P5.6—Evaluation, reliability, validity and confidence

Reliable data are repeatable; valid data measure the intended variable; confidence reflects how strongly the evidence supports the conclusion. Anomalies, narrow ranges, uncontrolled variables and unsuitable apparatus can weaken these separately.

Explain the direction of each weakness: temperature drift may raise or lower rate, poor resolution increases scatter, and measuring the wrong endpoint threatens validity. Replication improves reliability only when the method is valid.

Three concordant titres support repeatability, but a contaminated standard solution can make all of them systematically wrong.

“The results are reliable because they agree” ignores systematic error and validity.