The frequentist approach is most commonly used to test these hypotheses and weighs the evidence against the null hypothesis by the P-value that is defined as the probability of observing a stronger association than that estimated from the data, when there is indeed no association. For example, if the association is measured by the odds ratio for the disease between exposed (in this case carriers of the risk allele) and unexposed subjects, the P-value is the probability of observing an odds ratio more extreme than that estimated from the data, assuming that there is no association. If there is association, the P-value should be very small because estimating a large odds ratio would be unlikely to happen. Therefore, a small P-value is taken as evidence against the null hypothesis, or evidence of a significant association, and this is the rationale for the decision rule to reject the null hypothesis when the P-value is smaller than a fixed threshold. The fixed threshold is called the significance level. This decision rule is not error free and rejecting the null hypotheses when it
