Support #618

W5 mainstage: NS-SEC for all respondents?

Added by Dan Philo over 7 years ago. Updated over 7 years ago.

Derived variables
Start date:
% Done:




I am after a W5 derived variable giving ns-sec for all respondents, including retired people and other 'economically inactive' categories at e_jbstat. This would based on current job, or last job for those not currently in work. The two dvs e_jbnssec_dv and e_jlnssec_dv seem to cover only those in work at w5 or, for those not in work, last job only for new entrants to the survey.

Is the best way to create such a dv by using these two variables from waves 1-5? i.e. starting with e_jbnssec_dv and then populating missings with e_jlnssec, then d_jbnssec_dv & d_jlnssec_dv etc back to w1? (Looking at w1 data, I think this MIGHT still leave some missing data as there are 1190 who answer no to a_jbhas & a_jboff but are missing at a_jbhad, a_jbnssec5_dv, & a_jlnssec5_dv).

Many thanks,


Updated by Gundi Knies over 7 years ago

  • Assignee changed from Gundi Knies to Dan Philo
  • % Done changed from 0 to 90

Hi Dan,
we cannot comment on which is "the best way" to derive your desired indicator seeing as that is something only you can evaluate in light of your specific research project.

Having said that, you can of course use information collected in previous waves - that is one of the beauties of longitudinal research afterall - but the definition of the NSSEC will become fuzzier once you start focussing on anything but _jbnssec. It may be relatively straightforward in the first instance to recode some of those to whom _jbnssec was inapplicable to "long-term unemployed" or "never worked", based on information provided elsewhere in the same wave of the study.

When using _jlnssec or _jbnssec from previous waves you should consider the different definitions. E.g. whilst _jbnssec may include casual work currently done by students if they said _jbhas==1 or _jboff==1, a note to _jbhad mentions not to include holiday work or student work. Also, some of the jobs reported in _jlnssec may have been held a very long time ago, when there was no official SOC 2000 code for the occupation; the respondent may also remember a job which was not in fact the last job but the last job they remember (recall bias). Ultimately, the NSSEC would become something along the lines of "most recent job ever reported in the study", which may or may not be fine for the purpose of your research.



Updated by Dan Philo over 7 years ago

Hello Gundi,

These details are very helpful, thank you. Due to these limitations we've decided against constructing this dv.

Best wishes,


Updated by Victoria Nolan over 7 years ago

  • Status changed from New to Closed
  • % Done changed from 90 to 100
  • Private changed from Yes to No

Also available in: Atom PDF