Comparative or Predictive Study?
Comparative (e.g. difference in blood pressure among groups of patients with different diseases). Predictive (e.g. what is the effect of x on y).
Comparing what?
Means. Proportions.
Dependent or independent samples?
Dependent samples (e.g. studying one group of patients before and after a treatment). Independent samples (e.g. comparing different groups of patients).
How many treatments?
Two treatments. Three or more treatments.
T-test on two paired samples.
Repeated measures ANOVA, mixed models.
How many treatments?
Two treatments. Three or more treatments.
Parametric or non-parametric?
Parametric. Non parametric.
F-test or t-test on two independent samples.
Paired or independent samples?
Paired samples. Independent samples.
Wilcoxon rank sums test.
Mann-whitney U test.
Parametric or non parametric?
Parametric. Non parametric.
ANOVA (if significant, then do SNK, Sheffe's, Tukey's, or Bonferroni's.
Paired or independent samples?
Paired. Independent.
Friedman test.
Kruksal-wallis test (if significant do a Dunn's test).
Comparing what proportions?
Observed proportion and theoretical proportion. Two observed proportions. Multiple observed proportions to multiple theoretical proportions.
Test for one proportion (chi squared).
Chi squared test.
Multinomial goodness of fit test.
Dependent or independent relationship?
Dependent (e.g. x predicts y). Independent (e.g. x and y are correlated).