The importance of Experimental Control
One of the biggest challenges presented by rejecting the null is that there could be infinite number of plausible explanations why the null was rejected. There is no formal statistical procedure to arrive at the correct scientific alternative hypothesis. Therefore, researchers must carefully narrow the number of plausible explanations that account for rejection of the null hypothesis. This is best done by incorporating an experimental control in the study design wherever feasible. Since it is impossible to control for every possible confounding factor, good researchers must strive to control for as many extraneous factors as possible. The quality and extent of experimental control is proportional to the chance of inferring correct scientific alternative hypotheses.
Alternative hypotheses that are inferred without controls built into the study design may at best be plausible explanations as to why the null was rejected, and at worst, hypotheses the researcher seeks to endorse without appropriate scientific basis to do so.
Inferring the correct alternative hypothesis is far more important than rejecting a null- anyone can reject a null, but few can identify and infer a correct alternative.
Link to the previous article in this series:
Link to a previous article on Null Hypothesis: