0

My data is panel and its on employee attendance and weather. I observe employees daily attendance for about 2-3 years. So I have a panel at the daily level for each employee.

I have a binary X variable that measures the number of days the maximum temperature was in a particular bin in my sample. My Y variable is also binary and it measures absenteeism from the work place. I include a bunch of fixed effects (such as day of the week) in my regression and find no impact of the temperature bin variables on Y.

Now I wish to investigate whether this is because I don't have much variation in my temperature bin variables. However,I am not sure how I can do this. Any suggestions would be of great help.

Ridhima
  • 163
  • 1
  • 1
  • 6
  • 1
    The general consensus on this site is that binning your data is not helping (see e.g. [here](https://stats.stackexchange.com/a/230756/176202)). Construct a spline, or introduce non-linearity some other way, but binning is just throwing away information. – Frans Rodenburg Apr 14 '21 at 12:18

0 Answers0