6P.5—Evaluating results, uncertainty and errors

Evaluate a result by considering random variation, systematic bias, measurement resolution, repeatability and sample size. State how the limitation changes confidence in the conclusion.

Use ranges, standard deviation or error bars when appropriate, identify anomalous points with evidence, and propose an improvement that addresses the cause rather than adding detail at random.

If replicate oxygen readings overlap widely, the apparent treatment difference may be smaller than the uncertainty; more controlled repeats could clarify it.

A precise measurement can still be invalid if the method is biased, and an anomalous point is not automatically deleted.