Often when we are looking at categorical variables, we will find that there are too few people in some of the categories to feel that we have a representative sample. One way to deal with this is to combine categories. Pick one variable from among Ethnicity, SES, and Study, and think about how it might make sense to combine students so that you have only two groups. For Ethnicity and SES you will need to refer to the data dictionary to get the category labels.

Describe and summarize your new binary variable and your rationale for how you decided to combine responses.

I would just like to know how I would answer this question and what factors I would need to consider .