State the opposite of what you hope to find (e.g., the mean number of parent–child conflicts during intervention is the same or greater than the mean number of conflicts during baseline)....

State the opposite of what you hope to find (e.g., the mean number of parent–child conflicts during intervention is the same or greater than the mean number of conflicts during baseline). Statisticians refer to this as the null hypothesis. It traditionally is stated this way to put the burden of “proof ” on the intervention/evaluation. Paradoxically, this is the hypothesis that is tested. If the null hypothesis is rejected, your evaluation hypothesis is supported because it’s the opposite of the null hypothesis.