Project

General

Profile

Support #1504

Weights

Added by Cameron Harries 2 months ago. Updated about 2 months ago.

Status:
Feedback
Priority:
Normal
Start date:
02/08/2021
% Done:

90%


Description

Dear Team,

I am having a little trouble understanding how to implement the weights in my analysis. I have read the documentation and some of the questions posted on here, but I wanted to make 100% certain.

The first question I had was with regards to using the longitudinal weights and this comment from the documentation "For example if you are looking at waves 4 to 9, use the appropriate longitudinal weight from the last wave in your analysis (note wave 9 variables begin with i_)". In the data, I remove the wave tag (i.e., "i_") and append. So when selecting a weight for the longitudinal analysis I would be using the weight found in each relevant wave. Based on the quoted comment it would appear I should merge the weights from the last wave to each observation in previous waves?

The second question: given the above, if I am trying to run a regression at the individual level with a sample of self-employed workers in a specific SOC, that uses shares calculated at the ttwa year level, shall I collapse with the weight (i.e., indpxus_lw ?) and then include the same weight in the regression?

Thank you for your help.

Cameron

#1

Updated by Understanding Society User Support Team 2 months ago

  • Status changed from New to In Progress
  • Assignee set to Olena Kaminska
  • % Done changed from 0 to 10
  • Private changed from Yes to No
#2

Updated by Understanding Society User Support Team 2 months ago

Many thanks for your enquiry. The Understanding Society team is looking into it and we will get back to you as soon as we can.

We aim to respond to simple queries within 48 hours and more complex issues within 7 working days. While we will aim to keep to this response times due to the current coronavirus (COVID-19) related situation it may take us longer to respond.

Best wishes,
Understanding Society User Support Team

#3

Updated by Olena Kaminska 2 months ago

Cameron,

It sounds like you are using pooled analysis. I suggest you search for other similar questions on this forum as you may find previous discussions with other users useful.
If you want to come back to us please let me know which information you include in your analysis - I am interested in which waves and instruments the information comes from. So, are you interested in previous experience and its influence on current situation or are you studying data cross-sectionally, but pooling it over time?

Thank you,
Olena

#4

Updated by Cameron Harries 2 months ago

Dear Olena,

Thank you for your reply.

Yes, my mistake on the cross-sectional weight in the second question. My first question still stands in trying to understand the comment (if I was to use longitudinal weights for an analysis).

However, for my second question, I do believe I should be using cross-sectional weights. I am running a regression using all waves with at the individual level using the restriction of a SOC code. Thus, the individual may be present in some years and not others if they change SOC. I am using some individual controls such as age, education, gender, as well as some ttwa level controls, such as the share of workers that are female in a given ttwa year, and year and ttwa fixed-effects. Thus, when I calculate the shares I may not be using the same sample as the in the regression.

Thank you!
Cameron

#5

Updated by Olena Kaminska 2 months ago

Cameron,

Yes, from what you have said it sounds like you would need cross-sectional weights. As long as you don't use any previous information from other waves in your regression you should use a relevant cross-sectional weight for each wave (so weight b_ for wave 2, weight c_ for wave 3 etc.) Please remember to take into account clustering within PSU - this is especially important in pooled analysis.

I am not sure what you mean by 'shall I collapse with the weight (i.e., indpxus_lw ?) ' in your first question. Are you planning to include a weight in your analysis as a covariate?

Thanks,
Olena

#6

Updated by Understanding Society User Support Team about 2 months ago

  • Status changed from In Progress to Feedback
  • % Done changed from 10 to 90

Also available in: Atom PDF