Different Tests Flashcards
Difference in means tests
1 - Nominal DV
=> Chi-square tests//binomial sign for related
2 - Ordinal DV
=> unrelated : Mann-Whitney // Kruskal Wallis
=> related: Wilcoxon // Friedman
3 - Interval, Ratio DV
=> unrelated: unrelated sample T-test // randomized (one way) ANOVA
=> related : related T-test // ANOVA repeat measures
Correlational tests
1 - Nominal
=> no possible -> equivalent is Chi-square 2x2
2 - Ordinal
=> Spearman‘s rho correlation coefficient
3 - Interval, Ratio
=> Pearson‘s product moment correlation coefficient
=> r between -1 and 1 -> strength of association between two metric variables
Assumptions for correlation
- Metric variables
- Normality (histogram) -> Skewness (-1/1), Kurtosis (-3/3)
- Linearity
- No outliers (boxplot)
Assumptions for regression
1 - Identical to correlation
2 - Influential cases (Cook‘s distance <1, CVR ratio 1+-(3(k+1/n)) )
3 - Homoscedasticity -> Plot residual for predicted values
4 -Multicollinearity -> ViF (variance inflation factor) <5