We appreciate the statistical critique of our article and clarification for readers provided by Dr Petrini. However, statistical probability is only one source of evidence to be weighed in testing a priori hypotheses; others include consistency and strength of the data.
Our primary hypothesis, that the women from the original walking intervention group would report higher levels of walking than those from the control group, was clearly supported by the data (P = .01 adjusted for multiple comparisons). In addition, the amount of sport and recreational activities reported by the women from the walking group was apparently higher than those from the control group, although the difference was not statistically significant. However, as indicated by Petrini, we may have had limited statistical power to test our secondary hypotheses.
Pereira MA, Kriska AM, Kuller LH. Postmenopausal Women and Walking. Arch Intern Med. 1999;159(7):756-757. doi: