Effects of Sampling and Screening Strategies in an RDD Survey Anthony M. Roman, Elizabeth Eggleston, Charles F. Turner, Susan M. Rogers, Rebecca Crow,

Slides:



Advertisements
Similar presentations
Sampling Fundamentals
Advertisements

Strategies for Increasing Efficiency of Cellular Telephone Samples Kurt Peters 1, William Robb 1, Cristine Delnevo 2, Daniel A. Gundersen 2 March 2014.
Connect Nevada Residential Technology Assessment Results.
Survey Methodology Nonresponse EPID 626 Lecture 6.
Preliminary Results from the 2008 Oklahoma Health Care Insurance and Access Survey Presentation to the Oklahoma Health Care Authority Board November 13,
1 Sampling Telephone Numbers and Adults, and Interview Length, and Weighting in the California Health Survey Cell Phone Pilot Study J. Michael Brick, Westat.
STATISTICS FOR MANAGERS LECTURE 2: SURVEY DESIGN.
RTI International is a trade name of Research Triangle Institute th Street, NW ■ Suite 750 ■ Washington, DC, USA Phone
Kevin Deardorff Assistant Division Chief, Decennial Management Division U.S. Census Bureau 2014 SDC / CIC Conference April 2, Census Updates.
Sample Design (Click icon for audio) Dr. Michael R. Hyman, NMSU.
Topic 7 Sampling And Sampling Distributions. The term Population represents everything we want to study, bearing in mind that the population is ever changing.
Why sample? Diversity in populations Practicality and cost.
Aaker, Kumar, Day Ninth Edition Instructor’s Presentation Slides
Chapter 12 Sample Surveys
The Excel NORMDIST Function Computes the cumulative probability to the value X Business Statistics: A First Course, 5e © 2009 Prentice-Hall, Inc
A new sampling method: stratified sampling
Sampling Methods.
Copyright ©2005 Brooks/Cole, a division of Thomson Learning, Inc. How to Get a Good Sample Chapter 4.
~ Draft version ~ 1 HOW TO CHOOSE THE NUMBER OF CALL ATTEMPTS IN A TELEPHONE SURVEY IN THE PRESENCE OF NONRESPONSE AND MEASUREMENT ERRORS Annica Isaksson.
Methodologic Overview of Two National Data Sets Centers for Disease Control and Prevention National Center for Health Statistics Issues in Comparing Findings.
FINAL REPORT: OUTLINE & OVERVIEW OF SURVEY ERRORS
2014 MASSACHUSETTS HEALTH INSURANCE SURVEY KEY FINDINGS Prepared by: Laura Skopec, Sharon K. Long, and Thomas H. Dimmock, Urban Institute Susan Sherr,
Estimating Phone Service and Usage Percentages: How to Weight the Data from a Local, Dual-Frame Sample Survey of Cellphone and Landline Telephone Users.
Lecture 30 sampling and field work
A Tale of Two Methods: Comparing mail and RDD data collection for the Minnesota Adult Tobacco Survey III Wendy Hicks and David Cantor Westat Ann St. Claire,
RTI International is a trade name of Research Triangle Institute 3040 Cornwallis Road ■ P.O. Box ■ Research Triangle Park, North Carolina, USA
COLLECTING QUANTITATIVE DATA: Sampling and Data collection
Overview of the American Community Survey Sample Design Prepared for the Quarterly Meeting of the Occupational Information Development Advisory Panel Social.
U.S. Hispanic Entertainment and Consumer Electronics Usage From ICR HispanicEXCEL and ICR CENTRIS September, 2005 I N T E R N A T I O N A L C O M M U N.
Not a benefit … a necessity: What Paid Family Leave means for NYC’s low-income families Nancy Rankin, Vice President for Policy Research and Advocacy Apurva.
RTI International is a trade name of Research Triangle Institute Untreated chlamydial infection among adolescents and young adults in Baltimore,
Overview of error model for estimates of foreign-born immigration using data from the American Community Survey Mary H. Mulry U.S. Census Bureau 2011 International.
Research Methodology Lecture No :14 (Sampling Design)
How to Design a Sample and Improve Response Rates Alex StannardScottish Government Kevin RalstonUniversity of Stirling.
Who Needs RDD? Combining Directory Listings with Cell Phone Exchanges for an Alternative Sampling Frame Presented at AAPOR 2008 New Orleans, LA May 16,
Sampling. Sampling Can’t talk to everybody Select some members of population of interest If sample is “representative” can generalize findings.
PPA 501 – A NALYTICAL M ETHODS IN A DMINISTRATION Lecture 3c – Sampling.
Sampling Methods.
© 2013 Cengage Learning. All Rights Reserved. May not be scanned, copied or duplicated, or posted to a publicly accessible website, in whole or in part.
Panel Study of Entrepreneurial Dynamics Richard Curtin University of Michigan.
Statistics for Managers Using Microsoft Excel, 4e © 2004 Prentice-Hall, Inc. Chap 1-1 Statistics for Managers Using Microsoft ® Excel 4 th Edition Chapter.
Sampling Design.
Growing Challenges to State Telephone Surveys of Health Insurance Coverage: Minnesota as a Case Study Supported by a grant from the Minnesota Department.
1. Population and Sampling  Probability Sampling  Non-probability Sampling 2.
7.1Sampling Methods 7.2Introduction to Sampling Distribution 7.0 Sampling and Sampling Distribution.
5-4-1 Unit 4: Sampling approaches After completing this unit you should be able to: Outline the purpose of sampling Understand key theoretical.
SAMPLE DESIGN: WHO WILL BE IN THE SAMPLE ? (CONTINUED) Lu Ann Aday, Ph.D. The University of Texas School of Public Health.
Understanding Sampling
Basic Business Statistics, 10e © 2006 Prentice-Hall, Inc.. Chap 7-1 Chapter 7 Sampling Distributions Basic Business Statistics.
Testing for Coverage Bias when Combining Directory-Listed And Cellphone Samples T. M. Guterbock, A. Diop, J. M. Ellis, J. L. P. Holmes and K. T. Le, Center.
Finding low-income telephone households and people who do not have health insurance using auxiliary sample frame information for a random digit dial survey.
Chapter 6: 1 Sampling. Introduction Sampling - the process of selecting observations Often not possible to collect information from all persons or other.
Impact of T-ACASI on Estimates of Youth Smoking Prevalence: Results of UMASS Tobacco Study Lois Biener, 1 Charles F. Turner, 2 & Amy L. Nyman 1 1 Center.
Basic Business Statistics, 11e © 2009 Prentice-Hall, Inc. Chap 7-1 Chapter 7 Sampling and Sampling Distributions Basic Business Statistics 11 th Edition.
The Use of Random Digit Dialing in Household Surveys: Challenges and Changes Chris Chapman 2008 IES Research Conference Washington, DC June 11, 2008
1 Responsive Design and Survey Management in the National Survey of Family Growth (NSFG) William D. Mosher, NCHS FCSM Statistical Policy Seminar Washington,
 An observational study observes individuals and measures variable of interest but does not attempt to influence the responses.  Often fails due to.
2015 MASSACHUSETTS HEALTH INSURANCE SURVEY KEY FINDINGS Prepared by: Laura Skopec, Sharon K. Long, and Emily Hayes, Urban Institute Susan Sherr, David.
1 of 22 INTRODUCTION TO SURVEY SAMPLING October 6, 2010 Linda Owens Survey Research Laboratory University of Illinois at Chicago
Population vs. Sample. Population: a set which includes all measurements of interest to the researcher (The collection of all responses, measurements,
Chapter 5 Sampling and Surveys. Section 5.3 Sample Surveys in the Real World.
Journalism 614: Non-Response and Cell Phone Adoption Issues.
Environmental and Social Influences on Tobacco Use Among 18 to 24 Year-Olds in Idaho Dr. John Hetherington Clearwater Research, Inc. Influences on Young.
The Practice of Statistics Third Edition Chapter 10.3 (12.1): Estimating a Population Proportion Copyright © 2008 by W. H. Freeman & Company Daniel S.
The State and Local Area Integrated Telephone Survey Marcie Cynamon Chief, Survey Planning and Special Surveys Branch National Center for Health Statistics.
Population vs Sample Population = The full set of cases Sample = A portion of population The need to sample: More practical Budget constraint Time constraint.
MATH Section 6.1. Sampling: Terms: Population – each element (or person) from the set of observations that can be made Sample – a subset of the.
Sexual Behavior and Sexually Transmitted Infections in a Probability Sample of Adolescents in Baltimore, MD E. Eggleston 1, S.M. Rogers 1, C.F. Turner.
Research Methods: Concepts and Connections First Edition
Sampling and estimation
Presentation transcript:

Effects of Sampling and Screening Strategies in an RDD Survey Anthony M. Roman, Elizabeth Eggleston, Charles F. Turner, Susan M. Rogers, Rebecca Crow, Sylvia Tan What: RDD study conducted in Baltimore When: Sept – August 2009 Who: Target population: People aged Why: Measure risk behaviors and prevalence of 3 STI’s (Gonorrhea, Chlamydia and Trichonomiasis) RDD SAMPLE FRAME INEFFICIENT DUE TO AMOUNT OF CALLING REQUIRED TO SCREEN OUT NON- RESIDENTIAL TELEPHONE NUMBERS AND HOUSEHOLDS WITHOUT SOMEONE AGED CONCERN: High costs will require fewer completed interviews and lower power in statistical analyses REWORD ORIGINAL SCREENER QUESTION Original question: “How many people aged currently live in this household?” New screener questions: 1) 1)“How many people aged 36 or older currently live in this household?” 2) 2)“How many people aged currently live in this household?” Table 1: Results of Telephone Number Dialing by Stratum Rate ofRate at which Overall Connecting toHouseholds hadRate ResidentialAge EligibleColumn 1 x Sample Source:Households:Respondent:Column 2: Original RDD30.80%31.10%9.58% List with person List Age Unknown Combined lists and RDD with lists removed ** The dual frame design resulted in a 31% increase in dialing efficiency and a relevant decrease in survey cost. Additional increase in efficiency can be realized with higher reliance on the lists. List assisted RDD sample (GENESYS) Address matching for advance letters Phone interviewers screen for eligibility: Age Live in city of Baltimore Speak English Have touch-tone phone Parental permission when required Random selection of eligible respondent within household TACASI interview $20 for minute interview Additional $40 for providing urine specimen by mail BACKGROUND ORIGINAL DESIGN ATTEMPTED SOLUTION PROBLEM 1: ELIGIBILITY RATE Based on census estimates, lower than expected rate of households with someone aged (21.3% vs. 31.6%) Concern: Bias caused by missing households, cell phone only households, higher costs due to lower eligibility rates RESULT Wording change produced rate of 31.3% of households with someone aged15-35 CLOSELY CORRESPONDS WITH CENSUS ESTIMATE! ATTEMPTED SOLUTION MOVE TO DUAL-FRAME SAMPLE USING COMBINATION OF LISTS & RDD Four strata within sample frame: 1) 1)List households believed to have someone aged ) 2)List households believed NOT to have someone aged ) 3)List households with no age information on occupants 4) 4)RDD sample with all list households removed Sampling from strata at different rates, all households in Baltimore can exist in one and only one stratum, all probabilities of selection known. PROBLEM 2: COST PROBLEM 3: GETTING ELIGIBLE RESPONDENT ON PHONE Age group is known to spend less time at home  talking to eligible respondent requires many call attempts Concerns: 1) Extra call attempts = higher cost; 2) Inability to EVER get some respondents = lower response rates ATTEMPTED SOLUTION ALTER METHOD OF RANDOM SELECTION OF ELIGIBLE RESPONDENT Original method: 1/n for each of n eligible people within household New method: Increased probability of selection for person who answered screener questions if that person is eligible themselves Screener respondent has 2/(n+1) chance of selection All other eligible people in household 1/(n+1) chance Example: 2 eligible people in household and screener respondent one of them  Original method gave this person ½ chance of selection  New method gave screener respondent 2/3 chance and other eligible 1/3 chance NET RESULTS DECREASED EFFORT = INCREASED SAVINGS; INCREASED RESPONSE RATE BEFORE MODIFICATIONSAFTER MODIFICATIONSRESULTS 4.64 interviewer hours per complete interview 4.15 interviewer hours per complete interview  10.6% reduction in interviewer effort  Comparable reduction in data collection costs  Additional savings can be gained with higher reliance on lists in future sample Table 3: Response rate changes due to sampling modifications InterviewAgreed to receiveReturned Response Rate:Specimen Kit:Specimen: Original design55.04%84.20%78.26% After modifications ) 1) The increased response rate among identified eligible respondents from 55% to 59.7% we assume to be due to: Selecting screener respondents more often meant getting more interviews A higher % of households received advance letters due to lists Lists produced slightly higher response rates 2) Increased agreement to receive specimen cup due to selecting more screener respondents as they had an increased rapport with interviewers and agreed more often. 3) The increased rate of returning cups was due to one last modification and that was offering $100 instead of $40 to those who initially agreed to send in a cup and then failed to do so. NEXT STEPS Examine effects of sample design modifications on: Survey weights Estimated standard errors Use results to optimize sampling fractions across strata RESULT RESULTS Original method: averaged call attempts per interview. New Method: averaged 8.88 call attempts per interview. **27.9% reduction in call attempts with relevant cost savings *ACS = American Community Survey Distributions of Respondent Characteristics