Create beginner Works with: Any AI

Which Statistical Test Should I Use

Decision tree Has example

Pick a test from the data and question.

The Open buttons send the prompt text as currently shown below to that site only when you click. Fill in the placeholders first.

Help me choose a statistical test. My question in plain words: [QUESTION]. Variables and their types (continuous, count, binary, categorical, ordinal): [VARIABLES]. Design (independent groups, paired measurements, one sample, repeated over time): [DESIGN]. Sample sizes and any known problems (skew, outliers, small groups): [SAMPLE_NOTES].

Follow this decision tree and show each answer:
1. What is the outcome type? 2. How many groups or conditions? 3. Are observations independent or paired? 4. Are the usual assumptions plausible (roughly symmetric, similar spread, enough observations), and how can I check each with a plot or a simple diagnostic?

Then output: the recommended test and the null hypothesis in words; a fallback test if assumptions fail (for example Mann-Whitney in place of a t-test); the effect size to report alongside the p-value and how to compute it; the code in [LANGUAGE] to run it; and a template sentence for reporting the result ("Group A (mean 12.4, SD 3.1, n = 40) scored higher than group B..., difference 2.3, 95% confidence interval 0.9 to 3.7").
Add three warnings: multiple comparisons, choosing the test after seeing the data, and reading p < 0.05 as proof. If the question is better answered by an estimate than a test, say so.

Fill in before use

Optional: type your details and the prompt above updates. Nothing is saved or sent.

Replace each bracketed placeholder with your own details, then paste into Any AI or another assistant.

Question: Do customers using the new checkout spend more? Two groups, 400 and 380, spend is skewed.
Decision path: outcome continuous; two independent groups; check distribution with a histogram (skewed).
Options: Welch t-test on log spend, or Mann-Whitney (compares ranks, not means), or a bootstrap difference in means.
Recommendation: bootstrap or Welch on the log scale; report the difference with a 95% interval.
Assumptions to check: independence (one row per customer), no repeated orders per customer, outliers reviewed.
Report sentence: "Mean spend was 4.20 higher (95% CI 0.40 to 8.10)."
Do not choose the test after seeing which gives p < 0.05.

• test-choice