I am trying to identify the gender of an own household child aged 15-17 using ATUS-X data. I cannot use the activity-level files and the variable SEXW in the “who” file as that would restrict my identification to only children present at activities.
Under the “Constructed Child” category, IPUMS provides age groupings based on the household roster (e.g., DTRUND18, SONUND18 …). However, the groupings are too large (13-17 years old). Is there any way to disaggregate these groupings? Or is there any way to access the Roster file through IPUMS, or would I have to merge it using the original BLS data?