91Ó°ÊÓ

Understanding the null and alternative hypothesis is a foundational block in hypothesis testing. In simple terms, the null hypothesis (\(H_0\)) is a statement that there is no effect or no difference - it represents the status quo or a point of skepticism. For instance, in the exercise focusing on hormone replacement therapy (HRT) and dementia, the null hypothesis asserts that HRT does not affect the risk of dementia. This is articulated as \(p_h - p_p = 0\) where \(p_h\) and \(p_p\) are the proportions of dementia in the hormone and placebo groups, respectively.

The alternative hypothesis (\(H_1\) or \(H_a\)), on the other hand, is the claim you hope to support with evidence. In our exercise, the alternative hypothesis posits that HRT does influence dementia risk and that this risk is higher compared to the placebo, stated as \(p_h - p_p > 0\). Crafting these hypotheses clearly is crucial because they guide the entire hypothesis testing process.

Calculating the test statistic is a critical step that helps us to determine how far our sample result is from the null hypothesis. In hypothesis testing, the test statistic, often a Z-score in proportion comparison cases, is a way of quantifying the difference between groups. For the HRT study, the test statistic is computed using the formula:

\[\begin{equation} z = \frac{(p_h - p_p) - (\text{difference under }H_0)}{\sqrt{\frac{p(1 - p)}{n_h} + \frac{p(1 - p)}{n_p}}} \end{equation}\]\
with \(p\) as the pooled proportion of dementia cases over both groups, and \(n_h\) and \(n_p\) representing the sizes of the hormone and placebo groups. When the computed Z-score is extreme enough, which is determined by the context and the chosen level of significance, it suggests that the observed result is unlikely under the null hypothesis, leading us to consider the alternative hypothesis.

The p-value is a key concept in hypothesis testing, representing the probability of obtaining a result at least as extreme as the observed one, assuming the null hypothesis is true. A low p-value indicates that under the null hypothesis, our result is unusual. The significance level (\(\alpha\)), often set at 0.05 or 0.01, is a threshold to determine when a p-value is considered sufficiently low to reject the null hypothesis. In the HRT study, testing at the 1% level of significance (\(\alpha = 0.01\)) sets a high standard for evidence against the null hypothesis. The result is compared against a critical value from the Z-distribution, and if the test statistic exceeds this critical value, the p-value is considered small enough to reject the null hypothesis. This approach helps control the risk of incorrectly rejecting a true null hypothesis, known as Type I error.

Comparing proportions between groups is often essential in statistics to assess whether there is evidence for a difference. In the case of the HRT study, the objective is to compare the proportion of dementia cases between hormone users and non-users. Through a series of calculations—first identifying the sample proportions, then calculating the pooled proportion, and finally the test statistic—we can infer about the population proportions. By comparing this inference to the set significance level, we can judge whether the observed difference in sample proportions reflects a true difference in population proportions, or if it could be due to random chance in sample selection. When the groups' proportions differ significantly, it may indicate a real contrast in the rates of an outcome, such as the incidence of dementia in the medical study we examined.