stats lab thought

[carriewright11]

For the practice we have a logistic regression which I think is nice to have them practice, but it is generally better practice to leave continuous variables as is and I have seen people do this a lot when it isn't appropriate - I wonder if we could do something else with a binary outcome that isn't just binning based on a threshold or just change this to not be logistic regression and evaluate other variables?

[carriewright11]

Maybe we could just change P3 to make a new fake binary variable that is something like yes no added a new water cleaning method or something like that that is just a random sequence of yes and no and say it is for learning purposes for logistic regression?

[carriewright11]

• maybe change * Add the use = "complete.obs" argument. to * Add the use = "complete.obs" argument.

[carriewright11]

maybe add a cor example with just two vectors