Week 4 Lecture 4 - repeated measures 1-way ANOVA Flashcards
What contributes to BIVL in a RM 1W ANOVA?
- manipulation of IV
- experimental error (random and constant)
What contributes to WIVL in a RM 1W ANOVA?
- experimental error (random)
How is the F stat calculated in a RM 1W ANOVA?
f = (Variance BIVL) / (variance WIVL - individual differences)
In a RM 1W ANOVA, why do F stats tend to be larger compared to independent 1W ANOVAs?
because individual differences aren’t included so you are dividing by a smaller number
What are the assumptions of a RM 1W ANOVA?
- normality
- sphericity (homogeneity of covariance)
- equivalent sample size
What is sphericity?
the variance in different scores under each IV level pair should be reasonably equivalent
What test is done to determine sphericity?
Mauchly’s
want a none significant result
If sphericity is not homogeneous, what corrects for this?
Greenhouse-Geisser
If the assumptions of a RM 1W ANOVA are violated, what test is used instead?
non-parametric alternative
Friedman Test
On SPSS output, what lines do you read from to determine significance?
if Mauchly’s isn’t significant, use Sphericity assumed
if Mauchly’s is significant, use Greenhouse-Geisser
How is an F stat written for a RM 1W ANOVA?
F (dfM, dfR) = F-value, p = p-value
What is the DF for BIVL in a RM 1W ANOVA?
dfM = k-1
What is the DF for WIVL in a RM 1W ANOVA?
dfR = dfM x (n-1)
What post hoc test is used in a RM 1W ANOVA?
Bonferroni
- more conservative
What are the advantages of RM design?
- need fewer ppts
- error variance is reduced (WIVL)
- more power with the same no. of ppts
