Project

General

Profile

Support #1964

Universe for mglife and other home-ownership related questions

Added by Vikram Patil 8 months ago. Updated 5 months ago.

Status:
Resolved
Priority:
Normal
Assignee:
-
Category:
Data documentation
Start date:
09/01/2023
% Done:

100%


Description

Hello,

My question concerns the universe for the following home-ownership related variables viz. mglife, hsval, hscost, hsyr04 etc.

From the documentation on the website, I understand that mglife (for example) is asked to all households if hsownd = 2|hsownd = 3. hsval, is also asked to households if hsownd = 2| hsownd = 3. Yet, in many waves, despite the overlap in universe, variables such as mglife/hsyr04/hscost etc. will be listed as "inapplicable" whereas hsval will not.

My question is: are variables such as mglife, hscost, hsyr04 etc. only asked if the owner-occupier respondent is NOT being interviewed at their original address (i.e., has shifted to a new home)? Is hsval always asked irrespective of whether the owner-occupier household is at the same address as before or has changed addresses?

For example, in Wave 4, 16495 households have hsval>=0. Out of these, 15296 had the same tenure in the previous wave and thus hsowndchk 1, 428 changed tenure (hsowndchk 2), and 771 are maybe new entrants since ff_hsownd is missing/inapplicable.

Out of the above 16495 households, 8749 have hsownd 2|hsownd 3. Thus, going by the stated universe of mglife, we should have
8749 non-missing responses for mglife AND hsval. Yet, only 595 households have non-missing values for both variables. Out of these 16495 households, 695 are NOT interviewed at their original address. This still leaves about 100 households who are interviewed at their original address but have non-missing values for both variables...so maybe there is something else at play.

Is my understanding of the above correct?

Thanks and best wishes - Vikram

#1

Updated by Understanding Society User Support Team 8 months ago

  • Category set to Data documentation
  • % Done changed from 0 to 70
  • Private changed from Yes to No

Hi Vikram,

The issue arises because the Universe syntax doesn't match the text description of this Universe. For example, foe mglife in wave 4 the questionnaire says:

"If (HsOwnd = 2| HsOwnd = 3)" but the text says "mortgage/shared ownership at new address or mortgaged/shared ownership since previous wave" which means that the group of respondents who got that question was much smaller, the more precise condition being if (d_hsownd==2 | d_hsownd==3) & d_origadd==2 | ~(d_ff_hsownd==2 | d_ff_hsownd==3)&(d_hsownd==2 | d_hsownd==3) - it still doesn't fully match with those who answered this question but the discrepancy is now much smaller (only 24 people). For other questions, you would need to similarly reconstruct the universe using the word descriptions and ignoring the original syntax part. Please let us know in case any further questions arise after you've done that. See also this issue which is about a similar problem: https://iserredex.essex.ac.uk/support/issues/1905 Separately, I'll check if we can update the questionnaires to include the correct syntax condition.

Best wishes,
Piotr Marzec
UKHLS User Support

#2

Updated by Understanding Society User Support Team 8 months ago

  • Status changed from New to Feedback
#3

Updated by Understanding Society User Support Team 5 months ago

  • Status changed from Feedback to Resolved
  • % Done changed from 70 to 100

Also available in: Atom PDF