Skip to main content
∑SolveGridAll tools

STATISTICAL TEST CHOOSER

Which statistical test should I use?

Answer four questions about your research question and data. You get the recommended test, a safer alternative, the assumptions to check, and an APA-style sentence to report it.

YOUR STUDY

Describe your question and data

4 QUESTIONS
1. WHAT DO YOU WANT TO FIND OUT?
2. WHAT TYPE IS YOUR OUTCOME VARIABLE?
3. HOW MANY GROUPS OR TIME POINTS?
4. ARE THE GROUPS THE SAME PEOPLE OR DIFFERENT PEOPLE?

Not sure if your data are normal? Check with the Shapiro-Wilk test. Likert items, ranks, and small skewed samples count as "ordinal / skewed". Summed scale scores from many items are often treated as numeric.

QUICK REFERENCE

Statistical test flowchart in words

Comparing groups (numeric outcome)

  • 2 different groups → independent t-test (Welch) · skewed/ordinal → Mann-Whitney U
  • Same people twice → paired t-test · skewed/ordinal → Wilcoxon signed-rank
  • 3+ different groups → one-way ANOVA · skewed/ordinal → Kruskal-Wallis
  • Same people 3+ times → repeated-measures ANOVA · skewed/ordinal → Friedman

Relationships and categories

  • Two numeric variables → Pearson correlation · skewed/ordinal → Spearman
  • Two categorical variables → chi-square test of independence · small counts → Fisher's exact
  • One categorical variable vs expected proportions → chi-square goodness-of-fit
  • Yes/no measured twice on the same people → McNemar's test

Should I use ANOVA or a t-test?

Use a t-test for exactly two groups and ANOVA for three or more. Running several t-tests instead of one ANOVA inflates the chance of a false positive: with three groups and three t-tests at α = .05, the familywise error rate rises to about 14%.

Should I use ANOVA or chi-square?

It depends on the outcome. If the outcome is a number (a score, time, or scale total), use ANOVA. If the outcome is a category (passed/failed, chose A/B/C), use chi-square.

References

  • Field, A. (2018). Discovering Statistics Using IBM SPSS Statistics (5th ed.). SAGE.
  • American Psychological Association. (2020). Publication Manual of the American Psychological Association (7th ed.).
  • Delacre, M., Lakens, D., & Leys, C. (2017). Why psychologists should by default use Welch's t-test instead of Student's t-test. International Review of Social Psychology, 30(1), 92–101.

ABOUT THIS TOOL

Statistical Test Chooser

Pick the right statistical test for your thesis in four questions, then see the assumptions to check and a ready-to-edit APA-style sentence for your results chapter.

  • Choose between t-tests, ANOVA, and their non-parametric alternatives
  • Decide between chi-square, Fisher's exact, and McNemar's test
  • Get the APA-style reporting sentence for the chosen test

KEY FUNCTIONS

14 test outcomesParametric and non-parametric alternativesAssumption checklistCopyable APA report template
Browse research statistics & survey tools →

LEARN THE METHOD

Guides for this calculator workflow

GUIDEHow to choose a statistical testSeparate descriptive statistics, t-tests, correlation, and one-way ANOVA by the question and data type.Read →GUIDELikert scale interpretation: weighted mean rangesBuild the 4-, 5-, and 7-point mean range tables, interpret weighted means, and report Likert results in a thesis.Read →GUIDEHow to report ANOVA results in APA styleCopy-ready APA templates for one-way, Welch, repeated-measures, and two-way ANOVA with effect sizes and post hoc tests.Read →

KEEP SOLVING

More in this category →Browse all tools →
t

Advanced Statistics Lab

Use an inference-focused statistics workspace for t-tests, correlation, simple regression, and one-way ANOVA with visible assumptions.

→Σw

Likert Scale Weighted Mean Calculator

Turn questionnaire frequency counts into weighted means, standard deviations, rankings, and a verbal interpretation table ready for a thesis results chapter.

→n

Sample Size Calculator

Find the minimum sample size for a survey or thesis using Slovin/Yamane or Cochran's formula, with confidence level, margin of error, population size, response-rate adjustment, and a copy-ready methods sentence.

→