RMResearch Metrics AfricaPublic research toolsAll three tools

OBJECTIVE-FIRST · ASSUMPTIONS VISIBLE

Plan the method.
Then test the fit.

This planner gives a defensible starting point from the information you choose. It does not inspect your dataset or replace protocol review, diagnostics or analyst judgement.

STEP 1 · DEFINE THE DECISION

Choose what your study is trying to do.

RMA ANALYSIS PLANNING RECORD

Welch t-test for two groups or Welch ANOVA for 3+ groups

Proceed with limitations until diagnostics are checked

Objective
Compare groups
Outcome
Continuous
Predictors
Groups / categories
Design
Cross-sectional
WHY THIS IS A CANDIDATE

The outcome is continuous and the objective is a group comparison.

ASSUMPTION GATE
  • Observations are independent unless the design model accounts for dependence.
  • Residual shape, variance pattern and influential observations are reviewed.
  • Groups and contrasts were defined before interpretation.
ALTERNATIVES TO CONSIDER
  • Robust regression, transformation or a rank-based method when its estimand fits the question.
  • Bootstrap confidence intervals when justified and documented.
INTERPRETATION LIMITS
  • Report the estimated difference and confidence interval, not only a p-value.
  • A non-significant result is not proof of equality.