Support #1218

job change since the previous survey

Added by Daniel Tomlinson over 4 years ago. Updated over 1 year ago.

Data management
Start date:
% Done:



I want to know how many people have left their job between each pair of survey waves, and would like to do this in both Understand Society and the BHPS data. I can see lots of potential variables for doing this (E.g. maybe jbbgly in the BHPS and jbbgy/m/d in USoc) but I can't tell from reading the documentation whether this will work.
Please could you let me know how I might tackle this question?



Updated by Alita Nandi over 4 years ago

  • Status changed from New to In Progress
  • Assignee set to Daniel Tomlinson
  • % Done changed from 0 to 50
  • Private changed from Yes to No

In UKHLS, you can identify people who have experienced a job change between two waves by using this condition (w_empchk==1 & w_jbsamr==2)|(w_empchk==1 & w_jbsamr==1 & w_samejob==2). But this is for those who were interviewed directly in both waves (and not proxies). If someone is interviewed by proxy then you need to see when they were next interviwed directly to identify job changes. If you want to measure further job changes then you should look at the questionnaire and follow the question route to identify those questions (w_nextstatN and w_nextjobN and x_currjobN etc). If you want to look at people who were not employed in the last wave then follow the route via the variable w_notempchk.

For BHPS, it will be best to take a look at this paper by David C Maré "Constructing Consistent Work-life Histories: A guide for users of the British Household Panel Survey"

Best wishes,


Updated by Alita Nandi over 4 years ago

Specifically for between wave job changes look at the bw_johist files and the variable bw_jhstat variable.


Updated by Alita Nandi over 4 years ago

Note #1215 is a similar query. It may be useful to sign up for our JISC mail for data users to discuss analysis issues related to our study. To do that email


Updated by Alita Nandi over 4 years ago


Updated by Daniel Tomlinson over 4 years ago

Thanks for this helpful reply Alita. I've got a couple of follow-up questions, if that's ok.

Firstly, I'm finding understanding the variables guide pages quite confusing, for example I wanted to understand more about the 'nextstat' variable and so looked it up here The text of the question on this page is:

"Immediately following that NxtSt = 2period of NxtStElse]NxtSt = 2 and 2nd or subsequent loopperiod of NextStat(i-1)NxtSt = 1job, did you have a period of paid employment or did you do something else?"

Which is quite tough to decipher, is there a more user friendly version of the routing for questions than this? Or somewhere that explains thing like 'loopperiod' and '2period'?

Secondly, I'm also interested in tracking those people who:
a) had a job in the previous wave
b) left the job and had a spell out of work (with empchk==2)
c) but who then are now back in work for the current wave

Do you know how I might do that in Understanding Society?




Updated by Alita Nandi over 4 years ago

Hi Dan,

The best way to understand question routing is to look at the questionnaire.
Specifically, look at the "universe" below each question which tells you who was asked this question. Once you write out the routing tree it will become clear how to answer your other questions. A colleague had created this tree. I am trying to locate that and get to you.

But I think you have almost answered your second question ("Secondly, I'm also interested...."). If you look at the questionnaire you will see that empchk is asked of those who were (i) interviewed at prior wave or has been interviewed previously and (ii) were In paid employment at last interview or did no paid work at lastinterview but claims employment or self-employment. Additionally jbstat=2 means currently in paid employment, OR jbhas=1|jboff=1 means was in paid employment in the last 2 weeks prior to interview (in some cases there is a discrepancy between the two approaches as the former is how a person identifies themselves and the latter is about what they do). So, empchk=2 means employed at last wave but was not continuously employed since then, and jbstat=2 means currently employed - that is the group you want to identify.

Best wishes,


Updated by Alita Nandi over 4 years ago

Hi Dan,

I am attaching the employment annual events history questionnaire routing (two versions) that my colleagues have prepared. Unfortunately these don't include the question numbers, except for CJOB. We will be adding the question numbers but as you need this soon, we are sending these to you as is. If you find any problems with these or have further questions please let us know.

Best wishes,


Updated by Alita Nandi over 3 years ago

  • Status changed from In Progress to Feedback
  • % Done changed from 50 to 90

Updated by Understanding Society User Support Team over 1 year ago

  • Category set to Data management
  • Status changed from Feedback to Resolved
  • Assignee deleted (Daniel Tomlinson)
  • % Done changed from 90 to 100

Also available in: Atom PDF