sta199-s23-2 / project-statisfactory

https://sta199-s23-2.github.io/project-statisfactory/
0 stars 1 forks source link

Results #6

Open elignesin opened 1 year ago

elignesin commented 1 year ago

Overall, this is a good start, but some thoughts:

  1. For this one, when you say it seems that age had very little effect, this is again an indication that you could be using a hypothesis test here do determine whether age actually has no effect or not. I also like that you are relating your results to your prediction, but they should be related more to an overall hypothesis rather than individual predictions for an individual plot.
  2. This second one is pretty good, you do a good job about talking about what you see from your plots and how it fits in your assumptions, though again saying "no definitive effect" suggests you may want to consider a hypothesis test or some more concrete statistical analysis than evaluating a graph by eye, even if you look at a quantile comparison by race.
  3. This is for both graphs 4 and 5, but there isn't much of a justifiable reason here to be using logistic regression nor to be using a scatterplot here.
  4. For graph 6, this is a pretty good discussion, though again saying little correlation without demonstrating how you know that.

This is an overall note, but when you talk about whether you think something is correlated, simply using the "eye test" and then saying it seems like something is true is not sufficient. You can use the eye test and say it seems like this is true, now let us back it up with some sort of statistical analysis (at the very least give the conditional probabilities!), but you may want to include hypothesis testing for at least some of these variables to give concrete statistical evidence backing up your intuition.

Overall, you should also aim to include an overarching discussion of your results in the context of your hypothesis, as well as potential limitations of your analysis and where you might be able to go from here.