LUMINE ANALYTICS • RESEARCH & STATISTICS

Which Statistical Test Should I Use for My Research?

Choosing the right statistical test depends on your research question, outcome type, study design, number of groups, and whether observations are independent or paired.

Quick Guide

Research ObjectiveCommon Method
Compare two independent groupst-test / Mann-Whitney U
Compare two paired measurementsPaired t-test / Wilcoxon signed-rank
Compare three or more groupsANOVA / Kruskal-Wallis
Association between categorical variablesChi-square / Fisher's exact
CorrelationPearson / Spearman
Continuous outcomeLinear regression
Binary outcomeLogistic regression
Count outcomePoisson / Negative Binomial regression
Time-to-event outcomeKaplan-Meier / Cox regression
Repeated measurementsMixed-effects model / MMRM

Before Choosing a Test

  1. What is my research question?
  2. What type of outcome variable do I have?
  3. How many groups are being compared?
  4. Are observations independent or paired?
  5. Are measurements repeated over time?
  6. Do I need to adjust for other variables?

For example, comparing two independent groups may require a t-test, while repeated measurements from the same participants may require a mixed-effects model.

Need help with your research?

Lumine Analytics provides statistical analysis and research methodology consulting for Master's students, PhD researchers and academic teams.

Book a free consultation →