1 / 21

REPRESENTATIVE RESPONSE Some Examples

REPRESENTATIVE RESPONSE Some Examples. Peter Lundquist Statistics Sweden. Household Finances 2006 (HF). Design: Stratified network sample (18 years and older from RTP all household members are included) Sample size: 17013 households Mode: telephone (approx. 1/2 hour interview)

quang
Télécharger la présentation

REPRESENTATIVE RESPONSE Some Examples

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. REPRESENTATIVE RESPONSE Some Examples Peter Lundquist Statistics Sweden ITSEW2009

  2. Household Finances 2006 (HF) Design: Stratified network sample (18 years and older from RTP all household members are included) Sample size: 17013 households Mode: telephone (approx. 1/2 hour interview) Response rate: 69.6% Auxiliary variables are matched from external registers

  3. HF: Estimated Mean Income Divided by Consumption Units From Westling (2008)

  4. HF: Comparison Between HT and GREG-estimators From Westling (2008)

  5. Living Condition Survey 2007 (LCS) Design:srs, 16 years and older from RTP Sample size: 7694 Mode: telephone (approx. 1 hour interview) Response rate: 73.4% Auxiliariy variables are matched from external registers Missing telephone-number for 3.7% of the sample

  6. LCS: Response Rate

  7. LCS: Auxiliary Variables • Age • Gender • Country of Birth • Marital Status • Employment status • Region • Social allowance • Type of housing estate • Income • Education • Telephone (Sample based)

  8. LCS: Major Reasons for Nonresponse • Age [Refusal: 35-64 years, Noncont: -34 years, Other:74+ years] • Gender • Country of Birth [Noncont and Other: born outside Sweden] • Marital Status [Noncont: unmarried ] • Employment status [Other: unemployed] • Region [Noncont: living in big cities] • Social allowance [Noncont and Other: if having allowance] • Type of housing estate [Noncont: rented housing] • Income [Noncont: no income Other: low income] • Education [Nocont and Other: education code is missing] • Telephone (Sample based) [Noncont: no phone]

  9. LCS: LogitModel Response as Dependent *) sign. level 10% **) sign. level 5% ***) sign. level 1%

  10. LCS: Indicators Based on estimated response probabilities under srs the followingR-indicator (Schouten and Cobben 2007, 2008) is used and under srsthe q2-indicator (Särndal and Lundström 2008) is

  11. LCS: Indicators A measure of the estimated relative bias (in per cent) is computed by As variables ykSickness and activity allowance (yes/no), Income, Sickness pay (yes/no) and Employed (yes/no) are used to estimate the relative bias. All indicators computed for Resp5w and PCT100

  12. LCS: Rand q2 Indicators for Response Before and After Follow-up

  13. LCS: Conclusions • The representativity doesn’t increase with the follow-up • The indicators are estimates e.g. they are subject to variation (not computed) • The same goes for the relative bias • Active work with strategies for the group having social allowance and those with missing telephone • Found auxiliary variables could be used in the creation of a new estimator

  14. Labour Force Survey (LFS) Data from March-December 2007 • Annual salary 2006 according to the Swedish Tax Register • Process data from WinDati (CATI-system) Supplemented with: .

  15. Response Rates for the Reference Weeks in a LFS Month (Means Based on LFS March-December 2007) .

  16. Contact Days for the Reference Weeks in a LFS Month (Means Based on LFS March-December 2007)

  17. Total Number Contact Days, Reference Weeks, in a LFS Month (Means Based on LFS March-December 2007)

  18. Relative Bias for Income Accumulated on Contact Days, in a LFS Month (Means Based on LFS March-December 2007)

  19. LFS: Final Variability in Relative Bias for Income, Reference Weeks (Means Based on LFS March-December 2007)

  20. HF 2006, LCS 2007 and LFS 2007: Estimated Relative Bias of mean Income in SEK in Percent and Outcome Rates *) Based on 36,864 individuals **) Based on the average of 10 months LFS (21,500 individual each month)

  21. References Cobben, F. and Schouten, B. (2008). An empirical validation of R-indicators. Discussion paper 08006, CBS, Voorburg. Schouten, B. and Cobben, F. (2007). R-indexes for the comparison of different fieldwork strategies and data collection modes. Discussion paper 07002, CBS, Voorburg. Särndal, C.E. and Lundström, S. (2008). Assessing auxiliary vectors for control of nonresponse bias in the calibration estimator. Journal of Official Statistics, 24, 167-191. Westling, S. (2008). “Utvecklingför system avkontaktstrategieriintervjuundersökningar med individerochhushåll” Delrapport II, unpublished report, Örebro, Sweden: Statistics Sweden. [In Swedish]

More Related