Balance assay batches before processing samples

Map the sources of laboratory variation

An assay batch can differ because of reagents, operators, instrument state, storage time or processing order. If all samples from one comparison group are processed in one batch and all from another group in a different batch, the biological comparison becomes entangled with the laboratory process.

The NIST guidance on randomised block designs explains how blocking can account for nuisance variation while randomisation operates within blocks. Apply that design logic before the laboratory starts work, with a plan appropriate to the assay and specimen constraints.

Build a processing allocation table

List the specimens, comparison groups, repeated measurements and factors that must remain linked. Identify practical batch capacity and reserve positions for calibrators, controls and other required checks. Randomise or balance the relevant study groups within batches so the comparison is not identical to processing date or plate.

Decide whether paired samples should be processed together and how that decision affects the scientific contrast. Record the randomisation procedure and preserve the allocation file. Use laboratory identifiers that maintain blinding where feasible, with the identity mapping held under the study's access controls.

Hypothetical example

A fictional experiment has 24 specimens, 12 from each of two groups, and two runs that can each hold 12 study specimens. Putting one group in each run makes group and run perfectly aligned. A planned allocation of six specimens from each group in each run separates the group comparison from a simple run effect more effectively.

The six-per-group allocation is an illustration, not a universal assay requirement. Extra constraints such as paired specimens, plate positions and control wells may require a different randomisation design.

Preserve deviations and quality evidence

Record actual run, plate position, lot, operator and processing date. Document failed runs, repeat assays and reasons for rerunning a sample without selecting results based on the preferred biological finding. Keep original and repeated measurements linked.

Before analysis, inspect group balance across batches and quality-control patterns. A post hoc batch adjustment cannot reliably recover a comparison that is completely confounded with batch without additional assumptions. Use the analysis decision meeting worksheet to record justified departures from the processing plan. Report the allocation, quality checks and remaining batch limitations so readers can distinguish design protection from statistical correction after the fact.

Source verification: 6 November 2026. Examples are hypothetical.