Stats 2: ANOVA Flashcards
Assumptions for a simple, one way ANOVA
- Normality
- Homogeneity of Variance
- Independence of observation
How do we test homogeneity of variance?
Levene’s test
Error variance vs systematic variance
Error: random variance in the sample mean
Systematic: variance due to the action and interaction of independent variables
Variation vs variance
Variation: spread/dispersion of scores around the mean
Variation: average spread/dispersion
Values for variation and variance
Variation: SSgroup and SSerror
Variance: MSgroup and MSerror
Sum of Squares values
SStotal = total variation. SSgroup = between group variation, due to interactions SSerror = within group variation, due to differences within the groups
How do we work out if there is systematic variance present? (I.e. If the variables’ interaction is significant)
F = MSgroup/MSerror
When do we use a factorial ANOVA?
To see the effects of more than one independent variable
What are the three sets of hypotheses with a factorial ANOVA?
1 All factor A means are equal
- All factor B means are equal
- There is no interaction between A and B
Logic of the splitting of ANOVA factorial
The variance is split into between groups and within groups, only now the variance between groups is split further took look at the variance due to factor A, due to factor B and due to their interaction.
Kinds of cell mean plots
Ordinal: simple effects in the same direction. (Lines are not parallel but do not touch)
Disordinal: simple effects in different directions. (Lines cross)
When do we use Tukey’s?
When you’ve got three levels of an independent variable, it’s used to compare the means to see where the difference lies. It identifies if the difference between two means is greater than the expected standard error.
