Please enable JavaScript.

Coggle requires JavaScript to display documents.

NON PARAMETRIC ANALYSES (what you do when assumptions for parametric…

- - - - how many conditions?
        
        2
        
        Mann Whitney or Wilcoxon rank sum
        
        PARAMETRIC EQ: Independent measures T Test
        
        3+
        
        Kruskal-Wallis
        
        One way independent ANOVA
    - - how many conditions?
        
        2
        
        Wilcoxon signed Rank
        
        Repeated t tests
        
        3+
        
        Friedman's
        
        One way repeated ANOVA
- - - - ie : (D(20)=.24, p=.005) #
  - - - H(df)=chi square value, p= Asymp. sig
        
        (H(3)=37.31, p<.001
        
        there is a significant main effect of country on the number of beers per month (H(3)=37.1, p<.001
        
        which conditions are significantly different?
        
        get descriptive stats- analyse> compare means> means
        
        report the MEDIAN> report table
        
        mean, N, STD , Median
        
        tells you who drinks the most and the least
    - - comparing two independent conditions at a time - three comparisons - SIMPLE PLANNED COMPARISON
- - - - ie. D(20)=.27, p=.001
  - - - X2(df) = chi square value, p= Asymp.Sig.
        
        X2(3)=34.92, p<.001
        
        then descriptives- medians
    - - T=(smallest number of ranks under 'Mean Rank'), p= Asymp. sig. EG: (T= 6.63, p=.003)
      - look at the Ranks table and the Test Statistics table
      - Boneferroni correction here : (0.050/ no of comparison)
- - - - Kolmogorov-Smirnov: D(df)=statistic, p= significance
      - Levenes if not repeated : F(df1, df2) = Statistic, p=significance
  - - - give stats!- Kruskal-Wallis: H(df)= Chi square value, p= Asymp.Sig
        
        Mann whitney: U=Mann whitney U, z=Z, p= Asymp.sig
        
        Friedman's: X2 (df)= Chi Square Value, p=Asymp.sig
        
        Wilcoxin signed rank: T =(smallest number of ranks under 'Mean Rank'), p= asymp. sig.
        
        how did you break down the main effect? (3+levels) ?
        
        which test and why?
        
        Which comparisons? how many? Bonferroni correction!