© GfK Using ancillary information to stratify and target young adults and Hispanics in national ABS samples J. Michael Dennis, Charles DiSogra, Erlina Hendarwan AAPOR 2012 Orlando, Florida
© GfK Probability-based ABS recruitment Recruitment takes place throughout the year Representative of U.S. adults Includes: Adults with no Internet access (24% of adults) KN provides laptop and free ISP Cell phone only (30% of adults) Spanish-language Extensive profile data maintained on each member demographics, attitudes, behaviors, health, media usage, etc. Samples from the panel are assigned to projects invitations and a link to the online survey questionnaire 50,000+ members
© GfK Mail methodology: materials and schedule 3 Current Resident / Residente Actual 123 Your Street The City, State Day 7 Reminder PC Day 28 NR Letter Initial Mailing
© GfK Address-Based Sample (ABS) frame U.S. Postal Service Computerized Delivery Sequence File (CDSF) ~97% coverage of physical addresses Frequently updated including status of addresses, such as, seasonal homes, vacant houses, etc. Can be matched to available landline telephone numbers Can be geo-coded Can attach ancillary information from a variety of sources for purposes of: Non-response analyses Targeting demographic / geographic mailings Sample stratification 4
© GfK Research questions Will the demographic yields using ancillary information in the sample design be as good as or better than using traditional Census block data? Does the use of ancillary information improve the efficiency of recruiting the demographic groups of interest? If using ancillary information turns out to be a good strategy, can it be leveraged to maximize yields? 5
© GfK Sample stratification in 2010 and national mailings (~22,000 ea.) Objective: Increase Hispanics in a national sample Method used: Targeted Census blocks with ABS sample Criteria: CBs with 30% or more Hispanic population Design and sample distribution (vs. Frame): Stratum 1 Hispanic CBs 41% (vs. 14%) Stratum 2 Balance of CBs 58% (vs. 86%) Design Effect: 1.60
© GfK Sample stratification in 2010 and national mailings (~22,000 ea.) Objective: Increase Hispanics in a national sample Method used: Targeted Census blocks with ABS sample Criteria: CBs with 30% or more Hispanic population Design and sample distribution (vs. Frame): Stratum 1 Hispanic CBs 41% (vs. 14%) Stratum 2 Balance of CBs 58% (vs. 86%) Design Effect: national mailings (~26,000 ea.) Objective: Increase Hispanics Increase Young Adults (ages 18-24) Method used: Ancillary information with ABS sample Criteria: Hispanic surname for household Any in household and rents home Design and sample distribution (vs. Frame): Stratum 1 Hispanic % (vs. 0.3%) Stratum 2 Hispanic % (vs. 8%) Stratum 3 All Else % (vs. 2%) Stratum 4 All Else % (vs. 89%) Design Effect: 1.52 Conservative oversamples (approx. 3x)
© GfK Percent Raw Yield 8 Raw yield = number of HHs recruited / total HHs in sample Slightly LOWER yield with ancillary info design Reason: A function of HISPANIC ancillary info targeting LOWER
© GfK Percent Raw Yield of Hispanics 9 Raw yield = number of Hispanic HHs recruited / total HHs in sample NO DIFFERENCE in yield with ancillary info design A HIGHER Hispanic yield with ancillary info targeting HIGHER
© GfK Percent Raw Yield of Young Adults 10 Raw yield = number of YA HHs recruited / total HHs in sample NO DIFFERENCE in yield with ancillary info design A HIGHER AGE GROUP yield with ancillary info targeting HIGHER
© GfK Percent Race/Ethnicity of Recruited 11 HIGHER EFFICIENCY to locate Hispanics with ancillary info targeting * May not add to 100% due to rounding.
© GfK Percent Young Adults (18-24) of Recruited 12 HIGHER EFFICIENCY to locate young adults with ancillary info targeting * May not add to 100% due to rounding.
© GfK Conclusions The yields for Hispanics and young adults, using ancillary information in the sample design, is as good as traditional Census block data with our 2010 design The use of ancillary information improved the efficiency of recruiting Hispanics and young adults (ages 18-24) Improved efficiency was also reflected in a modest 1.7% lower cost per recruited young adult and 3.2% lower cost per recruited Hispanic in 2011 compared to 2010 Using ancillary information, with more aggressive over-sampling to take advantage of the higher efficiency, can be leveraged to maximize yields in future samples 13
© GfK Thank you! J. Michael Dennis