Statistical tests for experiments
Where the researcher manipulates the IV(s) under controlled conditions and measures changes in the DV.
For parametric tests:
|Two conditions||T test (Related)||T test (Unrelated)|
|Three or more conditions||One-way ANOVA (Related)||One-way ANOVA (Unrelated)|
For non-parametric tests:
Related tests have single names: Wilcoxon (two conditions) and Friedman (three conditions).
Unrelated tests have double names: Mann-Whitney (two conditions) and Kruskal-Wallis (three or more conditions).
||Wilcoxon test||Mann-Whitney test|
|Three or more conditions
||Friedman test (two-tailed only)
||Kruskal-Wallis test (two-tailed only)
The above tools allow us to check whether the differences in conditions are statistically significant when have 1 experimental variable.
Statistical tests for correlations
Where associations between variables are predicted.