Task1

For this task I’ve chosen the dataset WAIS, which contains information about the results of 4 subsets of the Wechsler Adult Intelligence Scale for two groups of people. First group contains 37 people who do not present a senile factor, second group contains 12 people presenting a senile factor.

Natural question would be whether the people without senile factor outperforms the second group. Simple boxplots shown below suggest that it could be the case even when we consider every subtest on it’s own. Interestingly, results in specific tests vary similarly for both groups except for results in picture completion, which are considerably more dispersed for senile group.

Let us explore the data structure a bit further. On the figure below, we can see pairwise relationships among the considered tests. As one might have expected, the data suggest positive correlation among all of the pairs. The highest one can be seen between information test and similarities test.

Interestingly, there are noticable differences in these statistics when we consider the groups seperately. The group without a senile factor shows much weaker correlations with picture completion test in comparison to the other group.

The biggest difference is in the relation of picture completion test with similarities test where the first group achieve almost zero correlation. You can observe this difference more properly on the figure below.