DC-AAPOR/WSS SUMMER CONFERENCE 2019 INCREASING REPRESENTATIVENESS - - PowerPoint PPT Presentation

dc aapor wss summer
SMART_READER_LITE
LIVE PREVIEW

DC-AAPOR/WSS SUMMER CONFERENCE 2019 INCREASING REPRESENTATIVENESS - - PowerPoint PPT Presentation

DC-AAPOR/WSS SUMMER CONFERENCE 2019 INCREASING REPRESENTATIVENESS THROUGH THE USE OF PREDICTIVE MODELING AND TARGETED OUTREACH A.L. DJANGALI, J. JOSEPH-DAVID, & L. TROFIMOVICH DC-AAPOR/WSS SUMMER CONFERENCE 2019 STUDY BACKGROUND


slide-1
SLIDE 1

INCREASING REPRESENTATIVENESS THROUGH THE USE OF PREDICTIVE MODELING AND TARGETED OUTREACH DC-AAPOR/WSS SUMMER CONFERENCE 2019

A.L. DJANGALI, J. JOSEPH-DAVID, & L. TROFIMOVICH

slide-2
SLIDE 2

DC-AAPOR/WSS SUMMER CONFERENCE 2019

STUDY BACKGROUND

▸ Kidney Disease Quality of Life Survey (KDQOL) ▸ Patient-reported outcomes on quality of life measures for Medicare

beneficiaries with end-stage renal disease (ESRD)

▸ Survey conducted yearly as part of a larger evaluation of an alternative

payment model

slide-3
SLIDE 3

DC-AAPOR/WSS SUMMER CONFERENCE 2019

SURVEY METHOD

▸ Conducted from 2016 to 2018 ▸ Two overlapping censuses in 2016 (≈13,511 each) ▸ Sample in 2017 (n=17,198) ▸ Sample in 2018 (n=18,122) ▸ Mail-CATI sequential mode design ▸ Conducted in English and Spanish

Pre-note letter

  • Toll-free RHL

1st Survey

  • Separate

English & Spanish materials 2nd Survey

  • Outbound

CATI with undeliverable mail cases Outbound CATI

  • Transfer to

bilingual staff as needed

slide-4
SLIDE 4

DC-AAPOR/WSS SUMMER CONFERENCE 2019

2016 RESPONSE ANALYSIS

After 2016 Census:

▸ Logistical regression model to identify which socio-demographic characteristics

would predict response

▸ African-Americans, Hispanics, 18-64 year olds: ▸ Less likely to respond overall ▸ More likely to complete by phone than mail ▸ More likely to complete over phone after making contact ▸ African-Americans and Hispanics disproportionally impacted by ESRD

Hubbard, F.A., Weiss, C., & Djangali, A.L. (2017). Modeling Contact and Cooperation in an Administrative Record Based Mail/CATI Survey. Paper presented at the 2017 American Association for Public Opinion Research (AAPOR) conference, New Orleans, Louisiana.

slide-5
SLIDE 5

DC-AAPOR/WSS SUMMER CONFERENCE 2019

CHANGES IMPLEMENTED FOR 2017 & 2018

▸ Started CATI follow-up 2 weeks earlier for hard-to-reach (HTR) groups ▸ Used Census tract data to identify sample members more likely to be

Hispanic

▸ Identified previous CATI completers

Djangali, A.L., Joseph-David, J., & Hubbard, F.A. (2018). Methods for increasing response rates for Medicare beneficiaries. Poster presented at the 2018 American Association for Public Opinion Research (AAPOR) conference, Denver, Colorado. DOI: 10.13140/RG.2.2.31912.67845

slide-6
SLIDE 6

DC-AAPOR/WSS SUMMER CONFERENCE 2019

CHANGES IMPLEMENTED FOR 2017 & 2018

▸ 2016 ▸ 2017

& 2018

Pre-note letter

  • Toll-free RHL

1st Survey

  • Separate

English & Spanish materials 2nd Survey

  • Outbound

CATI with undeliverable mail cases Outbound CATI

  • Transfer to

bilingual staff as needed

Pre-note letter

  • Toll-free RHL
  • Bilingual tickler

message

  • Outbound calls

to previous CATI completers 1st Survey

  • Bilingual tickler

message

  • Double-stuffed

English & Spanish materials 2nd Survey

  • Outbound

CATI with undeliverable mail cases

  • Bilingual tickler

message

  • Double-stuffed

materials

  • Early
  • utbound

calling for HTR Outbound CATI

  • Transfer to

bilingual staff as needed

  • Proactively

assign cases to bilingual staff

slide-7
SLIDE 7

DC-AAPOR/WSS SUMMER CONFERENCE 2019

RESPONSE RATE COMPARISONS†

38% 28% 34% 28% 43% 34% 40% 36% 42% 31% 38% 41%

0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5

Overall 18-54 Black/Afr. Amer. Hispanic

2016 2017 2018

† AAPOR RR3

slide-8
SLIDE 8

DC-AAPOR/WSS SUMMER CONFERENCE 2019

SIGNIFICANCE TESTING

▸ 2-sample t-tests of mean scores to control for sample size differences ▸ Significance (p<.05) across all subgroups

Demographics 2016 2017 2018 n Mean SE n Mean SE n Mean SE 18-54 3,588 28.3% 0.75% 4,627 33.9% 0.70% 4,601 30.9% 0.68% Black/Afr. Amer. 6,042 34.3% 0.61% 7,487 39.5% 0.57% 7,380 37.6% 0.56% Hispanic 909 27.9% 1.49% 1,019 35.6% 1.50% 1,079 40.6% 1.49% 18-54 Black/Afr. Amer. Hispanic 2016 vs 2017

* *** ***

2017 vs 2018

*** * ***

2016 vs 2018

* ** ***

* p<.05 ** p<.01 *** p<.001

slide-9
SLIDE 9

DC-AAPOR/WSS SUMMER CONFERENCE 2019

FINDINGS & RECOMMENDATIONS

▸ Providing preferred language can help to increase representation among

Hispanic respondents

▸ Increasing outreach attempts to traditional HTR respondents helps to

increase response rates for those groups

▸ Small, data-driven changes can make a big impact

slide-10
SLIDE 10

DC-AAPOR/WSS SUMMER CONFERENCE 2019

QUESTIONS?

Please feel free to contact me with questions or to receive a copy of this presentation: Amy L. Djangali Senior Research Director, IMPAQ adjangali@impaqint.com