FREE TOOL · RUNS IN YOUR BROWSER
Which statistical test
fits your question?
Answer a few plain-language questions about what you want to find out and what your data look like. You'll get a suggested analysis, the assumptions to check before you run it, an alternative, and the effect size to report.
AT A GLANCE
Common dissertation analyses.
The analyses dissertations use most, matched to the question each answers. Our benchmark of 112,903 dissertation abstracts shows how common each one is in your field.
| Your question | Outcome | Common analysis | Alternative |
|---|---|---|---|
| Compare two independent groups | Continuous | Independent-samples t test (Welch) | Mann–Whitney U |
| Compare the same people twice | Continuous | Paired-samples t test | Wilcoxon signed-rank |
| Compare three or more groups | Continuous | One-way ANOVA | Kruskal–Wallis |
| Compare the same people three or more times | Continuous | Repeated-measures ANOVA or linear mixed model | Friedman |
| Compare groups on a category | Categorical | Chi-square test of independence | Fisher’s exact test |
| Relate two variables | Continuous | Pearson correlation | Spearman correlation |
| Predict an outcome from several variables | Continuous | Multiple linear regression | Robust regression |
| Predict a yes/no outcome | Binary | Logistic regression | Exact logistic regression (small samples) |
| Test how or why a relationship works | Continuous | Mediation analysis | Structural equation modeling |
| Analyze people nested in groups | Any | Multilevel model (HLM) | Cluster-robust standard errors |
| Check a scale’s reliability | Item scores | Omega and Cronbach’s alpha | Test–retest ICC |
| Test a scale’s structure | Item scores | Confirmatory factor analysis | Exploratory factor analysis |
BEFORE YOU DECIDE
Three questions a chooser can't answer for you.
- What does your committee expect? Fields and programs have conventions. A defensible choice fits both your data and your readers.
- Is your sample large enough? Every analysis needs a sample-size justification. Our power analysis tools help you plan and document it.
- Does the design support your claim? Most tests show association, not cause. Mediation and group comparisons without random assignment need careful wording and, often, a design such as propensity score matching.
Once you have results, the APA 7 results formatter turns them into correctly formatted sentences. If you're still unsure which analysis fits, a free 30-minute consultation can help.