Population Assessment of Tobacco and Health

Sampling weights Tips for Using the PATH Study Data User Forum

Return to the Population Assessment of Tobacco and Health (PATH) Study Series page.

6 posts / 0 new
Last post
oosi
Sampling weights

If I want subsamples from the adult data (say ages 18 - 34) and the youth data (ages 12-17). Is it possible to merge the youth and adult data so I am able
to look at this subpopulation and if so, how do I go about dealing with the replicate weights provided for individual data sets so that my analysis is representative of the populations from where they were taken from (the youth has replicate weights: R01_Y_PWGT1 - R01_Y_PWGT100; and adults: R01_A_PWGT1 - R01_A_PWGT100). So in the merged
dataset, which do I use as the replicate weight?

Thank you.

Ola

kayemarz
Recode to combine the youth and adult weights into one variable

While specific guidance is not included in the User Guide, it is recommended to use the weights corresponding to the data files for which they are provided. So, in this particular case, the weight variables will need to be recoded to combine the youth and adult weights into one variable such that for a youth, their youth weight should be used and for an adult, their adult weight should be used.

--Kaye

hgc
Aged-up adolescents

I am looking at the transition of tobacco use between wave 1 and wave 2 among wave 1 adolescents. If I understand correctly, 17 year olds at wave 1 who became 18 year olds at wave 2 are included in the wave 2 adult dataset. I am wondering what the weighting and replication procedures I should apply for these "aged-up" adolescents. Thank you!

kayemarz
Weights for analyses on aged-up adolescents

Your analyses should use the Wave 2 adult weights for the adolescents that aged-up to adults in Wave 2. Whatever group the participants are in indicates the weights that should be used. Thanks for checking.

--Kaye

hgc
Weights for aged-up adolescens

Thank you for the reply, Kaye! I appreciate it.

Just want to check if I understand correctly. If I want to look at 12-17 year olds at wave 1 as a whole, I should combine the adolescent and adult wave 2 data and create a weight variable that takes the value of wave 2 adolescent weights for adolescents who stayed adolescents and adult weights for aged-up adolescents. Same for the replication variables. Is that correct?

Hui

kayemarz
Weights for aged-up adolescents

Hi, Hui. Your interpretation is correct. Thanks for checking.

--Kaye

Log in or register to post comments