Module 5 Flashcards
F distribution characteristics
- right skewed
- only positive values
- area = 1
- defined by df (increase less skewed)
F distribution df (2)
df1 = number of groups -1
df2 = number of observations - number of groups
F ratio
calculated by analyzing the amount of variation between and within groups
- between, spread between sample means
- within, spread within each group due to sampling variability (variation due to error)
- close to 1 no difference
ANOVA test use
when there are more than two groups, so the risk of type I error (false positive) isn’t inflated, if multiple t tests used less confident w every additional test
one way ANOVA + assumptions
one independent variable w three + groups and one dependent variable
assumptions:
- scale
- independent samples
- normality
- equal variance
calculating one way ANOVA
- calculate F ratio value
- look at df and find chart F value
- compare chart value and F value
- if F ratio greater than chart reject the null, less fail to reject
post hoc test
used to see where differences are, tukeys HDS (below 0.05 there is a difference)
one way ANOVA non parametric
Kruskal wallis
- uses median or mean ranks
- median has same shape distribution for groups
- mean ranked has different distributions for groups
one way repeated measure ANOVA + assumptions
same group over multiple times
assumptions:
- scale
- independent samples
- normality
- sphericity
finding sphericity
difference between all combinations must be equal, uses mauchly test
one way RM ANOVA non parametric
Friedman
two way ANOVA
- two independent factors
- 2x2 table created with means
- main effects calculated by looking at average across rows and column (marginal means)
- interactive effects if patterns of differences across rows and columns are different
