Project

General

Profile

Support #353

fed-forward (_ff_): what does this mean (doesn't always match previous wave)

Added by Phil Jones almost 9 years ago. Updated almost 9 years ago.

Status:
Closed
Priority:
Normal
Assignee:
Redmine Admin
Category:
Data inconsistency
Start date:
03/23/2015
% Done:

100%


Description

In wave 4 the fed-forward employment status (d_ff_jbstat) does not match the employment status provided in the previous wave (c_jbstat) in all cases.

I opened both c_indresp.tab and d_indresp.tab. I removed all variables except pidp (wave 4), pidp (wave 3), d_ff_jbstat, and c_jbstat. I merged these (on pidp) and tested to see if d_ff_jbstat and c_jbstat were equivalent.

In 456 cases they are not the same, although I assumed they would be.

R code below:

require("dplyr")
cind <- read.table("data/understandingSociety/UKDA-6614-tab/tab/c_indresp.tab",
                   header = T,
                   stringsAsFactors = F)
dind <- read.table("data/understandingSociety/UKDA-6614-tab/tab/d_indresp.tab",
                   header = T,
                   stringsAsFactors = F)
cindJbstat <- select(cind, pidp, c_jbstat)
dindJbstat <- select(dind, pidp, d_ff_jbstat)
dindJbstat <- inner_join(dindJbstat, cindJbstat)
# # joins on pidp
table(dindJbstat$d_ff_jbstat == dindJbstat$c_jbstat)
# FALSE  TRUE
# 456    39002

Have I misunderstood exactly what 'fed-forward' means, is this an error in the dataset, or is it the case that c_jbstat may have been incorrectly recorded in wave 3 and subsequently corrected in wave 4 by the interviewer, or something else entirely?

I've looked through the documentation to see if I can find an answer but haven't been able to, nor have I been able to find a similar open or closed issue.

Thank you.

#1

Updated by Redmine Admin almost 9 years ago

  • Status changed from New to In Progress
  • Target version set to X M
  • % Done changed from 0 to 50

The fed-forward data follow the current wave individual response data and are not available for proxy respondents (IVFIO). Smaller discrepancies have been explained as corrections taking place between the time of sample issue and the data return.
Jakob

#2

Updated by Redmine Admin almost 9 years ago

  • Status changed from In Progress to Closed
  • % Done changed from 50 to 100

Also available in: Atom PDF