comparing two means

why can't you use multiple t-tests?

increase chance of type one error

t-test assumes that values of outcome variables are independd of each other, which is not true when you compare 3+ groups (1v2, 2v3, 1v3, etc.)

what test is used for comparing groups with more than two means?

analysis of variance (ANOVA)

what is the statistic of ANVOA

f-statistic

compared to critical value of F

how to calculate F

variance between samples (difference from treatment) / variance within samples (difference from error)

what is inferred with a small F value

clinical insignificance

if variance between samples is small F will be _____________________

small

if variance within samples is small F will be ___________________

large

assumptions for ANOVA

same as the rest of the parametrics ;)))

how do you test for homogeneity of variance for ANOVA

levenes!

variation within groups

except w/ 3 groups

what does the f statistic measure

measures if observed difference in a set of means greater than the difference expected by error alone

types of anovas

lots-- only covering one way anova

one way anova (separate groups for each intervention)
one way repeated measures anova (one group for each intervention)

Levine's test is significant when

p is less than .05

ANOVA's comparison of group means

ANOVA looks at distance of each group mean from grand mean and if the distance is statistically significant

omnibus test

overall, nonspecific test

aka f test-- tells you a difference exists but not WHERE it exists

what are multiple comparison tests

a test that tells you where the difference is if the F test states that a difference exists

pros of comparing means

removes time as an independent variable!

effect sizes of ANOVA

eta squared (most common)
small - .01
medium- .06
large- .14

cohen's f (1/2 value of t-test)
small- .10
medium- .24
large- .40

are variances equal?

levine's determines

to say that there is a difference, p needs to be less than .05

failure to reject null = equal variances