Significant results (according to p values) are not a reason to act when effect sizes are low; with large administrative datasets and poor experimental design, it is easy to find a highly significant result of no research value whatsoever.
1
—
—
Significant results (according to p values) are not a reason to act when effect sizes are low; with large administrative datasets and poor experimental design, it is easy to find a highly significant result of no research value whatsoever.