When to use Anova over t-test?
Anova is used for 3+ groups and t-tests are only used for 2
Strengths of using ANOVA over multiple t-tests
reduces the risk of type I error (experiment wise alpha level)
Logic behind F-Ratio
determines weather the variability between group means is larger than the variability of observations within a group
Within Group variability
variability inside each treatment condition
Between group variability
differences between sample means
Alt. and Null hypothesis for ANOVA
H0= all mews equal to each other (treatment condition has no affect)
H1= there is at least one mean difference
F- Ratio calculation meaning
close to 1 = no evidence that suggested treatments has any effect
SIGNIFICANTLY HIGHER THAN 1 = suggestion of a treatment effect
Why are post hoc tests conducted?
Tell you which differences are significant and which ones are not (something ANOVA cannot provide)
effect size for ANOVA
n2= ss BTWN divided by ss TOTAL
ANOVA affect sizes
.02-.12: SMALL
.13-.25: MEDIUM
.26+: LARGE
