liud4 / rVMAP

Data Management code for VMAC-MAP study
Other
3 stars 0 forks source link

No factor variable for cdr related variables in eligibility data #29

Closed liud4 closed 4 years ago

liud4 commented 4 years ago

The following cdr variables from the eligibility database should have factor version. But I did not see it in the merged dataset: cdr, cdr_mem, cdr_orient, cdr_judg, cdr_affairs, cdr_hobbies, cdr_care. For example, I saw cdr.elig in the merged dataset, but not cdr.factor.elig.

In addition, cdr_boxes in the eligibility database (named cdr.boxes.elig in the merged dataset) should be a numeric variable instead of a character variable.

https://github.com/liud4/rVMAP/blob/8f722d8362e7e7e3b8aeffd842c18f9e9a88a663/R/process_eligibility.R#L45-L52

omair-a-khan commented 4 years ago

@liud4 In addition to cdr_boxes, are there any other variables in the eligibility data that should be numeric? We are manually setting the following to be character:

    cdr = as.character(as.numeric(cdr))
    cdr_boxes = as.character(as.numeric(cdr_boxes))
    cdr_mem = as.character(as.numeric(cdr_mem))
    cdr_orient = as.character(as.numeric(cdr_orient))
    cdr_judg = as.character(as.numeric(cdr_judg))
    cdr_affairs = as.character(as.numeric(cdr_affairs))
    cdr_hobbies = as.character(as.numeric(cdr_hobbies))
    cdr_care = as.character(as.numeric(cdr_care))
omair-a-khan commented 4 years ago

As per our phone call today, I will update process_eligibilty.R to use process_factor_variables.R. This should correct the values for the CDR substituents to have levels 0, 0.5, 1, 2, and 3.