Associations Between Variables
Tests and models that quantify how two or more variables move together
When the question is “do these variables go together?” rather than “do these groups differ?”, the analysis is an association: correlation, contingency, or regression. The choice between them depends on the scale level of each variable, the number of variables, and whether a directional model is intended.
Method pages
Two variables
- Chi-squared contingency test – two nominal variables, with Fisher’s exact test for small cells.
- Spearman rank correlation – monotonic association, ordinal or non-normal.
- Pearson correlation – linear association, continuous bivariate normal.
- Kendall’s tau – rank association, robust to ties and small samples.
- Simple linear regression – directed relationship from one continuous predictor.
More than two variables
- Multiple linear regression – continuous outcome with several predictors.
- Binary logistic regression – dichotomous outcome.
- Ordinal logistic regression – ordered categorical outcome.
- Multinomial logistic regression – unordered categorical outcome with three or more levels.
Structure inspired by the University of Zurich Methodenberatung (methodenberatung.uzh.ch). All text, examples, R code, and reporting sentences are independently authored in English.