91Ó°ÊÓ

When comparing male and female memorization abilities, statistical hypotheses help us set a framework for understanding our analysis. The *null hypothesis*, often denoted as \(H_0\), is a claim we assume to be true unless evidence suggests otherwise. In this case, the null hypothesis is that there is no difference in the average number of words recalled by males and females, mathematically expressed as \(H_0 : \mu_m = \mu_f\). This implies that any observed difference is due to random chance, not a systematic difference in performance. On the other hand, we have the *alternative hypothesis*, \(H_1\), which is the statement we want to test against the null. Here, it posits that there is a difference in recall ability between the two genders, represented as \(H_1 : \mu_m eq \mu_f\). This suggests that any difference in mean recall isn't just a fluke, but indicates a true distinction in memory abilities between males and females.

The difference in means is a simple yet powerful statistic in our analysis of memorization ability. To calculate this, we subtract the mean number of words recalled by one group from the other. In our scenario, the males recalled an average of 12 words, while the females averaged 14 words. Therefore, the difference in means is \(14 - 12 = 2\). This difference is our observed test statistic and serves as a crucial point in the permutation test. A permutation test leverages this difference by examining if such a statistic could have arisen simply by chance. By shuffling or "permuting" the data and subsequently calculating new mean differences from the rearranged data, we assess the likelihood that our observed difference could happen randomly. If we frequently see mean differences of 2 or more in these hypothetical scenarios, the original observed difference might not substantiate claims of differing memorization capacities. But if such results are rare, it lends weight to the evidence that there truly is a difference.

The exercise aims to determine whether there is a real difference between males and females in their ability to memorize words. The focus is on conducting a permutation test, which provides a non-parametric method to test the hypothesis. This means it's less reliant on assumptions about the data, such as normal distribution, making it versatile and robust. So, how do we go about this test? Follow these basic steps to perform the analysis:

• First, pool all test scores together: 10, 12, 14, 11, 14, 17.
• Next, randomly split this data into two groups of three, treating each split as a unique permutation.
• Calculate the mean for each group and find the difference in these means; this is your test statistic for that permutation.
• Repeat this process many times, often thousands, to create a distribution of test statistics under the assumption that the null hypothesis is true.

An example permutation might assign (10, 11, 12) to one group and (14, 14, 17) to the other, resulting in a test statistic of 4, since the means are 11 and 15. Compare all obtained test statistics to the observed difference of 2. The larger the number of times a computed difference meets or exceeds 2, the more support is drawn for random chance causing any differences; otherwise, if rare, it suggests a true discrepancy in memorization capabilities.