1 / 13

Response representativeness, proxy report and measurement error in the LFS

Response representativeness, proxy report and measurement error in the LFS. Barry Schouten & Kasper Leufkens Statistics Netherlands. ITSEW 2010, June 14-16. Goal. Investigate relation between response representativeness and measurement error as a function of design features

Download Presentation

Response representativeness, proxy report and measurement error in the LFS

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. Response representativeness, proxy report and measurement error in the LFS Barry Schouten & Kasper Leufkens Statistics Netherlands ITSEW 2010, June 14-16

  2. Goal • Investigate relation between response representativeness and measurement error as a function of design features • Generalize from survey item specific measurement error to measurement risk profiles

  3. Design features • Number of visits to address • Appointments • Proxy reporting Assumption: Interviewer strategies not affected by restriction to smaller number of visits, no appointments, self-report only

  4. Design features

  5. Definitions Response propensities: Measurement risk profile(s): Profile propensities Response error:

  6. Definitions Survey error quality indicators based on bias of response mean Maximal nonresponse bias Maximal interaction bias Profile rate

  7. Implementation for LFS Two profiles: • Socially desirable answers: employed according to respondent but not in registration, or not employed but subscribed employment office according to respondent but not in registration • Acquiescence: not employed according to respondent but in registration Response and profile propensities estimated using registrations and frame data: age, household composition, urbanization of residence area, job status, ethnicity, house value

  8. Results – Nonresponse bias Diamonds = no proxy, no appointment Traingles = no proxy, appointment Filled circles = proxy, no appointment Solid circles = proxy, appointment

  9. Results – Nonresponse bias Diamonds = no proxy, no appointment Traingles = no proxy, appointment Filled circles = proxy, no appointment Solid circles = proxy, appointment

  10. Results – profile 1 Diamonds = no proxy, no appointment Traingles = no proxy, appointment Filled circles = proxy, no appointment Solid circles = proxy, appointment

  11. Results – profile 2 Diamonds = no proxy, no appointment Traingles = no proxy, appointment Filled circles = proxy, no appointment Solid circles = proxy, appointment

  12. Conclusions • All design features have strong impact on response rate • Maximal nonresponse bias stable with respect to number of attempts and appointments • Maximal nonresponse bias smaller when proxy allowed • Profile 1: interaction bias and profile rate stable, small increase in interaction bias when proxy allowed • Profile 2: interaction bias and profile rate negatively affected by proxy reporting. No impact of other design features

  13. Discussion • Are assumptions about interviewer strategies reasonable? • Is concept of measurement risk profile useful and is it sensibly implemented in LFS? • If so, is it useful to monitor the proposed “survey item independent” indicators? • Are results conforming to literature/experience?

More Related