week 8: two way independent anova

two way anova = two independent variables

different ppts in all conditions = independent design

factorial design

one way anova = one iv, three conditions

• ppts take part in one condition = one way independent measures anova

• ppts take part in all conditions = one way repeated measures anova

factorial (two way)

• reporting = two way, 3 (x y and z) x 2 (male and female) anova was conducted on the independent variable

  • report each of main effects, then interaction effect

• analyses the effect of two ivs on a dependent variable

• each factor can have two or more levels

• examine effect of each factor on its own = main effect

• examine if one factor is dependent on another = interaction effect

• can be more powerful than examining the effects of just one iv 

  • more insight

  • reduces error term by accounting for variance which would otherwise be unexplained

total variance in data = variance explained by model vs unexplained variance error

• if experiment is successful, the model will explain more variance than it can’t

• see slides for diagram

• also see slides for worked example

use post hoc testing if significant anova

interaction effect usually most significant