Support #630

weight with fixed-effects

Added by Carolina Zuccotti over 6 years ago. Updated about 6 years ago.

Victoria Nolan
Start date:
% Done:




I want to study the relationship between subjective well-being and labour market situation.
I want to use fixed effects for this purpose.

I have set-up the data and I am working with the 5 waves. If I want to use the longitudinal weights proposed in the User Manual - I should use e_indscus_lw given that I am working with the adult self-response - then that means I will be working with a balanced panel right? I.e. people who participated in the 5 waves. As stated in the User Manual, I have assigned the weight in e_ to individuals in the first four waves.

What I wanted to know, given that xtset/xtreg do not work with svy, is whether this syntax is correct (although I am missing the complex design):

xtreg sclfsato i.status2 [pweight=my_lw], fe

Where my_lw is the e_ weight assigned to individuals who have cases in all waves (and 0 is assigned to individuals who have participated in less waves) - as stated in the User Manual.

Another question: If I wanted to include in the analysis individuals who have participated in less waves (at least 2), weighting gets more complicated right? I was just wondering if you thought of giving a course on advanced weighting in the UKHLS (just an idea!).

Many thanks in advance,



Updated by Victoria Nolan over 6 years ago

  • Status changed from New to In Progress
  • Assignee changed from Olena Kaminska to Victoria Nolan
  • % Done changed from 0 to 10
  • Private changed from Yes to No

Dear Carolina,

Many thanks for your enquiry and apologies for the delay in getting back to you. I have passed your query onto our weighting team and we will get back to you as soon as we can.

Best wishes, Victoria

On behalf of the Understanding Society User Support Team


Updated by Peter Lynn over 6 years ago


Firstly, apologies for the slow reply. One half of the weighting team had just started maternity leave when you posted your question; the other half has just returned from 2 weeks of paternity leave.

Yes, e_indscus_lw applies to the balanced panel (and only has non-zero values for the balanced panel).

Your specification of my_lw looks correct to me.

The idea of some training or guidance on advanced issues in weighting is certainly on our wish list, but we do not currently have the resources to implement this, unfortunately.



Updated by Carolina Zuccotti over 6 years ago

Many thanks Peter.
In the meantime, I got in contact with another researcher who had posted a similar question, and we all seem to agree :)
And count me in if you ever manage to get the funds for the weighting course!
Best wishes,


Updated by Victoria Nolan over 6 years ago

  • Status changed from In Progress to Resolved
  • % Done changed from 10 to 100

Updated by Victoria Nolan about 6 years ago

  • Status changed from Resolved to Closed

Also available in: Atom PDF