Mean Absolute Error of Rubric Items

If we had a rubric with two items, both worth one point, the error contribution of one submission would be:
For expert answer {1, 0},
0 if grader is {1, 0}
1 if grader is {0, 0} or {1, 1}
2 if grader is {0, 1}
It's recommended to normalize_full_credit() for eval1 and eval2 prior to using this function.

Usage

rubric_mae(eval1, eval2)

Arguments

eval1

first dataframe of Gradescope evaluations

eval2

second dataframe of Gradescope evaluations

Value

double for mean absolute error