Module 7: Factorial ANOVA (Two-Way Repeated Measures and Two-Way Mixed Designs)

0.0(0)
Studied by 0 people
call kaiCall Kai
learnLearn
examPractice Test
spaced repetitionSpaced Repetition
heart puzzleMatch
flashcardsFlashcards
GameKnowt Play
Card Sorting

1/49

flashcard set

Earn XP

Description and Tags

Comprehensive vocabulary flashcards covering two-way repeated measures ANOVA, two-way mixed ANOVA, variance partitioning, sphericity assumptions, Mauchly's test, main effects, interactions, simple effects, and SPSS output interpretation.

50 Terms

1
New cards

Factorial ANOVA

A statistical test of difference used to examine the influence of more than one independent variable or factor (each with two or more levels) on a continuous dependent variable.

2
New cards

Between-Subjects Factorial Design

A factorial design where all independent variables are between-subjects, meaning each participant takes part in only one condition.

3
New cards

Within-Subjects Factorial Design

A factorial design where all independent variables are within-subjects, meaning each participant takes part in all experimental conditions.

4
New cards

Mixed Factorial Design

A factorial design featuring a combination of between-subject and within-subject independent variables, meaning participants take part in more than one condition, but not all conditions.

5
New cards

Main Effect

The influence of any single factor considered independently of other factors, evaluated by examining differences in marginal means.

6
New cards

Marginal Means

The mean scores calculated for individual levels of one independent variable while collapsing across and ignoring the levels of the other independent variable(s).

7
New cards

Interaction Effect

An effect occurring when the influence of one factor on the dependent variable depends on the specific level of another factor, evaluated through the examination of cell means.

8
New cards

Cell Means

The mean scores calculated for each specific condition or unique combination of levels across all independent variables.

9
New cards

Stable Individual Differences

Person-specific variance that is subtracted from within-IV variance and excluded from the FF-ratio error term in repeated measures designs because each participant acts as their own control.

10
New cards

Error Term Partitioning (Repeated Measures ANOVA)

The separation of error variance into distinct error terms for each effect, allowing each main effect and interaction to be tested against its own specific error term.

11
New cards

F-ratio (Two-Way Repeated Measures ANOVA)

The test statistic formed by dividing the model variance for a given effect by its corresponding specific error variance (MSeffect/MSerror\text{MS}_{\text{effect}} / \text{MS}_{\text{error}}).

12
New cards

Repeated Measures Data Input Setup (SPSS)

The data entry format in SPSS where each participant occupies a single row and separate columns represent each individual condition (combination of IV levels).

13
New cards

Cell Means in Factorial Reporting Tables

The descriptive statistics reported in an APA summary table, displaying the mean (MM), standard deviation (SD\text{SD}), and 95%95\% confidence intervals for each condition, with one IV arranged in rows and the other in columns.

14
New cards

Normality Assumption (Two-Way Repeated Measures ANOVA)

The assumption that the difference scores between each pair of independent variable levels are approximately normally distributed in the population.

15
New cards

Sphericity Assumption

The assumption that the variances of the difference scores across all pairs of independent variable levels are approximately equal in the population.

16
New cards

Mauchly's Test of Sphericity

A hypothesis test evaluating whether pairwise difference scores have equal population variances; a statistically significant result (p≤.05p \le .05) indicates that sphericity is violated.

17
New cards

Automatic Sphericity

The condition where an independent variable has only 22 levels, causing sphericity to be automatically satisfied and eliminating the need for Mauchly's test.

18
New cards

Greenhouse-Geisser Correction

A degrees-of-freedom correction applied to the FF-test in repeated measures ANOVA when the assumption of sphericity is violated.

19
New cards

Paired Observations Assumption

The assumption that each participant included in a repeated measures analysis must have a valid score recorded in every single condition.

20
New cards

Independence of Observations Assumption (Repeated Measures)

The assumption that while repeated measurements from the same participant are related across conditions, measurements collected from different participants must be entirely independent.

21
New cards

Main Effect of Presence of Critters

The statistically significant main effect in the celebrity jungle study showing that critter presence altered heart rate: F(1,9)=58.023F(1, 9) = 58.023, p<.001p < .001, and partial η2=.866\text{partial } \eta^2 = .866.

22
New cards

Main Effect of Trial Type

The statistically significant main effect of task type in the celebrity jungle study: F(2,18)=11.517F(2, 18) = 11.517, p<.001p < .001 (Sphericity Assumed), and partial η2=.561\text{partial } \eta^2 = .561.

23
New cards

Presence of Critters * Trial Type Interaction

The statistically significant interaction in the celebrity jungle study showing the effect of critters on heart rate depended on the trial type: F(2,18)=18.082F(2, 18) = 18.082, p<.001p < .001, and partial η2=.668\text{partial } \eta^2 = .668.

24
New cards

Partial Eta Squared (partial η2\text{partial } \eta^2)

The effect size metric reported for main effects and interactions in factorial ANOVA, indicating the proportion of variance explained by a specific factor after excluding other model effects.

25
New cards

Post Hoc Tests for Main Effects

Pairwise comparisons conducted following a statistically significant main effect for a factor with more than two levels (such as trial type) to determine which pairs of levels differ significantly.

26
New cards
<p>Profile Plot</p>

Profile Plot

A visual line graph plotting estimated marginal means across factor combinations, used to inspect the pattern and direction of an interaction effect.

27
New cards

Tests of Simple Effects

Targeted statistical tests conducted only after finding a significant interaction to examine the effect of a primary independent variable separately at each level of a secondary independent variable.

28
New cards

Simple Effects (Jungle Critters Experiment)

Three paired tt-tests comparing heart rates between critter absence and presence separately within water (t(9)=4.93,p=.001t(9) = 4.93, p = .001), confinement (t(9)=8.62,p<.001t(9) = 8.62, p < .001), and height (t(9)=2.67,p=.026t(9) = 2.67, p = .026) trials.

29
New cards

Bonferroni-Adjusted Alpha Level

A corrected threshold (α/c\alpha / c) used to maintain the familywise Type I error rate when conducting multiple comparisons by dividing the original alpha level by the number of comparisons (cc).

30
New cards

Cohen's d (Factorial ANOVA Simple Effects)

A standardized measure of effect size that must be calculated and reported for every simple effect comparison, regardless of whether that comparison reaches statistical significance.

31
New cards

Two-Way Mixed ANOVA

A factorial analysis of variance used when one independent variable is a between-subjects factor and the other independent variable is a within-subjects factor.

32
New cards

Dating-App Study Design

A 2×22 \times 2 mixed factorial experiment evaluating attractiveness ratings across intelligence (between-subjects: high vs. low) and social interaction (within-subjects: before chat vs. after chat).

33
New cards

Mixed ANOVA Data Input Setup (SPSS)

The SPSS data setup requiring one dedicated column for the between-subjects grouping variable and separate columns for each level of the within-subjects factor, with each participant on their own row.

34
New cards

Homogeneity of Variance (Mixed ANOVA)

The assumption that error variances for the dependent variable are equal across between-subjects groups, evaluated separately at each level of the within-subjects factor using Levene's test.

35
New cards

Levene's Test of Equality of Error Variances

A test evaluating the equality of variances across groups; in the dating-app study, error variances did not significantly differ before chat (p=.173p = .173) or after chat (p=.083p = .083).

36
New cards

Tests of Between-Subjects Effects Table

The SPSS output table in a mixed ANOVA that displays the main effect, test statistic, significance level, and residual error term for the between-subjects independent variable.

37
New cards

Main Effect of Intelligence

The non-significant between-subjects effect in the dating study comparing attractiveness ratings between intelligence groups: F(1,18)=1.961F(1, 18) = 1.961, p=.178p = .178, and partial η2=.098\text{partial } \eta^2 = .098.

38
New cards

Tests of Within-Subjects Effects Table

The SPSS output table displaying within-subjects main effects, interaction effects involving within-subjects factors, and their corresponding error terms in a mixed ANOVA.

39
New cards

Main Effect of Social Interaction

The non-significant within-subjects effect in the dating study comparing attractiveness before and after chat: F(1,18)=.380F(1, 18) = .380, p=.545p = .545, and partial η2=.021\text{partial } \eta^2 = .021.

40
New cards

Social Interaction * Intelligence Interaction

The statistically significant interaction in the dating study showing that the impact of intelligence depended on chat status: F(1,18)=51.761F(1, 18) = 51.761, p<.001p < .001, and partial η2=.742\text{partial } \eta^2 = .742.

41
New cards

Shared Error Term in Mixed ANOVA

The error term (Error(social_interaction)\text{Error}(\text{social\_interaction}), MS=.592\text{MS} = .592) shared by the within-subjects main effect and the interaction, reflecting variance in participants' before-after difference scores within each group.

42
New cards

Simple Effects of Intelligence (Across Social Interaction)

Follow-up tests comparing high vs. low intelligence separately before chat and after chat using independent-samples tt-tests, executable without splitting the SPSS data file.

43
New cards

Simple Effect of Intelligence Before Chat

An independent-samples tt-test comparing attractiveness ratings between high and low intelligence prior to chatting, yielding a non-significant result: t(18)=1.06t(18) = 1.06, p=.302p = .302.

44
New cards

Simple Effect of Intelligence After Chat

An independent-samples tt-test comparing attractiveness ratings between high and low intelligence following chat, yielding a statistically significant result: t(18)=3.93t(18) = 3.93, p=.001p = .001.

45
New cards

Simple Effects of Social Interaction (Across Intelligence)

Follow-up comparisons testing before vs. after chat ratings separately for high and low intelligence using two paired-samples tt-tests after splitting the data file by intelligence.

46
New cards

SPSS Split File Command

The procedure in SPSS required when running simple effects across levels of a within-subjects factor separately for each subgroup of a between-subjects factor.

47
New cards

Independent t-test (Differences Decision Tree)

The experimental difference test used when an experiment features exactly one independent variable with two levels tested between participants.

48
New cards

One-Way Repeated Measures ANOVA (Differences Decision Tree)

The experimental difference test used when an experiment features exactly one independent variable with more than two levels tested within participants.

49
New cards

Two-Way Independent ANOVA (Differences Decision Tree)

The factorial ANOVA model used when an experiment tests two independent variables where all factors are between participants.

50
New cards

Campaign Against the Abusive Use of ANOVAs

An instructional reminder that statistical complexity is no substitute for good design, urging researchers to keep designs simple and plan analyses prior to data collection.