MethodVahti · Qualitative sampling
Defend the sample you can actually recruit.
A reviewer asks why twelve interviews were enough. An ethics committee wants the planned number. Your pilot data has made the original plan look optimistic. MethodVahti turns the study you have into a sampling range, a written rationale, and a record of the judgment behind both.
What is the correct N?What makes this N defensible?
Runs in your browser · nothing uploads · you make the research decision
Start with the problem you have
Choose the situation. MethodVahti asks for the design details that matter there, instead of making you translate your study into a generic calculator.
“I’m writing the protocol and need an N I can defend.”
- You get
- A planned N as a range from three models, a saturation expectation, and the methods paragraph for your protocol.
- It asks
- Your orientation, how varied the sample is, target analytic depth, and any pilot signal.
- Our stance
- A defensible range with the reasoning shown.
What you leave with
A methods argument you can edit, question, and take responsibility for.
A reasoned range
Three models shown together, with the assumptions and sensitivity visible.
Methods wording
Draft text for a protocol, ethics application, manuscript, or reviewer response.
A decision record
The study details, constraints, stopping rule, and human judgment behind the final choice.
COREQ and SRQR help structure what you report. They do not certify that the sample was adequate.
The defensibility profile
Sample size is one decision among several a reviewer will probe. The profile rates each design dimension Strong, Adequate, or Limited, and the overall judgement follows the least defensible dimension. Strengths do not offset weaknesses, the same way risk-of-bias tools work.
| Dimension | Judgement |
|---|---|
| Research question | Strong |
| Sampling | Limited |
| Data collection | Adequate |
| Analysis | Adequate |
| Reflexivity | Strong |
| Theorizing traceability | Adequate |
Overall defensibility: Limited — set by sampling. There is no overall number, on purpose: a single score would hide exactly the weakness a reviewer will find, and the distances between dimensions are not known.
The sixth dimension asks whether the path from surprise to claim is on record: what stopped you, what you conjectured, which analytic move tested it, and what you rejected. It rates the documentation, never the theory. The judgement of what your data means stays out of every dimension, this one included. How to keep that path in plain files.
Disagree with where the rule lands? Override it, in writing. A documented departure is defensible; a quiet adjustment is not. The dimension judgements are yours throughout, the tool orders them, it does not make them.
The profile template ships in the Sample Defence Pack (v1.2, a free update for existing buyers). The classification itself is open source in the methodvahti module.
Try the reasoning first
The free tools run locally in your browser. They are useful on their own and show how MethodVahti thinks before you buy anything.
€49 · one payment
The Sample Defence Pack: what is in the box
One zip, eight files, for the point where exploring is no longer enough and the reasoning has to survive a protocol, manuscript, or reviewer. Everything opens on your machine.
sample-defence-builder.html: the main tool. Six steps in your browser: study and methodology, sampling logic, adequacy and stopping rationale, coverage matrix, recruitment and stopping log, then review and export. The export is your methods wording and a reviewer defence sheet, in Markdown you can edit; your study saves as a plain file you can reopen and version.ai-use-audit.html: records what role AI was allowed at each stage of the study and what it actually did, so your disclosure matches your practice.adequacy-log.xlsx: recruitment, coverage, stopping decisions, and planned-versus-achieved reconciliation.methods-and-defence-templates.md: editable wording for the protocol sampling section, adequacy branches by tradition, a reviewer-response variant, the six-question reviewer defence sheet, and the design defensibility profile (new in v1.1).- Three PDFs: a quickstart, a worked example carried end to end, and a terminology guide that maps your analytic tradition to the adequacy and stopping concepts compatible with it.
It never states that saturation was reached for you. Fourteen-day refund if the pack does not give you a clearer sampling section.
Technical details
The exact facts, for the researcher, the IT review, or the AI agent evaluating the pack.
- Format: two self-contained HTML tools that open in any modern browser, one Markdown template file, one .xlsx log, three PDFs, and LICENSE.txt. Nothing to install.
- Network: zero network calls in both tools, counted: no fetch, no XMLHttpRequest, no external URLs. Study details never leave the browser.
- Data: the builder saves your study as a plain-text
.mvstudy.jsonyou can back up and version; exports are Markdown and plain text. - Delivery and licence: zip download through Polar checkout. One personal licence, 1.x updates free, no account, no subscription, no telemetry.
- Open method: the sampling models and the defensibility classification are open source in the methodvahti module; VALIDATION.md states what is and is not yet validated.
For doctoral schools and research teams
Use the same sampling record across a project, let supervisors review the reasoning, and keep each revision visible.
Ask about a team licenceWhat MethodVahti does not decide
It does not certify that your sample is adequate, declare saturation, infer causality, or judge the quality of the study. The modelled range is decision support. You decide the sample and defend the reasoning.
The sampling-heterogeneity model is a Vahtian hypothesis and is not yet externally validated. Design defensibility is a separate question: MethodVahti reports it as an ordinal judgement with the full dimension profile, never as a number: RoB 2 and AMSTAR 2 do not sum their domains, and neither do we. The assumptions, coefficients, and validation plan are open.
Read the method
The fuller explanations belong in Learn, where they can be read and cited without sitting inside a sales page.