Module 10 — Reproducibility

Question: six months from now, can you — or a reviewer — reproduce every number in your report? In ChakataStat the answer is built in: every module of this course was quietly writing a script.

The Syntax console

Open the Syntax tab. Everything you ran through the menus is there as commands — the same ones each module's "syntax trail" showed. This console is editable and executable:

1. Click Run all: the entire session replays, top to bottom, and the output regenerates.
2. Edit one line — change a variable, add one — select it, and Run selection to run just that.
3. Save syntax… writes the script as a .cks file next to your dataset.

It is a real editor, not a text box: commands and their parameters are highlighted, Ctrl+Space completes a command name or a variable, and a line the parser cannot read is marked before you run it rather than after. You are meant to write here, not only to read what the dialogs wrote.

Dialogs are how you find an analysis; syntax is how you keep it.

The reproducible bundle

A finished piece of work is three small files:

• study.ckd — data plus every variable definition (modules 1–2);
• study.cks — the ordered commands that produce the results;
• the exported output (File → Export Output…, HTML/Markdown/LaTeX) — what the numbers looked like when you reported them.

Anyone with the first two can regenerate the third — that is the standard your future self will thank you for. (Batch users can even run it headless: chakatastat --run study.cks study.ckd --out report.html.)

Capstone exercise

Reproduce a mini-report start to finish, in syntax:

1. Open the sample. In the Syntax tab, write (or paste from your trail):

   descriptives variables=[Age, BMI, SystolicBP]
   crosstabs row=SmokingStatus, column=Hypertensive, chi_square=true
   anova variables=[SystolicBP], factor=SmokingStatus, posthoc=true
   linear_regression dependent=SystolicBP, predictors=[Age, BMI]
   histogram variable=BMI
   

2. Run all, review the output, export it to HTML.

3. Save the .cks and the .ckd, close the app, reopen, run the script again — confirm the report reproduces exactly.

Where next

• Analyzing Data — the full procedure catalog, from nonparametrics to survival and time series.
• The AI assistant — plain-language analysis over the same engine, with the same syntax trail.
• Reproducibility in depth — batch runs, Python export, output formats.

You now know the whole loop: open → describe → visualize → test → transform → report — reproducibly. That is the course.