Nonresponse Bias in a Nationwide Dual-Mode Survey

Slides:



Advertisements
Similar presentations
1 Avalaura L. Gaither and Eric C. Newburger Population Division U.S. Census Bureau Washington, D.C. June 2000 Population Division Working Paper No. 44.
Advertisements

1 Sampling Telephone Numbers and Adults, and Interview Length, and Weighting in the California Health Survey Cell Phone Pilot Study J. Michael Brick, Westat.
Quantitative Methods in the Social Sciences (QMSS) Lugano August 2005 Increasing response rates Ineke Stoop SCP.
Why Do People Vote, and Does it Matter if They Don’t? Michael Alvarez PS 120.
Confronting the Challenges of Household Surveys by Mixing Modes Roger Tourangeau, Westat Note: This presentation accompanied the Keynote Address by Dr.
Fieldwork efforts  Monitoring fieldwork efforts  Monitoring fieldwork efforts: Did interviewers /survey organisations implement fieldwork guidelines.
Lesli Scott Ashley Bowers Sue Ellen Hansen Robin Tepper Jacob Survey Research Center, University of Michigan Third International Conference on Establishment.
1 Unobserved common causes of measurement and nonresponse error on the 2008 ANES Panel Survey International Total Survey Error Workshop Stowe, VT - June.
Building Wave Response Rates in a Longitudinal Survey: Essential for Nonsampling Error Reduction or Last In - First Out? Steven B. Cohen Fred Rohde and.
AP US Government & Politics Review Part II. II. Political beliefs and behaviors of individuals (10-20%) Beliefs that citizens hold about their government.
Voter ID Laws, Politics, and the Public Paul R. Brewer Director, Center for Political Communication University of Delaware.
Voting in the U.S. AP Government and Politics Akwete McAlister.
Polling Performance and Campaign Coverage in the
A. rule of law B. individualism C. equality of results D
The 2020 Census: A New Design for the 21st Century
Mesfin S. Mulatu, Ph.D., M.P.H. The MayaTech Corporation
Context for the experiment?
A Comparison of Two Nonprobability Samples with Probability Samples
Canadian Census E&I – Lessons Learned from 2006 with Plans for 2011
Erin Green, RD;  Alisha Gaines, PhD; Jamie S Dollahite, PhD
Table 1: NHBS HET3 Participant Characteristics
Voting Turnout and Behavior
PUBLIC OPINION, THE MEDIA, AND INTEREST GROUPS
UNECE Work Session on Gender Statistics Belgrade November, 2017
Demographic and Socio-Economic Profiles that Relate to Political Party Affiliation Examined in Massachusetts and Wyoming for the 2016 Presidential Election.
Colorado and Florida Target Counties for Clinton
American Government and Organization
Trust Translates Into Influence As Multi-Screen TV Is Much More Likely To Influence Voters’ Final Decision Than Any Other Media Survey Fun Fact: 55% of.
Update and Overview of Administrative Records for the 2020 Census
How Hispanics Are Changing the Face of Nevada
Chapter 2: The nonresponse problem
The European Statistical Training Programme (ESTP)
Nonresponse and Measurement Error in Employment Research
The Effect of Interviewer And Personal Visits on Response Consistency
American Government Chapter 6 Notes.
The European Statistical Training Programme (ESTP)
Voting October 31, 2017.
Referendum Baseline Public Opinion Poll June 2010
Chapter 7: Reducing nonresponse
Surveys of Consumers: Mixed Mode Experiments
Public Opinion and Political Socialization
The European Statistical Training Programme (ESTP)
Chapter 14: Mixed-mode datacollection
Chapter 12: Other nonresponse correction techniques
The European Statistical Training Programme (ESTP)
Chapter 1 The Where, Why, and How of Data Collection
Public Opinion and Political Action
How media cover election polls in the U.S.
Chapter 5: Political Participation
Who Wants to be an A Student?
Public Opinion and Political Action
Current conditions.
Chapter 1 The Where, Why, and How of Data Collection
Voting & Participation.
AP US Government & Politics Review Part II
New Techniques and Technologies for Statistics 2017  Estimation of Response Propensities and Indicators of Representative Response Using Population-Level.
Field procedures and non-sampling errors
Did not have a usual source of care Went without care because of cost
The European Statistical Training Programme (ESTP)
Non-sampling Errors in Polls and Surveys
Unit 5 – political participation
The Where, Why, and How of Data Collection
Chapter 2: The nonresponse problem
Chapter 5: The analysis of nonresponse
New Mexico Census outreach message/messenger testing survey
Voting GOVT 2305, Module 11.
Workshop on best practices for EU-SILC revision, −
New Mexico Census outreach message/messenger testing survey
Stratification, calibration and reducing attrition rate in the Dutch EU-SILC Judit Arends.
Chapter 1 The Where, Why, and How of Data Collection
Presentation transcript:

Nonresponse Bias in a Nationwide Dual-Mode Survey Matthew DeBell, Stanford University Natalya Maisel, Stanford University Ted Brader, University of Michigan Vanessa Meldener, Westat ITSEW 2018, Durham NC

American National Election Studies 2016 Time Series Study Survey of adult US citizens Seeks to explain voter turnout & candidate choice in presidential elections Dual mode ABS Face-to-face Mail-to-internet Clustered not clustered n=1180 n=3090 RR1=50% RR1=44%

This Talk What’s non-random about non-response? 1. Accuracy of estimates 2. “Easy to get” vs “hard to get” 3. Respondents vs non-respondents in a NRFU study What if we got more respondents from the Web sample? Conclusions or implications for… accuracy of estimates field effort field strategy & adaptive design

1. Accuracy of estimates Comparisons to benchmarks (Current Population Survey) Not necessarily non-response error, but indicative Age, gender, education, race/ethnicity, marital status, income, household size, home tenure, region, employment status, and nativity

Accuracy of estimates Characteristic Web error Face-to-face error Errors of 4 points or more (unweighted) Characteristic Web error Face-to-face error Voter turnout +18 +19 Edu: high school cr. -11 -7 Own home -6 -10 One-person HH +7 Age 18-29 -4 Income < $25K ns Married White +4

2. Hard to get? “Easy to get” Rs: make contact with few attempts & cooperate readily “Hard to get” Rs: require multiple contact attempts or refusal conversion Definitions of hard-to-get Web: if a refusal conversion letter was mailed or a refusal was recorded (39%) FtF: any refusal, >5 contacts, or late contact (41%) We assume non-respondents are more like hard-to-get Rs Differences between easy- and hard-to-get Rs would indicate field effort matters

How “hard to get” Rs differ from easier Rs Mail-to-web Younger Less: education, white, married More: Hispanic, renters Lower income More Southern, less NE Less likely to vote More Trump voters Face-to-face Younger No difference Higher income Less Southern, more NE No difference (ns) No difference (ns, opposite)

3. Non-Response Follow-Up (NRFU) Study Mail survey Both responding and non-responding households n=4,725 Study name & sponsorship differed from ANES One page paper questionnaire with 18 questions RR = 39%

NRFU Results: no differences Ethnicity (Hispanic) Talk to neighbors Like college professors Like news reporters Children under 18 Party ID Presidential vote choice

NRFU Results: differences Web Face-to-face Voter turnout (10 points) Like surveys (.11) Internet access (15 points) Worry personal privacy (.08) Interpersonal trust (.04) Free time (.02) Education Age (2.6 years) Voter turnout (10 points) Like surveys (.15) Like talking politics (.05)

4. What if we got NRFU Rs in the first place? Estimated effect on Web sample of adding 400 Rs by mail like NRFU Rs

4. What if we got NRFU Rs in the first place? Estimated effect on Web sample of adding 400 Rs by mail like NRFU Rs No effect (<1 percentage point) on observed demographics (age, sex, education, Hispanic ethnicity) No effect (<1 percentage point) on party ID

4. What if we got NRFU Rs in the first place? Estimated effect on Web sample of adding 400 Rs by mail like NRFU Rs No effect (<1 percentage point) on observed demographics (age, sex, education, Hispanic ethnicity) No effect (<1 percentage point) on party ID Trump vote error narrows from -2.1 to -1.7 Clinton vote error narrows from 1.2 to 1.0

4. What if we got NRFU Rs in the first place? Estimated effect on Web sample of adding 400 Rs by mail like NRFU Rs No effect (<1 percentage point) on observed demographics (age, sex, education, Hispanic ethnicity) No effect (<1 percentage point) on party ID Trump vote error narrows from -2.1 to -1.7 Clinton vote error narrows from 1.2 to 1.0 Turnout error narrows from 15.1 to 12.0

Conclusions: accuracy of estimates Both modes miss Non-voters* People who dislike surveys* Low education Home owners Younger respondents (18-29) Face-to-face misses Mid-higher income Married

Conclusions: field effort Full field effort Matters for web turnout & candidate choice Appears to matter less for FtF Sequential mail mode for Web sample would likely improve voter turnout accuracy

Conclusions: field strategy Finding Fieldwork implication NR bias for turnout and liking surveys De-emphasize voting and “surveys” in communication strategy

Conclusions: field strategy Finding Fieldwork implication NR bias for turnout and liking surveys De-emphasize voting and “surveys” in communication strategy Attitudes toward professors and news reporters not a source of NR bias References to these groups may be acceptable in our communications

Conclusions: field strategy Finding Fieldwork implication NR bias for turnout and liking surveys De-emphasize voting and “surveys” in communication strategy Attitudes toward professors and news reporters not a source of NR bias References to these groups may be acceptable in our communications Home-owners under-represented Target non-rental properties for greater recruitment effort (more contact attempts and rapidly escalated incentives)

Conclusions: field strategy Finding Fieldwork implication NR bias for turnout and liking surveys De-emphasize voting and “surveys” in communication strategy Attitudes toward professors and news reporters not a source of NR bias References to these groups may be acceptable in our communications Home-owners under-represented Target non-rental properties for greater recruitment effort (more contact attempts and rapidly escalated incentives) Younger and lower education (and married, in FTF) are under-represented Once HHs are screened, target these HHs for greater recruitment effort

in a Nationwide Dual-Mode Survey Thank you Nonresponse Bias in a Nationwide Dual-Mode Survey Matthew DeBell, Stanford University Natalya Maisel, Stanford University Ted Brader, University of Michigan Vanessa Meldener, Westat Contact: debell@stanford.edu