Dissertation statistics: CSV to Results chapter

Statistics for a thesis or dissertation come with one non-negotiable rule in CiteDash AI: numbers come from computation, not from a language model. Here's the path from raw data to a Results chapter whose every statistic traces to computed output.

1. Design before data

If you're pre-data, start with study design: variables, sampling, instruments, and validity threats are drawn from methodology literature with real citations, and you can generate ethics-application and pre-registration drafts from the design.

2. Upload your CSV and pick a test

Upload a CSV and choose from the named statistical tests: t-tests, correlation, ANOVA, regression, chi-square, Mann-Whitney and more. Each runs as deterministic code against your actual data. The AI's role is interpretation: it explains what the computed numbers mean, and that interpretation cites the computed output it interprets.

3. Advanced analysis in a sandbox

Need something beyond the named tests (a mixed model, a custom plot)? Advanced analysis writes Python for your request and runs it in an isolated cloud sandbox with hard timeouts. You get the code, the output, and the interpretation; the code never runs on your machine.

4. Qualitative data too

Upload interview transcripts and the coding workflow produces a codebook and themes. Every supporting quote is re-located verbatim in your transcript before it's shown. A quote that can't be found character-for-character is flagged, not trusted.

5. Flow findings into chapters

Once analyses are saved, Results, Discussion, and Conclusion sections can be drafted directly from them in Thesis Editor, grounded in your computed findings plus your library, and verified like everything else.