Compare before you combine
Check that investigations address comparable questions and use equivalent variables, units, samples, timing, controls, measurement rules and repeats. Then compare the direction, size and spread of findings.
Name the error mechanism
Replace “human error” with a specific cause and pathway: what differed, which measurement it affected, and whether it could shift values high, low or unpredictably.
Four useful error patterns
Random variation changes readings unpredictably; measurement or reading errors arise from tools or technique; inconsistent procedures make trials non-equivalent; confounding variables change alongside the intended factor.
Handle anomalies transparently
An anomaly is an unusual value, not automatic rubbish. Check records and equipment, repeat where possible, retain raw data and justify any exclusion using evidence.
Select relevant evidence
Use the measurements that link the changed condition to the outcome. Prefer matched groups, repeats, means and spread where useful. Do not cherry-pick a convenient trial or use evidence for a different question.
Write a reasoned conclusion
Answer the original question, cite the relevant pattern or values, limit the claim to the tested sample and conditions, and acknowledge a method limit. “The hypothesis was correct” is not enough.
Judge the strength of a claim
Small or biased samples, overlapping results, missing controls and systematic error weaken certainty. Imperfection does not make all evidence useless, but it should narrow the conclusion.
Ask what comes next
Pose a measurable question that tests an uncertainty, mechanism or purposeful new factor while preserving controls. A good extension follows logically from the findings or limitation.