research methods outline meta-analysis correlation quasi-experiments discuss questionnaire data...

23
Research Methods Outline Meta-analysis Correlation Quasi-Experiments Discuss Questionnaire Data Observational Research Activity: Observational Research

Upload: sophia-banks

Post on 02-Jan-2016

220 views

Category:

Documents


1 download

TRANSCRIPT

Research MethodsOutline Meta-analysis Correlation Quasi-Experiments Discuss Questionnaire Data Observational Research Activity: Observational Research

Meta-analysis A set of statistical procedures for combining the

results of a number of studies in order to provide a general assessment of the relationship between the variables

Tells us if there is a difference between groups

Tells us how strong the finding is, the effect size

Meta-Analysis Examples Example: Participation

rate & leader emergence (Mullen, Salas, & Driskell, 1989)

One of the strongest determinants of who emerges leader in a group is the member with the highest participation

Example: Social loafing (Karau & Williams, 1989)

Social loafing is greater for males than females and more pronounced as the size of the group increases

SOCIAL

Correlational Research Method Researchers gather a set

of observations about a group of people and test for associations between different variables

Tells you whether 2 variables are associated with each other systematically

Cannot prove cause-effect X could cause Y Y could cause X Z could cause X

and Y (third variable problem)

Correlation Measures the degree and

direction of linear relationship between 2 variables

Two parts to a correlation

Strength

Direction Positive or negative

The Pearson Correlation Coefficient (r)

Closer to 1 or –1 means stronger relationship

Positive value indicates positive relationship, negative value is negative relationship

Positive Correlations As one variable goes up the other one goes up as well

(or as one goes down the other goes down)

STRONG (r = 1.0) WEAKER ( r= .60)

Negative Correlations As one variable goes up the other one goes DOWN (or as one goes down the other goes up)

STRONG (r = -1.0) WEAKER ( r= -.60)

Quasi-Experiments Sometimes known as queasy-experiments

Can give experimental purists a queasy feeling

Researcher has only partial control over her/his independent variables Participants assigned to 1 or more conditions by means other

than random assignment

A correlational method in which real-world groups are compared on a DV.

Quasi-ExperimentResearch Example

Out-group homogeneity effect Perceive out-group as more

homogeneous (less diverse) than the in-group

In-group bias Tendency to more favorably

evaluate members of In-group

EXAMPLE: Park and Rothbart (1982)

3 sororities Rate own and other 2

Results: Clear evidence of out-group

homogeneity effect Clear evidence of in-group bias

Yet another example… Stereotype Threat (Steele)

Stereotyped group is aware of the stereotype

Creates apprehension and impaired performance

EXAMPLE: Steele & Aronson, 1994

Difficult verbal test: Black & White Stanford students

Two conditions: Testing intellectual ability Testing the test

Results: When testing the test: B & W

performed equally well When testing IQ= B poorer

than W

Class Questionnaire Data Develop Hypotheses with these variables:

Comfort with working in groups and cooperation in last student project

Years playing team sports and preference to work in a group Gender and teamwork attitude Gender and leadership in student project Gender and number of years in team sports

Class Questionnaire Data Correlation

Group Comfort and cooperation in group tasks

r= -.27 Years and preference for

groups r = + .33

Quasi-experiment Gender & teamwork attitude

No difference Women and men: 5.15

Gender & leadership in student project

Men report more leadership experience

Men: 5.85, Women: 4.88 Gender & number of year of

team experience Men report more years of

team experience Women: 10.15, men:12.77

Final Questionnaire Issues Social desirability bias

A bias resulting from participants giving responses that make them look good rather than giving honest responses

You completed 2 different scales: LIKERT scale: Items typically asking whether participants

strongly agree, agree, disagree, strongly disagree

SEMANTIC DIFFERENTIAL scale: Participants rate a given concept on a series of 7-point bipolar attitude scales

Summarizing open-ended question responses

Observational Research Observational methods

Nonexperimental methods in which observers watch groups to gather information

Methods vary with respect to Degree to which observer is part of a group Degree to which observations are structured

Observational Research

Gender, group size, and amount of beer consumed (Geller, Russ, & Altomari)

Observed 56 females 187 males at local bars (students 18-25 years)

Results Drank more in a group Males drank more than females

Function of container type

Unstructured Observational Methods

Unstructured: Observers offer impressionistic, descriptive

accounts of the group

Participant observation Researcher becomes a member of the group

being studied

Unstructured Observational Methods Participant observation examples:

Whyte (1943) Joined Italian American gang in Boston

Festinger, Reiken, & Schachter (1956) Joined ‘doomsday’ cult, pretended to believe in world

ending Ezekiel (1995)

Attended neo-Nazi and KKK groups to get an insider’s look Martel (2001)

Joined Jewish Harley Davidson riders Bock (2000)

Was member of ‘single moms by choice’ support group

Criticisms of Participant Observation By joining the group, they change it in some way Groups are unaware they are being observed

May be unethical People don’t liked to be spied on, feel betrayed,

foolish Observer bias Hard to quantify unstructured data

Structured Observational Methods Structured

Quantitative methods in which group behaviors are observed and recorded with objective system

Researcher must meticulously develop coding system Code things relevant to research Coding system is lens of research

Interrater reliability Extent to which the ratings of different observers are in

agreement

Interaction Process Analysis

Observational coding system developed by Bales to measure 6 task and 6 socioemotional activities in a group

Interaction Process Analysis

Example (Hutson-Comeaux & Kelly, 1996)

Investigated whether female and male groups of college students differed in their interaction style

Found that females engaged in more positive socioemotional behavior and males engaged in more task-oriented behavior

Class Activity: Studying Groups Communication and

content analysis

Castle building exercise

Observational Research Discussion Use your data to draw conclusions about the

groups structure and process Use the information from the communication analysis to

identify the group’s leader as well as other members who were more or less active in the activity.

Use the information from the content analysis to identify the overall content of the groups discussion. (Items 1-3 and 10-12 are socioemotional activities, 4-9 are task activites, see Forsyth, p.32)

Report your findings to the class