Making Sense Of Statistics For Family Practitioner: "p" ing with confidence

Abstract

It is common in the medical literature to be inundated with "p" values in publications. But what do they mean? To interpret p values, it is important to understand the "null hypothesis". The majority of statistical analyses involve comparisons between groups of study participants, and the comparison of interest is often called the "effect". In general, the null hypothesis states that the results observed in the particular study are no different from what might have been expected as a result of chance alone. lt is often the opposite of the research hypothesis that leads to the study. Having set up the null hypothesis we then evaluate the probability that the observed data could have resulted from chance alone. lt is this probability that is called the "p value". In other words, the p value gives probability that the observed difference could have occurred by chance alone, assuming that in reality there is no difference between the populations. The smaller the p value, the more untenable is the null hypothesis.That is to say, a small p value of 0.0025 means that there is only a 25 in 10 000 probability (0.25%) that the observed difference is due to chance alone, assuming that in reality there is no difference.