Today in class we reviewed linear models. We went over what different language in questions means when choosing which test to use. After that, we went on to review each other’s project proposals.
Some strengths that I saw in my classmates’ proposal were that they were very well written up. Where our proposal was list-oriented, the other two proposals I looked at were well written. After seeing this I will definitely make sure our next assignments are as well written as the proposals I viewed today. Something that I really struggled with, and that the group didn’t totally consider, is lack of knowledge on the subject. One of my partners is doing data on the NBA, and in his paper I didn’t really understand why he was proposing looking at certain predictors. His proposal definitely lacked explaination to the lay-man.
Something else that I saw that wasn’t totally thought through, on my proposal either, was the size of the categorical predictor. One of my partners had a categorical group with probably 6+ categories in it. This really made me think about our predictors as well, since we are looking at seasons.
To a group planning on doing a similar project in the future, I would definitely advising deciding on a response variable first. Our group decided quickly that we were going to look at fatal crashes, but spent a long time deciding how to analyze them. In hindsight, if we knew what we wanted to measure, we could have found data supporting that quicker.