Understanding Society User Support

Thank you for your help Alita & Victoria, that is extremely useful to know!

Additionally, concerning the youth questionnaire in wave 13 (myouth); would those individuals also be included in the larger sample within the indall data file, since I will need to know the number of siblings and birth position of the youths and therefore have to merge the data across the wave on STATA.

Thanks in advance.

Yes. Indall includes some basic information on every household member (all ages) of every responding household in that wave.

Since some variables in Wave 13 do not have enough responses, I am trying to merge variables from other waves into this wave since they are more completely answered.
For example the variable mmaju has many missing values whereas when I look back into previous wave 11 there are many more 'valid cases'.
However I am struggling to use the STATA 'merge' command to take the data from individual variable, kmaju and add it onto the data from Wave 13.
Is this possible and if so please could you give me some guidance on how to apply this change in STATA? I may wish to do the same with other variables which are not adequately answered or even included in wave 13.

I hope it is clear what I'm trying to ask, thanks again.

Hello,

To merge individual level files from different waves use pidp. For example,
use pidp kmaju using kindresp, clear
merge 1:1 pidp usng mindresp, keepusing(pidp mmaju)

But we have done this exercise for you. We provide this variable and other time invariant variables in the file xwavedat. It is an individual level file (each row uniquely identified by pidp). It contains all individuals ever present in a responding household in the 18 waves of the BHPS (N=43,272). So, after merging your analysis sample with this file, you should remove all those individuals who are in xwavedat but not in your analysis sample.

Hope this helps.

Error: For the BHPS the unique cross-wave individual identifier is "pid" and not "pidp" ("pidp" is for Understanding Society).

Hi thanks for the reply,

Unfortunately when I follow those commands for my data I end up only retaining the three variables pid, kmaju and mmaju. But I actually want to retain all my original data from Wave 13 and in effect replace mmaju information with applicable kmaju information(I understand I may lose some individuals).
I am unsure about your advice concerning xwavedat; do you mean I should merge any additional variables to that and treat it as my base data set? Howeve xwavedat does not contain siblings or birth position so I would have to add these in...

Thanks.

Hi Sharon,

Sorry about the confusion - the above Stata code was an example to show how to merge the different files. You will need to include the variables you need.

About xwavedat: Some questions such as parental info (when respondent was aged 14) is only collected the first time the person is interviewed. So, for example person A is interviewed for the first time in wave 1. Then in wave 2, B joins A's household. So, as B is interviewed for the first time in wave 2, B is asked these questions in wave 2. That is why you will find many cases with missing parental information at each wave. What BHPS data team has done, is picked up these variables from whichever wave it has been reported and puts them in one person level flat file called xwavedat. So, I was suggesting that instead of doing this exercise yourself you could just use the xwavedat.

If you would like to know more about data management for the BHPS take a look at the free online course:
https://www.understandingsociety.ac.uk/documentation/training/online/bhps

Hope this helps.