How it works

One path from a research question to a submitted dissertation. Every step feeds the next, so nothing is retyped and nothing is copied between files by hand.

1. LearnShort explanations of study design, sampling, and what each statistical test is for — free, and available before you pay for anything.
2. PlanWrite your research question in your own words. Chavery turns it into objectives, a study design and the list of variables you will record.
3. Calculate the sample sizeFor the design you chose, with the power, alpha and expected effect size recorded so you can report exactly what you assumed.
4. Build the master chartChavery creates the assessment schedule your design implies — a baseline and a follow-up for a before-and-after study, for example. You can change it.
5. Record your observationsDay by day as the study runs, on a phone at the bedside if that is where the data is, or entered directly if you already hold your findings. Values are checked as they are entered; a locked value can only be amended with a reason, and the amendment is kept.
6. AnalyseSay what you want to compare. Chavery checks normality and variance against your actual data, recommends the right test, names the alternative if an assumption fails, and explains why in plain language.
7. Draw the graphsThe figure that suits the test and the data, at publication resolution, in a palette that survives greyscale printing and colour-blind readers.
8. WriteWrite section by section with the writing checks marking problems where they occur — including statistical reporting mistakes an ordinary grammar checker cannot see.
9. Manage referencesOne stored record per work, rendered into Vancouver, NLM, APA, Harvard or MLA. Citations are checked against the list in both directions before you submit.
10. Send sections to your guideYour supervisor approves a section, sends it back with what needs changing, or comments on a passage — while there is still time to act on it.
11. Generate the documentOne file: title page, table of contents, roman front matter, numbered chapters, your tables and figures with captions, and the bibliography built from what you actually cited. Word, PDF and Excel.

Why it is not an AI tool

A dissertation is examined. If a number cannot be traced to the data that produced it, or a sentence cannot be traced to a rule, it cannot be defended in a viva. So Chavery does no generation: statistics come from established formulae, test selection from a published decision tree, writing checks from a fixed rule set, and documents from templates.

The practical consequences: the same input always gives the same output, every recommendation can be explained, and your unpublished research is never sent to an external language-model service.