Today we peer reviewed group project proposals. I came into it with a project proposal that I was a little concerned about: predicting Rotten Tomato ratings. I wasn’t sure if it was a good enough or original enough idea. My peer review group approved it and thought that it would be an interesting topic. They did warn me, however, that there is already some fivethrityeight data on this topic so I will have to take a look at that and make sure that our project is different enough to explore.
My classmates proposals were interesting as well. I saw group projects exploring NBA wins and fatal car accidents. I am a big basketball fan and sports statistics are always super interesting in my opinion. I will be looking forward to that presentation. The fatal car accidents project was interesting as well, we helped this group narrow in on a specific research question but fatal car accidents can be a good topic, very topical for a Minnesota winter. One group had some sample size issues, but we were able to work through ways that we could work around it. One group used the whole population as the sample so we were able to help narrow that down and also help the strength of the relationship between the variables.
I would recommend to my classmates to make sure that you have a project that you are passionate about and a project that you would actually be interested in the results. I did a group project last year about golf, and I am not a huge golf fan. This time around I will be more engaged, whether we stick with our current topic or not. Based on what I saw from other groups, I would be careful with independence and sample size. Your linear regression will not be reliable if you have dependent data with a small sample size. Find a way to expand your data, whether that be to other regions, or narrow down your response variable a little bit.
In class we worked with interaction models in multiple linear regression, conducting partial F-tests and such. This was mostly review from the last class, so I won’t go into the interpretation of code again. I appreciate taking the time to slow down on this topic, as it was very confusing to me at first.