One-way within-subjects ANOVA Flashcards
advantages of repeated measures design
increased statistical power
removes effect of individual differences
fewer ppts needed
less time and money
disadvantages of repeated measures design
practice effects
fatigue
contrast effects
demand characteristics
solutions to order effects
randomise testing order
counterbalancing
calculating F in repeated measures
f = BG variance/ WG variance
no BG V is due to individual differences as ppts are in each group
assumptions
DV is interval or ratio level
normally distributed population
homogeneity of variances
sphericity - equal variances between all possible pairs of conditions (replaces homogeneity of variances assumption in independent-measures ANOVA)
SPSS outputs we need
within-subjects factors
descriptive stats
Mauchly’s test of sphericity
tests of within-subjects effects
within-subjects factors
levels of factors
descriptive stats
mean and SD
Mauchly’s test of sphericity
want a non-significant result (p>.05) meaning assumption is met due to equal variances
report X^2( ) , p=
tests of within-subjects effects
main effects (sphericity assumed row)
F( , )= . ,p<.001
n^2= , s/m/l effect size
reporting results of one-way within subjects ANOVA
- present means and SD in table/text
- state ANOVA type, effect of IV on DV
- mention assumptions (inc. sphericity)
- report ANOVA results giving df, f-ration, p-value
- report effect size and what it means (n^2p)
- report comparisons
a priori comparisons
-paired samples t-test
- bonferroni adjustment
- divide significance by number of comparisons to get new significance
- paired samples t-test table to report t( )= ,p=
- effect size in descriptive stats table
reporting a priori comparisons for one-way within subjects ANOVA
a single planned comparison paired samples t-test was performed with no bonferroni adjustment
state whether there was a significant difference
t( ) = ,p= ,d= ,s/m/l
bonferroni post hoc test
- pairwise comparisons table (report comparisons with p values)
- effect size in descriptives table
- significance in pairwise table gives significance for effect size
what does it mean if sphericity is violated
variances not similar
