ANOVA

ANOVA

1. Independent (Between) Subjects

   • One Way
     • Total Variability
       • Residual Variability (Within Group Variation)
       • Model Variability (Between Group Variation)
     • Compare Means
       • One Way ANOVA
         • DV in D List, Group in Factor
           • Options: Descriptives, Homogeneity, Welch
             • Post Hoc: Tukey
               • Read from Tukey if homogeneity is broken
                 • Effect Size: SSModel(Between Grps) over SSTotal
   • Factorial
     • Total Variability
       • Variance explained by model
         • Effect 1
         • Effect 2
         • Interaction of 1&2
       • Unexplained (residual) varicance
     • General Linear Model
       • Univariate
         • IVs in Fixed Factors, DV in DV
           • Plot for Interaction!
             • Standard in Option
               • PLUS homogeneity test!
               • When there is interaction, use syntax!
                 • GLM "DV" by IV1 IV2 => /EMMEANS = TABLES(IV1*IV2) COMPARE(IV)

2. Within Subjects

   • One Way
     • Total Variability
       • Between Subject Variability
       • Within Subject Variability
         • Effect Variability
         • Unexplained Varabiility
     • General Linear Model
       • Repeated Measures
         • Define Variables
           • Standard in Options
   • Factorial
     • Total Variability
       • Between Subjects Variability
       • Within Subjects Variability
         • Effect of Experiment
           • Effect A
           • Effect B
           • Interaction AxB
         • Error (Residual) - Unexplained Variability
           • Error of A
           • Error of B
           • Error of Interaction AxB
     • General Linear Model
       • Repeated Measures
         • Define Variables
           • Plot for Interaction!
           • Standard Procedures in Options
           • When there IS an interaction, check in syntax!
             • GLM => /WSFACTOR => /EMMEANS = TABLES(*) COMPARE()

3. Mixed Design

   • Analyze -> General Linear Model
     • Repeated Measures
       • Define Variables
         • Don't forget to put between subject variable in appropriate column!
         • Plots -> Check for interaction!
         • In Options, check for effect size AND homogeneity!
         • Keep in mind that sphericity can only be tested when there are more than 2 levels in a variable!