Skip to contents

For each research question, record what you expect to find, which measurements you will use and how you will analyse them, before any responses arrive. surveyframe keeps this plan in the questionnaire file. You enter each research question under Plan on the Design screen, and the analyses run as written.

What each research question records

  • The research question, in words.
  • The hypotheses: the null hypothesis (H0), that there is no difference or relation, and the alternative hypothesis (H1), what you expect to find.
  • The method, such as an independent samples t-test or a binary logistic regression.
  • The variables, each in its role, such as the outcome you measure, the group you compare, or the predictors you use.
  • The significance level, alpha, where the method tests a hypothesis.
  • The decision rule: what result will count as support.
  • The objective of the study it serves.

Why before the data

A plan written after seeing the data can bend to fit it. Declared first, and sealed in the file’s hash, it shows a reader which analyses were planned and which were added.

The pre-analysis plan

render_report() with no data, or SurveyStudio before any responses are loaded, writes the plan as a document: the study, the questionnaire, each research question with its method and decision rule, and what a preregistration still asks for. See Register your plan on OSF.

When the plan must change

A change after the answers arrive is recorded as an amendment. Record what changed and why. The amendment record identifies the affected parts of the study, and the report shows each recorded change. Each amendment also carries a tier, which says how far the change reaches, from wording to the analysis itself. The amendment log stays in the file, so the change is on the record.