Survey research and design in psychology/Tutorials/ANOVA/Effect size

From Wikiversity
Jump to: navigation, search

Effect size

Gnome-settings-background.svg View the accompanying screencast: [1]

ANOVAs in SPSS can provide the partial eta-square (\eta^2_p; although its labelled as eta-square). \eta^2 (which is equivalent of R2 in MLR) needs to be calculated manually from the output. This will help to explain the overall effects.

If you wish to discuss the size of effects between two means, then Cohen's d for each pairwise contrast of interest should be provided.

Calculate and interpret all effect sizes that can enhance understanding of the data in relation to the hypothesis or research question.

Eta-square
Percentage of variance explained by one or more IVs
  1. Eta-square: Calculation of overall \eta^2 needs to be done manually.
  2. Partial eta-square: Calculation is via PASW Analyze - (Type) - Options - Effect size
Cohen's d
Standardised mean difference between two means
  1. Cohen's d: Calculate using a spreadsheet calculator
  2. Cohen's d.xls
See also
  1. Statistical power