development of an emotional processing scale

Journal of Psychosomatic Res

Development of an emotional processing scale

Roger Bakera, Sarah Thomasa,4, Peter W. Thomasa,b, Matthew Owensa

aDorset Research and Development Support Unit, Poole Hospital NHS Trust, Poole, Dorset, United KingdombInstitute of Health and Community Studies, Bournemouth University, Bournemouth, Dorset, United Kingdom

Received 10 November 2005; received in revised form 4 July 2006; accepted 14 September 2006

Abstract

Objective: The objective of this study was to report on the

development and preliminary psychometric evaluation of an

emotional processing scale, a 38-item self-report questionnaire

designed to identify emotional processing styles and deficits.

Methods: An initial item pool derived from a conceptual model

and clinical observations was piloted on clinical and community

samples (n=150). The resulting 45-item scale was administered to

patients with psychological problems, psychosomatic disorders,

and physical disease, and to healthy individuals (n=460).

Exploratory factor analysis was used to explore the underlying

factor structure. Results: Maximum likelihood factor analysis

0022-3999/07/$ – see front matter D 2007 Elsevier Inc. All rights reserved.

doi:10.1016/j.jpsychores.2006.09.005

4 Corresponding author. Dorset Research and Development Support

Unit, Poole Hospital NHS Trust, Longfleet Road, Poole, BH15 2JB Dorset,

United Kingdom.

yielded an eight-factor solution relating to styles of emotional

experience (Lack of Attunement, Discordant, and Externalized),

mechanisms controlling the experience and expression of emotions

(Suppression, Dissociation, Avoidance, and Uncontrolled), and

signs of inadequate processing (Intrusion). Internal reliability was

moderate to high for six of eight factors. Preliminary findings

suggested satisfactory convergent validity. Discussion: Overall, the

psychometric properties of this scale appear promising. Work is in

progress to refine the scale by incorporating additional items and

by conducting further psychometric evaluations on new samples.

D 2007 Elsevier Inc. All rights reserved.

Keywords: Assessment; Emotions; Emotional processing; Multifaceted; Psychometric; Scale

Introduction

Rachman [1] first introduced the concept of bemotional

processingQ in the context of anxiety disorders. Rachman [1]

defined emotional processing as ba process whereby emo-

tional disturbances are absorbed and decline to the extent

that other experiences and behavior can proceed without

disruptionQ (p. 51). He noted that while bmost people

successfully process the overwhelming majority of disturb-

ing events that occur in their lifeQ (p. 56), sometimes failures

in emotional processing occur.

Rachman argued that if emotional experiences were

incompletely absorbed or processed, then certain direct

signs of this failure would appear (e.g., the return of fears,

obsessions, and unpleasant intrusive thoughts). Further-

more, he proposed that excessive avoidance or prolonged

and rigid inhibition of negative emotional experiences

would prevent their reintegration and resolution. In a more

recent paper, Rachman [2] describes the symptoms of

posttraumatic stress disorder (PTSD) as being partly the

product of failures in emotional processing.

In addition to PTSD [2–4], the concept of emotional

processing may help to explain the emergence or main-

tenance of other psychological disorders, such as panic

disorder [5,6] and depression [7,8]. It may also contribute to

the psychological understanding and treatment of psycho-

somatic conditions, such as fibromyalgia, chronic fatigue,

chronic pain, inflammatory bowel disease, and functional

gastrointestinal disorders [9–16]. In recent years, there has

been an increasing acknowledgement in the literature that

excessive emotion regulation may be related to a number of

major physical illnesses, including cardiovascular disease,

cancer, and arthritis [17–22].

Although the concept of emotional processing seems

clinically useful and relevant, research has been impeded by

the lack of any psychometrically sound assessment instru-

ment that encompasses the different facets of emotional

earch 62 (2007) 167–178

R. Baker et al. / Journal of Psychosomatic Research 62 (2007) 167–178168

processing. There are scales originating from diverse

traditions, such as business and psychoanalysis, that

measure constructs related to emotional processing (e.g.,

emotional intelligence [23–25], psychological mindedness

[26], and alexithymia [27,28]). There are also scales that

measure elements of emotional processing or emotional

processing deficits, such as control [29,30], awareness [31],

ambivalence [32], expressiveness [33–35], regulation

[36,37], and schemas [38].

However, to our knowledge, there is no scale that draws

together various dimensions that may impede or disrupt

emotional processing. This paper describes the preliminary

development of a multifaceted scale, called the Emotional

Processing Scale (EPS). The aim of developing such a scale

was to produce a comprehensive measure that incorporates

Rachman’s original conceptualization of emotional process-

ing with other psychological mechanisms that may impede

emotional processing. The EPS was designed to identify

difficulties in the processing of emotions, not to measure

emotion states (cf., Profile of Mood States [39]) or intensity

of emotions.

The first step, therefore, in developing an assessment

instrument was to develop a model of various domains

underpinning emotional processing. The schematic model

of emotional processing depicted below provides a

conceptual starting point for the development of the EPS

[40,41].

In the model, a negative event is regarded as an input that

needs to be registered (either consciously or unconsciously)

by an individual as a prerequisite for emotional experience.

A negative input could refer to a discrete major event (e.g., a

car crash), a discrete minor event (e.g., an argument), or a

continuous series of events (stressful work environment or

deteriorating marital relationship). The cognitive appraisal

of the meaning of the event [42–44] shapes the nature of the

emotion experienced and is often unconscious and rapid.

Examples of problems that may occur at this stage include

Fig. 1. Model of the main domain

failure to register, misinterpretation, or active avoidance of

an input event.

The experience of emotions is regarded as the central

phase. Possible disruptions in this process include the

following: failure to experience an emotion as a psycho-

logical whole or bGestaltQ [45,46], lack of awareness of

emotional experiences, difficulties in identifying and label-

ing emotions or distinguishing them from bodily sensations

(alexithymia) [47], too much awareness of emotional

feelings [48,49], blocking or blunting of certain emotional

experiences [50], or an inability to link emotional feelings

with the event(s) that triggered them [51–54].

Expression of emotion is seen as an output. Examples of

problems that could arise at this stage include negative

values and beliefs held about expressing emotions [55] or,

conversely, an inability to control strong emotions [56].

Previous research on emotional processing and panic [5]

suggests that it may be important to separate the role of

controlling the experience of emotions and controlling the

expression of emotions. Trying to control, suppress, or

block the experience of an unpleasant emotion is regarded

as perhaps more fundamental and damaging than controlling

the overt expression of emotions [57–59].

Materials and methods

Initial item selection and format

Over a period of 12 years, a pool of approximately 300

draft items/ideas was assembled based on a combination of

the emotional processing model by Baker [40,41] (see

Fig. 1), clinical experience, case histories and autobio-

graphical studies [59], and literature from clinical psychol-

ogy and emotion research. Since this draft item pool was too

large to administer on a single occasion, it was condensed

by selecting those items deemed as the best examples of the

s of emotional processing.

R. Baker et al. / Journal of Psychosomatic Research 62 (2007) 167–178 169

domains specified in Baker’s model. The resulting prelimi-

nary item pool (n=152) was given to clinical, educational,

and research staff. The rationale of the scale was explained,

and suggestions and comments, particularly relating to

clarity, relevance, and redundancy, were encouraged. Feed-

back obtained during this phase led to the removal of 51

items. In addition, minor changes in wording and layout

were made.

Items in the pool related to the processing of general

emotions (bI could not tolerate unpleasant feelingsQ), as wellas of specific emotions (bWhen I was upset or angry, I

smothered my feelingsQ). Items encompassed different

classes of action, as follows: the experience of emotions

(bI tended to experience the same emotion repeatedlyQ);understanding emotions (bAt times I felt pleasant but could

not name the feeling in much detailQ); reactions to situations

(bI reacted too much to what people said or didQ); use of

coping mechanisms (bI harmed myself physically to avoid

thinking about emotional painQ); and behavioral reactions

(bI avoided watching unpleasant scenes on TV or reading

about unpleasant thingsQ). Items described both positive

emotional experience and negative emotional experience;

however, overall, the latter was more numerous. Both

negatively and positively worded items were included

(bWhen emotional or upset, I was able to express my point

of view constructivelyQ; bI could not express my emotionsQ).The keying of items was varied [bI have been able to link

my feelings to things that have happened to me in the last

weekQ (agree=healthy); bMy emotions felt blunt or dullQ(disagree=healthy)].

Scale construction

In order to permit sufficient change to be detected and to

increase the sensitivity of the scale in general, a 10-point

visual analogue format similar to that of the phobic rating

scale of Marks and Mathews [60] and the REHAB

questionnaire of Baker and Hall [61,62] was used. Five

attitudinal anchor statements joined by dotted lines were

placed under a visual analogue line to orient respondents

(completely disagree, disagree, in between, agree, and

completely agree). Respondents were asked to rate how

closely each item described their experiences during the

previous week by marking through the line at any point.

Thirty-two items incorporated a bnot applicableQ response

option to enable respondents to indicate their perceived

nonoccurrence of an emotion-related event during the

previous week (e.g., bLast week I felt angry and did not

know whyQ).Instructions for the scale were worded so as to encourage

respondents to focus specifically on emotional processing

during the last week, thus minimizing recall errors, the

likelihood of responses being based on an individual’s

generalized attitudes, or post hoc rationalizations. This

would permit ratings to be made on a repeated basis, thereby

facilitating the detection of subtle changes over time.

Statistical analysis

Data analysis was undertaken using Statistica ’99 Edition

and SPSS Version 12.0. Critical P was set at .05. Since a

variety of analyses have been used and in order to aid

clarity, they have been described alongside corresponding

results.

Study 1: Working version of the EPS

Participants

The 101-item version of the EPS was administered to

three groups that are expected to vary in their emotional

processing abilities. The first of these (n=37) comprised

patients with psychological problems who had been referred

by their medical practitioner to a clinical psychologist or

counselor (17 males and 20 females). This included patients

with anxiety disorder (n=19), mood disorder (n=8), person-

ality disorder (n=5), adjustment disorder (n=3), eating

disorder (n=1), and sexual disorder (n=1). The second

(n=53) consisted of an opportunistic sample of patients

waiting to see their medical practitioner (13 males and 40

females). Although no data relating to mental and physical

health were collected from this group, it was expected that

this group would contain a significant proportion of

individuals who somatize their problems (surveys estimate

18%) [63], in addition to those with physical conditions.

Finally, the third group consisted of nonpatient adults

(n=60), including staff at a university faculty, a research

center, and a clinical department (15 males and 39 females;

the gender of six participants was not recorded).

Preliminary item analysis

Multiple criteria were used to determine which items to

retain in the working version of the scale, including the

following:

1. Score distribution: Histograms were produced for

each item, and distributions were visually inspected.

To facilitate adequate levels of distinction between

different groups of respondents, items where scores

were distributed relatively normally and encom-

passed the majority of 10 rating points were

generally preferred. However, for certain items, a

moderately nonnormal distribution was expected

(e.g., I felt the urge to smash something).

2. One-way analysis of variance (ANOVA): One-way

ANOVA was used to make groupwise comparisons

for each item. No corrections for multiple signifi-

cance testing were made in this exploratory study to

help minimize type II errors (i.e., discarding useful

items).Thirty-four of 101 items had significant P

values (Pb.05); of these, 20 had Pb.01.


3. Preliminary principal components analysis: Using

Varimax rotation, a principal components analysis of

the 101 items was undertaken for all three groups

combined. Greater preference was given to items

with high factor loadings (z.6).

4. Missing data: Items with substantial levels of

missing data were discarded. Across items, levels

of missing data ranged from 5% to 13%.

5. Content: To ensure that the scale focused primarily

on general emotional processing dimensions rather

than on the processing of specific emotions, the

majority of items referring to particular emotions

(e.g., contentment or guilt) were removed as were

items focusing solely on the strength/frequency of

emotions.

6. bNot applicableQ response: Items requiring a bnotapplicableQ response option were discarded. It was

reasoned that a potential weakness of items requiring

a bnot applicableQ option is that endorsement of bnotapplicableQ could reflect a lack of emotional under-

standing/insight rather than the nonoccurrence of an

emotional event.

Based on these criteria, 45 items were selected for the

working version of the scale.

Table 1

Age breakdown of participants in Study 1

Age category (years)

Group [n (%)]

Patient (n=37) GP (n=53) Nonpatients (n=60)

18–20 1 (3) 3 (6) 6 (10)

21–30 4 (11) 11 (21) 11 (18)

31–40 12 (32) 8 (15) 10 (17)

41–50 13 (35) 8 (15) 15 (25)

51–60 4 (11) 12 (23) 5 (8)

61–70 3 (8) 4 (8) 1 (2)

71–80 – 4 (8) 4 (7)

N81 – 3 (6) 3 (5)

System missing – – 5 (8)

Percentages are rounded up to the nearest integer and, thus, may sum up to

N100%. GP=general practitioner.

Study 2: Validation of the 45 final items

Participants

A total of 460 participants (175 males and 284 females;

the gender of one participant was not recorded) were

recruited from a variety of settings to respond to the 45-

item version of the questionnaire. Participants were asked to

indicate the highest formal qualification obtained.

Participants included the following:

! Nonpatient older adult controls (n=73; male=34,

mean age=61.0 years, S.D.=10.7)

No formal qualifications, n=13

General Certificate of Secondary Education (GCSE)/

Ordinary (OV) levels/Advanced (AV) levels, n=21Degree or higher qualifications, n=23

bOtherQ qualifications (such as City & Guilds), n=16.

! University undergraduate students (n=100; male=12,

mean age=24.3 years, S.D.=7.8, range=18–58)

! Colorectal cancer patients (n=124; male=65, mean

age=68.7 years, S.D.=10.6)


GCSE/OV levels/AV levels, n=26

Degree or higher qualifications, n=12

bOtherQ qualifications, n=24! Individuals with chronic back pain (n=11; male=5,

mean age=48.6 years, S.D.=13.7, range=32–74)


GCSE/OV levels/AV levels, n=3

bOtherQ qualifications, n=5! Individuals with ankylosing spondylitis (n=5; male=

2, mean age=51.4 years, S.D.=16.7, range=33–70)

OV levels/AV levels, n=1

Degree or higher qualifications, n=1

bOtherQ qualifications, n=2Education data missing, n=1

! Individuals referred to a clinical psychologist or a

counselor for a range of mental health problems

(n=147; male=57, 1 gender not recorded, mean

age=37.4 years, S.D.=14.0, range=17–80; age data

for 7 participants missing).

Anxiety disorder, n=32

Adjustment disorder, n=30

Mood disorder, n=12

Sexual disorder, n=1

Eating disorder, n=1

Personality disorder, n=7

Somatoform disorder, n=2

Multiple diagnoses, n=21

Diagnosis missing, n=41

Item analysis

Negatively keyed items were reverse-scored prior to data

analysis so that higher scores represent greater levels of

emotional processing deficits. Two independent-samples

t tests were undertaken to determine which items discrimi-

nated between groups. As in Study 1, no correction was

made for multiple significance testing. In recognition of

differing age distributions in the groups (Table 1), pairwise

comparisons were made: (a) between the undergraduate

student sample and the mental health patient sample, and (b)

between the colorectal cancer sample and the older adult

control sample. Of the 45 items, all but 6 items (Items 15,

28, 31, 33, 40, and 43) showed statistically significant

differences on at least one of two comparisons. Five items

had corrected item–total correlations lower than .20 (Items

5, 15, 27, 31, and 33). Three items were removed prior to


factor analysis because they had low item–total correlations

and failed to show any between-group differences on either

of the independent-samples t tests (Items 15, 31, and 33).

Data suitability

After excluding participants who did not provide

complete responses to all 42 items, we were left with a

Table 2

Maximum likelihood factor analysis with Promax rotation

Item number Description

Factor I: Intrusion

8a So busy with life that I rarely thought about feelings

42 Experienced the same emotion repeatedly

6 Unwanted feelings kept intruding

10 Emotional reactions lasted more than a day

38 Hard to wind down

5 Tried to work out why I felt like I did

1a Night’s sleep dealt with upset feelings

39 Felt anxious/down without knowing why

Factor II: Suppression

44 Bottled up emotions

41 Kept quiet about feelings

20 Smothered feelings

36 Could not express feelings

Factor III: Lack of Attunement

34a Taking notice of feelings helped in making choices

27a Could link feelings to events from previous week

45a When upset could express point of view constructively

32a Allowed self to fully feel emotions

11a Could have described emotions during the week

Factor IV: Uncontrolled

37 Felt urge to smash something

29 Wanted to get back at someone

13 Difficult to control what I said when upset or angry

21 Reacted too much to what people said or did

Factor V: Dissociation

4 Switched off feelings

12 Detached self from feelings

3 Emotions felt blunt/dullb

35 Did not experience emotions when should have

Factor VI: Avoidance

26 Could not tolerate unpleasant emotions

25 Tried to talk only about pleasant things

9 Avoided looking at unpleasant things

Factor VII: Discordant

16 Feelings did not seem to belong to me

14 Feelings confused

18 Wished I could have removed emotions

23 Hard to work out if feeling ill or emotional

24 Seemed to be a big blank in feelings

22 Afraid of strong feelingsb

17 Talking about negative feelings made them worse

Factor VIII: Externalized

19 Emotions were due to biological changes in body

40 Food eaten responsible for emotions

2 Very aware of bodily sensations

Factor loadings in italics indicate the assignment of items to factors. Only loadin

Loadings�100 (decimal points have been omitted).a Reversed item.b Cross-loadingV0.1.

sample of 397 participants, giving a participant-to-item ratio

of almost 10:1, satisfying the criterion of Bryant and

Yarnold [64] that the ratio should be no lower than 5:1.

Prior to factor analysis, the matrix was assessed for

psychometric adequacy. The Kaiser–Meyer–Olkin measure

of sampling adequacy (MSA=.91) was well above the

minimally accepted level (0.50) [65], and Bartlett’s test of

sphericity [66] was highly significant, indicating that the

Factors

I II III IV V VI VII VIII

81

74

69

62 39

52

49

48

32

88

86

68

62

78

68

59

58

48

77

75

54

35

71

59

37 40

37

81

79

53

84

36 50

37 47

46

43

33 40

36

55

47

39

gs z.32 are shown.

Fig. 2. The eight-factor solution applied to the conceptual model.


items were interdependent (v2=6206.20, df=703, Pb.0001).

An anti-image covariance analysis showed that only 64

(4.6%) of the off-diagonal elements were N.09, indicating

that the matrix of covariance of individual items approached

a diagonal. Examination of individual item skew and

kurtosis characteristics (mean skew=0.28, S.D.=0.46,

range=�0.57 to 1.69; mean kurtosis=�0.83, S.D.=0.59,

range=�1.37 to 2.27) confirmed the suitability of the

maximum likelihood factor extraction procedure [67].

Given the suitability of data, we conducted a maximum

likelihood factor analysis [67,68] with Promax rotation

(j=4) to explore the underlying structure of the 42 items and

to identify emotional processing dimensions. An oblique

rotation was chosen to allow for correlation between factors.

The number of factors to retain was evaluated using: (a)

Kaiser’s eigenvalues exceeding unity extraction criterion

[69]; (b) scree plot analysis [70]; and (c) the interpretability

of resulting factor structures [71]. Although seven factors

emerged using the criterion eigenvalueN1, an eight-factor

solution accounting for 58.6% of cumulative variance was

chosen because it was more interpretable and conceptually

cohesive. A minimum loading of .32 was used as a selection

criterion [72]. Of the 42 items, there were four items with

loadings of b.32 (Items 7, 28, 30, and 43). These items were

discarded. Factor analysis was then recomputed. After this

second iteration, all 38 items loaded at z.32. There were

two items (Items 3 and 24) with cross-loadings (defined as

any item that loaded at z.32 on two or more factors with a

difference of V.1). Cross-loadings were not unexpected

given that we expected some conceptual overlap between

facets of the scale; thus, these items were retained (Table 2).

Results

Inspection of the items indicated that Factors 3, 7, and 8

appear to tap into emotional styles comprising attitudes to

emotions, understanding emotions, attributions about emo-

tions, awareness, and labeling. Factors 2, 4, 5, and 6 appear

more connected to the cognitive or behavioral control of

emotions. Factor 1 seems to reflect signs of incomplete or

inadequate processing. The discussion that follows will be

structured according to this conceptual grouping.

Factor 7 seems to relate to an impoverished style of

emotional processing whereby an individual poorly under-

stands his/her emotions (bMy feelings were pretty con-

fusedQ) and does not feel at ease with them (bI wished I

could have removed my emotionsQ). This factor was named

Discordant.

Factor 3 seems to reflect an emotional style whereby the

individual does not attend to his/her emotions [bI allowedmyself to fully feel any emotion that came alongQ; reverse-scored (R)] and is neither able to describe them [bIf you’dsaid to me in the week dWhat emotions are you feeling right

now?T I would’ve been able to tell youQ (R)] nor able to link

them to events [bI have been able to link my feelings to

things that have happened to me last weekQ (R)]. This factorwas named Lack of Attunement.

Factor 8 seems to relate to a style in which the individual

has a somatic orientation and the components of emotion are

attributed to external causes (bI was very aware of bodily

sensationsQ). This factor was called Externalized.

Factors 2, 4, 5, and 6 relate to the control of the input,

experience, and expression of emotions. Factor 2 seems to

Table 3

Cronbach’s a values, mean interitem correlations, and item–total correla-

tions for EPS subscales

Factor

Number

of items n Cronbach’s a

Mean

interitem

correlation

Mean

item–total

correlation

Intrusion 8 436 .83 .38 .55

Suppression 4 441 .82 .53 .64

Lack of

Attunement

5 433 .71 .33 .47

Uncontrolled 4 442 .76 .44 .56

Dissociation 4 435 .70 .38 .49

Avoidance 3 445 .66 .40 .46

Discordant 7 435 .88 .51 .66

Externalized 3 430 .42 .20 .26

EPS total 38 397 .92 .23 .46


involve excessive control of emotions and their expression,

and was named Suppression (bI bottled up my emotionsQ).Factor 4 seems to reflect difficulty in controlling the

expression of emotion (bWhen upset or angry, it was

difficult to control what I saidQ). This factor was labeled

Uncontrolled. Factor 5 appears related to dissociation or

detachment from emotional experience and was labeled

Dissociation. Some items refer to mechanisms by which

dissociation takes place (bI switched off my feelingsQ), andsome refer to the consequences of these (bMy emotions felt

blunt or dullQ). Factor 6 relates to the avoidance of negative

emotions or stimuli that elicit them (bI avoided looking at

unpleasant things, e.g., on TV/in magazinesQ) and was

labeled Avoidance.

Factor 1 reflects signs of incomplete processing accord-

ing to Rachman’s original conceptualization and was called

Intrusion (bI tended to experience the same emotion

repeatedlyQ; bMy emotional reactions lasted more than a

dayQ; bUnwanted feelings kept intrudingQ).Fig. 2 shows the eight-factor solution applied to the

emotional processing model of Baker [40,41].

Scores for the eight factors were computed by summing

the scores for those items comprising each factor and

dividing by the number of items. Items 1, 8, 11, 27, 32, 34,

and 45 were reverse-scored. Thus, for each factor, higher

scores represent poorer levels of emotional processing. An

EPS total score was calculated by summing scores for all

items and dividing by the number of items. Individual factor

scores and total EPS scores were produced so long as 60%

of the respective items were completed.

Demographics

A series of 2�2 ANOVA was conducted on nonpatient–

participant data from Study 2 (older adult controls vs.

students) to see whether there were gender differences on

EPS subscale and total scores. Overall, there were no main

effects of gender apart from the Externalized subscale [F(1,

166)=4.79, P=.03] (with females scoring more highly than

males); there were no main effects of group, with the

exception of the Intrusion subscale [F(1, 166)=15.24,

Pb.001] (with students scoring more highly than older

adults); and there were no Gender�Group interactions.

There was a low-magnitude negative correlation between

EPS total scores and age (r=�.23, Pb.001).

Internal reliability

The coefficient a value for the entire scale was .92 (Table

3). Corrected item–total correlations ranged from r=.18 to

r=.71. Twenty-nine of 38 items had corrected item–total

correlations of z.30. Internal consistency was high (N.80)

for three factors and was moderate for three (N.70). It was

poor for Avoidance (.66) and Externalized (.42); however,

this might be partly due to the fact that these factors contain

fewer items than the other six factors.

Correlations between subscales

Correlations between the eight subscale scores and the

EPS total score were high for most factors, with Avoidance

(.50), Lack of Attunement (.46), and Externalized (.40)

yielding comparatively lower correlations. Correlations

between the eight subscales ranged from moderate to high.

The Avoidance, Externalized, and Lack of Attunement

subscales tended to possess intersubscale correlations lower

than those of the other five.

Study 3: Correlationswith theoretically related constructs

Although emotional processing deficits are implicated in

the development and maintenance of psychiatric disorders,

the EPS was intended to tap a dimension different from

symptomatology. Thus, we hypothesized that the EPS

would be more highly correlated with emotion scales

[Toronto Alexithymia Scale (TAS-20) and Courtauld Emo-

tional Control Scale (CECS)] than with a symptom scale

[Personal Disturbance Scale (sAD)], where a more moderate

correlation was predicted.

We expected EPS subscales to correlate more highly with

Factors 1 and 2 of the TAS-20 than with Factor 3. Factor 3

of the TAS-20 is derived from psychoanalytic concepts

related to poverty of symbolic thinking, and this construct

was not used in the development of the EPS. The CECS

focuses on the control of emotions so we expected the EPS

Suppression factor to correlate significantly with it.

We assessed the convergent validity of the EPS by

computing Pearson Product–Moment Correlation Coeffi-

cients between EPS subscales and a range of measures of

theoretically related constructs.

Participants

Two-hundred eighty-four participants from Study 2

completed the TAS-20, the CECS, and the sAD. This

included 71 students (from the original 100), 124 patients

with colorectal cancer, 73 older adult controls, 11 patients

Table 4

Correlations between EPS scores with theoretically related constructs (n=248)a and between EPS subscales (n=426)b

Intrusion Suppression Lack of Attunement Uncontrolled Dissociation Avoidance Discordant Externalized Total

TAS-20

Factor 1 4044 4744 2144 3944 5244 3244 7144 4444 6944

Factor 2 3244 6744 3544 3644 5144 2744 5944 2844 6644

Factor 3 �1744 4144 4144 09 3444 2544 2344 09 2844

Total 2544 6544 4044 3644 5844 3644 6644 3644 7044

sAD

Anxiety 5144 2344 01 3844 2244 1844 4744 3344 5044

Depression 4244 2944 10 3144 2944 134 4944 2644 4844

Total 5244 2844 05 3944 2844 1844 5444 3344 5544

CECS

Anger 03 5244 1744 �10 1944 1944 154 08 2244

Anxiety 07 5744 2344 06 3144 134 2444 11 3244

Depression 02 4244 164 �05 1944 164 144 11 2144

Total 07 5344 1844 00 2344 1944 2244 154 2944

Intrusion –

Suppression 3344 –

Lack of Attunement 1544 3444 –

Uncontrolled 5844 3544 2744 –

Dissociation 3044 5744 3244 4144 –

Avoidance 2044 3444 03 2944 3744 –

Discordant 6644 5744 3744 6844 6344 4144 –

Externalized 2344 1744 03 2744 2644 2544 3444 –

Total 7344 6844 4644 7644 7144 5044 9244 4044 –

r�100 (decimal points have been omitted).a The sample size differs from 284 participants due to missing data across measures.b The sample size differs from the total sample of 460 participants due to missing data across measures.

4 Pb .05.

44 Pb .01.


with chronic back pain, and 5 patients with ankylosing

spondylitis.

Measures

The CECS

The CECS [29] is a 21-item self-report questionnaire that

assesses the extent to which individuals control their

reactions when angry, sad, or anxious. The scale has been

shown to have good face, discriminant, and construct

validity; to be free from social desirability bias; and to

possess good internal reliability [29,73]. It was selected as a

validity measure for the control/suppression aspects of

emotional processing.

The TAS-20

The TAS-20 [27] is a 20-item self-report scale that

measures three factors: (a) difficulty identifying feelings

(Factor 1); (b) difficulty describing feelings to others

(Factor 2); and (c) external concrete thinking (Factor 3).

The TAS-20 has been shown to have good internal

reliability, test–retest reliability, and construct, convergent,

discriminant, and criterion validity [28,74,75]. Alexithy-

mia is characterized by difficulties in describing or being

aware of one’s emotions, and the TAS-20 was chosen as

a measure for validating such aspects of emotional

processing.

The sAD of the Delusions Symptoms States Inventory (DSSI)

The sAD [76] is a 14-item self-report scale that

measures the severity of anxious and depressed symptoms.

It has been used with both clinical and nonclinical

populations and has been shown to have good construct

and concurrent validity [77,78]. It was chosen to explore

the relationship between emotional processing and affective

symptomatology.

Results

sAD symptom scores correlated moderately with Dis-

cordant [r=.47 (sAD anxiety); r=.49 (sAD depression);

r=.54 (sAD total score)] and Intrusion [r=.51 (sAD

anxiety); r=.42 (sAD depression); r=.52 (sAD total score)]

subscales (Table 4). Correlations with other EPS subscales

were relatively low. Modest correlations between most EPS

subscales and measures of anxiety and depression supported

our expectation that the EPS is not merely a proxy measure

of psychiatric disturbance.

As predicted, the EPS was most strongly correlated

with Factors 1 and 2 of the TAS-20, but less so with

Factor 3. In particular, the Discordant factor correlated

highly with Factor 1 of the TAS-20 (r=.71). Other factors

had much lower correlations with TAS-20 subscales (e.g.,

Avoidance, r=.25–.32; Intrusion, r=�.17–.40; Uncon-

trolled, r=.10–.39), suggesting that the EPS encompasses


a broader spectrum of emotional phenomena than the

TAS-20.

The CECS was, as predicted, more strongly correlated

(r=.53) with the Suppression factor than with other factors

of the EPS.

Study 4: Between-group differences

One would hope that a valid measure of emotional

processing would detect differences between individuals

with psychiatric problems and healthy controls, as was

found in the context of panic disorder [5]. We expected that

a group of mental health patients (n=147) who were part of

our sample from Study 2 would score higher (i.e., exhibit

more deficits) than a group of undergraduate students

(n=100).

Results

Table 5 shows that the mental health sample scored

significantly higher than the students on all EPS subscales.

All but one (Externalized) of these differences remained

when a Bonferroni-corrected critical P (.05/9) was applied

even with age and gender held constant using ANCOVA.

Study 5: Test–retest reliability

Participants

Test–retest reliability was assessed over a 4- to 6-week

period and was based on a subsample of 17 undergraduate

students from Study 2 (mean age=27.1 years, S.D.=8.1,

range=21–56; 14 females).

Table 5

Mean EPS subscale scores for mental health subgroup and student

subgroup (n=241)a

Factor subscale

Group [mean (S.D.)]

t(df) P

Mental health

subgroup

(n=142)

Student

subgroup

(n=99)

Intrusion 6.2 (1.4) 4.8 (1.2) 8.2 (239) b.0001

Suppression 5.2 (2.3) 3.2 (1.9) 7.1 (231.3)b b.0001

Lack of

Attunement

4.9 (1.8) 3.6 (1.2) 6.6 (239.0)b b.0001

Uncontrolled 4.5 (2.3) 2.8 (1.7) 6.4 (237.8)b b.0001

Dissociation 4.0 (1.9) 2.4 (1.4) 7.5 (238.9)b b.0001

Avoidance 4.1 (2.1) 3.0 (1.7) 4.4 (234.3)b b.0001

Discordant 4.86 (2.0) 2.9 (1.6) 8.5 (236.6)b b.0001

Externalized 3.4 (1.9) 2.9 (1.3) 2.2 (239.0)b .03

EPS total 4.9 (1.1) 3.4 (1.0) 10.5 (239) b.0001

a The sample size differs from the total sample of participants (N=247)

due to missing data across measures.b Levene’s test suggested that variances were not equal; modified t test

was used.

Results

The test–retest correlation coefficient obtained for the

entire scale was .79 [95% confidence interval (95%

CI)=0.49 to 0.92; Pb.001]. Test–retest reliabilities for

individual subscales were found to range from moderate

(Externalized: r=.51; 95% CI=0.04 to 0.80; P=.04) to high

(Discordant: r=.88; 95% CI=0.69 to 0.96; Pb.001), with

the exception of the Uncontrolled (r=.38; 95% CI=�0.12to 0.73; P=.13) and Intrusion subscales (r=.30; 95%

CI=�0.22 to 0.68; P=.25). However, the confidence

intervals for these correlations are wide because of the

small sample size.

Study 6: Treatment sensitivity

Participants

The treatment sensitivity of the EPS was evaluated in a

subsample from Study 2 comprising individuals referred to

an adult mental health team (n=23; mean age=40.3 years,

S.D.=12.4, range=17–67; 15 females). Seventeen of the

participants were seen by clinical psychologists, and six

were seen by counselors. Individuals presented with a range

of diagnoses, including anxiety disorder (n=8), adjustment

disorder (n=6), mood disorder (n=2), somatoform disorder

(n=1), multiple diagnoses (n=4), and diagnosis missing

(n=2).

Results

Paired t tests were conducted to measure treatment

sensitivity. The t tests indicated significant changes from

pretreatment to posttreatment for the EPS total score, and for

five of eight subscales. Overall, the most significant changes

were found in the EPS total score [pre x̄=4.7 (S.D.=1.3);

post x̄=3.8 (S.D.=1.5)], Suppression factor [pre x̄=5.5

(S.D.=2.2); post x̄=4.1 (S.D.=2.6)], and Intrusion factor

[pre x̄=5.8 (S.D.=2.0); post x̄=4.6 (S.D.=1.8)] (all Pb.01).

When a Bonferroni-corrected critical P (.05/9) was applied,

significant effects remained in the case of Suppression,

Intrusion, and the EPS total score. However, the small

sample size means that these results should be interpreted

cautiously.

Discussion

This paper describes the development and initial vali-

dation of a new scale called the EPS. It is a 38-item self-

report scale that comprises eight factors related to emotional

processing deficits. The factor structure supported the

emotional processing model reasonably well, particularly

concerning the control of input, experience, and expression

of emotion.


Three of the factor subscales describe styles of emo-

tional experience (Discordant, Lack of Attunement, and

Externalized). Discordant is a style in which the individual

poorly understands his/her emotions and does not feel at

ease with them. It would appear to be related closely to

the first two factors in the TAS-20: ability to identify

emotions and ability to describe emotions (r=.71 and .59,

respectively).

The Lack of Attunement factor represents an attitude to

emotions in which they are not seen as useful and their

meaning is not understood. This may conceptually overlap

with the bpsychological mindednessQ construct [79]. How-ever, it is possible that this factor may be a product of a social

desirability response bias or a methodological artifact due to

a reversed-item response set [80], rather than a reflection of a

processing dimension per se. In future analyses, we intend to

utilize a measure of social desirability, such as the Marlowe–

Crowne Social Desirability Scale, to establish whether the

EPS is free from social desirability bias [81].

The Externalized factor is an orientation in which the

person is very aware of bodily aspects of emotion but

attributes them to external causes. Although this factor

requires further work, it could capture an important aspect

of somatization and relate to the development and reporting

of medical symptoms [17,82].

The Intrusion factor is comprised of items that provide an

index of inadequate emotional processing, according to

Rachman’s original framework, including intrusive emo-

tional experiences, the persistent and repeated nature of

negative emotional experiences, and the slow rate of decay

of unpleasant affective feelings.

There were four factors describing mechanisms related to

the control of the input, experience, and expression of

emotions (Suppression, Dissociation, Avoidance, and

Uncontrolled). Much of the literature on emotional control

and dysregulation refers to a single concept of emotional

suppression. A distinction between these four types of

control may help elucidate the underlying mechanisms.

These factors will have particular relevance to the growing

research literature on intrusive thoughts and the paradoxical

effects of deliberate suppression [83].

Internal reliability was moderate to high on most factors,

but was poorer for the Avoidance and Externalized

subscales. In an attempt to improve the internal reliability

of the Externalized and Avoidance factors, additional items

have been devised.

Most of the subscales showed satisfactory test–retest

reliability and sensitivity to change, although caution is

warranted given the small sample sizes used. Further

evaluations of test–retest and sensitivity to change using

larger samples are required.

Evidence for the convergent validity of the EPS was

obtained from significant correlations with self-report

measures of constructs related to emotional processing. As

would be expected, the EPS correlated most highly with the

most conceptually similar scale (the TAS-20) and less so

with measures of emotional control (CECS) and anxiety and

depression (sAD).

Scores on the EPS distinguished between groups that one

might expect to differ in terms of emotional processing

deficits. Mental health clients scored more highly than

students on all subscales.

In sum, these preliminary findings suggest that the 38-

item EPS holds promise as a measure of emotional

processing. It appears to assess a range of emotional

processes broader than that of other emotion scales and

appears to have potential application for both psychological

and physical disorders. Data are now being collected for a

revised version of the scale with 13 additional items in

preparation for further psychometric evaluation.

Acknowledgments

We would like to thank Jane Holloway (formerly of the

Dorset Research and Development Support Unit) for her

help during the early stages of the project. We would like to

thank and acknowledge the support of Mr. Tim Hollingbury

and the clinical psychologists and counselors at the Depart-

ment of Psychological Therapies, Dorset HealthCare NHS

Trust; the clinical psychologists at the Royal Cornhill

Hospital and Grampian HealthCare NHS Trust; the research

staff at the Institute of Health and Community Studies and

Andy King and Rosalie Lasby, Academic Services Bourne-

mouth University; the Department of Psychology, Aberdeen

University; Mrs. Sandra Horne and Dr. Sharon Lothian

(University of Southampton); Dr. Paul Thompson and Dr.

Selwyn Richards (Consultant Physicians, Department of

Rheumatology, Poole Hospital NHS Trust); Dr. Tamas

Hickish (Consultant Physician, Royal Bournemouth and

Poole NHS Trusts); Dr. Simon Pennell (Southbourne

Surgery, Boscombe); The Dorset Primary Care Trusts; and

the Southwest Regional Health Authority.

References

[1] Rachman S. Emotional processing. Behav Res Ther 1980;18:51–60.

[2] Rachman S. Emotional processing, with special reference to post-

traumatic stress disorder. Int Rev Psychiatry 2001;13:164–71.

[3] Brewin CR, Dalgleish T, Joseph S. A dual representation theory of

post traumatic stress disorder. Psychol Rev 1996;103:670–86.

[4] Foa EB, Riggs DS. PTSD following assault: theoretical considerations

and empirical findings. Curr Dir Psychol Sci 1995;4:61–5.

[5] Baker R, Holloway J, Thomas PW, Thomas S, Owens M. Emotional

processing and panic. Behav Res Ther 2004;42:1271–87.

[6] Berg CZ, Shapiro N, Chambless DL, Ahrens AH. Are emotions

frightening? II An analogue study of fear of emotion, interpersonal

conflict and panic onset. Behav Res Ther 1998;36:3–15.

[7] Hunt MG. The only way is through: emotional processing and recovery

after a depressing life event. Behav Res Ther 1998;36:361–84.

[8] Teasdale JD. Emotional processing, three modes of mind and the

prevention of relapse in depression. Behav Res Ther 1999;37:S53–77.

[9] Bucci W. Symptoms and symbols: a multiple code theory of

somatisation. Psychoanal Enquiry 1997;12:151–72.


[10] Brosschot JF, Aarsse HR. Restricted emotional processing and somatic

attribution in fibromyalgia. Int J Psychiatry Med 2001;31:127–46.

[11] Raleigh J. Fibromyalgia and emotional processing. [master’s dis-

sertation]. Southampton (UK): University of Southampton, 2004.

[12] Keefe FJ, Lumley M, Anderson T, Lynch T, Carson KL. Pain and

emotion: new research directions. J Clin Psychol 2001;57:587–607.

[13] Porcelli P, Zaka S, Leoci C, Centonze S, Taylor GI. Alexithymia in

inflammatory bowel-disease—a case–control study. Psychother Psy-

chosom 1995;64:49–53.

[14] Verı́ssimo R, Mota-Cardosa R, Taylor G. Relationships between

alexithymia, emotional control, and quality of life in patients with

inflammatory bowel disease. Psychother Psychosom 1998;67:61–70.

[15] Porcelli P, Bagby RM, Taylor GJ, De Carne M, Leandro G,

Todarello O. Alexithymia as predictor of treatment outcome in

patients with functional gastrointestinal disorders. Psychosom Med

2003;65:911–8.

[16] Coughlin Della Selva P. Emotional processing the treatment of

psychosomatic disorders. J Clin Psychol 2006;62:539–50.

[17] Pennebaker JW. The psychology of physical symptoms. New York7

Springer, 1982.

[18] Todarello O, Taylor GJ, Parker JDA, Fanelli M. Alexithymia in

essential hypertensive and psychiatric outpatients: a comparative

study. J Psychosom Res 1995;39:987–94.

[19] Jensen MR. Psychobiological factors predicting the course of breast

cancer. J Pers 1987;55:317–42.

[20] Temoshok L. Personality, coping style, emotion and cancer: towards

an integrative model. Cancer Surv 1987;6:544–67.

[21] Gross JJ. Emotional expression in cancer onset and progression. Soc

Sci Med 1989;28:1239–48.

[22] Udelman HD, Udelman DL. Emotions and rheumatologic disorders.

Am J Psychother 1981;35:576–87.

[23] Bar-on R. Bar-On Emotional Quotient Inventory (EQ-i): technical

manual. Toronto7 Multi-Health Systems, 1997.

[24] Mayer JD, Salovey P, Caruso DR. MSCEIT item booklet, version 20.

Toronto7 Multi-Health Systems, 2000.

[25] Tapia M. Measuring emotional intelligence. Psychol Rep

2001;88:353–64.

[26] Conte HR, Ratto R, Karasu TB. The Psychological Mindedness Scale:

factor structure and relationship to outcome of psychotherapy. J

Psychother Pract Res 1996;5:250–9.

[27] Bagby RM, Parker JDA, Taylor GJ. The 20-item Toronto Alexithymia

Scale: I Item selection and cross-validation of the factor structure. J

Psychosom Res 1994;38:23–32.

[28] Bagby RM, Taylor GJ, Parker JDA. The 20-item Toronto Alexithymia

Scale: II Convergent, discriminant, and concurrent validity. J

Psychosom Res 1994;38:33–40.

[29] Watson M, Greer S. Development of a questionnaire measure of

emotional control. J Psychosom Res 1983;27:299–305.

[30] Roger D, Nesshoever W. The construction and preliminary validation

of a scale for measuring emotional control. Pers Individ Differ

1987;8:527–34.

[31] Lane RD, Quinlan DM, Schwartz GE, Walker PA, Zeitlin SB. The

levels of emotional awareness scale: a cognitive–developmental

measure of emotion. J Pers Assess 1990;55:124–34.

[32] King LA, Emmons RA. Conflict over emotional expression—

psychological and physical correlates. J Pers Soc Psychol

1990;58:864–77.

[33] King LA, Emmons RA. Psychological, physical, and interpersonal

correlates of emotional expressiveness, conflict, and control. Eur J

Pers 1991;5:131–50.

[34] Gross JJ, John OP. Revealing feelings: facets of emotional expressiv-

ity in self-reports, peer ratings and behavior. J Pers Soc Psychol

1997;72:435–48.

[35] Stanton AL, Kirk SB, Cameron CL, Danoff-Burg S. Coping through

emotional approach: scale construction and validation. J Pers Soc

Psychol 2000;78:1150–69.

[36] Mayer JD, Stevens AA. An emerging understanding of the reflective

(meta-) experience of mood. J Res Pers 1994;28:351–73.

[37] Gratz KL, Roemer L. Multidimensional assessment of emotion

regulation and dysregulation: development, factor structure, and initial

validation of the difficulties in emotion regulation scale. J Psychopa-

thol Behav Assess 2004;26:41–54.

[38] Leahy RL. A model of emotional schemas. Cogn Behav Pract 2002;9:

177–90.

[39] Albrecht R, Ewing SJ. Standardizing the administration of the Profile

of Mood States (POMS): development of alternative word lists. J Pers

Assess 1989;53:31–9.

[40] Baker R. An emotional processing model for counselling and

psychotherapy: a way forward? Couns Pract 2001;7:8–11.

[41] Baker R. Model of emotional processing, 2004. Available from:http://

www.emotionalprocessing.org.uk/What%20is%20Emotional%20Pro-

cessing/EP%20Model/EP%20Model_files/frame.htm [accessed July

25, 2005].

[42] Kelly GA. A theory of personality. New York7 Norton, 1963.

[43] Lazarus RS. Emotion and adaptation. New York (NY)7 Oxford

University Press, 1991.

[44] Scherer KR. Appraisal theory. In: Dagleish T, Power M, editors.

Handbook of cognition and emotion. Chichester (UK)7 John Wiley

and Sons, Ltd., 1999. pp. 637–63.

[45] Frijda NH. The emotions (studies in emotion and social interaction). .

New York7 Cambridge University Press, 1986.

[46] Gendlin ET. Focusing-oriented psychotherapy A manual of the

experiential method. . New York7 The Guilford Press, 1996.

[47] Taylor GJ, Bagby RM, Parker JDA. The alexithymia construct—a

potential paradigm for psychosomatic-medicine. Psychosomatics

1991;32:153–64.

[48] Barsky AJ. Amplification, somatisation and the somatoform disorders.

Psychosomatics 1992;33:28–34.

[49] Salkovskis PM, Clark DM. Panic disorder and hypochrondriasis. Adv

Behav Res Ther 1993;15:23–48.

[50] Weinberger DA. The construct validity of repressive coping. In:

Singer JL, editor. Repression and dissociation: implications for

personality theory, psychopathology, and health. Chicago7 University

of Chicago Press, 1990. pp. 337–86.

[51] Conway MA, Bekerian DA. Situational knowledge and emotions.

Cogn Emotion 1987;1:145–91.

[52] Epstein S. The ecological study of emotions in humans. In: Pliner P,

Blankstein KR, Spigel IM, editors. Advances in the study of

communication and affect: Perception of emotions in self and others,

vol 5. New York7 Plenum, 1979. pp. 47–83.

[53] Leventhal H. The integration of emotion and cognition: a view from

the perceptual–motor theory of emotion. In: Clarke MS, Fiske ST,

editors. Affect and cognition: the 17th annual Carnegie Symposium on

cognition. Hillsdale (NJ)7 Erlbaum, 1982. pp. 121–56.

[54] Rogers CR. A process conception of psychotherapy. In: Rogers CR,

editor. On becoming a person: a therapist’s view of psychotherapy.

London7 Constable and Robinson, 1961. pp. 125–59.

[55] Kennedy-Moore E, Watson JC, editors. Expressing emotion: myths

realities and therapeutic strategies. New York7 Guilford Press, 2002.

pp. 16.

[56] Tavris C. On the wisdom of counting to ten: personal and social

dangers of anger expression. In: Shaver P, editor. Review of

personality and social psychology: Emotions, relationships, and

health, vol 5. Beverly Hills (Calif)7 Sage, 1984. pp. 170–91.

[57] Gross JJ. Antecedent- and response-focused emotion regulation:

divergent consequences for experience, expression, and physiology.

J Pers Soc Psychol 1998;74:224–37.

[58] Parkinson B, Totterdell P. Classifying affect—regulation strategies.

Cogn Emotion 1999;13:277–303.

[59] Baker R. Personal accounts of panic. In: Baker R, editor. Panic

disorder: theory, research and therapy. Chichester7 John Wiley, 1989.

pp. 66–88.

http://www.emotionalprocessing.org.uk/What%20is%20Emotional%20Processing/EP%20Model/EP%20Model_files/frame.htm


[60] Marks IM, Mathews AM. Brief standard self-rating for phobic

patients. Behav Res Ther 1979;17:263–7.

[61] Baker R, Hall JN. REHAB: a new assessment instrument for chronic

psychiatric patients. Schizophr Bull 1988;14:97–111.

[62] Baker R, Hall JN. A review of the applications of the REHAB

assessment system. Behav Cogn Psychother 1994;22:211–31.

[63] Bridges K, Goldberg D, Evans B. Determinants of somatisation in

primary care. Psychol Med 1991;21:473–83.

[64] Bryant FB, Yarnold PR. Principal components analysis and explor-

atory and confirmatory factor analysis. In: Grimm LB, Yarnold PR,

editors. Reading and understanding multivariate statistics. Washington

(DC)7 APA Press, 1995. pp. 99–136.

[65] Kaiser HF, Rice J. Little jiffy, Mark IV. Educ Psychol Meas 1974;34:

111–7.

[66] Dziuban CD, Shirkey EC. When is a correlation matrix appropriate for

factor analysis? Psychol Bull 1974;6:358–61.

[67] Fabrigar IW, Wegener DT, Maccallum RC, Strahan EJ. Evaluating the

use of exploratory factor analysis in psychological research. Psychol

Methods 1999;3:272–99.

[68] Costello AB, Osborne JW. Best practices in exploratory factor

analysis: four recommendations for getting the most from your

analysis. Pract Assess Res Eval 2005;10: Available from: http:// www.

pareonline.net/getvn.asp?v=10&n=7.

[69] Kaiser HF. A note on Guttman’s lower bound for the number of

common factors. Multivariate Behav Res 1961;1:249–76.

[70] Cattell R. The scree test for the number of factors. Multivariate Behav

Res 1966;1:245–76.

[71] Gorsuch RL. Factor analysis. Mahwah7 Lawrence Erlbaum, 1983.

[72] Tabachnick BG, Fidell LS. Using multivariate statistics. Boston7 Allyn

and Bacon, 2001.

[73] Walker LG. The measurement of anxiety. Postgrad Med J 1990;66:

S11–7.

[74] Parker JDA, Bagby RM, Taylor GJ, Endler NS, Schmitz P. Factorial

validity of the 20-item Toronto Alexithymia Scale. Eur J Pers

1993;7:221–32.

[75] Parker JDA, Taylor GJ, Bagby RM. The 20-item Toronto Alexithymia

Scale: III Reliability and factorial validity in a community population.

J Psychosom Res 2003;55:269–75.

[76] Henry JD, Crawford JR, Bedford A, Crombie C, Taylor EP. The

Personal Disturbance Scale (sAD): normative data and latent structure

in a large non-clinical sample. Pers Individ Differ 2002;33:1343–60.

[77] Bedford A, Foulds GA, Sheffield BF. A new Personal Disturbance

Scale (DSSI/sAD). Br J Soc Clin Psychol 1976;15:387–94.

[78] Bedford A, Deary IJ. The Personal Disturbance Scale (DSSI/sAD):

development, use and structure. Pers Individ Differ 1997;22:493–510.

[79] Appelbaum SS. Psychological-mindedness: word, concept, and

essence. Int J Psychoanal 1973;54:35–46.

[80] Knight RG, Chisholm BJ, Marsh NV, Godfrey HP. Some normative,

reliability, and factor analytic data for the revised UCLA Loneliness

Scale. J Clin Psychol 1988;44:203–6.

[81] Crowne DP, Marlowe D. A new scale of social desirability

independent of psychopathology. J Consult Clin Psychol 1960;24:

349–54.

[82] De Gucht V, Willem H. Alexithymia and somatisation: quantitative

review of the literature. J Psychosom Res 2003;54:425–34.

[83] Rassin E. Thought suppression. London7 Elsevier, 2005.

http://www.pareonline.net/getvn.asp?v=10&n=7

development of an emotional processing scale

Documents