Repeated Measures Anovas

Question 1

Background

Answer 1

Ps have multiple data points as same ps used in diff conditions. So data not independent.

Question 2

Benefits / weaknesses

Answer 2

Unsystematic variance reduced as individual diffs taken out but more sensitive to experimental effects. Cheaper as fewer ps

Question 3

Theory

Answer 3

Explained variance is within, unexplained between.for groups to be sig diff, need large explained variance. Have total variance, split between within and between, within split into effect of experiment and error. In this test, don’t measure variance between ps. F stat same as last (variance explained over left over)

Question 4

Independence and sphericity

Answer 4

Same ps means scores correlated so not independent., means sphericity. Measure using mauchlys test, needs to be non sig (over 0.05). Only applies when more than 2 levels of iv

Question 5

If sphericity violated

Answer 5

Just look at diff section of table, must correct df by multiplying by a correction factor: greenhouse giesser estimate (conservative), huynh feldt (lib) and lower bound (extreme/don’t find effect). If sphericity violated, report greenhouse instead of uncorrected, will be in decimals (don’t look at wrong thing)

Question 6

Reporting outputs

Answer 6

Look at tests of within subject effects table, if mauchlys sig, ignore sphericity assumed row and read from correction row for f ratio, p value and df

Question 7

In spss

Answer 7

When dragging data into factors, need to do it in order, do either contrasts or post hoc, not both. Test takes a long time,, ignore multivariate test table

Question 8

Mixed design

Answer 8

Between p variance split into effect of group and between p variance in each group, within split into effect of manipulation, interaction between group and influence of group on between p variance. Now have multiple variables and one between subject factor, diff subjects but at same level. Mixed anova gives main effects and two way interactions