Project

General

Profile

Actions
Copy link

Support #2221

open

Longitudinal Weights for Ethnicity Representation across the 13 waves of the UKHLS

Added by Sebastian Ascui 17 days ago. Updated 13 days ago.

Status:
Feedback
Priority:
High
Assignee:
Understanding Society User Support Team
Category:
Weights
Start date:
03/09/2025
% Done:

50%

Description

Dear Staff member,

I am reaching out to seek advice on the correct application of weights within the Understanding Society: The UK Household Longitudinal Study (UKHLS) dataset, particularly in relation to representing ethnic minorities. My research is using the first thirteen waves of the UKHLS dataset for a longitudinal life course analysis. But I am concerned that some of my regression outputs could be biased, potentially due to the underrepresentation of certain groups.

Despite reviewing the available documentation, I find myself in need of a clearer understanding of how to appropriately use the dataset's weights to mitigate this issue. Given the complexity of longitudinal analysis and the importance of accurately representing ethnic minorities, a more deep advice is required. I would also like to stress that I live in Hong Kong and assisting to any course or training regarding the construction of usage of available weights in the dataset in rather impossible at the moment.

Warm regards,

Sebastian

Actions
Copy link
 #1

Updated by Understanding Society User Support Team 13 days ago

  • Category set to Weights
  • Status changed from New to Feedback
  • % Done changed from 0 to 50
  • Private changed from Yes to No

Hello Sebastian

The best way to determine which weight to use is by following the advice in the “Selecting the Correct Weight” section of the 6614 Main Survey Guide - Weighting Guidance (https://www.understandingsociety.ac.uk/wp-content/uploads/documentation/user-guides/6614_main_survey_user_guide_weighting_guidance.pdf) which is also available at https://www.understandingsociety.ac.uk/documentation/mainstage/user-guides/main-survey-user-guide/selecting-the-correct-weight-for-your-analysis/

The key is to identify:
  • Who you are analysing: individuals aged 16+, all respondents, youth, etc.
  • The type of variables you are using: whether they come from a questionnaire covering the entire sample, only respondents, only proxies, etc.
  • The sample you are working with: GPS and EMBS (Wave 1), BHPS sample (Wave 2), GPS, EMBS, and BHPS (Wave 2), or GPS, EMBS, BHPS, and IEMBS (Wave 6), etc.
  • Whether your analysis is cross-sectional or longitudinal.

Based on what you mentioned, you would need to use a longitudinal weight (_lw) from Wave 13, assuming you are conducting longitudinal analysis. For example, if you are analysing individuals aged 16+ who completed the self-completion questionnaire using the combined GPS, EMB, BHPS & IEMB sample, you would use m_indscui_lw. This weight accounts for factors such as the oversampling of ethnic minorities.

The guide mentioned before also includes examples related to migrants, which may be helpful.

Additionally, you may find the FAQ section useful, as it addresses various issues that might be relevant to your analysis. Here is the link https://www.understandingsociety.ac.uk/wp-content/uploads/documentation/user-guides/6614_main_survey_user_guide_weighting_faqs.pdf

Lastly, we offer a course, “Introduction to Understanding Society: Self-Paced Moodle”, which includes a dedicated section on weights. It also contains practical exercises in Stata, R, SPSS, and SAS that could be beneficial. You can access it here https://www.understandingsociety.ac.uk/help/training/introduction-to-understanding-society-self-paced-moodle/

I hope this information is helpful.

Best wishes,
Roberto Cavazos
Understanding Society User Support Team

Actions
Copy link

Also available in: Atom PDF