Building an Experiment Review Sheet

Standardize post-run audits with structured checklist protocols for data entry validation, outlier screening, and residual diagnostic verification.

2026-11-05
Martin Locke
4 min read
Process Verification4 min read
Published: 2026-11-05
Building an Experiment Review Sheet

Contents Overview

Article Sections

Why Use a Review Sheet

A review sheet is a short, structured record that travels with an experiment. It makes the plan, the execution and the analysis checkable by someone who was not in the room.

It also helps the experimenter catch problems such as missing runs, changed settings or data entry errors before they affect conclusions.

Key Principle

The effect of one factor may depend on the level of another.


What to Include

Record the objective, the response and how it is measured, each factor with its levels and the reason for them, the design type, the number of replicates, and the randomized run order actually used.

After the runs, note deviations from the plan, unusual observations, and the results of residual checks for normality, constant variance and independence.


Using the Sheet in Review

Reviewers can compare the planned and actual run order, confirm that the analysed data match the recorded data, and see whether the conclusions stay within the tested ranges.

Keep the sheet with the worksheet and analysis output so the full chain from plan to conclusion is preserved.

Swipe to view all columns

Review itemWhat to record
Objective & responseQuestion being asked; response and measurement method
Factors & levelsEach factor, its levels and the reason for the range
DesignType, number of runs, replicates, center points
ExecutionRandomized run order used; deviations from plan
Analysis checksResidual plots reviewed; conclusions within tested range

Key Takeaways

  • Document objective, factors, levels, design and run order.
  • Record deviations and unusual observations as they happen.
  • Check residual diagnostics before drawing conclusions.
Martin Locke

Martin Locke

Contributor

Educational article for FactorInteraction Notebook. It explains planning logic and does not report experimental results or statistical significance.

Editorial Method