Flashcards
Logic of ANOVA: hypotheses, F statistic and effect size
Cards in this deck (8)
- 1What are the reasons ANOVAs are run?Generalisability
Interaction effects
Informing theory
Efficiency - 2Logic of ANOVA: how do we test the null hypothesis?1. We assume the null hypothesis
2. Create a model around an alternative hypothesis
3. Calculate p-value of the extent it will fit
4. If p-value is less than .05, we assume the null hypothesis is false. - 3Differences in variances if means difer and if they do not differh1- means differ (alternative hypothesis)
Variability between groups is high but within groups is low.
h0- means do not differ- within group is high, between groups is low - 4What is the logic of the comparisons of two variances.the marginal means are the variability between groups (Ssbetween). Anything around those means, seen in the cell means, creates error variances (Sswithin).
If H0 is true, there should be low variance between groups, hence homogeneity of variance. - 5How if the F statistic calculated?Ssbetween/(k-1)
--------------------------
Sswithin(N-k)
where K= number of groups
N= number of scores - 6How is effect size for anova calculated?eta squared (n2) is the between/treatment variance proportionate to the total variance
n2= Ssbetween/ Sstotal - 7How is an interaction expressed in equation terms?Ssinteraction= Ssbetween-Ssfactor1-Ssfactor2
- 8—
Information
- Description
- Flashcards on Untitled, created by chantel Smith on 07/03/2019.
- Created by
- chantel Smith
- Views
- 1
- Rating
More on Analysis of Variance (ANOVA)