The Data, Evidence and Conclusions
What you must understand
Data become evidence only through interpretation. Candidates must distinguish qualitative and quantitative data, accuracy (closeness to true value) from precision (closeness of repeated measurements), and random from systematic error. Reliability is improved by repetition and standardisation; validity by controlling confounders and choosing appropriate measures. Descriptive statistics - mean, median, mode, range, standard deviation - summarise; inferential reasoning asks whether an observed difference could plausibly arise by chance. Understand significance levels, p-values and confidence intervals conceptually: a p-value is not the probability the hypothesis is true, and statistical significance is not practical importance. Effect size and sample size both matter. Conclusions must be proportionate to the evidence, acknowledging limitations, generalisability and alternative explanations.
Traps in questions
Frequent traps include treating a small p-value as proof of a large or important effect; confusing precision with accuracy when a systematic offset shifts all readings; generalising from an unrepresentative sample; and ignoring outliers or discarding them without justification. Beware p-hacking and selective reporting, and the claim that a non-significant result proves no effect - absence of evidence is not evidence of absence. Questions often ask what additional data would strengthen a conclusion; look for replication, larger samples or controlled designs.
Rule of thumb
Before accepting any conclusion, ask three questions: How strong is the effect? How uncertain is it? Could something else explain it? If any answer is missing, the conclusion is overstated.