H. and D.G. Ferguson (1988): Location and survical in the National Hockey League. The Journal of Industrial Economics. 36. 443-457

Kahane. Leo H. (2005): Production Efficiency and Discriminatory Hiring Practices in the National Hockey League: A Stochastic Frontier Approach. Review of Industrial Organization. 27. 47-71

Keller. Kevin Lane (1993): Conceptualizing. Measuring and Managing Customer-Based Brand Equity. Journal of Marketing. 57. 1-22

Last. Anne-Kathrin & Heike Wetzel (2010): The efficiency of German public theaters: a stochastic frontier analysis approach. Journal of Cultural Economics. 34. 89-110

Simmons. Rob (2006): The demand for spectator sports. In Handbook on the Economics of Sport. edited by Wladimir Andreff & Stefan Szymanski. 77-89. Edward Elgar

Wakefield. Kirk L. & Hugh J. Sloan (1995): The Effects of Team Loyalty and Selected Stadium Factors on Spectator Attendance. Journal of Sport Management. vol. 9. 153-172

Winfree. Jason A.. Jill J. McCluskey. Ron C. Mittelhammer & Rodney Fort (2004): Location and attendance in major league baseball. Applied Economics 36.2117-2124.

Source of the statistical data: official statistics (www.stat.fi) or The Finnish National Hockey League (www.sm-liiga.fi). Estimation made with NLogit 4.0 (www.limdep.com)

130

Table 27: (Table 2.7) Average attendance, home games, regular seasons

Jokerit HIFK TPS Ilves Tappara

Kärpät Blues Ässät JYP SaiPa HPK Pelicans

KalPa Lukko Total

2005-2006

8850 6821 6444 5867 5866 5791 5111 4183 3808 3724 3555 3455 3302 3183 69960

2006-2007

8928 6629 6441 5660 5619 5697 4763 4391 3351 3439 3556 4042 3120 3769 69405

2007-2008

8591 6573 5979 5914 5712 6055 4838 4235 4055 3558 3282 4253 3388 3733 70166

2008-2009

8456 6324 5139 5672 5138 5741 4651 4120 4016 3437 3780 4081 4599 3708 68862

share. %2005-2006

12.7 % 9.7 % 9.2 % 8.4 % 8.4 % 8.3 % 7.3 % 6.0 % 5.4 % 5.3 % 5.1 % 4.9 % 4.7 % 4.5 %

2006-2007

12.9 % 9.6 % 9.3 % 8.2 % 8.1 % 8.2 % 6.9 % 6.3 % 4.8 % 5.0 % 5.1 % 5.8 % 4.5 % 5.4 %

2007-2008

12.2 % 9.4 % 8.5 % 8.4 % 8.1 % 8.6 % 6.9 % 6.0 % 5.8 % 5.1 % 4.7 % 6.1 % 4.8 % 5.3 %

2008-2009

12.3 % 9.2 % 7.5 % 8.2 % 7.5 % 8.3 % 6.8 % 6.0 % 5.8 % 5.0 % 5.5 % 5.9 % 6.7 % 5.4 %

131

change%05/06 ->

06/070.9 % -2.8 % 0.0 % -3.5 % -4.2 % -1.6 % -6.8 % 5.0 % -12.0

%-7.7 % 0.0 % 17.0 % -5.5 % 18.4 % -0.8 %

06/07 -> 07/08

-3.8 % -0.8 % -7.2 % 4.5 % 1.7 % 6.3 % 1.6 % -3.6 % 21.0 % 3.5 % -7.7 % 5.2 % 8.6 % -1.0 % 1.1 %

07/08 -> 08/09

-1.6 % -3.8 % -14.0 %

-4.1 % -10.0 %

-5.2 % -3.9 % -2.7 % -1.0 % -3.4 % 15.2 % -4.0 % 35.7 % -0.7 % -1.9 %

Medalists

2000-2001

1. 3. 2.

2001-2002

1. 2. 3.

2002-2003

1. 2. 3.

2003-2004

3. 2. 1.

2004-2005

2. 1. 3.

2005-2006

3. 2. 1.

2006-2005

2. 1. 3.

132

2007-2008

3. 1. 2.

2008-2009

2. 1. 3.

Appendix 1: Source: www.sm-liiga.fi

133

Fan loyalty in Finnish Ice Hockey

Seppo SuominenHaaga-Helia University of Applied SciencesMalmi campus, Hietakummuntie 1 AFIN-00700 Helsinki, Finlande-mail: seppo.suominen@haaga-helia.fi

The study studies fan loyalty in Finnish men’s ice hockey during the regular

season 2008-2009 using stochastic frontier analysis. Fan loyalty is measured

as inefficiency score of the stochastic frontier model explaining games’

attendance. There were 14 teams playing in the highest men’s ice hockey

league in Finland with 406 games, i.e. all teams had 29 games at home

stadium and 29 games as visitor. The error term of the stochastic frontier

model has two components and the other of these can be considered as the

inefficiency term or inversely as the fan loyalty term. The fan loyalty measure

is reasonable and negatively correlated with the distance between home

stadium and the nearest stadium of the other team. The distance is a proxy for

local provenance or monopoly position. The fixed effects model is plausible

and the fan loyalty terms are reasonable while the random effects model is not

efficient.

134

4. Spectators of performing arts – who is sitting in the auditorium?

Seppo SuominenHaaga-Helia University of Applied SciencesMalmi Campus, Hietakummuntie 1 A, FIN 00700 Helsinki, Finland

4.1 Introduction

Approximately 5 or 6 per cent of Finns go to see performing arts or to an art

exhibition several times per month and every sixth do not go at all. Most of

Finns occasionally go10. The purpose of this study is to find out in more detail

the differences in visitor density. What kind of person is a heavy user of

performing arts (art exhibition, opera and theatrical performances) and

correspondingly what kind of people do not go at all? Are there any

differences across areas or provinces when the effect of a person’s education

is taken into account?

10 According to a survey on cultural participation that was conducted in 1999 (“Kulttuuripuntari”), a typical opera visitor is a 50 to 64 year- old female with university education living in Uusimaa region (in Southern Finland including the capital, Helsinki). The opposite person is a young male with low education living in a sparsely populated area in Northern Finland. Correspondingly an art exhibition visitor is typically a 50-64 year- old female living in the city centre and a non-visitor is young male with vocational education and living in a sparsely populated area or an adult man with low education. A typical theatre visitor is more than 50 year- old working female with university education, living in a city in Southern or Western Finland, while the non-visitors are on average unemployed 20-24 year- old men without vocational education living in Northern Finland. Kulttuuripuntari 1999, Opetusministeriön Kulttuuripolitiikan osaston julkaisusarja 9:1999, readable: http://www.minedu.fi/OPM/Julkaisut/1999/kulttuuripuntari_1999__raportti_kulttuuripalvelujen_kaytosta_ja?lang=fi, read 31.5.2010

135

Fairly many good reports have been drawn up on the audiences of cultural

events, but a great majority of the results have been presented as descriptive

statistics and virtually there are no studies that have used multivariate

analysis. Leisure activities: Culture and sport in 1991 and 1999 (Statistics

Finland 2001) clarifies among others what the activity of more than 10-year-

old Finns is to go to the theatre, dance performances, concerts and opera and

as comparison go to sport events during the years in question. The results

(table 1) are equal to the 1999 survey (Kulttuuripuntari): women go more

often than men, highly educated are more active and the differences among

provinces are substantial.

(table 1 about here)

Based on the statistics above, the theatre visitor density did not change

essentially during the 1990’s except that the youngest and citizens living in

eastern Finland may have increased the visitor density. The visitor density of

the dance performances was roughly half of the theatres and no substantial

changes took place during that decade. Opera and concert performances

increased especially among the 25-44 and 45-64 year-old cohort and in the

Uusimaa region (both the Helsinki metropolitan area and eastern Uusimaa

province). In comparison, the spectator frequency of the sport events and

competitions decreased. The majority of the sport spectators are men. The

audience survey of two regional and occasional operas at Pori and Tampere

(Kivekäs 1991) reveal that the visitors have on average a higher education

than those of the theatres but it is to some extent lower than that of the

136

National opera and the Savonlinna opera festival. During the autumn season

in 1999 the Pori opera was distinctively an opera of its own area, while the

audience at Tampere opera was more national. The audience of the

Savonlinna opera festival has a high education: more than 55 per cent have a

university degree. There are more women than men, the 60—69 year-old

cohort is the largest if the audience is classified with ten- year spaces

(Mikkonen and Pasanen 2009). By far the most of the spectators come from

other cities during the summer holidays since the opera festival is kept in July

and August. The music festival in Kangasniemi, which is similar to a music

species event, has a similar audience to that of Savonlinna opera festival

(Mikkonen, Pasanen and Taskinen 2008).

In the early 1990’s one of the most severe economic recession in the

economic history of Finland occurred. Unemployment rose to the record and

that might have had an impact on the visitor density. When the years 1981,

1991 and 1999 are compared (table 4) it can be seen that practically all

groups (women, men, different age cohorts) had the lowest density figure in

1991. Only the pensioners did not cut down going to the theatre, on the

contrary they increased visits.

The figures in the Culture barometer (Kulttuuripuntari in table 2) are

substantially higher than those presented above. The reason for these higher

figures is not known. The leisure activities: culture and sport statistics are

based on interviews on the use of leisure time made by the Statistics Finland.

The 1999 figures were collected between March 1999 and February 2000,

137

while the Culture barometer data was collected in connection with the labour

study. One commercial research institution (Taloustutkimus) has conducted

several surveys on the visitor density of theatre, opera or ballet performances

on the basis of assignments by the association of Finnish theatres (Suomen

teatterit). The sample size has been around 1000 in the surveys in 1995, 1998,

2001, 2004 and 2007 (table 3).

(tables 2 and 3 about here)

The results in table 3 show that women go more often to the theatre, opera or

ballet than men, especially the difference is biggest among those that visit

twice or 3 – 5 times per year. It is noticeable that the amount of those that

visit 3-5 times per year has gradually declined throughout the years 1985 –

2007. The 45-64 year-old are the most active, especially among those that go

twice or 3 – 5 times per year. The effect of the visitor’s formal education is

clear; higher education and a bigger frequency are related. The effect is

stronger among those that go to the theatre, opera or ballet more than 6 times

per year. Persons living in southern Finland go most often and those in

northern Finland go least. Furthermore, the visitor density of the people in

southern Finland has been growing during the period 1985 – 2007, while

people in northern Finland have decreased visitor activity. The most often

mentioned reasons for not going were the lack of interest (56 %) or the lack of

time (26 %). Furthermore, 12 % informed that there is no theatre on the

locality.

138

The number of opera visitors in the Eurobarometer 56.0 –study (table 4) is

substantially lower than the figure in the Culture barometer survey (1999).

The Eurobarometer study was made two years later than the Culture

barometer survey. It is not known how e.g. the composition of the sample

could explain that difference. Also the visitor density in the theatres seems to

have declined: for women 68.8 % 49.5% and for men 47.1 % 30.4%. Any

macroeconomic variable, like the unemployment rate (that declined by

approximately one per cent) or the income level index (that increased by 8 per

cent) cannot explain that drop.

(table 4 about here)

On the basis of preliminary statistical examination and based on earlier

studies a hypothesis can be set:

H1: the visitor density on cultural events depends on gender, person’s age and

education. Furthermore, the regional supply has an effect.

In the Helsinki metropolitan area there are more visitors than elsewhere.

Women go more often than men. The audience composition of the theatrical or

opera performances and classical music concerts is different than that of the

sport events. Moreover, the composition of sport spectators also depends on

gender, person’s age and education in contrast to the performing arts

audiences. The second hypothesis is set:

139

H2: the spectators of sport events are reversely related to the composition of

performing arts audiences.

4.2 Method and sample

The most recent data, ISSP 2007 (International Social Survey Programme) is

based on a mail survey that was carried out by the Statistics Finland Autumn

2007 (18th September – 11th December 2007). The sample unit is a person, age

between 15 and 74. The sample method was a systematic random sample from

the population register, the sample size was 2500 but only 1354 answers were

returned, i.e. the response rate was 54.2%. The key words in the ISSP 2007

are the following: use of time, physical condition, hobbies, organisations,

board games, physical education, holiday, games, social relations, sports,

leisure. As background information the following data among others was

collected: gender, year of birth, size of the household, education, participation

in working life, profession, source of livelihood or branch, regular weekly

working hours, professional status, employer (private or public sector),

membership of a trade union, voting behaviour, religiousness, income and

some information concerning the place of residence.

There are at least three suitable statistical methods that can be used with that

data: 1) the analysis of variance (ANOVA), the multivariate analysis of

variance (MANOVA) or the covariance analysis (MANCOVA), 2) the

multinomial logit and 3) the bivariate probit. The analysis of variance is a

140

suitable method for comparing the difference of means of two groups (e.g. the

heavy users and the rest). With the MANOVA it is possible to have more than

one explanatory variable (e.g. gender, income, province, and age). In the

MANCOVA the values of the explanatory variables are corrected with the

information from the covariate. The purpose of this covariate is to reduce the

heterogeneity of the variable to be explained: for example most of the

audience at the opera live in the Uusimaa region, therefore it is reasonable to

use the place of residence as covariate. If there are many explanatory

variables, both MANOVA and MANCOVA give results that can be divided into

the separate and joint effects of each variable. The total sum of squared

deviations about the grand mean is partitioned into a sum of squares due to

many sources and a residual sum of squares. However, the direction of the

effect remains open, e.g. it is not known whether higher incomes increase or

decrease opera visits.

The deviation of the individual from the grand mean Xij – GM in the analysis of

variance can be divided into two parts:¿) + ( X j−GM ). The first part is the

deviation of the individual from its own group’s mean and the second part is

the deviation of the group mean from the grand mean. When the deviation is

calculated to all observations, the total sum of squares ∑i=1

n

∑j=1

k

(X ij –GM )2 = SStotal

can be partitioned into two parts: SSwithin and SSbetween, i.e. the internal (within)

sum of squares and the sum of squares between the groups (between). When

the sums of squares are divided by their degrees of freedom (within = N – k,

between = k – 1, where N is the sample size and k in the number of groups),

141

the mean squares are obtained. The mean squares of the parts (i.e. within and

between) are compared with the F-test.

The test statistics F=

SSbetween(k−1)SSwithin(N−k )

is distributed according to the F-distribution. If

the difference between groups is significant, the difference can be evaluated

with ή2 = SSbetween/SStotal which tells how much of the variation of the variable

to be explained can be explained by the grouping variables (Metsämuuronen

2009, 785-789).

The second possible statistical method is a logistic regression analysis or

multinomial logit. An equation explaining the visitor density of performing

arts must be formulated to find out the impact of each explanatory variable.

Furthermore, it is possible to predict behaviour because the effect and

direction of explanatory variables are found out. The variable to be explained

is either a binary variable (binary logistic) or multinomial but rather often also

ordered variable (multinomial logistic). In the ISSP 2007 data the question is:

“How often during the past 12 months on your leisure did you go to converts,

theatrical performances, art exhibitions, etc.?” The answer alternatives were:

1 = daily, 2 = several times per week, 3 = several times per month, 4 = less

often, 5 = never. When a binary logistic method is used, the alternatives could

be reclassified for example so that one alternative is a combination of 1,2 and

3 and the second alternative is a combination of 4 and 5. If the probability of

the first choice is p and the probability of the second is 1-p, then

142

logit ( p )=log p1−p

=log (p )−log (1−p )=Xβ+u where X includes all explanatory

variables and β is the vector of coefficients, u is the error term. The statistical

significance of β can be evaluated with a suitable test. Usually it is assumed

that the error term is distributed according to logistic (Weibull) distribution or

to normal distribution. In the last case, the model is probit. Both logit and

probit give more information compared with the analysis of variance because

both the coefficients of the explanatory variables and the direction of the

effect are found out: positive or negative and its statistical significance.

One step forward is to simultaneously study the visitor density of different

leisure activities, e.g. “performing arts” on the one hand and “at the movies”

or “physical exercise activity”. If the unobserved person’s preference for

performing arts is y1* and the preference for movies is y2* and the

corresponding explanation models are y1¿=Xβ+u1 and y2

¿=Xβ+u2 where the error

terms u1 and u2 are jointly bivariate distributed N(0,1). If the u1 and u2 are

linearly independent, the correlation coefficient ρ measures the relation of

different leisure activities’ visitor density. Under the null hypothesis that ρ

equals zero, the model consists of two independent probit equations (Greene

2008, 820). If the correlation coefficient equals zero, the performing arts

consumption and movies at the cinema consumption are unrelated (Prieto-

Rodriguez and Fernandez-Blanco 2000). The estimation of the equations could

be based on classification y1 = (“daily” or “several times per week” or “several

times per month”) = 1 if y1¿>0 and y2 = (“less often” or “never”) = 0, if y1

¿≤0. In

the above example a difference is made between “several times per month”

143

and “less often” but this separation point could be another. With the probit

model the marginal effects of each variable could be evaluated (Greene 2008,

821). The coefficients in the probit model are difficult to interpret since they

present what the effect of the variables is on the unobserved dependent

variable y*1. However, the marginal effects of the explanatory variables are on

the observed variable y1. The total marginal effect could be partitioned into

two parts: the direct marginal effect and the indirect marginal effect. The

latter part is formed through the correlation coefficient of the error terms.

All the variables of ISSP 2007 are not used in this study. Only those that are

related to the concerts, theatre and exhibitions are used according to the

purpose of the study. The questionnaire has the following question: How

often during the past 12 months on your leisure did you go to converts,

theatrical performances, art exhibitions, etc.?” The descriptive statistics are

presented in table 5.

(table 5 about here)

It is most reasonable to divide from the point of view of further analysis the

visit activity into three classes: regularly (daily, several times per week,

several times per month), occasionally (less often) and never. The results of

the analysis of variance are presented in table 6.

144

(table 6 about here)

The statistical programme (PASW 18) available did not conduct the

multivariate analysis of variance (MANOVA) with four explanatory variables

(gender, year of birth, place of province and education). On the basis of the

results it is clear that the first hypothesis is supported: performing arts visitor

density depends on gender, person’s age and education. Moreover, the

regional supply has an effect. The variance analysis shows that every

explanatory variable (gender, year of birth, place of province, and education)

would alone separate into classes: regularly, occasionally and never. The joint

effect of the explanatory variables is nearly always significant if the education

variable is present. In the table 7 different genders have been examined

separately. Both for women and men, education would seem to be the

crucially important variable.

(table 7 about here)

Even if the ISSP 2007 data would make it possible to use other explanatory

variables, these are not used since, based on rather high values of ή2, the

variables are adequate to explain consumers’ performing arts behaviour. Any

single variable alone is not good enough, but a combination of the variables

explains more.

Next the multinomial logit model results will be presented. It is assumed in

the binary logit model that the dependent variable has two classes, whose

145

probabilities are p and (1-p). In this case, the equation to be estimated is:

logit ( p )=log p1−p

=log (p )−log (1−p )=Xβ+u where X contains the explanatory

variables and β is a vector of coefficients. In the multinomial logit there are

more than two alternatives or classes, e.g. “often”, “occasionally” and “never”.

The probability of the choice “often” is Prob (Y i¿ 'often '|w i )=Pi ' often '=

exp (w iβ i)

1+∑k=1

j

exp(wk βk)

where “i” stands for the person’s i choice between different alternatives k (1,

…,) on the condition wi These conditions are characteristics that have an

impact on the person’s choice, like gender, year of birth or education. In the

multinomial logit model one alternative is the zero alternative and the other

alternatives are compared in relation to zero. The variable which describes

age has been recoded more roughly due to a more rational way of presenting

the results: age15_24, age25_34, age35_44, age45-54, age55_64, age64_.

Unfortunately the place of residence (province) cannot be used since the

number of observations on some provinces is too small.

(table 8 about here)

The often visiting group has graduated from an upper secondary school, from

a university of applied sciences or from a university. Others with different

educational background do not significantly belong to the group “often”. The

age cohorts 45-54 and 55-64 are the most active. The “occasionally” group

consists of those with elementary school, vocational school or comprehensive

146

school education or they are still at school (zero alternative, constant in the

equation). Women belong significantly more often than men to the groups

“often” or “occasionally”. The results are in line with the earlier studies

(Kivekäs 1991) and support the first hypothesis: the person’s education,

gender and age have a significant effect on the performing arts visiting

density. The province variable is not used in the estimations presented in table

8 since the number of provinces is too big. To a few groups there would have

been too few observations and the results would not have been credible. Due

to that the models in the table 9 are estimated so that the provinces have been

regrouped into bigger entities. The entities have mainly been formed

according to the NUTS2 classification.

(table 9 about here)

The first hypothesis is verified by the results of the multinomial logit analysis:

consumption of cultural events depends on gender, person’s age and

education and there are substantial differences across regions. The effect of

gender is clear between all the groups: “often”, “occasionally” and “never”.

Women are more active. As the groups “occasionally” and “never” are

compared, the effect of the person’s education is significant already when the

education is either upper secondary school or vocational school. As the

previous groups and the “often” group are compared, the effect of education

is significant when the education is upper secondary school or any university

degree. The age cohort 45-54 is most active. The impact of the region is the

following: the Uusimaa region, other southern Finland and western Finland

are different than other areas in Finland among those that “never” go to

147

cultural events. The null area (in constant) in the estimations is northern

Finland. To conclude, it can be argued that crucial educational level is upper

secondary school. All education after the upper secondary school seems to

increase cultural consumption. The separating points between the groups

“occasionally” and “often” seems to be university degree and regionally

Uusimaa region or eastern Finland.

One must keep in mind that not all persons with a university degree have

visited cultural events. Approximately 4 per cent of those who have completed

a university degree (either a bachelor’s or a master’s degree) have not

participated at all. Correspondingly 13 per cent of those belong to the group

“often”.

The second hypothesis proposes that there is a reverse relationship between

cultural events and sport events. Montgomery and Robinson (2006) show

with American data (USA 2004) that these events are exclusionary. They have

a separate public. Descriptive statistics of the sport audiences is presented in

table 10.

(table 10 about here)

Since the number of those that visit “daily” or “several times per week” is so

small, these groups and “several times per month” are combined. Hence there

are three groups: “often”, “occasionally” and “never” as above was done in

connection with the cultural events. The correlation of the three valued

148

(“often”, “occasionally”, “never”) participation into cultural events and

participation into sports events is 0.09.

The sport events consumption is classified into three groups: “often”,

“occasionally” and “never” in the multinomial logit model (table 11). Gender

classifies into groups so that men are more active. The most active sports

events consumers are those with elementary school (edu2) or comprehensive

school (edu3) education, men and younger than 45. The results are rather

opposite with the visitor density of cultural events. However, the cultural

events participation variable (cult3) has a positive coefficient in the equations

showing that the activity in the cultural events is positively related with the

activity in the sport events. These events are complements although the

audiences are rather separate. The separating point of those that have been

consuming occasionally is the age cohort 35-44. After that age consumers

choose either cultural events or a passive relationship, i.e. neither going to

see cultural events and nor sport events. The behaviour is most visible among

the 45-64 years-old.

(table 11 about here)

Finally the connection of culture events and sport events is examined with the

bivariate probit analysis. First the depending variable is binary such that 0

equals “daily”, “several times per week” or “several times per month” and 1

equals “less often” or “never”. The results are presented in table 12. It must

149

be noticed that the depending variable values of the multinomial logit analysis

are the opposite than they are in table 12.

(table 12 about here)

The classification in table 12 shows that only some explanatory variables are

significant: the age cohorts 45-54 and 55-64. They are classified significantly

less often than the group “less often” + “never”. The error term u1 in the

cultural events participation equation and the error term u2 in the sport events

participation equation are correlated: ρ = 0.463, meaning that there is a

latent visitor density factor. The classification of the depending variables in

table 13 has been formed as follows: y1 = 0 if the response is “daily”, “several

times per week”, “several times per month” or “less often” and y2 = 1 if

“never”. This classification (yes/no) is for the both depending variables: “How

often during the past 12 months on your leisure did you go to concerts,

theatrical performances, art exhibitions, etc.?” and “How often during the

past 12 month on your leisure did you go to sport events (ice hockey, football,

athletics, motor racing, etc.?”

(table 13 about here)

The results in table 13 reveal that the effect of gender is clear: women go

more often to cultural events while men are more active sport events

consumers. All education above the level 5 (5 = upper secondary school, 6 =

college, 7 = university of applied sciences, 8 = bachelor’s degree, university,

9 = master’s degree) are statistically significant in the culture participation

model. The direct marginal effect is larger the higher the education except for

the level 7 (university of applied sciences). The age cohorts 35-44, 45-54 and

150

55-64 are significantly more active in culture participation and less active in

sport events. The vital segregation point is around the age 35. At that age

consumers choose cultural events in the expense of sporting events. The place

of residence (in comparison with northern Finland which is the constant in the

equation) is significant in the cultural participation model, while in the

sporting events model only other southern Finland and western Finland are

different from the other areas. Typically the unemployment rate in northern

and eastern Finland is higher than elsewhere and this might be the reason for

the higher sporting events consumption. Moreover, the higher unemployment

in combination with the lower educational level in these regions explains that

sporting events are favoured in northern and eastern Finland.

4.3 Conclusions and evaluation

The purpose of this study is to analyse the properties of the audiences of

cultural events with the Finnish 2007 data. Mainly these events are concerts,

art exhibitions and theatrical or opera performances. The preliminary method

is the analysis of variance (ANOVA or MANOVA). The results indicate that

gender, person’s education, age and the place of residence are important

factors to classify visitor density. By definition the analysis of variance does

not expose the well-known fact that women are more active visitors. Another

method, the multivariate logit analysis is more useful since it reveals both the

direction of the effect and the statistical significance of each explanatory

variable. The results show that women are significantly more active visitors

than men even when the impact of other explanatory variables is taken into

account. The other significant factors are education (higher than upper

151

secondary school) and age (between 35 and 64) and the place of residence

(compared with the northern Finland). The most active (“often”) group and

the less active group (“occasionally”) can be separated with the factors

mentioned above: gender, education, age (in this case, often vs. occasionally:

45 – 64 years old) and Uusimaa region and eastern Finland. The first

hypothesis is verified.

An alternative way to spend leisure is to go to sporting events instead of

cultural (concerts, exhibitions, performances) events. As the impact of sport

consumption is taken into account, a suitable method is a bivariate probit

analysis. The results of the bivariate probit model show that the visitor density

of these two alternatives is positively correlated meaning that there is a

common activity factor. No previous studies have been made with Finnish

cultural consumption data using either multivariate logit or bivariate probit

model. These models give similar results, but the latter enables evaluation of

the marginal effects, both direct and indirect marginal effects. The indirect

marginal effect is significantly negative and thus reducing the cultural events

(concerts, exhibitions, performance) participation in the following cases:

education is comprehensive school (edu3), vocational school (edu4), college

(edu6), university of applied sciences (edu7) or a master’s degree (edu9). To

the contrary the direct marginal effect is positive and larger than the negative

indirect effect in the following cases: education is upper secondary school

(edu5), college (edu6), university of applied sciences (edu7), a bachelor’s or a

master’s degree (edu8 or edu9). The second hypothesis proposed is not

152

verified. The audiences in the cultural events and sport events are not

separate.

The results are in line with those of Borgonovi (2004), Montgomery and

Robinson (2006) or Masters (2007). Montgomery and Robinson showed with

data from the USA that arts’ spectators are not young, better educated and

mostly women. A similar conclusion can be made with the Finnish data.

Frateschi and Lazzaro (2008) found out with Italian data that the spouse’s,

especially husband’s high education is an important factor to explain art

(museums, concerts, theatre) consumption if they go together. However,

international comparisons must be made with caution since the international

differences in culture consumption are large (Seaman 2005 or Virtanen 2007).

The impact of regions is substantial. In this study the regions have been

formed mainly based on the NUTS2-classification. The marginal effects in the

bivariate probit model are the strongest in southern Finland: Uusimaa region

(-0.149), western Finland (-0.130), eastern Finland (-0.111) and the rest of

southern Finland (excluding Uusimaa, -0.107) when the groups “never” and

“yes” are compared.

References

Borgonovi, Francesca (2004) Performing arts attendance: an economic approach. Applied Economics 36: 1871-1885

153

Eurobarometer 56.0, http://www.fsd.uta.fi/aineistot/luettelo/FSD0099/meF0099.html

Fratesch, Carlofilippo ja Elisabetta Lazzaro (2008) Attendance to cultural events and spousal influences: the Italian case. Universita degli studi di Padova, “Marco fanno” Working Paper 84

Greene, William H. (2008) Econometric analysis, 6th Edition. Pearson International Edition

ISSP 2007: vapaa-aika ja urheilu: Suomen aineisto [elektroninen aineisto]. FSD2330, versio 1.0 (2008-03-18). Blom, Raimo (Tampereen yliopisto. Sosiologian ja sosiaalipsykologian laitos) & Melin, Harri (Turun yliopisto. Sosiologian laitos) & Tanskanen, Eero (Tilastokeskus. Haastattelu- ja tutkimuspalvelut) [tekijä(t)]. Tampere: Yhteiskuntatieteellinen tietoaineisto [jakaja], 2008

Kivekäs, Kyllikki (1991) Alueoopperat ja yleisöt. Taiteen keskustoimikunta, Tilastotietoa taiteesta 5A

Kulttuuri- ja liikuntaharrastukset 1991 ja 1999 (2001) Tilastokeskus. Kulttuuri ja viestintä 2001:5

Kulttuuripuntari 1999 (1991) Opetusministeriön Kulttuuripolitiikan osaston julkaisusarja 9: 1999. http://www.minedu.fi/OPM/Julkaisut/1999/kulttuuripuntari_1999__raportti_kulttuuripalvelujen_kaytosta_ja?lang=fi

Masters, Tristan Andrew (2007) Supply and demand of creative arts in regional Victoria, Australia. Ph.D. –thesis. Royal Melbourne Institute of Technology

Metsämuuronen, Jari (2009) Tutkimuksen tekemisen perusteet ihmistieteissä 4. Methelp 2009

Mikkonen, Jenni ja Katja Pasanen (2009) Selvitys Savonlinnan oopperajuhlien asiakasprofiileista ja aluellisesta vaikuttavuudesta. East side story – Puhtia

154

itäsuomalaiseen tapahtumamatkailuun – hankkeen tutkimusraportti. Itä-Suomen yliopisto

Mikkonen, Jenni, Katja Pasanen ja Heidi Taskinen (2008) Itäsuomalaisten tapahtumien asiakasprofiilit ja aluetaloudellinen vaikuttavuus. Joensuun yliopisto, Matkailualan opetus- ja tutkimuslaitoksen julkaisuja 1.

Montgomery, Sarah S. ja Michael D. Robinson (2006) Take Me Out to the Opera: Are Sports and Arts Complements? Evidence from the Performing Arts Research Coalition Data. International Journal of Arts Management 8:2, 24-39

Prieto-Rodríguez, Juan ja Víctor Frenández-Blanco (2000) Are Popular and Classical Music Listeners the Same People? Journal of Cultural Economics, 24: 147-164

Seaman, Bruce A. (2005) Attendance and Public Participation in the Performing Arts: A Review of the Empicical Literature. Georgia State University, Working Paper 05-03, August 2005

Suomen teatteriliitto (2001) Suomalaisten teatterissa käynti. Taloustutkimus Oy

Suomen teatteriliitto (2004) Suomalaisten teatterissa käynti. Taloustutkimus Oy

Suomen Teatterit ry (2007) Suomalaisten teaterissa käynti. Taloustutkimus Oy

Virtanen, Taru (2007) Across and beyond the bounds of taste on cultural consumption patterns in the European Union. Turun kauppakorkeakoulu Sarja A-11: 2007

Statistical estimations:

ANOVA ja MANOVA: PASW 18 (www.spss.com)

Multinomial logit and bivariate probit: NLOGIT 4.0 (www.limdep.com)

155

Table 28: (Table 4.1) Culture and physical education hobbies 1981, 1991 and 1999

Has visited during the past 12 months, %

Year Theatre Dance performa

nce

Concert Opera Sport event

Women 1981/1991/1999

52/46/47 31/22/22 40/39/42 8/6/9 -/40/30

Men 1981/1991/1999

36/28/29 19/14/16 29/27/33 4/3/5 -/57/49

Ageage 10-14 1981/1991/1

99948/35/40 25/17/21 42/31/33 5/2/3 -/79/57

age 15-24 1981/1991/1999

46/36/31 26/21/22 48/48/52 5/3/5 -/71/52

age 25-44 1981/1991/1999

49/36/39 27/18/22 31/34/41 6/4/7 -/54/48

age 45-64 1981/1991/1999

46/43/44 26/22/21 34/32/36 8/6/10 -/36/34

age 65- 1981/1991/1999

29/31/35 19/11/9 22/22/22 4/3/4 -/14/13

education:Lower basic 1991/1999 28/28 13/12 18/20 2/2 28/24

Upper basic 1991/1999 36/34 19/21 41/39 3/5 63/42

Middle 1991/1999 36/34 19/18 36/39 3/6 52/41

Lower Higher 1991/1999 49/55 24/27 47/50 8/11 55/46

Higher 1991/1999 63/60 28/29 58/61 16/22 51/46

Region:Metropolitan area

1991/1999 47/48 21/25 43/50 11/19 48/42

Rest of Uusimaa

1991/1999 33/41 18/19 34/40 4/8 46/36

Southern Finland

1991/1999 41/40 18/20 33/38 3/5 44/39

Eastern 1991/1999 30/35 17/18 34/33 3/4 43/34

156

Finland

Central Finland

1991/1999 35/36 17/16 30/31 2/3 47/39

Nortern Finland

1991/1999 26/25 17/14 26/30 3/5 48/29

Sample 4677 households in year 1999

157

Table 29: (Table 4.2) Kulttuuripuntari (culture barometer) 1999:

Has visited during the past 12 months, % 1999

Theatre Concert Opera Sport event

Women 68,8 75,8 15,1 42,9

Men 47,1 65,3 8,6 62,7

age 20-24 48,4 80,1 7,5 55,0

age 30-34 52,1 71,3 9,2 55,5

age 40-44 63,8 69,7 10,7 54,7

age 50-54 62,6 66,3 17,4 42,3

age 60-64 60,2 56,8 13,1 37,0

education:

No vocational 52,9 63,0 6,7 49,0

Vocational 49,1 63,5 6,3 54,8

University 85,0 88,9 39,3 53,2

Province (etc.):

Southern Finland

61,7 74,9 17,8 54,1

Western Finland

60,3 69,9 7,5 55,4

Eastern Finland

50,0 67,3 9,4 48,4

Oulu 49,1 67,0 8,5 51,1

Lappi 45,5 52,6 6,1 39,2

Sample size 1810 carried out in August 1999 in connection with labour study as telephone interview

158

Table 30: (Table 4.3) Suomen Teatterit (Taloustutkimus), survey on visits to theatre, opera or ballet during the past 12 months, years 1985, 1998, 2001, 2004 and 2007

Has visited during the past 12 months

1994n = 956

1998n = 1013

2001n = 994

2004n = 984

2007n = 999

once/2/3-5/6- times = total

16/12/13/3 = 44

19/12/11/3 = 45

19/11/9/4 = 43

23/11/10/2 = 42

21/14/8/4 = 47

gender:

women: 1/2/3-5/6-

17/16/18/4 = 55

20/14/16/5 = 55

22/13/14/5 = 54

25/12/14/4 = 55

22/17/11/4 = 54

men: 1/2/3-5/6- 16/9/7/3 = 38

18/10/5/1 = 34

15/9/5/3 = 36

21/10/6/2 = 39

19/10/4/3 = 36

age:

15-24 y..: 1/2/3-5/6-

19/10/10/1 = 40

22/10/7/2 = 41

23/9/6/3 = 41

25/10/4/1 = 40

17/6/5/1 = 29

25-44 y.: 1/2/3-5/6-

19/11/10/2 = 42

19/11/10/3 = 43

22/10/9/4 = 45

26/9/8/3= 46

22/12/7/3= 44

45-64 y: 1/2/3-5/6-

13/13/17/5 = 48

19/15/14/3 = 51

14/13/10/4 = 41

22/13/15/3 = 53

25/17/10/4 = 56

65-79 y: 1/2/3-5/6-

12/16/12/3 = 43

12/9/10/5 = 36

15/14/13/3 = 45

15/12/9/3 = 39

14/18/8/5 = 45

education

Elemen/compr.: 1/2/3-5/6-

15/12/10/2 = 39

17/10/5/2 = 42

16/9/7/2 = 34

19/9/5/0 = 33

18/9/7/2 = 36

Voc/tech/busi: 1/2/3-5/6-

16/11/7/3 = 37

18/11/9/1 = 39

15/8/6/1 = 30

20/8/8/1 = 37

17/13/4/2 = 36

Upper sec: 1/2/3-5/6-

17/14/18/2 = 51

26/12/9/5 = 52

27/8/11/5 = 51

20/14/11/2 = 47

25/7/8/1 = 41

Coll/polytech: 1/2/3-5/6-

16/13/22/3 = 54

20/13/17/3 = 53

24/17/12/4 = 57

36/11/12/5 = 64

23/24/12/3 = 62

Univers: 1/2/3-5/6-

24/16/15/14 = 69

18/16/22/12 = 68

19/14/17/19 = 69

17/21/21/11 = 70

29/12/11/10 = 62

Region:

159

Southern F: 1/2/3-5/6-

17/14/14/5 = 50

19/14/12/4 = 49

21/12/10/5 = 48

25/12/11/5 = 53

23/16/11/5 = 55

Central F: 1/2/3-5/6-

14/13/13/2 = 42

19/9/9/1 = 38

15/14/9/2 = 40

23/10/11/1 = 45

18/11/3/1 = 33

Northern F: 1/2/3-5/6-

17/9/7/1= 34

17/8/9/2 = 36

12/5/7/4 = 28

14/9/6/0 = 29

14/9/2/1 = 26

160

Table 31: (Table 4.4) Eurobarometer 56.0: August-September 2001, n = 1024.

Visited, % Ballet/Dance

Theatre

Concert Concert:

classical

Concert:

opera

Concert:

rock/pop

Sport event

Women 17,8 49,5 41,2 30,4 15,4 27,5 30,4

Men 10,4 30,4 32,8 16,9 7,7 45,8 59,9

ageage: 15-24 14,8 36,3 50,8 12,6 2,1 68,8 57,8

age: 25-34 12,2 36,6 39,7 13,5 7,7 48,0 55,7

age: 35-44 21,6 41,7 43,2 25,7 13,5 41,9 45,6

age: 45-54 16,9 44,7 32,6 35,3 14,7 11,8 38,9

age: 55-64 13,2 47,8 34,8 29,2 20,8 6,3 40,0

age: 65- 8,3 41,2 25,5 44,7 25,5 0,0 23,0

Region:Uusimaa 17,5 43,6 36,9 24,7 18,0 37,0 46,0

Rest southern F

15,0 44,9 40,6 23,1 13,4 34,8 39,9

Eastern F 11,2 35,3 34,1 25,6 7,0 32,6 45,2

Central F 11,6 43,8 35,4 30,4 8,7 43,5 49,3

Northern F 12,2 28,0 34,3 27,0 5,4 16,2 37,4

n = 382

n = 379

n = 382

If Concert = 1, then

http://www.fsd.uta.fi/aineistot/luettelo/FSD0099/meF0099.html

161

Table 32: (Table 4.5) ISSP 2007, ”How often in your leisure do you go to concerts, exhibitions, theatre etc.?”

Daily Several times per week

Several times per month

Less often Never Missing

Frequency

0 4 71 1040 209 30

% 0 0,3 5,2 76,8 15,4 2,2

% of responses

0 0,3 5,4 78,5 15,8 --

Women, %

0 0,3 6,6 81,2 11,3 n = 741

Men, % 0 0,3 3,9 74,3 21,5 n = 568

Women, %

15 < age < 24

0 1,1 4,5 78,7 15,7 n = 89

25 < age < 34

0 0 1,9 85,5 15,5 n = 103

35 < age < 44

0 0 0,8 88,9 10,3 n = 117

45 < age < 54

0 0 6,7 85,2 8,1 n = 135

55 < age < 64

0 0 10,2 80,8 9,0 n = 167

65 < age 0 0,7 12,3 74,6 12,3 n = 130

Men, %

15 < age < 24

0 0 5,2 58,6 36,2 n = 58

25 < age < 34

0 1,3 2,7 78,7 17,3 n = 75

35 < age < 44

0 0 4,2 80,2 15,6 n = 96

162

45 < age < 54

0 0 3,5 76,3 20,2 n = 114

55 < age < 64

0 0,8 4,7 73,2 21,3 n = 127

65 < age 0 0 3,1 73,5 23,5 n = 98

163

Table 33: (Table 4.6) Visitor density, ANOVA (significance in parenthesis)

Grouping variable

F-value (sig.)

ή2 Grouping variable

F-value (sig.)

ή2

ANOVAGender (S) 26,218

(0,000)0,019 Year of birth(Y) 1,319

(0,055)0,062

Province (A) 2,624 (0,000)

0,037 Education (E) 10,175 (0,000)

0,060

MANOVAGender (S) 20,068

(0,000)Gender (S) 12,695

(0,000)

Year of birth (Y)

1,366 (0,036)

Province (A) 2,612 (0,000)

S*Y 1,025 (0,426)

0,130 S*A 0,663 (0,857)

0,064

Gender (S) 25,716 (0,000)

Year of birth (Y) 1,099 (0,291)

Education (E)

9,638 (0,000)

Province (A) 1,755 (0,022)

S*E 2,115 (0,032)

0,089 Y*A 1,029 (0,363)

0,523

Year of birth (Y)

1,655 (0,002)

Province (A) 1,188 (0,260)

Education (E)

9,394 (0,000)

Education (E) 3,594 (0,000)

E*Y 1,127 (0,102)

0,374 A*E 1,054 (0,328)

0,193

Gender (S) 7,485 Gender (S) 11,581

164

(0,006) (0,001)

Year of birth (Y)

1,295 (0,078)

Province (A) 1,429 (0,104)

Province (A) 1,569 (0,060)

Education (E) 3,329 (0,001)

S*Y 1,171 (0,193)

S*A 1,221 (0,231)

S*A 0,681 (0,813)

S*E 1,598 (0,121)

Y*A 1,123 (0,100)

A*E 1,144 (0,137)

S*Y*A 1,178 (0,126)

0,688 S*A*E 1,403 (0,010)

0,314

Gender (S) 15,496 (0,000)

Year of birth (Y) 1,764 (0,002)

Year of birth (Y)

1,510 (0,010)

Province (A) 1,775 (0,027)

Education (E)

8,314 (0,000)

Education (E) 4,735 (0,000)

S*Y 1,233 (0,120)

Y*A 1,162 (0,101)

S*E 1,575 (0,129)

Y*E 1,324 (0,018)

Y*E 1,226 (0,018)

A*E 1,174 (0,188)

S*Y*E 1,149 (0,152)

0,535 Y*A*E 0,975 (0,532)

0,861

Education: 1 = pupil, student, 2 = elementary school, 3 = comprehensive school, 4= vocational school or course, 5= upper secondary school, 6 = college 7= university of applied sciences, 8 = bachelor, university, 9 = master, university

165

Table 34: (Table 4.7) Visitor density, Anova and Manova, Women and Men separately

Grouping variable

F-value (sig.)

ή2 Grouping variable

F-value (sig.)

ή2

ANOVA Men ANOVA Women

Year of birth 1,007 (0,465)

0,112 Year of birth

1,454 (0,017)

0,116

Province 1,870 (0,014)

0,061 Province 1,202 (0,249)

0,031

Education 6,501 (0,000)

0,087 Education 5,186 (0,000)

0,055

MANOVA

Men Women

Year of birth (Y)

1,007 (0,473)

Year of birth (Y)

1,198 (0,169)

Province (A) 1,208 (0,203)

Province (A) 0,655 (0,861)

Y*A 1,238 (0,066)

0,746 Y*A 1,002 (0,495)

0,621

Year of birth (Y)

0,829 (0,807)

Year of birth (Y)

2,281 (0,000)

Education (E) 5,062 (0,000)

Education (E)

5,888 (0,000)

Y*E 1,035 (0,395)

0,527 Y*E 1,335 (0,006)

0,525

Province (A) 1,413 (0,116)

Province (A) 1,021 (0,434)

Education (E) 2,724 (0,006)

Education (E)

1,819 (0,071)

A*E 1,447 (0,006)

0,370 A*E 0,924 (0,699)

0,232

Year of birth 1,575 Year of 1,612

166

(Y) (0,051) birth (Y) (0,018)

Province (A) 2,209 (0,013)

Province (A) 0,756 (0,752)

Education (E) 3,817 (0,001)

Education (E)

3,111 (0,004)

Y*A 1,533 (0,044)

Y*A 1,098 (0,303)

Y*E 1,573 (0,051)

Y*E 1,324 (0,078)

A*E 0,944 (0,514)

A*E 0,927 (0,566)

Y*A*E -- 0,948 Y*A*E -- 0,885

167

Table 35: (Table 4.8) Multivariate logit analysis

n = 1269 ’less often, occasionally’ ’often’

β (s.e.) β (s.e.)

Gender 0,643 (0,157)*** 0,657 (0,253)**

school=2 -0,611 (0,375) -0,352 (0,654)

school=3 0,124 (0,384) 0,269 (0,698)

school=4 0,154 (0,315) 0,023 (0,616)

school=5 0,768 (0,390)* 1,471 (0,672)*

school=6 1,155 (0,364)** 0,716 (0,655)

school=7 1,035 (0,430)* 1,446 (0,740)(*)

school=8 1,744 (0,779)* 2,554 (0,977)**

school=9 2,131 (0,591)*** 2,894 (0,801)***

Age25_34 0,366 (0,334) -0,418 (0,660)

Age35_44 0,604 (0,316)(*) 0,647 (0,561)

Age45_54 0,743 (0,317)* 1,558 (0,534)**

Age55_64 0,882 (0,366)* 2,112 (0,577)***

Age65_ 0,479 (0,311) 0,284 (0,573)

constant -0,330 (0,353) -3,165 (0,686)***

LL = -794.9, AIC = 1.30, BIC = 1.42, HQIC = 1.35

Pseudo-R2 (McFadden) = 0,083

Depending variable =” How often in your leisure do you go to concerts, exhibitions, theatre etc.?”, Null (constant) is ’never’, education = pupil of student, age = age15_24. The province variables are not used. Gender: 1 = man, 2 = woman. ***,**,*,(*) significant at 0.1, 1, 5, 10 per cent level

168

Table 36: (Table 4.9) Multivariate logit analysis

n = 1269 ’less often, occasionally’

’often’

β (s.e.) β (s.e.)

Gender 0,645 (0,159)*** 0,645 (0,255)**

school=2 -0,575 (0,381) -0,362 (0,662)

school=3 0,084 (0,387) 0,240 (0,702)

school=4 0,199 (0,318) 0,044 (0,623)

school=5 0,703 (0,393)(*) 1,366 (0,675)*

school=6 1,164 (0,366)** 0,696 (0,660)

school=7 1,038 (0,434)* 1,408 (0,743)(*)

school=8 1,693 (0,784)* 2,482 (0,985)*

school=9 2,160 (0,594)*** 2,840 (0,805)***

Age25_34 0,366 (0,338) -0,392 (0,664)

Age35_44 0,617 (0,319)(*) 0,664 (0,567)

Age45_54 0,725 (0,319)* 1,563 (0,539)**

Age55_64 0,842 (0,254)* 2,115 (0,584)***

Age65_ 0,480 (0,314) 0,320 (0,578)

Uusimaa 0,842 (0,254)*** 1,045 (0,427)*

Rest southern F 0,672 (0,246)** 0,471 (0,448)

Eastern F 0,716 (0,294)* 1,142 (0,474)*

Western F 0,888 (0,260)*** 0,570 (0,464)

constant -0,976 (0,394)* -3,801 (0,754)***

LL = -784.4, AIC = 1.30, BIC = 1.45, HQIC = 1.35

Pseudo-R2 (McFadden) = 0,095

Depending variable =” How often in your leisure do you go to concerts, exhibitions, theatre etc.?”, Null (constant) is ’never’, education = pupil of student, age = age15_24, province = Norhern Finland - Regions: Rest souther Finland = Itä-Uusimaa, Varsinais-Suomi, Kanta-Häme, Päijät-Häme, Kymenlaakso, Etelä-Karjala, Ahvenanmaa. Eastern Finland = Etelä-Savo, Pohjois-Savo, Pohjois-Karjala, Kainuu. Western Finland = Satakunta, Pirkanmaa, Keski-Suomi, Etelä-Pohjanmaa, Pohjanmaa. Norhern Finland = Keski-Pohjanmaa, Pohjois-Pohjanmaa, Lappi, i.e. NUTS2-regions except. Ahvenanmaa and Uusimaa. Gender: 1 = man, 2 = woman

169

170

Table 37: (Table 4.10) ISSP 2007, ”How often on your leisure do you go to see sport events on the location (ice hockey, football, athletics, motor racing etc.)? n = 1355

Daily Several times per week

Several times per month

Less often Never Missing

Frequency

4 17 82 691 526 35

% 0,3 1,3 6,1 51,0 38,8 2,6

% of responses

0,3 1,3 6,2 52,3 39,8

Women, %

0,1 0,7 5,0 43,4 49,0 1,7

Men, % 0,5 2,1 7,6 63,3 26,5 3,4

Women, %

15 < age < 24

0 0 9,1 48,5 42,4 0

25 < age < 34

0 3,1 6,9 46,6 43,5 0

35 < age < 44

0 0,8 6,1 42,7 49,6 0,8

45 < age < 54

0,6 0 2,4 40,5 54,8 1,8

55 < age < 64

0 0 1,9 28,7 63,0 6,5

65 < age 0 0 5,2 54,3 38,8 1,7

Men, %

15 < age < 24

0 1,6 8,1 67,7 21,0 1,6

25 < age < 34

0 4,2 9,4 69,8 16,7 0

35 < age < 44

0,8 0,8 9,2 68,9 17,6 2,5

171

45 < age < 54

0 0,8 6,8 56,4 30,8 5,3

55 < age < 64

1,5 1,5 1,5 45,5 43,9 6,1

65 < age 0,9 3,6 7,3 56,4 27,3 4,5

Table 38: (Table 4.11) Multivariate logit analysis

n = 1269 ’less often, occasionally’

’often’

β (s.e.) β (s.e.)

Culture* 0,747 (0,145)*** 1,361 (0,234)***

Gender -0,999 (0,130)*** -1,273 (0,204)***

school=2 0,162 (0,350) 1,007 (0,564)(*)

school=3 0,628 (0,350)(*) 1,250 (0,553)*

school=4 0,489 (0,292)(*) 0,321 (0,505)

school=5 0,377 (0,335) 0,237 (0,568)

school=6 0,743 (0,301)* 0,807 (0,510)

school=7 0,862 (0,351)* 0,622 (0,587)

school=8 0,356 (0,410) -0,159 (0,779)

school=9 0,569 (0,336)(*) -0,250 (0,627)

Age25_34 -0,415 (0,280) -0,091 (0,453)

Age35_44 -0,476 (0,269)(*) -0,500 (0,440)

Age45_54 -0,797 (0,268)** -1,188 (0,440)**

Age55_64 -1,288 (0,302)*** -1,463 (0,507)**

Age65_ -0,536 (0,271)* -0,378 (0,440)

Uusimaa 0,183 (0,205) -0,199 (0,352)

Rest southern F 0,342 (0,211) 0,284 (0,346)

Eastern F 0,019 (0,246) 0,559 (0,367)

Western F 0,463 (0,213)* 0,256 (0,354)

constant 1,030 (0,351)** -0,723 (0,599)

LL = -1106.6, AIC = 1.81, BIC = 1.97, HQIC = 1.87

Pseudo-R2 (McFadden) = 0,078

172

Depending variable =”How often on your leisure do you go to see sport events on the location (ice hockey, football, athletics, motor racing etc.)? Null (constant) is: education = pupil or student, age = age15_24, region = Northern Finland. Culture* = previously defined culture participation (0=’never’, 1=’occasionally’ and 2=’often’), Gender: 1 = man, 2 = woman.

173

Table 39: (Table 4.12) Bivariate probit analysis,

Cult2 Cult2: direct marginal effect

Cult2: indirect marginal effect

Sport2

Gender -0,048 (0,128) -0,005 (0,013) -0,006

(0,002)*

0,293

(0,106)**

school2 -0,077 (0,308) -0,008 (0,032) 0,008 (0,006) -0,428 (0,287)

school3 -0,095 (0,317) -0,010 (0,033) 0,009 (0,006) -0,463 (0,278)

(*)

school4 0,006 (0,287) 0,001 (0,030) 0,000 (0,005) -0,026 (0,252)

school5 -0,440 (0,307) -0,045 (0,031) 0,002 (0,005) -0,122 (0,279)

school6 0,083 (0,283) 0,009 (0,029) 0,005 (0,005) -0,243 (0,261)

school7 -0,305 (0,320) -0,031 (0,033) 0,003 (0,005) -0,135 (0,280)

school8 -0,543 (0,353) -0,056 (0,036) 0,000 (0,007) -0,004 (0,373)

school9 -0,480 (0,310) -0,049 (0,032) -0,003 (0,006) 0,148 (0,304)

age25-34 0,288 (0,304) 0,030 (0,031) 0,002 (0,004) -0,098 (0,223)

age35-44 -0,044 (0,236) -0,005 (0,024) -0,001 (0,004) 0,056 (0,219)

age45-54 -0,412

(0,229) (*)

-0,042

(0,024)(*)

-0,005 (0,005) 0,269 (0,228)

age55-64 -0,671

(0,250)**

-0,069

(0,026)**

-0,005 (0,005) 0,254 (0,252)

age65- 0,089 (0,246) 0,009 (0,025) 0,000 (0,004) -0,012 (0,228)

Uusimaa -0,216 (0,197) -0,022 (0,020) -0,002 (0,003) 0,088 (0,174)

Rest southern

F

-0,004 (0,206) -0,000 (0,021) 0,002 (0,003) -0,088 (0,174)

Eastern F -0,267 (0,227) -0,027 (0,023) 0,007 (0,004) -0,359

174

(*) (0,189)(*)

Western F 0,059 (0,216) 0,006 (0,022) 0,001 (0,003) -0,044 (0,176)

constant 1,922

(0,375)***

0,976

(0,273)***

ρ = 0,463

(0,086)***

(standard errors in parenthesis). Cult2: ’0 = daily, several times per week or several times per month’ and ’1 = less often or never’, Sport2 classified in the same way.Log Likelihood = - 701,7, AIC = 1,167, BIC = 1,326, HQIC = 1,227, (*), *, **, *** = significant at 10,5,1,0.1 %

175

Table 40: (Table 4.13) Bivariate probit analysis,

Cul2 Cul2: direct marginal effect

Cul2: indirect marginal effect

Sport2

Gender -0,351

(0,083)***

-0,107

(0,026)***

-0,044

(0,009)***

0,568

(0,067)***

school2 0,167 (0,194) 0,051 (0,059) 0,010 (0,015) -0,127 (0,188)

school3 -0,120 (0,204) -0,037 (0,062) 0,031 (0,016)

(*)

-0,398

(0,191)*

school4 -0,238 (0,173) -0,073 (0,053) 0,022 (0,013)

(*)

-0,282

(0,159)(*)

school5 -0,505

(0,206)*

-0,154

(0,063)*

0,021 (0,015) -0,266 (0,187)

school6 -0,705

(0,184)***

-0,215

(0,057)***

0,040

(0,015)**

-0,508

(0,165)**

school7 -0,661

(0,029)**

-0,202

(0,064)**

0,049

(0,017)**

-0,631

(0,192)***

school8 -1,064

(0,373)**

-0,325

(0,113)**

0,027 (0,018) -0,352 (0,226)

school9 -1,205

(0,270)***

-0,368

(0,084)***

0,028 (0,015)

(*)

-0,358

(0,186)(*)

age25-34 -0,105 (0,177) -0,032 (0,054) -0,019 (0,013) 0,248 (0,159)

age35-44 -0,324

(0,173)(*)

-0,099

(0,053)(*)

-0,020 (0,12)

(*)

0,260 (0,152)

(*)

age45-54 -0,363

(0,169)*

-0,111

(0,051)*

-0,038

(0,014)**

0,484

(0,151)**

age55-64 -0,541 -0,165 -0,057 0,729

176

(0,194)** (0,059)** (0,017)*** (0,165)***

age65- -0,214 (0,160) -0,065 (0,049) -0,026

(0,013)*

0,339 (0,149)*

Uusimaa -0,488

(0,136)***

-0,149

(0,042)***

0,012 (0,009) -0,154 (0,112)

Rest southern

F

-0,349

(0,133)**

-0,107

(0,041)**

0,019 (0,010)

(*)

-0,240

(0,115)*

Eastern F -0,364

(0,155)*

-0,111

(0,048)*

0,012 (0,011) -0,160 (0,134)

Western F -0,425

(0,143)**

-0,130

(0,044)**

0,024 (0,010)* -0,312

(0,117)*

constant 0,482

(0,207)**

-1,006

(0,188)***

ρ = 0,382

(0,050)***

(standard errors in parenthesis). Cul2: ’0 = daily, several times per week or several times per month or less often’ and ’1 = never’, Sport2 classified in the same way.Log Likelihood = - 1404,9 AIC = 2,038 BIC = 2,183 HQIC = 2,092, (*), *, **, ***

=significant at 10,5,1,0.1 %

177

5. Are the spectators of performing arts and the spectators of movies the same?

Seppo SuominenHaaga-Helia University of Applied SciencesMalmi campus, Hietakummuntie 1 A, FIN-00700 Helsinki, Finlande-mail: seppo.suominen@haaga-helia.fi

5.1 Introduction

The purpose of this paper is to study performing arts consumption and movies

at the cinema consumption using the ISSP 2007 survey data. A number of

different socioeconomic variables are used to explain cultural consumption.

The bivariate probit approach to studying performing arts and movies at the

cinema consumption together is useful because it reveals substantially new

evidence on the average profile of culture consumption. It is expected that

females go more often to art exhibitions, opera or theatrical performances and

this was supported. The results of the bivariate probit analysis also reveal that

gender is important also in explaining movie attendance. Females go more

often to see movies at the cinema. There is a significantly positive correlation

between these two audiences indicating that there is a common background

between both groups. The approach also allows finding the other relevant

socioeconomic characteristics explaining cultural consumption. However,

bivariate probit approach classifies consumption into two categories: yes or

no. Roughly 5 percent of the consumers in the sample could be classified as

heavy users, and another approach must be used to study three groups:

178

heavy, occasional and not at all. A multivariate logit analysis is one approach

to classify these groups. Using both bivariate probit analysis and multivariate

logit analysis results in new evidence in cultural consumption. It is widely

known that gender, age and educational level are significant variables to

explain cultural consumption. It is shown with the Finnish ISSP 2007 data that

besides these variables, also educational level of the spouse and the number

of children significantly classify cultural consumption. Naturally the place of

residence matters since in Southern and Western Finland the residential

density is higher and there are more cultural institutions than elsewhere in

Finland. There is only one permanent opera in Helsinki but also some opera

associations that are more provisory and have performances outside Helsinki.

The theatre institutions are located mostly in bigger cities, but the number of

traveling theatre groups makes is possible for citizens in the countryside to

go and see performing arts.

Recently roughly 60 percent of the adult population (age between 15-79) in

Finland have seen a movie at the cinema during the last year (Kotimaisen

elokuvan yleisöt – tutkimus 2010). Ten percent of the adult population are

heavy users: they go to the cinema one to three times a month. More than a

fourth of young audience (age: 15-24) are heavy users. However, the results of

that survey may be misleading since the interviews were made during January

– February 2010 and it is well known that the Christmas season is the prime

time. Another recent study (ISSP 2007)11 reveals that only 1.9 % of the 11 International Social Survey Programme 2007, sample size 2500 with 1354 valid results, respondents age between 15-74, interviews made between 18th September – 11th December 2007

179

population are heavy users and 17.9 % have not seen a movie at the cinema

during the last year. The figure is comparable with the spectator number of

performing arts (concert, theatrical performance, art exhibition) where the

corresponding numbers are: 5.9 % are heavy users and 15.7 % have not been

at all. A third recent survey (2006)12 claims that 3 % are heavy users and 45 %

have not seen a movie at the cinema at all during the last year. This survey is

based on interviews made during March – June 2006. The figures in the

European Cultural Values study are somewhat different as seen in table 1.

Hence, it seems that the timing of the interviews has a big impact on the

results.

(Table 1 about here)

Studies related to the spectators of performing arts are rather common:

females go more often and the audience is composed of middle-aged people

with high educational and income levels (Baumol and Bowen 1966, Liikkanen

1996, Kracman 1996, Bihagen and Katz-Gerro 2000, Borgonovi 2004, Seaman

2005, Montgomery and Robinson 2006, Vander Stichele and Laermans 2006).

Spectators of movies in cinemas are usually young students but there are no

gender differences (Austin 1986 or F & L Research 1999). Recently Redondo

and Holbrook (2010) showed that the family-audience profile (i.e. middle-aged

with children) and the family-movie profile (various genres) are strongly

associated, while young men seem to favour action, mystery, thriller and

violence genres. It is also known that young males prefer action and 12 Adult education survey 2006, sample size 6800 with 4370 valid results, respondents age between 18 and 64, interviews made between May – June 2006

180

excitement on the screen and women tend to favour emotional dramas

(Kramer 1998). Typically the ticket price is substantially higher for performing

arts than for movies, and this may explain the difference between age-groups:

e.g. in 2009 the average movie ticket price was € 8.3in Finland and the ticket

revenue per spectator was € 32.62 in the Finnish national opera. In 2009 the

average ticket price in Finnish big- and medium-size theatres was €16.21 in

top 30 theatres13. The performing arts are heavily subsidized by the state

(ministry of education: state aid) and municipalities since the share of the

ticket revenues was only 15 % for the Finnish national opera and 20 % for the

top 30 theatres14.

13 According to ticket revenue in top 10 theatres the unweighted mean was €19.48, in the next 10 (11th – 20th) €14.07 and the next 10 (21st – 30th) €15.10. The weighted average price in big- and medium-size theatres was €18.63

14 These numbers have remained fairly stable recently: e.g.in 2007 the average movie ticket price at the cinema was €7.8, in the Finnish national opera €33.19 and €17.63 in big- and medium-size theaters.

181

In 2007 there were 316 cinema screens in Finland15. The number of films in

spreading was 410 and there were 163 premieres. The total number of

spectators was 6.5 Million (i.e. 1.2 per capita). Correspondingly, there were

46 drama theatres subsidized by law with 12,361 performances and 2,446,500

spectators, 16 summer theatres with 821 performances and 35,1473

spectators, and 51 theatre groups outsize the law subsidies with 4,139

performances and 465997 spectators. Overall this means 103 theatres and

16,695 performances and 3,066,530 spectators, i.e. 184 spectators per

performance or 0.57 per capita. Moreover, the Finnish national opera16 and

other operas (13 local operas with only few performances) had 285

performances with 182,728 spectators (641 per performance). Furthermore,

39 dance theatres (including the National Ballet) gave 2,377 performances

with 523,620 spectators (220 per performance).17 The total number of 15 Top towns based on admissions 2007. Source: The Finnish Film Foundation, Facts & Figures 2008. www.ses.fi

Town Admissions Admissions/capita Screens Seats Seats/Screens

Helsinki 2188094 3.84 37 7327 198.02

Tampere 667205 3.21 17 2663 156.64

Turku 542398 3.09 17 2602 153.06

Oulu 329533 2.50 13 1762 135.54

Jyväskylä 237075 2.77 8 1097 137.13

Lahti 174234 1.75 7 1036 148.00

Espoo 170354 0.71 5 825 165.00

Pori 169364 2.22 6 748 124.67

Kuopio 123835 1.35 6 1191 198.50

Joensuu 123515 2.14 4 659 164.75

16 The main stage of The Finnish National Opera was closed 6 months in 2007 due to renovation.

17 16 Circus companies had 804 performances with 279544 spectators.

182

different plays performed in the drama theatres during the season 2006 –

2007 was 357 and there were 118 premieres. A large majority (203/357) of the

plays were written by a Finnish writer (e.g. Saisio, Nopola, Wuolijoki,

Krogerus.). English (e.g. Shakespeare, Pownall, Russell), American (e.g.

Woolverton, Quilter, Williams), Swedish (e.g. Nordqvist, Lindgren), French

(e.g. Duras, Molière) and Russian (e.g. Gogol, Tshehov) plays were the most

performed foreign ones. Practically all dance theatre performances except The

Finnish National Ballet were of domestic origin, whereas the ballet and opera

plays were mostly of foreign origin. In top 10 towns according to the movie

spectator number, the admissions per capita for movies and drama theatre

performances18 are highly correlated (0.81), hence the supply conditions for

both cultural events are fairly equal. Urban citizens have better access both to

the cinema and to the theatres and concerts than people living in the rural

areas.

18 Top towns based on movie admissions 2007: * some smaller drama theatres regularly made tours

Town (m. adm.) Theatre Adm. Admissions/capita Drama theatres* Performances Adm/Perform

Helsinki (2188094) 753233 1.33 10 3132 240.50

Tampere (667205) 325335 1.58 5 902 360.68

Turku (542398) 166442 0.95 3 991 167.95

Oulu (329533) 66725 0.51 1 351 190.10

Jyväskylä (237075) 110251 1.30 2 443 248.87

Lahti (174234) 84498 0.86 1 266 317.66

Espoo (170354) 57444 0.24 3 516 111.33

Pori (169364) 48450 0.64 1 282 171.81

Kuopio (123835) 49871 0.55 1 322 154.88

Joensuu (123515) 30986 0.54 1 280 110.66

183

However, it is not known whether the spectators of movies and performing

arts are the same. Especially middle-aged high-income highly educated

women seem to favour performing arts. Are they also movie lovers? A

bivariate probit model is a nice method to study this question since the model

enables to evaluate the marginal effects, both direct and indirect. In the table

1 four recent surveys have been compared. The International Social Survey

Programme (ISSP 2007) study is most useful since the variables in that study

are suitable.

5.2 Literature review

This paper is closely related to the sociological literature of performing arts

participation. The classical work is Bourdieu (1979). The relation of social

positions to cultural tastes and practices is structurally invariable. There are

two interrelated spaces: the space of social space (positions) and the space of

lifestyles. The social space has three dimensions: economic, social and cultural

capital. Bourdieu argued that there is a structural correspondence between

social space and cultural practices and the habitus serves as a mediating

mechanism. Therefore the tastes, knowledge and practices are class-based.

The “highbrow” cultural consumption is typical for the dominant classes.

Bourdieu argued that cultural capital or social statuses are symptoms of social

exclusion, cultural dominance and inequality. Bourdieu’s claims have been

criticized substantially, since the taste of the dominant class has lost its

exclusiveness (Purhonen, Gronow and Rahkonen 2010). The dominant class

184

has changed its cultural participation pattern. They are more omnivore. The

audience segmentation has changed from elite and mass to omnivore and

univore (Peterson 1992, Peterson and Kern 1996). In the European context

Finland and the Nordic countries in general are the leading countries in the

proportion of omnivores in the population (Virtanen 2007).

The omnivorousness in cultural taste has been measured according to the

number of cultural participation areas and/or by number of genres in one

specific area. A person is omnivore if she has seen a ballet, a theatre

performance, a movie at the cinema, reads books, goes to a sport event and so

on. Correspondingly, she is univore If she prefers e.g. only sport events and is

active in that field but not in the other areas of culture (Sintas and Álvarez

2004, Chan and Goldthorpe 2005). On the other hand, she is omnivore if she

reads books of different genres: thrillers, scifi, fantasy, romances, biographies,

modern literature, classical literature, poetry, plays, religious books, leisure

books (Purhonen, Gronow and Rahkonen 2010). Omnivores have a high

probability of participating in everything, from the unpopular (e.g. classical

music) to the popular (e.g. cinema attendance), whereas paucivores engage in

intermediate levels of cultural consumption across a range of activities, and

inactives have a low probability of participating in any of the activities

(Alderson, Junisbai and Heacock 2007). Omnivores have usually higher levels

of education and higher incomes than univores (Chan and Goldthorpe 2005).

Using a multinomial logit analysis, Alderson, Junisbai and Heacock (2007)

show that social status, having a bachelor’s degree and family incomes

significantly classify inactive and the two other groups (omnivore and

paucivore), while having a graduate degree classifies omnivores and the other

185

groups (paucivore and inanctive). Age is important to categorize paucivore

from omnivore and inactive. Unexpectedly gender is not a significant variable

to classify. The omnivore consumption pattern is typical among the upper

social classes, univore among the upper-middle and middle classes and

fragmental consumption among the lower social classes (Sintas and Álvarez

2004).

The sociology of cultural participation has shown than consumers can be

classified into three groups: omnivore, paucivore and inactive (Alderson,

Junisbai and Heacock 2007). The omnivore group is active in all cultural

consumption, from cinema to classical music. The concept of cultural capital is

associated with the lowbrow/highbrow consumption styles. Arts consumption

is a form of cultural capital (DiMaggio 1987). Cultural capital is the

accumulated amount of past consumption of cultural goods and the initial

endowment of cultural capital (Stigler and Becker 1977). The accumulation

function is related to human capital, i.e. formal education. The human capital

argument is based on the idea that cultural behavior is constrained somehow,

i.e. differences in cultural consumption are related to differences in cultural

capital endowments, differences in budget, time, social and physical

constraints (Frey 2000). Since cultural capital endowment is related both to

formal education and age, these are proper explanatory variables. Moreover,

it has been shown that gender and marital status are important to explain

cultural consumption. Time constraints are related to place of residence

(province) and finally budget constraints are measured by incomes (c.f.

186

Ateca-Amestoy 2008). However, there is some evidence showing that

economic wealth (net incomes, material wealth) is not a significant variable

explaining cultural participation (c.f. Vander Stichele and Laermans 2006).

Alderson, Junisbai and Heacock (2007) argue that gender is not a significant

variable to classify cultural consumption pattern classes (in the USA, 2002),

but Bihagen and Katz-Gerro (2000) show with Swedish data (1993) that

gender is important. Women are more active in highbrow consumption (opera,

dance performance, theatrical performance) and men in lowbrow television

(entertainment, sport) watching. Highbrow television (documentary, culture,

news) and lowbrow culture (movies, rent a video) are less connected to

gender, class and education, but these are strongly related to age. Younger

seem to favour lowbrow culture and older highbrow television watching.

Lizardo (2006) shows using cluster analysis with pooled data from the 1998

and 2002 United States General Social Survey that four genres fall on to the

highbrow cluster: arts consumption, going to the ballet, going to a theater and

attending a classical music or opera concert. The lowbrow cluster consists of

going to a popular music live concert, going to see a movie in cinema or

reading a novel, poem or play. Gender matters but only with those that are

active in the labour force. Among those that are not active in labour force,

there is no gender difference in highbrow cultural consumption. Purhonen,

Gronow and Rahkonen (2009) present similar results with Finnish data. Warde

and Gayo-Cal (2009) find also mixed evidence concerning the gender effect on

omnivorousness with British 2002-2003 survey data. Women seem to be more

187

active in ‘legitimate’ culture. Different terminologies have been used to rank

tastes, like: highbrow – middlebrow – lowbrow, or high – popular, or legitimate

– vulgar. Bourdieu defines legitimate as being connected with dominant

classes, powerful social groups and being aesthetically the most valuable. The

top quartile omnivores are associated with legitimate taste, while the lowest

quartile in omnivorousness is least related with legitimate cultural

consumption. Omnivorousness increases with age up to around 50 and

strongly diminishes among those over 70 (Warde and Gayo-Cal 2009). Family

background as a whole matters, since parents’ cultural participation seems to

be related with cultural consumption (van Eijck 1997), while participating in a

culturally orientated course at school does not have any or only slight impact

on cultural consumption (Nagel, Damen and Haanstra 2010).

5.3 The method and sample

The ISSP 2007 survey was carried out between 18th September and 11th

December 2007 by means of a mail questionnaire in Finland. The ISSP is a

continuous programme of cross-national collaboration on social science

surveys. The surveys are internationally integrated. In Finland the ISSP

surveys are carried out together by three institutions: Finnish Social Science

Data Archive, The Department of Social Research at the University of

Tampere and the Interview and Survey Services of Statistics Finland19. The

19 http://www.fsd.uta.fi/english/data/catalogue/series.html#issp, cited 24.9.2010. The observation unit is a person 15-74 – years old, the sampling method is a systematic random sample from the population register, the sample size was 2500 but the 1354 answers were obtained, in other words response rate was 54.2%. The index terms of ISSP 2007 are: use of

188

other surveys mentioned in table 1 did not collect e.g. marital status, which

has been shown to have an impact on the attendance of cultural events

(Upright 2004, Frateschi & Lazzaro 2008).

The cultural participation questions in The ISSP survey were: “How many

times in the last twelve months have you seen an art exhibition, opera or

theatrical performance?” Or “How many times in the last twelve months have

you been to the cinema?” Five alternatives were given: ‘ Every day’, ‘Several

times a week’, ‘Several times a month’, ‘Less often’ or ‘Never in the last

twelve months’. A conventional method to study this is to use some discrete

choice model, like probit or logit. A Poisson model is more suitable to study

count data, which is not the case here. The normal distribution for the binary

choice (no = 0 / yes = 1) has been used frequently generating the probit

model.

Prob (Y=1|x )=∫−∞

x ' β

∅ ( t )dt=Φ(x ' β )

The functionΦ (x' β) is the commonly used notation for the standard normal

distribution (Greene 2008, 773) and x is a vector of explanatory variables and

β is the corresponding vector of parameters. The logistic distribution which is

mathematically convenient has been very popular.

time, physical condition, hobbies, organisations, board games, physical education, holiday, games, social relations, sports, leisure. Among others, gender, year of birth, size of the household, education, participation in the working life, profession, source of livelihood or branch, regular weekly working hours, professional station, employer (the private/public sector), the membership of the trade union, voting behaviour, religiousness, incomes and residential were collected as background information.

189

Prob (Y=1|x )= ex ' β

1+ex' β=Λ (x ' β )

The function Λ( x' β) is the logistic cumulative distribution function. If the

responses are coded 0,1,2,3 or 4 (‘ Every day’, ‘Several times a week’, ‘Several

times a month’, ‘Less often’ or ‘Never in the last twelve months’) the ordered

probit or logit models have been very common. The models begin with y* = x’β

+ ε in which y* is unobserved and ε is random error. The discrete choices y

are observed by the following way:

y = 0, if y* ≤ 0

y = 1, if 0 < y* ≤ µ1

y = 2, if µ1 < y* ≤ µ2

y = 3, if µ2 < y* ≤ µ3

y = 4, if µ3 ≤ y*

The µ’s are unknown parameters to be estimated with β. If ε is normally

distributed with zero mean and variance equal to one [ε~N(0,1)], the

following probabilities ensue (Greene 2008, 831-832):

Prob ( y=0|x )=Φ (−x ' β)

Prob ( y=1|x )=Φ (μ1−x' β)−Φ (−x ' β )

Prob ( y=2|x )=Φ (μ2−x' β )−Φ (μ1−x' β )

Prob ( y=3|x )=Φ (μ3−x ' β )−Φ(μ2−x' β)

Prob ( y=3|x )=1−Φ (μ3−x' β)

The parameters of the multivariate probit model, β’s, are not necessarily the

marginal effects that describe the effects of the explanatory variables on

cultural participation since the model is not linear. The multivariate probit

190

model is useful to evaluate the cultural participation and influences of

different explanatory variables. However, it is widely known that the

categories “every day” or “several times a week” or “several times a month”

get a small number of respondents and it is reasonable to combine these

categories with “less often” (e.g. Vander Stichele and Laermans 2006). One

step further is to assume that the error terms of two explanatory models are

correlated. One model is estimated for highbrow (ballet, dance performance,

opera) and another for cinema (lowbrow). If the disturbances are correlated,

both the direct marginal effects and the indirect marginal effects can be

evaluated. With this method the omnivore group of people can be found. The

general specification for a two-equation model assuming binary choice is then

(Greene 2008, 817)

y1¿=x1

' β1+ε 1 , then y1=1 if y1¿>0 ,∧ y1=0otherwise

y2¿=x2

' β2+ε2, then y2=1 if y2¿>0 ,∧ y2=0otherwise

E [ε 1|x1 , x2 ]=E [ ε2|x1 , x2 ]=0

Var [ε1|x1 , x2 ]=Var [ε2|x1 , x2 ]=1Cov [ ε1 , ε2|x1 , x2 ]=ρ

If ρ equals zero, the two spectator groups are independent, and two

independent probit models could be estimated and it could be claimed that the

highbrow attenders are different from cinema attenders (Prieto-Rodríguez and

Fernández-Blanco 2000).

191

Naturally, consumption depends on the ticket price, but since data available

does not include price variable, it is not considered here.

The cultural consumption y* thus depends on the following variables:

y* = f(education, age, gender, marital status, province, incomes)

Since it has been shown that middle-aged are among the most active in

highbrow cultural consumption, a suitable method is to classify age into age

groups. The observation unit in the ISSP 2007 survey is a person 15-74 –years

old and for the purpose of this study persons have been classified into 12

subsets: 15-19 –years old, 20-24 –years old, and so on with the last consisting

persons of 70-74 –years old.

(table 2 about here)

Descriptive statistics of the explanatory variables reveal that age (age group)

and education are related. Most of the youngest in the sample were pupils or

students (at a comprehensive, an upper secondary, a vocational school, of

course, or at a college) and correspondingly the oldest had a rather low

education (elementary or comprehensive school). A college level education

was mainly replaced by bachelor’s degree education in the early 1990’s and,

therefore, persons having a bachelor’s degree from a polytechnic (university

of applied sciences) are somewhat younger than persons having a college

diploma. Persons less than 50 –years old on average have a (better and)

longer education than persons older than 50. Age and education are related 192

with household or personal incomes. Middle-aged and high-educated seem to

have the highest incomes (including all social security contributions, e.g. child

benefit that may explain why the age group 30-34 has the highest incomes,

see table 3). There are some differences in education between genders. Men

are somewhat less educated than women.

(table 3 about here)

Since the income variable in the sample includes all social security

contributions (e.g. child benefit), the number of children is used as an

explanatory variable. There are two different variables: the number of less

than 6-year-old children and the number of 7-17-year-old children. This leads

to the following relation explaining cultural consumption. Since the number of

children is considered as explanatory variable, the marital status is also

added.

y* = f(education, age, gender, marital status, province, incomes,

number of children)

Since the cultural participation variables are recoded conversely into binary

variables: Art-consumption01234 (‘every day’ = 0, ’several times a week’ = 1,

‘several times a month’ = 2, ‘less often’ = 3, ‘never in the last 12 months’ = 4)

art1234_5 (‘no’ = 0, ‘yes’ = 1), some information is lost. However, the

correlation of the original and the recoded variables is high: r = -0.937.

Respectively the correlation of the original movie consumption variable and

193

the recoded variable is also high: r = -0.844. The correlation of the recoded

art participation and the movie consumption variables is positive: r = 0.397.

Therefore, there are good arguments to study these sectors of culture jointly.

In the sample there are more females (57%) than males (43%). Most are

married (50%) and the two other large groups according to the marital status

are: single (20%) and common-law marriage (17%). Separated or widowed are

considered as the reference group (constant) in further analysis as well as

Northern Finland and Ahvenanmaa.

(table 4 about here)

Figure 4: Nuts areas

194

Area1 = FI18 (Southern Finland), Area2 = FI19 (Western Finland), Area3 = FI13 (Eastern Finland), FI1A (Northern Finland) + FI20 (Ahvenanmaa –the South-Western archipelago) are considered as reference value.

5.4 Results

Table 5 presents the results of bivariate probit analysis when age group 50-54

and elementary school (edu2) are considered as reference value (i.e. the

constant in the equation). The two spectator groups are not independent since

ρ = 0.625. Hence the hypothesis that spectators of movies and arts belong to

independent groups can be rejected. There are common characteristics, a

common background which could be called as an intrinsic culture orientation.

If a person likes art exhibitions, opera and theatrical performances, she also

likes to see movies at cinema. Those that are inactive and culture orientated

195

do not go to exhibitions or performances and to the cinema. However, there

are some particular effects that are related with exhibitions and performances

or with movies.

The importance of gender is very strong: females are more active in both arts

(highbrow) and movies. The direct marginal effect of gender (female) is

positive, but the indirect marginal effect is negative. Both the direct and

indirect marginal effects have been reported only for the highbrow art (art

exhibition, opera and theatrical performances. The negative indirect effect

describes the preference of seeing a film in the cinema. These leisure time

activities are to some extent substitutes. Marital status matters: married or

common-law married citizens go more often to see highbrow art. The

coefficient of common-law marriage in the probit equation for the movies is

negative, indicating that they prefer more highbrow arts than movies in the

cinema. Unmarried or single citizens on the contrary go to the cinema more

often. The gender effect on art consumption found here is in line with the

results of Bihagen and Katz-Gerro (2000). Females are more omnivore

compared with males who are more paucivore.

(table 5 about here)

The effect of age on cultural consumption in table 5 is relative to the age

group 50-54. All younger cohorts prefer movies and only the oldest (70-74)

seem to go less often to the cinema than the reference group. The indirect

marginal effect of age on highbrow art is negative for each younger age

group. The direct marginal effects of cohorts are not significant. The results

196

indicate that age is not a relevant variable to classify highbrow art

consumption into active and inactive groups. Education seems to be very

important to classify culture consumption structures. When the reference

level is elementary school (edu2), citizens with any other education level are

significantly more active in culture consumption, in both directions: highbrow

art and movies. The highest marginal (direct + indirect) effect is for those that

have the best education (edu9 = master’s degree): 0.160 = 0.195 – 0.035.

However, those with a bachelor’s degree (edu8) have the largest direct and

largest (negative) indirect effect: 0.150 = 0.250 – 0.100. They seem to be the

most omnivore group. They are most active in highbrow art consumption as

well as in movies at the cinema consumption. Consumers with college level

education (edu6) are third most active group. The results confirm the well-

known hypothesis that omnivores have higher levels of education (Chan and

Goldthorpe 2005, Alderson, Junisbai and Heacock 2007). Spouse’s education

in some cases is relevant to explain consumption of movies at the cinema. If

the spouse has either a master’s degree or upper secondary diploma, the

person is more active to go to the cinema and that indirectly reduces

highbrow art participation.

The effect of domicile on culture consumption is selective. In Southern and

Western Finland (Area1 or Area2) people are more active in both highbrow art

and movies at the cinema consumption. In Eastern Finland (Area3) people are

less active in highbrow art consumption but significantly more active in movie

attendance than in Northern Finland or in the Ahvenanmaa archipelago

197

(reference areas). Household’s size matters only indirectly to highbrow art

consumption since bigger families seem to favour movies. The number of

small children (less than 7) or older children (7-17) significantly reduces both

culture consumption segments. Household incomes (or personal incomes – not

reported here) are not significant.

Table 6 presents the results of bivariate probit model explaining

simultaneously highbrow art (art1234_5) and movie (mov1234_5) consumption

when age-cohort 40-45 and upper secondary school level education (edu5) are

the reference values. The gender effect is similar than in table 5. Females are

more omnivore. Married citizens are more active in highbrow art

consumption, but the dummy variable ‘common-law marriage’ is not

significant. In the previous estimation (table 5) when the reference was age-

cohort 50-54 and elementary school education that variable was significant.

This indicates that the effect of ‘common-law marriage’ is related to age or

educational level. Age-cohorts younger than 40 are significantly more active in

movie consumption but there are no differences in highbrow art consumption.

Elementary education (edu2) in relation to upper secondary (graduate) level

education (edu5) lowers significantly both highbrow art and movies at the

cinema consumption. College (edu6) or bachelor’s degree (edu8) educated are

more active in highbrow art participation. The results in tables 5 and 6

indicate that the most omnivore citizens are those with a bachelor’s degree.

They go to see art exhibition, opera or theatrical performances and also

198

movies at the cinema. Spouse’s education in relation to upper secondary level

education is significantly lowering cinema activity if the spouse has either

elementary school (edu2) or vocational school or course (edu4) education. The

effects of the area as well as the family size or the number of children are

similar in table 6 and in table 5.

The marginal effects of age-cohorts in tables 5 and 6 are different since the

reference value is different: the age-cohort 50-54 in table 5 and the age-cohort

40-44 in table 6, but only the level is different. Otherwise they reveal the same

information. In figure 1 there are direct (DirME5 and DirME6) and indirect

(IndME5 and IndME6) marginal effects of age-cohorts in tables 5 and 6. The

values are highly correlated: ρDirME5, DirME6, age = 0.958, ρIndME5, IndME6,age = 0.981.

The direct and indirect marginal effects of age-cohorts are not significantly

correlated. The marginal effects in tables 5 and 6 are not all significantly

different from zero, but still it is worth noticing that age-cohort 20-24 has the

most negative attitude towards highbrow arts and they favour movies at the

cinema. Figure 1 reveals that the largest amplitudes from the negative

indirect marginal effect to the positive direct marginal effect is by the age-

cohorts 30-34 and 35-39. The amplitude for the cohort 30-34 is (-0.073, 0.023)

↔ 0.096 and for the cohort 35-39 (-0.045, 0.058) ↔ 1.003.

199

Figure 5: (Figure 5.1) Direct and indirect marginal effect of age-cohorts on highbrow art consumption

age15_1

9

age20_2

4

age25_2

9

age30_3

4

age35_3

9

age40_4

4

age45_4

9

age50_5

4

age55_5

9

age60_6

4

age65_6

9

age70_7

4

-0.08

-0.06

-0.04

-0.02

0

0.02

0.04

0.06

0.08

DirME5IndME5DirME6IndME6

The age-cohorts 30-34 and 35-39 are most omnivore but this indication is

unreliable to some extent. The marginal effects of education (Figure 2) are

more reliable since mainly they are significantly different from zero.

Figure 6: (Figure 5.2) Direct and indirect marginal effects of education on highbrow art consumption

edu1 edu2 edu3 edu4 edu5 edu6 edu7 edu8 edu9

-0.15

-0.1

-0.05

0

0.05

0.1

0.15

0.2

0.25

0.3

DirME5IndME5DirME5IndME6

The marginal effects of education in tables 5 and 6 are highly correlated:

ρDirME5, DirME6, edu = 0.977, ρIndME5, IndME6, edu = 0.993. The direct and indirect

marginal effects are highly negatively correlated (ρDirME5, IndME5, edu = -0.859 and

200

ρDirME6, IndME6, edu = -0.884) indicating that those active in highbrow art

consumption are active also in cinema consumption.

The results with the Finnish data are in harmony with the results of Kracman

(1996), Bihagen and Katz-Gerro (2000) or Vander Stichele and Laermans

(2006) who show that educational level, gender and age are related with

performing arts consumption. However, the effect of education is not linear. It

is true that better and longer education leads to higher probability of

consuming performing arts, but those with master’s level education are not

necessarily more active than people with a bachelor’s degree (obtained from

university). If the bachelor’s degree is obtained from a university of applied

sciences (polytechnic), the marginal effect is positive but less positive than for

those with a university level bachelor’s degree. Incomes do not seem to

explain cultural participation but the number of children significantly reduces

cultural participation, both performing arts and movies.

For the purpose of analyzing cultural participation using bivariate probit

analysis, the original data was recoded and reclassified into two categories:

yes vs. no. However, about 5 percent of the respondents in the sample could

be classified to the category ‘often’ (‘every day’ + ‘several times per week’ +

‘several times per month’) in participating in performing arts events. With

multinomial logit analysis, the three groups can be studied but the indirect

effects (between performing arts and movies) that could be evaluated by using

bivariate probit model could not be obtained. Still, this classification into three

201

groups is reasonable. The results of the MNL analysis to explain performing

arts consumption are presented in tables 7 and 8. In table 7 the reference

values of the age-cohort and educational levels are 50-54 years old and

elementary school (edu2).

(table 7 about here)

The results of the MNL analysis confirm the importance of gender. Females

are more active to go to an arts exhibition, opera and/or theatrical

performances. Both the marginal effects of the gender variable or over

individuals show that females most often belong to the group ‘less often’

(occasionally). The only marital status variable to classify into three groups is

‘married’. There are no differences if the person is single or living in

common-law marriage. Married persons most often belong to the group ‘less

often’.

The age-cohort 25-29 is most passive in going to see performing arts.

Surprisingly the older age-cohorts (55-59, 65-69 and 70-74) are most active.

The oldest seem to strongly classify into totally not-going and actively going

groups, but the probability of belonging into ‘less often’ –group is lowest.

Education is very important to classify performing arts consumption.

Consumers with a bachelor’ degree (edu8) are most active. The following

groups in activity are those that have either a college level (edu6) or a

master’s degree (edu9), but the probability of ‘less often’ is bigger if the

202

education level is college, while the probability of ‘often’ is bigger if the

consumer has a master’s degree. The spouse’s education is significant only

when the spouse has a master’s degree. As expected in Southern Finland

(Area1), the category ‘often’ has the biggest probability since the biggest

cities with the largest number of performing arts institutions are in Southern

Finland. Area2 ( Western Finland) is also a significant variable to classify

between ‘never’ and the two other categories ‘less often’ and ‘often’ but there

is a difference between these two last categories.

The results of the other MNL analysis with other reference values for the age-

cohort (40-44) and educational level (edu5 = upper secondary) are presented

in the table 8.

(table 8 about here)

The results in the table 8 are similar than in the table 7 but it shows that there

are no differences between educational levels 1 (pupil or student), 4

(vocational school), 5 (upper secondary) or 7 (bachelor’s degree, university of

applied sciences, polytechnic). Interestingly, if the spouse is a pupil or

student (spouse-edu1), the respondent most probably is active (‘often’) in

performing arts consumption. As expected, if the respondent has children, it

significantly lowers the probability of going to an art exhibition, opera and/or

theatrical performance.

203

5.5 Conclusions

The purpose of this paper is to study performing arts consumption and movies

at the cinema consumption. A number of different socioeconomic variables are

used to explain cultural consumption. The bivariate probit approach to

studying performing arts and movies at the cinema consumption in bundle is

useful because it reveals substantially new evidence on the average profile of

culture consumption. It is expected that females go more often to an art

exhibition, opera or theatrical performances and this was supported. The

results of the bivariate probit analysis also reveal that gender is important to

explain also movie attendance. Females go more often to see movies at the

cinema. There is a significantly positive correlation between these two

audiences indicating that there is a common background between both

groups. The approach also allows finding the most relevant socioeconomic

characteristics explaining cultural consumption.

It is widely known that gender, age and educational level of the consumer

have an impact on cultural consumption (e.g. Kracman 1996, Borgonovi 2004

or Montgomery and Robinson 2006). The novelty of the results here indicates

that also the educational level of the spouse matters. If the spouse has high

education (master’s degree), it significantly increases highbrow cultural

consumption. The probability of being classified into heavy user group

increases. The analysis shows that when the effects of other socioeconomic

variables have been controlled, the gross income level does not significantly

204

explain cultural consumption. Younger people prefer movies and their

incomes are typically low and this explains why incomes do not explain movie

attendance. However the effect of incomes on highbrow performing art

consumption is also zero. Education matters but incomes do not. Married

consumers seem to prefer highbrow arts but the more informal partnership,

‘common-law marriage’, seems to have a negative impact on movie attendance

but no effect on highbrow art consumption.

The sociology of cultural participation classifies consumers into three groups:

omnivore, paucivore and inactive (Alderson, Junisbai and Heacock 2007).

Omnivores are active in all cultural consumption and paucivores are less

active. Female age-cohorts 30-34 and 35-39 with a bachelor’s degree

(university) are most omnivore and the oldest male age-cohorts with the

lowest education (elementary school) are most inactive.

205

References

Alderson, Arthur S., Azamat Junisbai and Isaac Heacock (2007): Social status and cultural consumption in the United States. Poetics 35, 191-212

Ateca-Amestoy, Victoria (2008): Determining heterogeneous behavior for theater attendance. Journal of Cultural Economics 32, 127-151

Austin, Bruce A. (1986): Motivations for Movie Attendance. Communication Quarterly, 34, 115-126

Bihagen, Erik and Tally Katz-Gerro (2000): Culture consumption in Sweden: The stability of gender differences. Poetics 27, 327-349

Borgonovi, Francesca (2004): Performing arts attendance: an economic approach. Applied Economics 36: 1871-1885

Bourdieu, Pierre (1979): La Distinction: Critique sociale du jugement. Les Editions de Minuit, Paris

Chen, Tak Wing and John H. Goldthorpe (2005): The social stratification of theatre, dance and cinema attendance. Cultural Trends 14, 193-212

DiMaggio, Paul (1987): Classification in art. American Sociological Review 52, 440-455

Eijck, Koen van (1997): The impact of family background and educational attainment on cultural consumption: A sibling analysis. Poetics 25, 195-224

F & L Research (1999): Suomalaisen elokuvan markkinat ja kilpailukyky. (In Finnish: The market of the Finnish film and competitive ability)

Frey, Bruno S. (2000): Arts & economics. Analysis & Cultural Policy. Springer Verlag

Kotimaisen elokuvan yleisöt –tutkimus 2010 (in Finnish, Audience of domestic movies)

Kracman, Kimberly (1996): The effect if school-based arts introduction on attendance at museums and the performing arts. Poetics 24, 203-218

206

Kramer, Peter (1998): Women first: ‘ Titanic’ (1997), action-adventure films and Hollywood’s female audience. Historical Journal of Film, Radio & Television, 18, 599-618

Liikkanen, Mirja (1996): Culture Consumption in Finland. Distinctive Characteristics. Nordicom Review 1, 1-17

Lizardo, Omar (2006): The puzzle of women’s “highbrow” culture consumption: Integrating gender and work into Bourbieu’s class theory of taste. Poetics 34, 1-23

Montgomery, Sarah S. ja Michael D. Robinson (2006): Take Me Out to the Opera: Are Sports and Arts Complements? Evidence from the Performing Arts Research Coalition Data. International Journal of Arts Management 8:2, 24-39

Nagel, Ineke, Marie-Louise Damen & Folkert Haanstra (2010): The arts course CKV1 and cultural participation in the Netherlands. Poetics 38, 365-385

Peterson, Richard A. (1992): Understanding audience segmentation: From elite and mass to omnivore and univore. Poetics 21, 243-241

Peterson, Richard A. and Roger Kern (1996): Changing highbrow taste: From snob to omnivore. American Sociological Review 61, 900-907

Prieto-Rodríguez, Juan and Víctor Fernández-Blanco (2000): Are Popular and Classical Music Listeners the Same People? Journal of Cultural Economics 24, 147-164

Purhonen, Semi, Jukka Gronow and Keijo Rahkonen (2009): Social differentiation of musical and literary taste patterns in Finland. Research on Finnish Society 2, 39-49

Purhonen, Semi, Jukka Gronow and Keijo Rahkonen (2010): Nordic democracy of taste? Cultural omnivorousness in musical and literary taste preferences in Finland. Poetics 38, 266-298

207

Redondo, Ignacio and Morris B. Holbrook (2010): Modeling the appeal of movie features to demographic segments of theatrical demand. Journal of Cultural Economics, 34, 299-315

Seaman, Bruce A. (2005): Attendance and Public Participation in the Performing Arts: A Review of the Empicical Literature. Georgia State University, Working Paper 05-03, August 2005

Sintas, Jordi López and Ercilia García Álvarez (2004): Omnivore versus univore consumption and its symbolic properties: evidence from Spaniards’ performing arts attendance. Poetics 32, 463-483

Suomalaisen elokuvan markkinat ja kilpailukyky 1999

Stigler, George and Gary S. Becker (1977): De Gustibus Non est Disputandum. American Economic Review 67, 76-90

Upright, Craig Barton (2004): Social capital and cultural participation: spousal influences on attendance at arts events. Poetics 32, 129-143

Van Eijck, Koen (1997): The impact of family background and educational attainment on cultural consumption: A sibling analysis. Poetics 25, 195-224

Vander Stichele, Alexander and Rudi Laermans (2006): Cultural participation in Flanders: Testing the cultural omnivore thesis with population data. Poetics 34, 45-64

Warde, Alan and Modesto Gayo-Cal (2009): The anatomy of cultural omnivorousness: The case of the United Kingdom. Poetics 37, 119-145’

Virtanen, Taru (2007): Across and beyond the bounds of taste: On cultural consumption patterns in the European Union. Turku School of Economics A-11

208

Table 41: (Table 5.1) Spectators of movies at the cinema and performing atrs (concert, theatre, art exhibition) in Finland, recent surveys

Survey Often Occasionally Never Interviews made

Sample size

NotesCinema

Arts Cinema

Arts Cinema

Arts

2006: Adult education study

8.2 %

10.5 %

46.7 %

53.3 %

45.0 %

36.2% March – June 2006

4370 Often = more than 7 times/year

2007: European Cultural Values

3.0%

bdo1%t3%c3%

49% bdo22%t47%c48%

48% bdo77%t52%;c49%

February – March 2007

1041 Often = more than 5 times/year

2007: ISSP

1.9%

5.6 %

80.2%

78.9 %

17.9%

15.5 %

September – December 2007

1354 Often = more than 12 times/year

2010: Kotimaisen elokuvan yleisöt

10% 70% 20% January – February 2010

504 Often = more than 12 times/year

In the European Cultural Values study: bdo = a ballet, a dance performance or an opera; t = the theatre; c = a concert

209

Table 42: (Table 5.2) descriptive statistics of age-group and education variables

Table 2: edu1 edu2 edu3 edu4 edu5 edu6 edu7 edu8 edu9

5.5% 10.6%

7.9% 22.1%

7.2% 24.6%

8.1% 4.1% 9.9%

age15_19 6.2% 84.0%

12.1%

age20_24 5.4% 11.6%

26.4%

age25_29 7.4% 4.3% 28.4%

13.6%

age30_34 6.0% 13.7%

age35_39 8.0% 12.9%

17.7%

14.4%

age40_44 8.7% 13.5%

17.6%

age45_49 10.0%

11.6%

12.1%

15.2%

15.4%

age50_54 8.7% 20.2%

age55_59 11.0%

19.2%

11.6%

13.5%

age60_64 11.2%

23.9%

15.1%

14.7%

13.5%

age65_69 6.4% 24.6%

age70_74 6.1% 23.9%

100% Three largest age-groups according to the education, e.g. 84% of the youngest are pupils/students and 23.9% of the oldest have only elementary school background.

edu1 = pupil or student (comprehensive, upper secondary, vocational school or course, college: 5.5% in the sample are pupils or studentsedu2 = elementary schooledu3 = comprehensive schooledu4 = vocational school or courseedu5 = upper secondary, secondary school graduateedu6 = college

210

edu7 = bachelor’s degree(polytechnic or university of applied sciences)edu8 = bachelor (university)edu9 = master’s degree

211

Table 43: (Table 5.3) Average monthly household and personal gross incomes

Table 3: Household income

Personal income

age15_19 2083 90age20_24 1629 859age25_29 3653 1948age30_34 6400 3310age35_39 5175 2496age40_44 4901 2996age45_49 5469 2663age50_54 4911 2483age55_59 3684 1931age60_64 2759 1770age65_69 2578 1687age70_74 2291 1449edu1 2323 134edu2 1759 1166edu3 2564 1382edu4 3063 1924edu5 3081 1374edu6 4905 2492edu7 5158 2764edu8 3885 2285edu9 7072 3579including taxes and social security contributions by age and by education groups

212

Table 44: (Table 5.4) Descriptive statistics of some explanatory variables

female: 57 % male: 43 % n = 1232marital status: single

18.3% 23.0% 20,3%

married or registered pair relation

48.6% 51.9% 50.0%

common-law marriage

17.0% 17.3% 17.1%

judicial separation*

0.3% 0.7% 0.5%

separated* 11.0% 5.2% 8.4%widow(er)* 4.9% 1.9% 3.6%Province: Area1 53.0% 49.3% 51.4%Area2 25.9% 25.7% 25.8%Area3 12.2% 13.6% 12.8%Rest of Finland*: 8.8% 11.5% 10.0% * = reference groups (constant) in probit or logit analysis

213

Table 45: (Table 5.5) Bivariate probit analysis

Art123

4_5

Art1234_5: total marginal effect

Art1234_5: direct marginal effect

Art1234_5: indirect marginal effect

Mov1234_5

gender (male=1, female=2)

0.423(0.097)***

0.064(0.016)***

0.074(0.017)***

-0.010(0.005)(*)

0.189(0.096)*

Marital status: single

-0.004(0.188)

0.009(0.031)

-0.001(0.033)

0.010(0.010)

-0.180(0.188)

Marital status: married

0.319(0.203)

0.070(0.034)*

0.056(0.036)

0.014(0.011)

-0.264(0.198)

MS: common-law marriage

0.198(0.217)

0.057(0.036)

0.034(0.038)

0.022(0.012)(*)

-0.403(0.225)(*)

age15_19 -0.038(0.310)

-0.063(0.053)

-0.007(0.054)

-0.056(0.026)*

1.027(0.458)*

age20_24 -0.242(0.270)

-0.087(0.045)*

-0.042(0.047)

-0.044(0.018)*

0.806(0.327)*

age25_29 0.056(0.259)

-0.033(0.043)

0.010(0.045)

-0.043(0.015)**

0.786(0.254)**

age30_34 0.133(0.297)

-0.049(0.048)

0.023(0.052)

-0.073(0.021)***

1.321(0.369)***

age35_39 0.331(0.249)

0.013(0.041)

0.058(0.044)

-0.045(0.015)**

0.826(0.268)**

age40_44 0.069(0.244)

-0.017(0.041)

0.012(0.043)

-0.029(0.013)*

0.531(0.229)*

age45_49 0.158(0.205)

-0.007(0.034)

0.028(0.036)

-0.034(0.012)***

0.626(0.212)***

age50_54 = con

214

age55_59 0.064(0.204)

0.001(0.034)

0.011(0.036)

-0.011(0.009)

0.193(0.169)

age60_64 0.035(0.189)

-0.007(0.032)

0.006(0.033)

-0.013(0.009)

0.234(0.168)

age65_69 0.053(0.233)

-0.006(0.037)

0.009(0.040)

-0.016(0.012)

0.284(0.212)

age70_74 -0.168(0.236)

-0.009(0.039)

-0.029(0.041)

0.020(0.011)(*)

-0.367(0.201)(*)

edu1 0.687(0.290)*

0.077(0.049)

0.120(0.052)*

-0.043(0.023)(*)

0.783(0.432)(*)

edu2 = con edu3 0.424

(0.213)*

0.053(0.035)

0.074(0.037)*

-0.022(0.011)*

0.395(0.196)*

edu4 0.550(0.167)***

0.075(0.028)**

0.096(0.030)**

-0.021(0.009)*

0.386(0.160)*

edu5 0.841(0.221)***

0.108(0.039)**

0.147(0.040)***

-0.039(0.014)***

0.714(0.248)***

edu6 1.037(0.191)***

0.145(0.033)***

0.182(0.035)***

-0.037(0.010)***

0.675(0.165)***

edu7 0.863(0.258)***

0.104(0.043)*

0.151(0.046)***

-0.047(0.016)***

0.865(0.280)***

edu8 1.425(0.463)***

0.150(0.082)(*)

0.250(0.081)***

-0.100(0.037)***

1.821(0.665)***

edu9 1.111(0.340)***

0.160(0.058)**

0.195(0.062)***

-0.035(0.014)*

0.635(0.249)**

spouse-edu1

-0.226(1.327)

-0.049(0.335)

-0.040(0.233)

-0.009(0.112)

0.169(2.030)

215

spouse-edu2 = Cspouse-edu3 -0.129

(0.257)-0.034(0.043)

-0.023(0.045)

-0.011(0.014)

0.208(0.252)

spouse-edu4 -0.225(0.185)

-0.039(0.030)

-0.040(0.032)

0.000(0.009)

-0.004(0.172)

spouse-edu5 0.197(0.329)

-0.004(0.063)

0.035(0.058)

-0.039(0.023)(*)

0.711(0.414)(*)

spouse-edu6 0.030(0.227)

-0.011(0.039)

0.005(0.040)

-0.016(0.011)

0.299(0.199)

spouse-edu7 0.135(0.333)

0.001(0.055)

0.024(0.058)

-0.022(0.018)

0.410(0.328)

spouse-edu8 -0.333(0.388)

-0.055(0.074)

-0.058(0.068)

0.004(0.022)

-0.066(0.407)

spouse-edu9 0.597(0.458)

0.069(0.079)

0.105(0.078)

-0.036(0.018)*

0.648(0.317)*

Area1 0.331(0.149)*

0.026(0.025)

0.058(0.026)*

-0.032(0.009)***

0.576(0.143)***

Area2 0.356(0.166)*

0.025(0.028)

0.062(0.029)*

-0.037(0.010)***

0.682(0.163)***

Area3 0.176(0.180)

0.003(0.030)

0.031(0.032)

-0.028(0.100)***

0.511(0.176)***

Household’s size

0.033(0.062)

-0.003(0.011)

0.006(0.011)

-0.009(0.004)*

0.162(0.076)*

Children <7 -0.221(0.098)*

-0.031(0.017)(*)

-0.039(0.017)*

0.007(0.006)

-0.133(0.111)

Children 7-17

-0.292(0.134)*

-0.032(0.021)

-0.051(0.024)*

0.019(0.009)*

-0.343(0.154)*

Household Incomes

0.191D-5(0.582D-5)

-0.294D-6(0.966D-6)

0.335D-6(0.102D-5)

-0.629D-6(0.492D-6))

0.114D-4(0.901D-5)

Constant -0.759 -1.033

216

(0.252)***

(0.270)***

ρ = 0.625 (0.048)***

(standard error in parenthesis.). Art1234_5: 0 =’ Never in the last twelve months’, 1 = Less often’ or ‘Several times per month’ or ‘Several times per week’ or ‘Every day’ - Mov1234_5 classified in the same way.Log Likelihood = - 985.15, AIC = 1.633, BIC = 1.953, HQIC = 1.754, (*), *, **, *** = significance level 10%,5%,1%,0,1% .

217

Table 46: (Table 5.6) Bivariate probit analysis ,

Art1234

_5

Art1234_5: marginal effect

Art1234_5: direct marginal effect

Art1234_5: indirect marginal effect

Mov1234_

5

gender (male=1, female=2)

0.422(0.092)***

0.063(0.016)***

0.074(0.016)***

-0.011(0.005)*

0.196(0.095)*

Marital status: single

0.032(0.182)

0.013(0.030)

0.006(0.032)

0.008(0.010)

-0.141(0.185)

Marital status: married

0.409(0.224)(*)

0.058(0.038)

0.071(0.039)(*)

-0.014(0.014)

0.249(0.255)

MS: common-law marriage

0.286(0.239)

0.043(0.040)

0.050(0.042)

-0.007(0.015)

0.125(0.269)

age15_19 -0.014(0.293)

-0.046(0.050)

-0.002(0.051)

-0.044(0.025)(*)

0.782(0.425)(*)

age20_24 -0.172(0.260)

-0.064(0.043)

-0.030(0.045)

-0.034(0.018)(*)

0.612(0.324)(*)

age25_29 0.061(0.252)

-0.019(0.042)

0.011(0.044)

-0.030(0.014)*

0.534(0.252)*

age30_34 0.109(0.278)

-0.038(0.046)

0.019(0.049)

-0.057(0.021)**

1.015(0.361)**

age35_39 0.335(0.238)

0.027(0.039)

0.058(0.042)

-0.031(0.014)*

0.561(0.254)*

age40_44 = C

-- -- -- -- --

age45_49 0.123(0.195)

0.002(0.033)

0.021(0.034)

-0.019(0.012)

0.342(0.220)

age50_54 0.011(0.214)

0.012(0.035)

0.002(0.037)

0.010(0.011)

-0.174(0.189)

218

age55_59 0.026(0.207)

0.007(0.034)

0.005(0.036)

0.003(0.010)

-0.046(0.187)

age60_64 0.009(0.196)

0.001(0.033)

0.002(0.034)

-0.001(0.011)

0.014(0.189)

age65_69 0.012(0.242)

-0.002(0.039)

0.002(0.042)

-0.005(0.013)

0.083(0.234)

age70_74 -0.253(0.236)

-0.009(0.039)

-0.044(0.041)

0.035(0.013)**

-0.627(0.218)**

edu1 0.020(0.279)

-0.003(0.047)

0.036(0.049)

-0.007(0.022)

0.119(0.398)

edu2 -0.646(0.208)***

-0.077(0.036)*

-0.113(0.037)**

0.035(0.013)**

-0.636(0.224)**

edu3 -0.198(0.225)

-0.025(0.038)

-0.035(0.039)

0.010(0.013)

-0.172(0.235)

edu4 -0.065(0.178)

-0.002(0.031)

-0.011(0.031)

0.009(0.011)

-0.169(0.201)

edu5 ccc ccc ccc ccc edu6 0.428

(0.198)*

0.067(0.034)*

0.075(0.035)*

-0.007(0.012)

0.132(0.209)

edu7 0.229(0.253)

0.023(0.042)

0.040(0.044)

-0.017(0.017)

0.308(0.297)

edu8 0.806(0.465)(*)

0.071(0.081)

0.141(0.081)(*)

-0.070(0.038)(*)

1.256(0.681)(*)

edu9 0.492(0.348)

0.080(0.059)

0.086(0.062)

-0.006(0.014)

0.103(0.258)

spouse-edu1

-0.273(1.375)

-0.029(0.347)

-0.048(0.241)

0.019(0.116)

-0.341(2.088)

spouse-edu2 0.044(0.242)

0.038(0.041)

0.008(0.042)

0.030(0.015)*

-0.540(0.256)*

spouse-edu3 -0.156(0.270)

-0.013(0.046)

-0.027(0.047)

0.014(0.016)

-0.250(0.294)

spouse-edu4 -0.270(0.200)

-0.021(0.035)

-0.047(0.035)

0.027(0.014)(*)

-0.476(0.235)*

219

spouse-edu5 ccc ccc ccc cccspouse-edu6 -0.012

(0.245)0.009(0.043)

-0.002(0.043)

0.011(0.014)

-0.200(0.254)

spouse-edu7 0.072(0.326)

0.017(0.055)

0.013(0.057)

0.004(0.020)

-0.079(0.360)

spouse-edu8 -0.374(0.384)

-0.036(0.073)

-0.065(0.067)

0.029(0.025)

-0.526(0.436)

spouse-edu9 0.575(0.469)

0.091(0.081)

0.100(0.080)

-0.009(0.019)

0.170(0.337)

Area1 0.359(0.146)*

0.030(0.025)

0.063(0.026)*

-0.033(0.009)***

0.593(0.141)***

Area2 0.368(0.164)*

0.026(0.027)

0.064(0.029)*

-0.038(0.010)***

0.678(0.161)***

Area3 0.203(0.179)

0.004(0.030)

0.035(0.031)

-0.031(0.010)**

0.556(0.174)**

Household’s size

0.035(0.060)

-0.002(0.011)

0.006(0.011)

-0.009(0.004)*

0.153(0.076)*

Children <7 -0.217(0.094)*

-0.033(0.016)*

-0.038(0.016)*

0.005(0.006)

-0.087(0.104)

Children 7-17

-0.295(0.130)*

-0.034(0.021)(*)

-0.051(0.023)*

0.018(0.008)*

-0.317(0.150)*

Household Incomes

0.278D-5(0.563D-5)

-0.234D-6(0.906D-6)

0.486D-6(0.989D-6)

-0.720D-6(0.499D-6

0.129D-4(0.900D-5)

Constant -0.186(0.271)

-0.280(0.301)

ρ = 0.631 (0.047)***

(standard error in parenthesis.). Art1234_5: 0 =’ Never in the last twelve months’, 1 = ‘Less often’ or ‘Several times per month’ or ‘Several times per week’ or ‘Every day’ - Mov1234_5 classified in the same way.Log Likelihood = - 985.15, AIC = 1.633, BIC = 1.953, HQIC =

220

1.754, (*), *, **, *** = significance level 10%,5%,1%,0,1% .

221

Table 47: (Table 5.7) Multinomial logit (MNL) analysis

y = 1 y = 2 marginal effects: y = 0

marginal effects: y = 1

marginal effects: y = 2

Marginal effects averaged over individuals:y = 0

Marginal effects averaged over individuals:y = 1

Marginal effects averaged over individuals:y = 2

Averages of Individual Elasticities of Probabilities:y = 0

Averages of Individual Elasticities of Probabilities:y = 1

Averages of Individual Elasticities of Probabilities:y = 2

gender (male=1, female=2)

0.771(0.159)***

1.156(0.313)***

-0.090(0.018)***

0.078(0.019)***

0.013(0.008)(*)

-0.098 0.076 0.023 -1.042 0.153 0.750

Marital status: single

-0.075(0.311)

0.497(0.572)

0.007(0.036)

-0.021(0.038)

0.015(0.014)

0.006 -0.032 0.026 0.005 -0.011 0.112

Marital status: married

0.562(0.343)(*)

-0.499(0.746)

-0.061(0.040)

0.087(0.043)*

-0.026(0.018)

-0.063 0.109 -0.045 -0.211 0.062 -0.454

MS: common-law marriage

0.275(0.380)

0.108(0.789)

-0.031(0.044)

0.035(0.047)

-0.003(0.019)

-0.033 0.039 -0.006 -0.037 0.009 -0.019

age15_19

-0.056(0.522)

-0.505(1.093)

0.008(0.060)

0.004(0.065)

-0.012(0.027)

0.010 0.011 -0.021 0.004 0.000 -0.028

222

age20_24

-0.443(0.432)

-0.615(0.797)

0.052(0.050)

-0.046(0.052)

-0.006(0.019)

0.056 -0.046 -0.011 0.020 -0.005 -0.014

age25_29

0.166(0.427)

-2.048(1.201)(*)

-0.011(0.049)

0.069(0.055)

-0.058(0.030)(*)

-0.007 0.108 -0.101 -0.009 0.003 -0.161

age30_34

0.244(0.467)

0.321(0.887)

-0.028(0.054)

0.026(0.056)

0.003(0.021)

-0.031 0.026 0.005 -0.013 0.002 0.007

age35_39

0.602(0.437)

0.544(0.817)

-0.069(0.050)

0.069(0.052)

0.001(0.019)

-0.075 0.073 0.001 -0.043 0.005 0.000

age40_44

0.029(0.385)

-0.902(0.942)

0.329D-4(0.044)

0.024(0.048)

-0.024(0.023)

0.002 0.040 -0.043 -0.001 0.002 -0.080

age45_49

0.192(0.359)

0.427(0.690)

-0.023(0.041)

0.016(0.043)

0.007(0.016)

-0.025 0.013 0.012 -0.018 0.001 0.025

age50_54,C

---

age55_59

-0.016(0.333)

1.037(0.632)(*)

-0.002(0.038)

-0.026(0.040)

0.028(0.015)(*)

-0.005 -0.044 0.048 -0.008 -0.009 0.107

age60_64

-0.049(0.329)

1.221(0.623)*

0.001(0.038)

-0.034(0.040)

0.033(0.015)*

-0.002 -0.056 0.058 -0.009 -0.014 0.128

age65_69

-0.015(0.395)

0.476(0.845)

-0.870D-4(0.045)

-0.013(0.049)

0.013(0.020)

-0.001 -0.021 0.022 -0.000 -0.001 0.030

age70_74

-0.412(0.385)

1.648(0.698)*

0.040(0.044)

-0.093(0.047)*

0.053(0.018)**

0.038 -0.130 0.092 0.003 -0.023 0.104

223

edu1 1.155(0.493)*

1.730(1.225)

-0.136(0.057)*

0.116(0.064)(*)

0.019(0.030)

-0.147 0.114 0.034 -0.046 0.016 0.046

edu2 C

edu3 0.667(0.331)*

0.733(0.866)

-0.077(0.038)*

0.073(0.044)(*)

0.004(0.022)

-0.083 0.076 0.007 -0.040 0.013 0.018

edu4 0.869(0.269)***

1.678(0.665)**

-0.103(0.031)***

0.079(0.035)*

0.024(0.017)

-0.113 0.071 0.043 -0.155 0.034 0.210

edu5 1.376(0.410)***

3.184(0.779)***

-0.166(0.048)***

0.113(0.051)*

0.053(0.020)**

-0.182 0.091 0.092 -0.098 0.001 0.132

edu6 1.805(0.321)***

2.551(0.689)***

-0.211(0.037)***

0.185(0.040)***

0.026(0.017)

-0.229 0.184 0.046 -0.400 0.028 0.205

edu7 1.392(0.418)***

3.063(0.862)***

-0.167(0.049)***

0.118(0.052)*

0.049(0.022)*

-0.184 0.098 0.085 -0.106 0.008 0.145

edu8 2.478(0.796)***

4.609(1.066)***

-0.294(0.090)***

0.229(0.090)**

0.065(0.022)**

-0.321 0.208 0.113 -0.100 -0.008 0.071

edu9 1.921(0.518)****

3.733(0.839)***

-0.229(0.059)***

0.174(0.061)**

0.055(0.020)**

-0.250 0.155 0.095 -0.196 -0.010 0.166

spouse- -0.549 2.398 0.053 -0.128 0.076 0.050 -0.182 0.132 0.000 -0.003 0.015

224

edu1 (0.987) (1.516) (0.113) (0.120) (0.036)*

spouse-edu2

C

spouse-edu3

-0.259(0.433)

0.694(0.926)

0.026(0.050)

-0.051(0.054)

0.024(0.023)

0.026 -0.069 0.042 0.007 -0.005 0.039

spouse-edu4

-0.363(0.324)

0.378(0.738)

0.039(0.037)

-0.057(0.041)

0.018(0.018)

0.041 -0.072 0.032 0.047 -0.018 0.116

spouse-edu5

0.393(0.539)

1.268(1.083)

-0.049(0.062)

0.024(0.066)

0.024(0.026)

-0.054 0.012 0.043 -0.015 0.001 0.035

spouse-edu6

0.119(0.391)

0.613(0.791)

-0.015(0.045)

0.002(0.048)

0.013(0.019)

-0.018 -0.006 0.023 -0.018 -0.002 0.069

spouse-edu7

0.164(0.516)

0.909(0.993)

-0.022(0.059)

0.001(0.062)

0.020(0.023)

-0.025 -0.010 0.035 -0.010 -0.001 0.041

spouse-edu8

-0.514(0.562)

-1.171(1.345)

0.062(0.065)

-0.043(0.070)

-0.019(0.033)

0.068 -0.035 -0.033 0.013 -0.002 -0.021

spouse-edu9

1.152(0.690)(*)

2.539(0.967)**

-0.138(0.078)(*)

0.097(0.078)

0.041(0.020)*

-0.152 0.081 0.071 -0.101 -0.010 0.099

Area1 0.533(0.248)*

1.402(0.609)*

-0.065(0.029)*

0.040(0.032)

0.025(0.015)(*)

-0.072 0.029 0.043 -0.256 0.011 0.447

Area2 0.583(0.273)*

1.288(0.649)*

-0.070(0.031)*

0.049(0.034)

0.021(0.016)

-0.077 0.041 0.036 -0.134 0.015 0.195

Area3 0.280(0.299)

0.588(0.739)

-0.033(0.035)

0.024(0.038)

0.009(0.018)

-0.037 0.021 0.016 -0.029 0.007 0.046

Househol 0.062 0.157 -0.008 0.005 0.003 -0.008 0.004 0.005 -0.130 0.015 0.238

225

d’s size (0.104) (0.233) (0.012) (0.013) (0.006)

Children <7

-0.423(0.147)**

-0.334(0.350)

0.049(0.017)**

-0.049(0.019)**

0.001(0.001)

0.052 -0.054 0.002 0.138 -0.039 -0.002

Children 7-17

-0.488(0.213)*

-2.359(1.038)*

0.063(0.025)**

-0.012(0.031)

-0.051(0.022)*

0.072 0.017 -0.089 0.068 -0.014 -0.324

Household Incomes

0.129D-4(0.173D-4)

-0.431D-4(0.517D-4)

-0.128D-5(0.025)

0.271D-5(0.226D-5)

-0.142D-5(0.128D-5)

0.000 0.000 0.000 -0.034 0.015 -0.199

Constant -1.339(0.428)***

-6.840(1.107)***

0.175(0.050)***

-0.025(0.059)

-0.150(0.035)***

Explanatory variable: y = “How many times in the last twelve months have you seen an art exhibition, opera or theatrical performance?” = 0 (never), 1 (less often) or 2 (daily, several times per week or several times per month).McFadden pseudo R2 = 0.146, χ2 = 246.006***, AIC = 1.261, BIC = 1.569, HQIC = 1.371(*), *, **, *** = significance level 10%,5%,1%,0,1% .Partial derivatives of probabilities with respect to the vector of characteristics are computed at the means of the Xs. Probabilities at the mean vector are Prob(y=0) = 0.133, Prob(y=1) = 0.840, Prob(y=2) = 0.027

226

Table 48: (Table 5.8) Multinomial logit (MNL) analysis,

y = 1 y = 2 marginal effects: y = 0

marginal effects: y = 1

marginal effects: y = 2

Marginal effects averaged over individuals:y = 0

Marginal effects averaged over individuals:y = 1

Marginal effects averaged over individuals:y = 2

Averages of Individual Elasticities of Probabilities:y = 0

Averages of Individual Elasticities of Probabilities:y = 1

Averages of Individual Elasticities of Probabilities:y = 2

gender (male=1, female=2)

0.761(0.156)***

1.126(0.310)***

-0.090(0.018)***

0.077(0.019)***

0.013(0.008)(*)

-0.098 0.076 0.022 -1.027 0.152 0.718

Marital status: single

-0.023(0.303)

0.547(0.558)

0.001(0.035)

-0.016(0.037)

0.016(0.014)

-0.001 -0.025 0.026 -0.003 -0.008 0.113

Marital status: married

0.701(0.390)(*)

-0.162(0.836)

-0.078(0.045)(*)

0.099(0.049)*

-0.021(0.022)

-0.082 0.117 -0.035 -0.274 0.067 -0.352

MS: common-law marriage

0.418(0.415)

0.456(0.868)

-0.049(0.048)

0.046(0.052)

0.003(0.022)

-0.053 0.048 0.005 -0.058 0.012 0.018

age15_19

0.032(0.497)

-0.033(1.059)

-0.003(0.057)

0.005(0.063)

-0.002(0.027)

-0.004 0.006 -0.003 -0.001 0.001 -0.003

227

age20_24

-0.257(0.421)

-0.135(0.795)

0.029(0.049)

-0.032(0.052)

0.002(0.020)

0.031 -0.035 0.004 0.010 -0.004 0.003

age25_29

0.201(0.420)

-1.884(1.205)

-0.015(0.049)

0.072(0.055)

-0.057(0.031)(*)

-0.011 0.107 -0.095 -0.011 0.004 -0.151

age30_34

0.239(0.454)

0.350(0.889)

-0.028(0.053)

0.024(0.056)

0.004(0.022)

-0.031 0.024 0.007 -0.013 0.002 0.008

age35_39

0.640(0.423)

0.646(0.799)

-0.074(0.049)

0.072(0.051)

0.003(0.019)

-0.080 0.076 0.005 -0.046 0.005 0.005

age40_44

C C C C C

age45_49

0.148(0.349)

0.466(0.668)

-0.018(0.040)

0.009(0.043)

0.009(0.016)

-0.021 0.005 0.016 -0.015 0.000 0.032

age50_54

0.003(0.356)

-0.760(0.918)

0.003(0.041)

0.019(0.046)

-0.021(0.024)

0.005 0.031 -0.035 0.001 0.001 -0.065

age55_59

-0.064(0.348)

0.902(0.661)

0.004(0.040)

-0.030(0.043)

0.027(0.017)

0.002 -0.046 0.045 -0.002 -0.010 0.097

age60_64

-0.080(0.0345)

1.145(0.649)(*)

0.005(0.040)

-0.038(0.042)

0.034(0.017)*

0.002 -0.058 0.056 -0.005 -0.014 0.123

age65_69

-0.057(0.412)

0.281(0.866)

0.005(0.048)

-0.014(0.051)

0.009(0.022)

0.005 -0.020 0.015 0.002 -0.002 0.020

age70_74

-0.540(0.399)

1.308(0.708)(*)

0.056(0.046)

-0.105(0.049)*

0.049(0.019)*

0.055 -0.138 0.082 0.010 -0.023 0.090

edu1 0.124 -0.863 0.011 0.038 -0.027 -0.009 0.054 -0.045 -0.003 0.003 -0.049

228

(0.460) (1.076) (0.053) (0.059) (0.028)

edu2 -0.998(0.358)**

-2.125(0.741)**

0.120(0.042)**

-0.085(0.046)(*)

-0.035(0.020)(*)

0.133 -0.074 -0.059 0.069 -0.034 -0.150

edu3 -0.276(0.359)

-1.495(0.788)(*)

0.037(0.042)

-0.002(0.046)

-0.035(0.021)(*)

0.043 0.016 -0.059 0.019 -0.002 -0.097

edu4 -0.074(0.295)

-0.645(0.554)

0.011(0.034)

0.005(0.037)

-0.016(0.014)

0.013 0.014 -0.027 0.018 0.002 -0.122

edu5 C C C C C C

edu6 0.868(0.344)*

0.248(0.581)

-0.098(0.039)*

0.112(0.041)**

-0.014(0.014)

-0.105 0.128 -0.023 -0.182 0.023 -0.124

edu7 0.417(0.416)

0.584(0.732)

-0.049(0.048)

0.043(0.051)

0.006(0.018)

-0.053 0.043 0.011 -0.031 0.004 0.017

edu8 1.515(0.800)(*)

2.190(0.975)*

-0.179(0.091)*

0.154(0.091)(*)

0.025(0.018)

-0.194 0.153 0.042 -0.057 -0.001 0.024

edu9 0.959(0.520)(*)

1.273(0.717)(*)

-0.113(0.060)(*)

0.100(0.060)(*)

0.012(0.015)

-0.122 0.101 0.021 -0.092 0.001 0.032

spouse-edu1

-0.651(0.987)

2.137(1.543)

0.065(0.114)

-0.140(0.121)

0.075(0.039)(*)

0.063 -0.188 0.125 0.001 -0.003 0.014

spouse-edu2

0.093(0.432)

0.334(0.999)

-0.012(0.050)

0.005(0.055)

0.007(0.026)

-0.013 0.001 0.012 -0.006 0.001 0.018

spouse- -0.293 0.570 0.031 -0.054 0.023 0.031 -0.069 0.038 0.008 -0.005 0.035

229

edu3 (0.463) (0.997) (0.054) (0.059) (0.026)

spouse-edu4

-0.422(0.355)

0.169(0.825)

0.047(0.041)

-0.062(0.045)

0.015(0.021)

0.049 -0.074 0.025 0.057 -0.019 0.087

spouse-edu5

C

spouse-edu6

0.070(0.419)

0.462(0.870)

-0.010(0.049)

-0.001(0.052)

0.011(0.022)

-0.011 -0.007 0.019 -0.012 -0.002 0.054

spouse-edu7

0.074(0.527)

0.740(1.039)

-0.011(0.061)

-0.008(0.065)

0.019(0.026)

-0.014 -0.018 0.031 -0.005 -0.001 0.037

spouse-edu8

-0.575(0.579)

-1.252(1.389)

0.069(0.067)

-0.048(0.073)

-0.021(0.036)

0.077 -0.041 -0.035 0.014 -0.002 -0.021

spouse-edu9

1.134(0.699)(*)

2.515(1.021)*

-0.137(0.079)(*)

0.094(0.080)

0.043(0.023)(*)

-0.151 0.080 0.072 -0.100 -0.010 0.098

Area1 0.573(0.246)*

1.433(0.603)*

-0.070(0.029)*

0.044(0.032)

0.026(0.016)(*)

-0.078 0.034 0.044 -0.273 0.014 0.446

Area2 0.593(0.271)*

1.256(0.643)*

-0.071(0.031)*

0.051(0.035)

0.021(0.017)

-0.079 0.044 0.035 -0.136 0.016 0.185

Area3 0.322(0.299)

0.630(0.733)

-0.039(0.035)

0.029(0.039)

0.010(0.019)

-0.042 0.026 0.017 -0.033 0.008 0.047

Household’s size

0.063(0.105)

0.151(0.232)

-0.008(0.012)

0.005(0.013)

0.003(0.006)

-0.009 0.004 0.005 -0.132 0.017 0.220

Children <7

-0.423(0.143)**

-0.443(0.343)

0.049(0.017)**

-0.047(0.018)**

-0.002(0.009)

0.053 -0.049 -0.004 0.140 -0.037 -0.046

Children -0.497 -2.415 0.065 -0.010 -0.055 0.075 0.017 -0.092 0.069 -0.014 -0.332

230

7-17 (0.213)* (1.032)* (0.025)** (0.032) (0.023)*

Household Incomes

0.152D-4(0.177D-4)

-0.299D-4(0.483D-4)

-0.160D-5(0.205D-5)

0.279D-5(0.229D-5)

-0.120D-5(0.125D-5)

0.000 0.000 0.000 -0.044 0.015 -0.159

Constant -0.466(0.460)

-4.558(1.028)***

Explanatory variable: y = “How many times in the last twelve months have you seen an art exhibition, opera or theatrical performance?” = 0 (never), 1 (less often) or 2 (daily, several times per week or several times per month).McFadden pseudo R2 = 0.139, χ2 = 233.029***, AIC = 1.263, BIC = 1.579, HQIC = 1.381(*), *, **, *** = significance level 10%,5%,1%,0,1% .Partial derivatives of probabilities with respect to the vector of characteristics are computed at the means of the Xs. Probabilities at the mean vector are Prob(y=0) = 0.134, Prob(y=1) = 0.837, Prob(y=2) = 0.029

231

232