Following a Non-Significant Finding (1 of 3)

---

An experimenter wishes to test the hypothesis that sleep deprivation increases reaction time. An experiment is conducted comparing the reaction times of 10 people who have missed a night's sleep with 10 control subjects. although the sleep-deprived subjects react more slowly, the difference is not significant, p = 0.10, two tailed. Should the experimenter be more or less certain that sleep deprivation increases reaction time than he or she was before the experiment was conducted. The naive approach is to argue that there was no significant difference between the sleep-deprived and the control group so the experimenter should now be less confident that sleep deprivation increases reaction time. This argument implicitly assumes that the null hypothesis should be accepted when it is not rejected. A more straightforward and more correct approach is to consider that the experimenter expected the sleep-deprived group to have slower reaction time, and they did. The experimenter's prediction was correct. It is just that the difference was not large enough to rule out chance as an explanation.