Support #1907
Weighting advice for information from previous wave on wave 12 respondents
Added by Peter Humphreys about 1 year ago.
Updated 8 months ago.
Description
Hi there
So I'm using the indresp dataset and looking at adult respondents. My population is respondents who are classed as economically inactive in wave 12 of the UKHLS (I recoded some variables to create a definition of economic inactivity) and I'm looking to calculate some descriptive statistics on variables for when these respondents were last in work (sector of last job worked, hours worked per week, average wage, etc.)
To do this analysis, would the correct weight be the indinus_lw value from wave 12?
- Status changed from New to Feedback
- Assignee changed from Understanding Society User Support Team to Olena Kaminska
- % Done changed from 0 to 80
- Private changed from Yes to No
Hello,
if you are using only variables from Wave 12 indresp file (l_indresp) to compute the statistics then you can use the cross-sectional weight, l_indinui_xw
But if you are using information from past waves as well then the weight variable to use is
- l_indinus_lw :if at least some of the past information is from Wave 1
- l_indinub_lw :if at least some of the past information is from Wave 2-6
- l_indinui_lw :if at least some of the past information is from Wave 7-11
Hope this helps. If not, please let us know.
Alita
Understanding Society User Support Team
I'm looking to estimate the probably that a survey respondent is classified as economically inactive in wave 12 based on variables from when the same respondent was previously in work. To note that I'm using R.
For example, a wave 12 respondent that was classed as inactive was previously classed as active in wave 9. So I would want to look at variables from wave 9 (e.g., hours worked per week, overtime hours) as predictors of their wave 12 inactive status. Would a logistic regression be the way to model this?
For the example you have described, the correct weight is l_indinui_xw.
Due to staff time constraints, we only advise on data and survey related issues. But just to confirm, as the outcome variable (whether the person was inactive or not) is a binary variable, a logistic regression is suitable.
One more thing, if you are using at least one self-completion variable in your analysis model then the correct weight is the self-completion weights: l_indscui_lw
You will need a longitudinal weight for your analysis if you are using information from more than 1 wave.
Olena
- Status changed from Feedback to Resolved
- % Done changed from 80 to 100
Also available in: Atom
PDF
Updated by 
Updated by 
Updated by