GatesProject_WM_NKA_Spring2020_Practice_Cleaned

This dataset has been preprocessed as follows:
1) Test accounts (non-actual student accounts) have been removed;
2) Data from two schools (WM, NKA) are combined;
3) Only practice assignments are extracted;
4) Invalid `Duration (sec)` and empty `Outcome` or `Step Name` rows have been removed;
5) Tutor-performed action rows (e.g., those with `tutor-performed` value in the `Student Response Subtype` column) and setIncorrect action rows (e.g., those with keyword `setIncorrect` in the `Step Name` column) have been removed;
6) Demographic data has been added when available (i.e., empty values mean the information is unavailable).

Note:
1) NKA students used iPad version (with physical keyboard) and this version is case insensitive in grading, while WM students used laptop version which is case sensitive. However, there's only a very small proportion (27/118831=0.02% rows) that could be considered as correct in the laptop version if we allow case insensitivity. Yun didn't do any regrading regarding this in practice assignments.
2) Not all the students in this dataset have both pretest and posttest activities for analyzing learning gains.