Issue with cohen's kappa calculation in Part 3

[timothyLeeXQ]

Hi Dr Lang,
I was working on Part 3 of Assignment 7 and got a kappa metric that is more than 1 using the code that was already in the assignment:
kappa(matrix2, exact = TRUE)/kappa(matrix2)

This should be impossible, so I calculated kappa manually and got numbers that made sense. I also tried psych::cohen.kappa() and got numbers that agreed with my calculations.

The documentation for base::kappa() suggests that the function is supposed to conduct QR decomposition of a matrix rather than cohen's kappa. As far as I can tell, these two concepts aren't related, and dividing kappa(x, exact = TRUE) by kappa(x) just gives a metric of how 'off' the estimate (computed when exact = FALSE) was from the actual decomposition (computed when exact = TRUE), rather than cohen's kappa.

Am I right about this? If so, I think it could be worth flagging to the class. I've submitted a pull request so you can replicate the issue - the relevant portions are within lines 151 and 260 of the .Rmd file.

Thanks,
Timothy