There is a task that doesn't appear to make much sense to me in a textbook. I have a dataset of categorical variables (the features and target are categorical) . I am asked to determine what are the most important variables that explain the target variable and whether they are statistically significant. I have always thought that variable ranking can only be achieved by applying a supervised learning algorithm such as logistic regression or random forests, boosting etc but that is explicitly not allowed in this task. How can I otherwise do this?