Discussion of AnneTheAgile's comment: Source: My summary of the Conclusion+Skim+My
ideas.
Summary:
1.Motivates and describes the use of a metric
for quantifying experimental results.
2.Effect size is normally a
highly semantically meaningful measure of differences found in an
experiment, so it should be included in all results, with preference
over tests of statistical significance, which emphasize the sample
size's significance.
3.Raw 'unstandardised' mean differences,
together with a confidence interval, might be better than the Effect
Size metricwhen a sample has restricted range or does not come from a
Normal distribution, or if the measurement from which it was derived
has unknown reliability.