"Repeat your readings" becomes uncertainty propagation

Repeat and take a mean; identify anomalous results; distinguish accuracy from precision informally; mention "human error".

Every measurement carries a stated absolute uncertainty (±half the smallest division for analogue, ± the stated tolerance for digital/glassware). Convert to percentage uncertainty. Then propagate: for quantities multiplied or divided, add percentage uncertainties; for quantities added or subtracted, add absolute uncertainties. The final answer's uncertainty determines the appropriate number of significant figures. Random error (reduced by repetition, shows as scatter) is rigorously distinguished from systematic error (not reduced by repetition, shows as a consistent offset, detected by comparison with a literature value / percentage error). "Human error" is not an acceptable phrase.

Error analysis moves from a paragraph of prose to arithmetic that must appear in the working. It also becomes high-stakes: the IA is 20% of the final grade and its analysis and evaluation criteria are largely about this.

GCSE practical assessment is qualitative in this respect.

Quoting an answer to 8 significant figures from a calculator; using "human error" in an evaluation; not recording uncertainties at the point of measurement (unrecoverable later); confusing precision with accuracy; not noticing that a burette reading involves two readings and therefore double the absolute uncertainty.