10. Repeated Measures ANOVA

Question 1

What would be the most appropriate statistical analysis for a study design with one IV with 4 repeated measures and one DV?

Answer 1

A One-Way
Repeated Measures (Within groups)
ANOVA

Question 2

In a independent groups ONE WAY ANOVA what is the variance associated with?

Answer 2

SST (Total) =
SSB (explained)
& SSW (Residual)

Question 3

In an independent groups (between subjects) FACTORIAL ANOVA what is the variance associated with?

Answer 3

SST (Total) =
SSW (Residual) 
SSB comprised of 3 parts:
Variance explained by IV1 
Variance explained by IV2
Variance explained by IV1*IV2

Question 4

In a WITHIN subjects (repeated measures) FACTORIAL ANOVA what is the variance associated with?

Answer 4

SST (Total) =
SSM (SSB/M/C: Between/ Model/ Conditions)
SSW (Within: Error- Two parts)

Question 5

In a WITHIN subjects (repeated measures) FACTORIAL ANOVA the variance associated with SSW can be broken into?

Answer 5

SSW =
SS Subjects (How each participant varies their answers- Within Participant Variance)
& SS Error / Residual

Question 6

Where is the model variance found in a repeated measures factorial ANOVA? (SSM)

Answer 6

• The Tests of Within Subjects Effects Box
• IV (Specific)
• Type II Sum of Squares
• Sphericity Assumed

Question 7

What are the df for SSWithin?

Answer 7

n - 1

n = number of participants within each group

Question 8

What are the df for SSModel?

Answer 8

k - 1

k = number of groups

Question 9

Where do you read the SS Error / Residual from?

Answer 9

• The Tests of Within Subjects Effects Box
• Error(IV)
• Type II Sum of Squares
• Sphericity Assumed

Question 10

df of SST for repeated measures ANOVA?

Answer 10

N - 1

N = total number of scores

Question 11

df of SSW (participants) for repeated measures ANOVA?

Answer 11

n (k - 1)
n = number of participants in each group
k = number of groups

Question 12

df of SSM for repeated measures ANOVA?

Answer 12

k - 1

k = number of groups

Question 13

df of SSE/R (error/ residual) for repeated measures ANOVA?

Answer 13

(n - 1) (k - 1)
n = number of participants in each group
k = number of groups

Question 14

(Partial) Eta sq is equivalent to ?

Answer 14

R sq

But eta sq can’t be generalised

Question 15

How do you interpret eta sq?

Answer 15

Convert to a percentage and then use cohens r sq conventions

Question 16

What does the observed power mean?

Answer 16

How likely it is that you will achieve the same results with the same sample again

Question 17

What do you have to do if you find a sig. effect in a WITHIN subjects Mixed Measures ANOVA and when would you use them?

Answer 17

Planned comparisons/ contrasts (specific hypothesis testing)

Pairwise comparisons/ Post-hoc tests (non specific hypothesis testing)

Question 18

Why would you use pairwise comparisons?

Answer 18

Data snooping and non specific hypothesis testing

Question 19

What are the 5 rules of planned contrasts?

Answer 19

k - 1 number of contrasts run

only appear in 1 contrast

If not involved = 0

compare only 2 chunks
sum of weights = 0

Question 20

What are the 5 contrasts you can do in spss?

Answer 20

deviation 
simple 
repeated 
helmert
difference

Question 21

What happens in a deviation contrast?

Answer 21

compare each group (except first) to SST

Question 22

What happens in a simple contrast?

Answer 22

compare first group to each other group

Question 23

What happens in a repeated contrast?

Answer 23

compare each level to the next level 
e.g. 
1  v. 2 
2 v. 3 
3 v. 4

Question 24

What happens in a helmert contrast?

Answer 24

compare each group to the total of the next groups
1 v. 2,3,4T
2 v. 3,4T
3 v. 4

Question 25

What happens in a difference contrast?

Answer 25

compare each group to previous groups starting from top 4 v. 3,2,1T 3 v. 2,1T 2 v. 1

Question 26

When we don't have any priori predictions what are we doing in analysis?

Answer 26

data mining/ snooping | post-hoc tests

Question 27

In a pairwise/ post-hoc test what are we comparing and how?

Answer 27

all possible pairs | t-tests using Bonferroni correction

Question 28

What does the bonferroni correction do?

Answer 28

Control the family wise error rate

Question 29

What assumption is used in a WITHIN (repeated measures) ANOVA instead of independence?

Answer 29

Sphericity

Question 30

What is sphericity?

Answer 30

the relationship between pairs of experimental conditions is similar (epsilon) aka homogeneity of variance between groups or compound symmetry

Question 31

Assuming compound symmetry means?

Answer 31

(VARIANCES) Similar variation between groups and (COVARIANCES) Groups not dependent on each other

Question 32

To meet the assumption of compound symmetry in the Correlation matrix the diagonal figures should be?

Answer 32

Equal

Question 33

To meet the assumption of compound symmetry in the Correlation matrix the bottom triangle (half) figures should be?

Answer 33

Equal

Question 34

Which is less stringent... compound symmetry or sphericity?

Answer 34

Sphericity

Question 35

How do you calculate sphericity?

Answer 35

Calculate differences between pairs Calculate the variances between these differences If equal (not sig. different)= sphericity

Question 36

What statistic determines the sig. of sphericity?

Answer 36

Mauchly (W) If not sig. then not sig. sphericity = good

Question 37

What do you do if sphericity is not assumed (Mauchly is sig.)?

Answer 37

Check epsilon Greenhouse Geisser IF .7 read from tests WITHIN sub effects: Greenhouse Geisser

Question 38

What are the assumptions for a mixed model BETWEEN sub factorial ANOVA?

Answer 38

- Normality - Homogeneity of Variance - Independence - Proportionality

Question 39

What are the assumptions for a mixed model REPEATED measures factorial ANOVA?

Answer 39

- Normality - Compound Symmetry/ Sphericity - Dependence - Proportionality