Can the AI correctly distinguish assumptions behind different statistical methods?

For classic methods (t-test, ANOVA) it can state the conditions clearly, but for real data (non-normal distributions, small samples) its advice can be rigid. Before any key decision, describe the data characteristics (sample size, distribution shape) and have it list candidate methods with each one's risks, far more reliable than just asking "which method should I use."

Can it run actual statistical analysis for me?

It can't run anything, but it can write executable R, Python, or SPSS code. Describe your data structure and analysis goal, have it output code, and run it locally. Numerical results the AI hands you directly (like p-values) are almost always invented. Don't cite them.

How do I use this prompt?

Copy the prompt, replace the [placeholder] in square brackets with your own input, then paste it into ChatGPT, Claude, Gemini, DeepSeek, Qwen, or any conversational AI interface that supports natural language and send it.