The DHS Program User Forum
Discussions regarding The DHS Program data and results
Home » Data » Dataset use (other programs) » Potential bias of DHS data set (Anticipatory bias)
Potential bias of DHS data set [message #16747] Tue, 26 February 2019 20:10 Go to next message
rkchettri is currently offline  rkchettri
Messages: 18
Registered: November 2017
Dear experts in DHS data,
We analysed and drafted paper using two recent DHS datasets. We faced direct rejection our paper from the editor without sending for peer review process on the basis of anticipatory analysis(the analysis is strongly biased by anticipatory analysis (see Hoem, 2006), procedures that seek to explain current behaviour by future outcomes (anticipatory analysis) constitute "a widespread but problematic approach because they disturb the temporal order of events"). Is this important bias to consider while analysing DHS data, how can we handle this?

Kind regards, Rk
Re: Potential bias of DHS data set [message #16838 is a reply to message #16747] Thu, 07 March 2019 13:21 Go to previous message
Bridgette-DHS is currently offline  Bridgette-DHS
Messages: 1619
Registered: February 2013
Senior Member

Following is a response from Senior DHS Stata Specialist, Tom Pullum:

It is difficult (some people would say impossible) to use DHS data for causal inferences. You may have gone beyond the limitations of the data. I knew Jan Hoem and had/have the greatest respect for him, but am not familiar with that 2006 publication. I doubt that he was arguing against the use of DHS-type data, but rather was arguing against over-interpretation. With a cross-sectional survey it is hard to assess the role of intentions, and there are several instances of temporal sequences being out of order. I would suggest that you fall back to a more descriptive analysis, talking more about association than about causation, and submit to another journal.... An appeal to the same journal, after rejection before review, is less likely to succeed.....

Previous Topic: Opening dataset in SAS studio
Goto Forum:

Current Time: Mon May 20 21:28:31 Eastern Daylight Time 2019