Prerequisites
Sampling, confidence intervals, study design, and critical reading.
Identify the population, comparison, outcome, and design before interpreting a numerical result. A narrow interval concerns sampling uncertainty under a model; it does not remove confounding, measurement error, or limited transportability.
Worked example
An observational difference can be estimated precisely while remaining noncausal. Precision answers a different question from whether the groups were otherwise comparable.
Model and Assumptions
Fictional report: 120 volunteers chose either Workshop A or B. At follow-up, A's mean rubric score exceeded B's by 4 points, with a model-based 95% confidence interval from 1 to 7. No baseline scores were recorded. The abstract says A caused better writing for all students.
Further inquiry
Write a 200-word peer-review note separating internal validity, statistical uncertainty, and transportability. Propose an improved design and identify one limitation that it would still leave.
Review criteria
- Match each criticism to a specific claim.
- Explain why sample size alone does not remove confounding.
- Retain what the original study can legitimately show.