GatesProject_DL_OC_Fall2019_Practice_Cleaned_AddGaming

-- This datasets adds a column ("CF (IsGaming)") indicating whether a transaction is part of a gaming pattern, and multiple columns ("CF (GPat1-13)") indicating which pattern it is. It is based on Paquette, de Carvalho, & Baker (2014)'s gaming detector implemented by Steven Dang, bug fixed and refined (e.g., similarity computation) by Yun Huang (confirmed by both Paquette and Dang). In Yun's local labels, this version of the gaming detector is labeled as "V2a". The pattern columns after GPat13 (e.g., GPat2i) are added by Yun by decomposing "attempt" into "correct" and "incorrect" in a gaming pattern in Paquette's 13 patterns.
-- In addition to the preprocessing in "GatesProject_DL_OC_Fall2019_Practice_Cleaned" dataset (with multiple attempts per step), this dataset further removes students that didn't access both pre- and post-test. The final 129 students are the students used in Huang et al. (2021) LAK paper.
-- Key preprocessing in pre/posttest include filling zeros for missing problems and missing steps, keeping only first attempts per step if pre/posttest provided hints, or keeping the last attempts otherwise, allowing case insensitive answers. Details can be found here: https://docs.google.com/document/d/1E_GhzGAdPg6qktzrSLhsFu6Vp-VOKqY_asKCl03JvH0/edit#heading=h.53zg521j5j2b
-- Any further questions, please contact Ken first and then feel free to Yun if needed ([e-mail])