Unlock document.
This document is partially blurred.
Unlock all pages and 1 million more documents.
Get Access
18
around a large room reading posters and talking to the authors about their poster. On the day of the final
exam we will have a mini-conference at which each of you will display your posters and each of you will
respond in writing to some questions about several posters you will read. We hope you will have fun doing
this assignment.
19
Evaluation Sheet for Poster Session
1. What question is being asked or what hypothesis is being tested in the experiment described on this
poster?
2. Are the hypotheses and expected results clearly explained? Why or why not?
3. What conclusion does the author of this poster reach on the basis of the proposed results of this
experiment?
4. Identify two strengths of the proposed experiment or positive aspects of the poster that you would like to
tell the author of the poster.
6. Please make any other comments or suggestions you think would be helpful to the author of the poster.
20
INSTRUCTORS’ LECTURE/DISCUSSION AIDS
The following pages may be used to facilitate class discussion or lecture. Most of the pages elaborate
Kassin, Goldstein, and Savitsky=s (2003) study on interrogations of guilty and innocent suspects. Because
this study is described several times in Chapter 8, elaborating on their findings in class may help students to
understand main effects and interaction effects.
1. Complex Designs: This page identifies the key features of complex designs and extends an
independent groups design from Chapter 6.
2. Complex Designs, continued: This page defines main effect and interaction effects.
3. Introduction to Research Example: Key features of the Kassin, Goldstein, and Savitsky (2003) study
are identified on this page.
4. Research Example, continued: This page illustrates the factorial combination of the two independent
variables in the Kassin et al. design.
5. Main Effects of Independent Variables: Two main effects in the Kassin et al. study are analyzed on this
page.
6. Interaction Effects: This page details ways to identify interaction effects.
7. Guidelines for Analysis of Two-Factor Experiment: This page has a flow diagram for decisions
regarding the analysis of a complex design, including simple main effects and comparison of two
means.
8. Simple Main Effects: This page uses figures to illustrate simple main effects in the Kassin et al.
experiment.
9. Simple Main Effects, continued: The last two simple main effects in the 2 2 design are identified.
10. Interaction Effects and External Validity: This page outlines the different decisions made regarding
external validity for when an interaction is or is not present.
11. Interaction Effects and Ceiling/Floor Effects: The problem associated with interpreting an interaction
effect when ceiling or floor effects are present is described on this page.
12. Interaction Effects and the Natural Groups Design: This page outlines the steps for using complex
designs to make a causal inference regarding a natural groups variable.
21
Complex Designs
! Experiments that involve
2 or more independent variables studied simultaneously
at least 1 dependent variable
! Simplest complex design
2 independent variables, each with 2 levels
1 dependent variable
! Factorial combination of independent variables (IVs)
Pair each level of one IV with each level of other IV(s)
! Example: Dittmar, Halliwell, & Ive (2006)
1st IV: Version of picture book (Barbie, Emme, neutral)
2nd IV (natural groups): Grade in school (Kindergarten, 1st, 2nd)
a 3 3 complex design B 2 IVs, each with 3 levels
9 conditions
Version of Picture Book
Barbie
Emme
Neutral
Kindergarten
1
2
3
Grade
1st
4
5
6
2nd
7
8
9
! Factorial combination allowed Dittmar et al. to examine
Overall effect of Version of Picture Book
$ Barbie images caused greater body dissatisfaction than Emme and
neutral images.
Overall effect of Grade
$ Body dissatisfaction increased as grade in school increased (a
22
correlation due to natural groups variable)
Combined effect of both IVs together
$ Interesting effects for combination of grade and exposure to the
images
Complex Designs, continued
! Guidelines for identifying complex designs
Experiment with at least 2 independent variables (IVs)
IVs can be independent groups designs
$ random groups, natural groups, matched groups
IVs can be repeated measures designs
$ Amixed design@ combines independent groups design and
repeated measures design
! Main effect
Overall effect of an independent variable in a complex design
$ Effect on DV as if only that IV was studied
! Interaction effect
Combined effect of IVs considered simultaneously
Effect of an independent variable differs depending on the level of a
second independent variable
23
Introduction to Research Example
! Kassin, Goldstein, and Savitsky (2003): Psychology and law
Research questions
$ Do interrogators= expectations about a suspect=s guilt or
innocence influence the interrogation tactics they use?
$ Do interrogators have a confirmation bias in which their initial beliefs
about a suspect=s guilt cause them to interrogate more
aggressively?
! 2 2 complex design
Interrogator Expectation (random groups)
$ Guilty expectation
$ Innocent expectation
Suspect Status (random groups)
$ Actual guilt
$ Actual innocence
! Students participated as interrogators or suspects in a laboratory Amock
crime.@
! Dependent Variables (there were many)
$ Number of guilt-presumptive questions the interrogator chooses for
interview with suspect
$ Number of persuasive interrogation techniques used during interview
with suspect
$ Ratings of amount of effort interrogator used to obtain a confession
24
Research Example, continued
! Factorial combination of 2 2 design
Interrogator Expectation
Guilty
Innocent
Suspect
Status
Actual
Guilt
Interrogators
believed suspect
was guilty and
suspect actually
committed the crime
Interrogators
believed suspect
was innocent and
suspect actually
committed the crime
Actual
Innocence
Interrogators
believed suspect
was guilty and
suspect did not
actually commit the
crime
Interrogators
believed suspect
was innocent and
suspect did not
actually commit the
crime
25
Main Effects of Independent Variables
! Main effects: The effect of one independent variable collapsing across the
effect of another independent variable
! Two possible main effects in the Kassin et al. (2003) study
$ Main effect of Interrogator Expectation
$ Main effect of Suspect Status
! Main effect of Interrogator Expectation
$ Compare Guilty Expectation and Innocent Expectation
$ DV: Mean number of guilt-presumptive questions selected
Interrogator Expectation
Guilty
Innocent
Suspect
Status
Actual
Guilt
3.54
2.54
Actual
Innocence
3.70
2.66
Means for
Interrogator Expectation
3.62
2.60
! Main effect of Suspect Status
$ Compare Actual Guilt and Actual Innocence
$ DV: Mean number of persuasive techniques
Interrogator Expectation
Guilty
Innocent
Means for
Suspect
Status
26
Suspect
Status
Actual
Guilt
7.71
6.59
7.15
Actual
Innocence
11.96
10.88
11.42
Interaction Effects
! Effect of one independent variable differs depending on the level of a
second independent variable
! Subtraction method: An initial approach to examining an interaction
! Interaction effect for Interrogator Expectation Suspect Status
$ DV: Mean rating of effort to obtain confession
Interrogator Expectation
Guilty
Innocent
Suspect
Status
Actual
Guilt
5.64
5.56
Actual
Innocence
7.17
5.85
Difference
Between Means
-1.53
-0.29
$ These mean differences are not the same, suggesting the presence of
an interaction effect.
! Graphs (AFigures@) can be used to observe interaction effects
$ Plot means for IV conditions
27
$ Interaction effect is likely if
$ Lines are not parallel (i.e., they converge, diverge, cross)
$ Interaction effect is not likely if
$ Lines are parallel
! Use NHST to determine if interaction and main effects are statistically
significant.
$ Analysis of Variance (ANOVA)
Guidelines for Analysis of Two-Factor Experiment
! Interaction: Effect of one IV differs depending on the level of 2nd IV
! Main effect: Overall effect of an IV, collapsed across 2nd IV
Is the interaction effect
statistically significant?
No
Are the main effects
statisically
significant?
No
Stop.
Yes
Compare
two means.
Yes
Are the simple main
effects statisically
significant?
No
Stop.
Yes
Compare two means.
28
! Simple main effect: Effect of one IV at one level of 2nd IV
! Comparison of two means: Identify specific source of interaction or main
effect (with more than 3 levels) by comparing means two at a time
29
Simple Main Effects
! Identify the source of an interaction effect
! Examine effect of IV at each level of 2nd IV
$ Simple main effects should differ
! Simple main effect of Suspect
Status for the Guilty
Expectation condition
$ Statistically significant simple
main effect
Mean
rating
for effort
to obtain
confession
! Simple main effect of Suspect
Status for the Innocent
Expectation condition
$ simple main effect not
statistically significant
Interrog ator Expe ctation
Gu ilty Innocent
1
2
3
4
5
6
7
8
9
10
Suspect Statu s
Actu al Guilt
Actu al Innoce nce
Interrog ator Exp ectatio n
Gu ilty Innocent
4
5
6
7
8
9
10
Suspect Status
Actu al G uilt
Actu al Inn ocence
30
Mean
rating
for effort
to obtain
confession
Simple Main Effects,
continued
! Two more simple main
2 design effects in this 2
$ Simple main effect of
Interrogator Expectation for
Guilty Suspects
$ Simple main effect of
Interrogator Expectation for
Innocent Suspects
$ Which is likely to be statistically significant?
Mean
rating
for effort
to obtain
confession
Interro gator Expectatio n
Gu ilty Inno cent
1
2
3
4
5
6
7
8
9
10
Suspect Status
Actual Guilt
Actual Inno cence
31
Interaction Effects and External Validity
! When interaction effect is not present
$ Generalize findings across conditions of experiment
! When interaction effect is present
$ Sets limits on generalizing a finding
$ Conditions of the experiment identify the limits
32
Interaction Effects and Ceiling/Floor Effects
! Sometimes an interaction effect is statistically significant Aby mistake@
! Means for one condition(s) reach, on average, near
$ Highest possible value: Ceiling
$ Lowest possible value: Floor
! Interaction effect is uninterpretable.
! Example: Interaction effect between Test
Difficulty (Easy, Hard) and Study Hours (10,
15)
$ Simple main effect of Study Hours for
Hard test, but not Easy test.
$ How to interpret this effect if maximum score is 50?
! If there’s enough Aroom@ to show effect of IV, the interaction effect
disappears.
$ Change test characteristics (e.g., more items)
$ Two main effects
Study Hours
10 15
0
10
20
30
40
50
Test Difficulty
Easy
Hard
10 15
0
20
40
60
80
100 Test Difficulty
Easy
Hard
33
Interaction Effects and the Natural Groups Design
! Test causal inferences for natural groups variables
! How? Natural groups variables are correlational.
! Test a theory for why the natural groups differ.
! Use complex design
! Steps for making a causal inference with a natural groups variable in
complex design
$ State your theory.
$ Why do the groups differ? What is the theoretical process?
$ Identify an independent variable to manipulate.
$ This IV should influence the likelihood the theorized process will
occur.
$ Look for an interaction effect.
$ The natural groups variable and manipulated IV should produce a
statistically significant interaction in the hypothesized direction.
$ This interaction effect allows a causal inference for why individuals in
the natural groups differ.
