1 / 16

Same vs. Different Interviewer: Does it Make a Difference?

Solving the Problem of Attrition in Longitudinal Surveys: Effects of Interviewer Continuity Peter Lynn, Olena Kaminska University of Essex and Harvey Goldstein University of Bristol. Same vs. Different Interviewer: Does it Make a Difference?. Theoretical evidence:

elin
Download Presentation

Same vs. Different Interviewer: Does it Make a Difference?

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Solving the Problem ofAttrition in Longitudinal Surveys:Effects of Interviewer ContinuityPeter Lynn, Olena KaminskaUniversity of Essex and Harvey Goldstein University of Bristol

  2. Same vs. Different Interviewer: Does it Make a Difference? • Theoretical evidence: • Trust (not a stranger, not a sales person, not a criminal) • Consistent behaviour • Interviewer may tailor the introduction / door-step conversation • Possibly harder to refuse to an acquaintance • Empirical evidence shortcomings: • Interviewer change doesn’t happen randomly – e.g. bad interviewers quit • Comparing same vs. different interviewers isn’t straightforward – e.g. same interviewers are promoted within a year.

  3. Data • Sample: participants of the March Omnibus 2008, who agreed to be revisited and provided a contactable name • Re-contacted in 2009 as part of Omnibus Follow-up 2009 • 1188 respondents reissued (1097 eligible for analysis excluding movers and ineligible) • Questionnaire topic: Perceptions of Safety on Public Transport • Response rate: 77% (refusals – 14%; noncontact – 3.5%)

  4. Experimental 2-way Balanced Design: interviewer change crossed with grade change

  5. Multilevel Analysis: Multiple Membership Model 85% of variance within 2009 interviewers 15% of variance within 2008 interviewers Interviewers 2008 2008 2009 2009 Respondents

  6. Results: Interviewer Continuity Influence on Response Rate Base- line

  7. Results: Interviewer Continuity Influence on Response Rate Base- line

  8. Results: Interviewer Continuity Influence on Response Rate Base- line

  9. Further Development of the Model Main effects, random effects and interactions of interviewer continuity with the following variables: • Respondent Characteristics • Age • Gender • Education • Marital status • Number of adults in the HH • Kids • Tenure • Employment • General health • Disability • Interviewer Observations • Area condition • House condition in comparison to neighbouring houses • House type

  10. Final Model Response Rate = intercept + interviewer continuity + age + employment + tenure + house condition + area condition + education + education*house ownership + interviewer continuity*employment + random effect of age 95% of variance within 2009 interviewers 5% of variance within 2008 interviewers

  11. Predicted Likelihood to Respond as a Function of Interviewer Continuity*Employment Status Employed Unemployed

  12. Predicted Likelihood to Respond as a Function of Interviewer Continuity and Age (random effect)

  13. Conclusions • Same interviewer performs the same or better than a different interviewer in obtaining an interview in wave 2 of the study • There is no evidence that sending a different interviewer with higher grade improves the chances of obtaining an interview in comparison to keeping the same interviewer • There is no evidence that changing interviewer increases attrition rate in wave 2 if a respondent was initially approached by an interviewer of low grade

  14. Conclusions • Conditional on an interview by the highest grade interviewer in the first wave, interviewers of the highest grade are 2.8 times more likely to obtain an interview in the 2nd wave if they interviewed the same respondent in previous wave • Same interviewer with the highest grade has the lowest attrition. It is a positive point considering that initially a number of hard refusals agreed to participate when approached by an interviewer of highest grade • Response in the second wave is more influenced by the interviewer in the second wave than by the interviewer in the first wave (ratio ranging from 85/15 to 95/05).

  15. Thoughts • Future studies of interviewer continuity should look at interviewer change controlled by interviewer experience • Interaction of interviewer characteristics with respondent characteristics should be explored for potential tailoring • Same or different interviewers have effect on other aspects of survey, for example on consistency of responses or measurement error. This needs to be taken into account when making a decision. • It is of interest to look at specific conditions where interviewer change may be beneficial

  16. Thank youCorresponding author:Olena Kaminska, olena@essex.ac.uk

More Related