notes.txt

Set up final and replicate weights assuming no difference between different modalities in HINTS 5, Cycle 3

Cycle2
Import cycle 2 data
select PersonID, PERSON_FINWT0 through PERSON_FINWT50
mutate survey = 1
rename PERSON_FINWT0 to Merged_NWGT0
*rename PERSON_FINWT1-PERSON_FINWT50 to Merged_NWGT1-NWGT50
*copy PERSON_FINWT0 to Merged_NWGT51-NWGT100
inner join with cycle 2 data by PersonID

Cycle3
Import cycle 3 data
select PersonID, TG_all_FINWT0 through TG_all_FINWT50
mutate survey = 2
*rename TG_all_FINWT0 to Merged_NWGT0
*copy TG_all_FINWT0 to Merged_NWGT1-NWGT50
rename TG_all_FINWT1-TG_all_FINWT50 to Merged_NWGT51-NWGT100
inner join with cycle 3 data by PersonID

bind_rows(Cycle2, Cycle3)
mutate_at "survey" to factor with distinct names for cycles

* work out how to do with mutate_at()

914 variables appear in either or both of cycle 2 and cycle 3
combined file adds 102 variables: cycle ID, final sample weight, 100 replicate
weights, for 1,016 variables
final sample sizes are 3,504 for cycle 2 and 5,438 for cycle 3; total 8,942

Richard Moser, NCI
moserr@mail.nih.gov