Marital and Cohabitation History Dataset
for a project, I want to look at the year couples got into a relationship so using the marriage datafile to obtain this information. I then want to add on characteristics of each individual from the latest wave of understanding society indresp file to assess whether couples characteristics are similar. I have cut the marriage dataset down to people who started getting sampled from wave 1 of understanding society and who have an ongoing indicator. Should I then merge on the indresp dataset for all waves 1-9 in understanding society or would it be okay to just merge on the latest available characteristics so wave 9 only?
Is the marital and cohabitation dataset going to be updated from 2015? how come the dataset stops here but I have the same people in the indresp datafile whose information has been collected since 2015 and they therefore appear in wave 9? I'm just concerned that I'm looking at w1-w9 marital status with 2019 characteristics but the marital status of the individual could have changed since then?
If I wanted to run weighted analysis what weight would you suggest I use as the marital dataset is longitudinal and if I only match on Wave 9 for example that's cross-sectional data? Also the marital dataset doesn't include the IEMBS as that boost came in wave 6 - do I need to account for this?
I'm just looking for advice on whether it is wise to merge these datasets or whether its best to avoid? your help would be much appreciated or alternatively if you have a better way to achieve this type of dataset it would be really helpful
#2 Updated by Alita Nandi 2 months ago
Yes you are right marital status since 2015 could have changed for some respondents. The updated version of this file will be released soon (expected by the end of February, if not sooner). This will include IEMBS and additional information collected in Waves 7-9.
Understanding Society User Support Team
#6 Updated by Olena Kaminska 2 months ago
Let's assume you will analyse wave 6 with the marriage datafile. For this you can use ub_xw weight. This will exclude IEMB sample for you (I understand you don't have this information for this sample).
When you have a new release for wave 9, you will be able to use wave 9 cross-sectional weight with the marriage datafile.
Conveniently, by weighting you will also exclude all the people for whom you have out of date information on marital status as they would have needed to respond in wave 6 or 9 and therefore this information will be up-to-date.
Hope this helps,