The limiting element of descriptives is they do not control for other factors. This may result in spurious results as it appears one factor is important, when in fact an entirely other one is. A classical example of that is that ice cream sales appear to drive murder rates in the US (are highly correlated with it technically). In fact heat drives both ice cream sales and murder rates. Regression is a relatively simple way to deal with this type of problem. For income, the regression results reflect the difference between being a customer in one of the categories (such as such as being a female in the female variable) and not being in the category (being male in this example). If the results are positive then you earn more than those not in your category, if they are negative you earn less. For instance, if you are female in the variable female and the result is 100 then you earn a hundred dollars more, if the result is negative 100 you would earn 100 dollars less. For variables that have more than two levels, like age, one of the levels is not calculated this is the reference level in the tables.

For employment, both two and four quarters you are predicting whether someone is employed or not. The results show the percent increase (if the results are positive) or decrease (if the results are negative) in the chances of being employed. So for factor female, if the result is .1 then being female increases the chance of being employed by ten percent. If they are -.1 it shows females are ten percent less likely than males to be employed. The data for this analysis includes as before customers with cases closed between July 2017 through December 2019.