yuqiliao/InternationalAssessment_MachineLearning

Rewrite the code to clean PIRLS data

Closed this issue · 2 comments

  • Missing rule
  • Recode plausible values to achievement level dummy variables
  • Choose student and teacher variables (using grepl) Line 40-45

@yuqiliao Why do we use asribm01 which is on scale 1-5 for Y? Can we not use rrea and then average it and get achievementLevel?

that's an option! I did that simply to follow the TIMSS paper. In the TIMSS/PIRLS, besides 5 PV, the datasets also provide 5 proficiency levels (such as asribm01) based on 5 PVs, i thought maybe it is "more correct" to use these 5 proficiency levels. you are right, we could totally do what you said.