Statistical Test Selector

Answer a few questions to get a recommended statistical test.

What are you trying to do?
What type of data do you have?
Are the observations…

Complete the three questions to see a recommendation.

This is general statistical guidance, not a substitute for statistical expertise. Consider your study design, the assumptions of each test, and your discipline's requirements before choosing a final test.

How the test selector works

Choosing the right statistical test depends on your research goal, the type of data, and whether the observations are independent or paired. This selector walks through those questions and recommends a starting point.

The recommendation is general statistical guidance. It does not replace statistical expertise: study design, assumptions, disciplinary conventions and software requirements all matter when choosing a final test.

Where a PanelRoster calculator implements the recommended test, the selector links to it directly. Tests that are not built in are named so you can run them in your own software.

The decision logic

Goal + data type + independence → recommended test

For example: two independent groups with a continuous outcome suggests Welch's t-test; three or more independent groups suggests one-way ANOVA; categorical-by-categorical association suggests a chi-square test; two continuous variables suggest Pearson correlation.

Worked example

Comparing two independent groups on a continuous measure

Goal: compare two groups. Data type: continuous. Observations: independent. Recommendation: Welch's t-test — with a link to the PanelRoster t-test calculator.

Frequently asked questions

Is this selector a substitute for statistical advice?

No. It is a starting point for choosing an appropriate test. Always confirm the choice with the assumptions, your study design, and disciplinary or journal requirements.

Why might I still need a different test?

Assumption violations (non-normality, unequal variances, small samples), complex designs, and specific research questions can all point to a different test than the initial recommendation.

What if the recommended test is not built in?

The selector names the test and its alternatives so you can run it in a statistics package. Only the tests with a PanelRoster calculator link directly to this site.

Built by PanelRoster

PanelRoster is an enterprise operations management platform for organizations coordinating distributed teams, assignments, workflows, quality assurance and operational execution.

Explore PanelRoster