Project

General

Profile

Support #1307

Inconsistency between releases for certain variables

Added by Chetan Jethwa 3 months ago. Updated 2 months ago.

Status:
Feedback
Priority:
High
Assignee:
Category:
Data releases
Target version:
-
Start date:
01/24/2020
Due date:
01/31/2020
% Done:

80%

Estimated time:

Description

Good afternoon,

I am recently looking at pension income (Variable name: a_fimnpen_dv) for wave found that income values for certain individuals changes between releases (For example, the wave 1 value for a certain individual in a wave 8 release of the datasets is different to the wave 1 value for the same individual in the wave 9 release of the dataset).

I have looked at ‘6614_waves1_to_8_revisions_nov_2019’ PDF and it does not mention any changes to the variable.
Would you have any idea why this inconsistency exists?

Thank you,

Chetan

History

#1 Updated by Stephanie Auty 2 months ago

  • Private changed from Yes to No
  • % Done changed from 0 to 80
  • Assignee changed from Stephanie Auty to Chetan Jethwa
  • Status changed from New to Feedback

Dear Chetan,

This is a derived variable, as indicated by the _dv suffix, and is based on some variables which may be imputed if the data is missing in the survey responses. We update our estimates of imputed variables each wave, as more information becomes available. This means that the values could change, and the data from the most up to date release should be the most reliable. This is one of the reasons we suggest updating data to the latest release, even if you are only using data from earlier waves.

Best wishes,
Stephanie

#2 Updated by Chetan Jethwa 2 months ago

Stephanie Auty wrote:

Dear Chetan,

This is a derived variable, as indicated by the _dv suffix, and is based on some variables which may be imputed if the data is missing in the survey responses. We update our estimates of imputed variables each wave, as more information becomes available. This means that the values could change, and the data from the most up to date release should be the most reliable. This is one of the reasons we suggest updating data to the latest release, even if you are only using data from earlier waves.

Best wishes,
Stephanie

Thank you!

Also available in: Atom PDF