I would like to convert the average daily minutes into weekly hours using pre-ATUS survey years (1975 and 1998-2001, in particular). I have been separately using both the MTUS and AHTUS datasets.
Would I simply do this by dividing the weighted means (using propwt for MTUS and recwght for AHTUS) by 60 (minutes) and multiplying by 7 (days), or are there other considerations/methods to do this?
(I know that, of course, the caveat is that this technically represents one day as opposed to a whole week, but I am trying to replicate previous studies that have used hours per week with the historical time use data.)
Your approach sounds reasonable to me (with your stated caveat that a single day is not necessarily representative of an entire week), but it is beyond the scope of IPUMS User Support to provide analytical recommendations. This 1994 Robinson & Bostrum paper from Monthly Labor Review uses some of the samples you are interested in and constructs synthetic weeks. My best recommendation is to look at other papers comparing 24 hour time diary data to either longer time diary recall periods or stylized questions. I also encourage you to review the IPUMS AHTUS sample characteristics page (notably that the 1975-1976 data may include multiple days for respondents).