Cover
Title Page
Notes on Contributors
Preface
Section 1: The Concept of TSE and the TSE Paradigm
1. 1 The Roots and Evolution of the Total Survey Error Concept
  1. 1.1 Introduction and Historical Backdrop
  2. 1.2 Specific Error Sources and Their Control or Evaluation
  3. 1.3 Survey Models and Total Survey Design
  4. 1.4 The Advent of More Systematic Approaches Toward Survey Quality
  5. 1.5 What the Future Will Bring
  6. References
2. 2 Total Twitter Error
  1. 2.1 Introduction
  2. 2.2 Social Media: An Evolving Online Public Sphere
  3. 2.3 Components of Twitter Error
  4. 2.4 Studying Public Opinion on the Twittersphere and the Potential Error Sources of Twitter Data: Two Case Studies
  5. 2.5 Discussion
  6. 2.6 Conclusion
  7. References
3. 3 Big Data
  1. 3.1 Introduction
  2. 3.2 Definitions
  3. 3.3 The Analytic Challenge: From Database Marketing to Big Data and Data Science
  4. 3.4 Assessing Data Quality
  5. 3.5 Applications in Market, Opinion, and Social Research
  6. 3.6 The Ethics of Research Using Big Data
  7. 3.7 The Future of Surveys in a Data‐Rich Environment
  8. References
4. 4 The Role of Statistical Disclosure Limitation in Total Survey Error
  1. 4.1 Introduction
  2. 4.2 Primer on SDL
  3. 4.3 TSE‐Aware SDL
  4. 4.4 Edit‐Respecting SDL
  5. 4.5 SDL‐Aware TSE
  6. 4.6 Full Unification of Edit, Imputation, and SDL
  7. 4.7 “Big Data” Issues
  8. 4.8 Conclusion
  9. Acknowledgments
  10. References
Section 2: Implications for Survey Design
1. 5 The Undercoverage–Nonresponse Tradeoff
  1. 5.1 Introduction
  2. 5.2 Examples of the Tradeoff
  3. 5.3 Simple Demonstration of the Tradeoff
  4. 5.4 Coverage and Response Propensities and Bias
  5. 5.5 Simulation Study of Rates and Bias
  6. 5.6 Costs
  7. 5.7 Lessons for Survey Practice
  8. References
2. 6 Mixing Modes
  1. 6.1 Introduction
  2. 6.2 The Effect of Offering a Choice of Modes
  3. 6.3 Getting People to Respond Online
  4. 6.4 Sequencing Different Modes of Data Collection
  5. 6.5 Separating the Effects of Mode on Selection and Reporting
  6. 6.6 Maximizing Comparability Versus Minimizing Error
  7. 6.7 Conclusions
  8. References
3. 7 Mobile Web Surveys
  1. 7.1 Introduction
  2. 7.2 Coverage
  3. 7.3 Nonresponse
  4. 7.4 Measurement Error
  5. 7.5 Links Between Different Error Sources
  6. 7.6 The Future of Mobile web Surveys
  7. References
4. 8 The Effects of a Mid‐Data Collection Change in Financial Incentives on Total Survey Error in the National Survey of Family Growth
  1. 8.1 Introduction
  2. 8.2 Literature Review: Incentives in Face‐to‐Face Surveys
  3. 8.3 Data and Methods
  4. 8.4 Results
  5. 8.5 Conclusion
  6. References
5. 9 A Total Survey Error Perspective on Surveys in Multinational, Multiregional, and Multicultural Contexts
  1. 9.1 Introduction
  2. 9.2 TSE in Multinational, Multiregional, and Multicultural Surveys
  3. 9.3 Challenges Related to Representation and Measurement Error Components in Comparative Surveys
  4. 9.4 QA and QC in 3MC Surveys
  5. References
6. 10 Smartphone Participation in Web Surveys
  1. 10.1 Introduction
  2. 10.2 Prevalence of Smartphone Participation in Web Surveys
  3. 10.3 Smartphone Participation Choices
  4. 10.4 Instrument Design Choices
  5. 10.5 Device and Design Treatment Choices
  6. 10.6 Conclusion
  7. 10.7 Future Challenges and Research Needs
  8. Appendix 10.A: Data Sources
  9. Appendix 10.B: Smartphone Prevalence in Web Surveys
  10. Appendix 10.C: Screen Captures from Peterson et al. (2013) Experiment
  11. Appendix 10.D: Survey Questions Used in the Analysis of the Peterson et al. (2013) Experiment
  12. References
7. 11 Survey Research and the Quality of Survey Data Among Ethnic Minorities
  1. 11.1 Introduction
  2. 11.2 On the Use of the Terms Ethnicity and Ethnic Minorities
  3. 11.3 On the Representation of Ethnic Minorities in Surveys
  4. 11.4 Measurement Issues
  5. 11.5 Comparability, Timeliness, and Cost Concerns
  6. 11.6 Conclusion
  7. References
Section 3: Data Collection and Data Processing Applications
1. 12 Measurement Error in Survey Operations Management
  1. 12.1 TSE Background on Survey Operations
  2. 12.2 Better and Better: Using Behavior Coding (CARIcode) and Paradata to Evaluate and Improve Question (Specification) Error and Interviewer Error
  3. 12.3 Field‐Centered Design: Mobile App for Rapid Reporting and Management
  4. 12.4 Faster and Cheaper: Detecting Falsification With GIS Tools
  5. 12.5 Putting It All Together: Field Supervisor Dashboards
  6. 12.6 Discussion
  7. References
2. 13 Total Survey Error for Longitudinal Surveys
  1. 13.1 Introduction
  2. 13.2 Distinctive Aspects of Longitudinal Surveys
  3. 13.3 TSE Components in Longitudinal Surveys
  4. 13.4 Design of Longitudinal Surveys from a TSE Perspective
  5. 13.5 Examples of Tradeoffs in Three Longitudinal Surveys
  6. 13.6 Discussion
  7. References
3. 14 Text Interviews on Mobile Devices
  1. 14.1 Texting as a Way of Interacting
  2. 14.2 Contacting and Inviting Potential Respondents through Text
  3. 14.3 Texting as an Interview Mode
  4. 14.4 Costs and Efficiency of Text Interviewing
  5. 14.5 Discussion
  6. References
4. 15 Quantifying Measurement Errors in Partially Edited Business Survey Data
  1. 15.1 Introduction
  2. 15.2 Selective Editing
  3. 15.3 Effects of Errors Remaining After SE
  4. 15.4 Case Study: Foreign Trade in Goods Within the European Union
  5. 15.5 Editing Big Data
  6. 15.6 Conclusions
  7. References
Section 4: Evaluation and Improvement
1. 16 Estimating Error Rates in an Administrative Register and Survey Questions Using a Latent Class Model
  1. 16.1 Introduction
  2. 16.2 Administrative and Survey Measures of Neighborhood
  3. 16.3 A Latent Class Model for Neighborhood of Residence
  4. 16.4 Results
  5. 16.5 Discussion and Conclusion
  6. Appendix 16.A: Program Input and Data
  7. Acknowledgments
  8. References
2. 17 ASPIRE
  1. 17.1 Introduction and Background
  2. 17.2 Overview of ASPIRE
  3. 17.3 The ASPIRE Model
  4. 17.4 Evaluation of Registers
  5. 17.5 National Accounts
  6. 17.6 A Sensitivity Analysis of GDP Error Sources
  7. 17.7 Concluding Remarks
  8. Appendix 17.A: Accuracy Dimension Checklist
  9. References
3. 18 Classification Error in Crime Victimization Surveys
  1. 18.1 Introduction
  2. 18.2 Background
  3. 18.3 Analytic Approach
  4. 18.4 Model Selection
  5. 18.5 Results
  6. 18.6 Discussion and Summary of Findings
  7. 18.7 Conclusions
  8. Appendix 18.A: Derivation of the Composite False‐Negative Rate
  9. Appendix 18.B: Derivation of the Lower Bound for False‐Negative Rates from a Composite Measure
  10. Appendix 18.C: Examples of Latent GOLD Syntax
  11. References
4. 19 Using Doorstep Concerns Data to Evaluate and Correct for Nonresponse Error in a Longitudinal Survey
  1. 19.1 Introduction
  2. 19.2 Data and Methods
  3. 19.3 Results
  4. 19.4 Discussion
  5. Acknowledgment
  6. References
5. 20 Total Survey Error Assessment for Sociodemographic Subgroups in the 2012 U.S. National Immunization Survey
  1. 20.1 Introduction
  2. 20.2 TSE Model Framework
  3. 20.3 Overview of the National Immunization Survey
  4. 20.4 National Immunization Survey: Inputs for TSE Model
  5. 20.5 National Immunization Survey TSE Analysis
  6. 20.6 Summary
  7. References
6. 21 Establishing Infrastructure for the Use of Big Data to Understand Total Survey Error
  1. Overview
  2. Part 1 Big Data Infrastructure at the Institute for Employment Research (IAB)
  3. References
  4. Part 2 Using Administrative Records Data at the U.S. Census Bureau: Lessons Learned from Two Research Projects Evaluating Survey Data
  5. References
  6. Part 3 Statistics New Zealand’s Approach to Making Use of Alternative Data Sources in a New Era of Integrated Data
  7. References
  8. Part 4 Big Data Serving Survey Research: Experiences at the University of Michigan Survey Research Center
  9. References
Section 5: Estimation and Analysis
1. 22 Analytic Error as an Important Component of Total Survey Error
  1. 22.1 Overview
  2. 22.2 Analytic Error as a Component of TSE
  3. 22.3 Appropriate Analytic Methods for Survey Data
  4. 22.4 Methods
  5. 22.5 Results
  6. 22.6 Discussion
  7. Acknowledgments
  8. References
2. 23 Mixed‐Mode Research
  1. 23.1 Introduction
  2. 23.2 Designing Mixed‐Mode Surveys
  3. 23.3 Literature Overview
  4. 23.4 Diagnosing Sources of Error in Mixed‐Mode Surveys
  5. 23.5 Adjusting for Mode Measurement Effects
  6. 23.6 Conclusion
  7. References
3. 24 The Effect of Nonresponse and Measurement Error on Wage Regression across Survey Modes
  1. 24.1 Introduction
  2. 24.2 Nonresponse and Response Bias in Survey Statistics
  3. 24.3 Data and Methods
  4. 24.4 Results
  5. 24.5 Summary and Conclusion
  6. Acknowledgments
  7. Appendix 24.A
  8. Appendix 24.B
  9. References
4. 25 Errors in Linking Survey and Administrative Data
  1. 25.1 Introduction
  2. 25.2 Conceptual Framework of Linkage and Error Sources
  3. 25.3 Errors Due to Linkage Consent
  4. 25.4 Erroneous Linkage with Unique Identifiers
  5. 25.5 Erroneous Linkage with Nonunique Identifiers
  6. 25.6 Applications and Practical Guidance
  7. 25.7 Conclusions and Take‐Home Points
  8. References
Wiley Series in Survey Methodology
Index
End User License Agreement

List of Tables

Chapter 02
1. Table 2.1 Characteristics of Twitter data error.
Chapter 04
1. Table 4.1 WSSM‐calculated Kullback–Leibler divergences of sample (with no measurement error), unit respondents (with no measurement error), final responses (with measurement error), and post‐SDL data to the original population.
2. Table 4.2 Number of records that violate edit rules across the 20 replications (or single realizations for Mic and MMic) after implementing SDL.
3. Table 4.3 Left: original tabular dataset O. Right: masked tabular dataset M, after primary and secondary cell suppression.
Chapter 05
1. Table 5.1 Coverage rates and response rates for two survey approaches.
2. Table 5.2 Parameters varied in simulations.
3. Table 5.3 Description of scenarios.
Chapter 06
1. Table 6.1 Overall and Internet response rates, by type of area and experimental condition.
2. Table 6.2 Overall response rates, by mode sequence and study.
3. Table 6.3 Estimated percentage of voters in 2006 election, by experimental variables.
Chapter 07
1. Table 7.1 Response rates for PC and mobile web.
2. Table 7.2 Breakoff rates (%) for PC and mobile web users.
3. Table 7.3 Measurement error in mobile web surveys: main findings and possible directions for further research.
Chapter 08
1. Table 8.1 Phase‐One weighted screener, main, and combined response rates.
2. Table 8.2 Phase‐One weighted main interview response rates by demographic subgroup.
3. Table 8.3 Phase Two and final‐weighted screener, main, and combined response rates by Phase‐One incentive.
4. Table 8.4 Final‐weighted main interview response rates.
5. Table 8.5 Selected key statistics by sex.
6. Table 8.6 Unweighted case counts by Phase-One incentive.
7. Table 8.7 Selected effort indicators.
8. Table 8.8 Reporting discrepancies between CAPI and ACASI by incentive treatment.
9. Table 8.9 $60 incentive predicting ordinal discrepancy outcome with and without covariates.
10. Table 8.10 p‐Values for quintiles of the main interview response propensity as a predictor of the ordinal discrepancy variable in the ordinal logistic regression model.
Chapter 10
1. Table 10.1 Prevalence (%) of survey starts by device and treatment.
2. Table 10.2 Prevalence (%) of survey completions by switching behavior and treatment.
3. Table 10.3 Description of experimental conditions.
4. Table 10.4 Estimated odds ratios for logistic regression models predicting top‐box or top‐two‐box score.
5. Table 10.B.1 Smartphone prevalence in web surveys—customer list samples.
6. Table 10.B.2 Smartphone prevalence in web surveys—institutional rosters.
7. Table 10.B.3 Smartphone prevalence in web surveys—probability‐based panels.
8. Table 10.B.4 Smartphone prevalence in web surveys—probability–based panels.
Chapter 13
1. Table 13.1 Nonresponse and coverage error over the course of the LISS panel.
2. Table 13.2 Flow through employment section of the 1996 SIPP questionnaire.
Chapter 14
1. Table 14.1 Properties of mobile voice and text communication, originally appeared in Schober et al. (2015).
2. Table 14.2 Absolute coverage bias estimates of texter population relative to U.S. mobile phone owners (2013) for six sociodemographic statistics.
Chapter 15
1. Table 15.1 Sum of absolute errors in observations of invoiced values found by SE and estimated sum of absolute errors in remaining observations (million SEK, standard errors in parenthesis).
2. Table 15.2 Net change by SE and estimated remaining error in total estimates of invoiced values (million SEK, standard errors in parenthesis).
Chapter 16
1. Table 16.1 Some official statistics on neighborhoods of Amsterdam (2011–2014).
2. Table 16.2 Cross tabulation of the observed survey measures with the administrative measure.
3. Table 16.3 Model fit measures for different specifications of the latent class model.
4. Table 16.4 Classification rates for the three observed measures (Model 6).
Chapter 17
1. Table 17.1 Sources of error considered by product for ASPIRE at Statistics Sweden.
2. Table 17.2 Ratings matrix for the TPR.
3. Table 17.3 Ratings matrix for the quarterly GDP.
4. Table 17.4 The final supply and use of product (CPA62) in 1,000,000 SEK for the years 2008–2012.
5. Table 17.5 The impact on GDP if the bias in the price of product CPA62 is −5, −10, +5, or +10% for all components in Table 17.4.
6. Table 17.6 The final supply and use of the product CPA291.
7. Table 17.7 The impact on GDP if the bias in the price of product CPA291 is −2.5, −5, +2.5, or +5% for production (S1) only in Table 17.4.
Chapter 18
1. Table 18.1 Fit statistics for MLC models of less serious crimes against an individual (models by sets of waves included).
2. Table 18.2 Fit statistics for MLC models of crimes against a household (models by sets of waves included).
3. Table 18.3 Estimated classification errors by wave and sets of waves from MLC models for less serious crimes against an individual.
4. Table 18.4 Estimated classification errors by wave and sets of waves from MLC models for crimes against a household.
5. Table 18.5 Evidence rating of classification error differences for demographic variables.
6. Table 18.6 Average estimated classification error rates for demographic variables with a strong evidence rating.
Chapter 19
1. Table 19.1 Concerns groups assignment with sample sizes and proxy measures of reluctance, using Wave 2 data.
2. Table 19.2 Unit response rates and efforts by concerns group.
3. Table 19.3 Mean number of “Don’t Know” and “Refused” answers by concerns group.
4. Table 19.4 Total and component nonresponse biases for five expenditure estimates.
5. Table 19.5 Total and component nonresponse biases in housing and education expenditure estimates.
6. Table 19.6 Total and component nonresponse biases estimates for five expenditures.
7. Table 19.7 Expenditure estimates in dollars (and standard errors) with and without weighting adjustments for nonresponse using doorstep concerns data.
Chapter 20
1. Table 20.1 Vaccines monitored by the National Immunization Survey (NIS), recommended doses, and 2012 national‐level vaccination coverage rates.
2. Table 20.2 2012 national‐level vaccination coverage rates by domains defined by important sociodemographic characteristics of the child collected in the household interview (2012 National Immunization Survey (NIS)).
3. Table 20.3 Total survey error model inputs by stages (2012 National Immunization Survey).
4. Table 20.4 Estimated mean TSE and 95% credible interval for TSE using the three‐stage model for national‐level vaccination coverage rates corresponding to 4 : 3 : 1 : 3 : 3 : 1, 4+DTaP, and 1+MMR (2012 National Immunization Survey).
5. Table 20.5 Means and 95% credible intervals for four component errors in national‐level vaccination coverage rates (2012 National Immunization Survey).
Chapter 21
1. Table 21.4.1 Proportion of housing units (HU) by commercial list substratum with one or more persons and actual number of persons in age ranges 18–44 and 65+ based on NSFG screener responses.
2. Table 21.4.2 Proportion of teachers identified as eligible by commercial educational database compared to MTTS screener results for schools with matched NCES IDs.
Chapter 22
1. Table 22.1 Variables coded for each of the 100 peer‐reviewed articles in the study.
2. Table 22.2 Chi‐square tests for bivariate associations of each of the estimation errors with whether or not the journal has a statistical consultant on the editorial board.
3. Table 22.3 Chi‐square tests for bivariate associations of software errors with whether or not the journal has a statistical consultant on the editorial board.
Chapter 23
1. Table 23.1 Multinomial regression predicting mode (face‐to‐face is reference category).
2. Table 23.2 Results of measurement equivalence testing, without and with adjustment.
3. Table 23.3 Estimates of means and variances of the Parenthood factor, unweighted and weighted data (scalar model).
4. Table 23.4 Means and variances of observed Parenthood scores with and without adjustment.
5. Table 23.5 Estimated means and variances of latent Parenthood factor unadjusted and with adjustment for mode measurement effect.
6. Table 23.6 Means and variances of observed Parenthood scores unadjusted and with adjustment for mode measurement effect.
7. Table 23.7 Means and variances of Parenthood score unadjusted and with adjustment for mode measurement effect.
8. Table 23.8 Estimated means and variances of latent Parenthood factor unadjusted and with adjustment for mode measurement effect.
Chapter 24
1. Table 24.1 Response rates across modes of data collection.
2. Table 24.A.1 Wage regression: coefficients, absolute and relative deviations (coverage web sample).
3. Table 24.A.2 Wage regression: coefficients, absolute and relative deviations (coverage tel. sample).
4. Table 24.A.3 Wage regression: coefficients, absolute and relative deviations (coverage pooled sample).
5. Table 24.B.1 Wage regression: coefficients, absolute and relative deviations (web sample).
6. Table 24.B.2 Weighted wage regression: coefficients, absolute and relative deviations (web sample).
7. Table 24.B.3 Wage regression: coefficients, absolute and relative deviations (telephone sample).
8. Table 24.B.4 Weighted wage regression: coefficients, absolute and relative deviations (telephone sample).
9. Table 24.B.5 Wage regression: coefficients, absolute and relative deviations (pooled sample, unweighted).
Chapter 25
1. Table 25.1 A selection of large‐scale surveys that link survey records to administrative records.

List of Illustrations

Chapter 01
1. Figure 1.1 The generic statistical business process model.
2. Figure 1.2 Subjective sample of events in the evolution of the concept of TSE.
Chapter 02
1. Figure 2.1 Theoretical spaces of Twitter data error.
2. Figure 2.2 Percent of online U.S. adults who use Twitter as of 2014‥
3. Figure 2.3 Marijuana legalization tweet volume by month, worldwide, U.S.A. and Colorado, basic query (2011–2014).
4. Figure 2.4 Pro‐choice/life tweet volume by month, worldwide, U.S. and Colorado, basic query (2011–2014).
5. Figure 2.5 Marijuana legalization tweet volume by month, worldwide, basic query, including and excluding RTs, 2011–2014.
6. Figure 2.6 Pro‐choice/life tweet volume by month, worldwide, basic query, including and excluding RTs, 2011–2014.
7. Figure 2.7 Marijuana legalization tweet volume by month, worldwide, basic vs. expanded query (2011–2014).
8. Figure 2.8 Pro‐choice/life tweet volume by month, worldwide, basic vs. expanded query (2011–2014).
9. Figure 2.9 Marijuana legalization Twitter and Gallup sentiment by month, worldwide, basic query, 2011–2014.
10. Figure 2.10 Pro‐choice/life Twitter and Gallup sentiment by month, worldwide, basic query, 2011–2014.
Chapter 03
1. Figure 3.1 Word cloud based on Dutcher (2014).
2. Figure 3.2 Infographic from James (2014).
3. Figure 3.3 Sample data visualization‥
4. Figure 3.4 Simplified view of data linkage.
Chapter 04
1. Figure 4.1 The role of SDL within the TSE paradigm, modeled on Groves (1989).
2. Figure 4.2 Forms of dissemination of confidential official statistics data.
3. Figure 4.3 Schematic representation of risk‐utility paradigms. Candidate releases on the efficient frontier are connected by dotted lines. The large point is optimal for the value function shown.
4. Figure 4.4 Risk‐distortion scatterplots for 108 candidate releases from the CPS‐8D database. Three swap rates (1%—circles, 2%—triangles, and 10%—plus signs) are shown, and for each, there are 36 candidate releases representing all choices of one or two swap variables.
5. Figure 4.5 Edit violations created by four SDL methods. Two variables are depicted: SL = number of hours of skilled labor and SW = wages paid to skilled labor. The dotted lines represent range and ratio edit constraints. Top left: original data, from the 1991 Columbian manufacturing survey. Top right: SDL method = additive noise. Bottom left: SDL method = rank swapping. Bottom right: SDL method = microaggregation followed by additive noise.
6. Figure 4.6 Risk‐utility map for the Colombian manufacturing survey data‐based simulation. The solid line indicates the risk‐utility frontier. The open symbols represent edit‐after‐SDL approaches, and the solid symbols represent edit‐preserving SDL approaches. Smaller values of PL1 and U_prop represent the higher levels of data protection and data utility. PL1 is the risk arising from linkage to the “nearest” record being correct; U_prop is the propensity score utility.
7. Figure 4.7 For simulated data, scatterplot of correlations among nine log‐transformed variables following edit–imputation against correlations in true data. Abbreviations: AAI, all active items; BE, Bayes edit; BE‐min, Bayes edit applied to minimal number of fields; FH, Fellegi–Holt.
8. Figure 4.8 For the CM data, scatterplot of correlations among 27 variables following edit–imputation against correlations in edit passing records. Abbreviations: BE, Bayes edit; BE‐min, Bayes edit applied to minimal number of fields; FC, final released data.
9. Figure 4.9 The effect on indexed microaggregation on stratum‐weight relationships. Left: beforehand, weight discloses stratum completely. Right: afterward, weight only places stratum in one of two classes.
Chapter 05
1. Figure 5.1 Diagram of relationship between CPs and RPs and survey variable Y.
2. Figure 5.2 Simulation results: coverage and response rates.
3. Figure 5.3 Simulation results: coverage rates and bias. See Table 5.3 for descriptions of the scenarios.
Chapter 06
1. Figure 6.1 Costs per household (in 2010 dollars) and mail return rates for the U.S. Decennial Census since 1970. The dashed line represents costs; the solid line, the return rate‥
Chapter 08
1. Figure 8.1 The ratio of standard errors under the $40 and $60 treatments across 16 key statistics for males, females, and age, race, and ethnicity subgroups.
Chapter 09
1. Figure 9.1 TSE (a) representation and (b) measurement in a cross‐cultural context.
2. Figure 9.2 The translation; review; adjudication; pretesting; and documentation model.
Chapter 10
1. Figure 10.1 Device and design choices and associated survey errors.
2. Figure 10.2 Prevalence of smartphone starts from the Decipher platform’s highest volume accounts.
3. Figure 10.B.1 Prevalence of smartphone starts ordered by prevalence within year: customer list samples.
4. Figure 10.B.2 Prevalence of smartphone starts: institutional roster samples.
5. Figure 10.B.3 Prevalence of smartphone starts: probability panel samples.
6. Figure 10.B.4 Prevalence of smartphone starts: opt‐in panel samples.
7. Figure 10.C.1 Smartphone/Legacy (screen captures from an iPhone 5).
8. Figure 10.C.2 Smartphone/New (screen captures from Blackberry Torch).
9. Figure 10.C.3 Smartphone/Drop‐down (screen captures from iPhone 4s).
10. Figure 10.C.4 Smartphone/Numeric (screen captures from iPhone 4s, Blackberry Torch).
11. Figure 10.C.5 Smartphone/Slider (screen captures from iPhone 5).
12. Figure 10.C.6 PC/Legacy (screen capture Windows 7 PC).
13. Figure 10.C.7 PC/New (screen capture Windows 7 PC).
Chapter 12
1. Exhibit 12.1 Example of a question‐level CARI coding screen.
2. Exhibit 12.2 Fictionalized screenshots for MyCases in the mFOS app.
3. Figure 12.1 Differences between recorded time and actual time on laptops.
4. Figure 12.2 Time of electronic record of contacts (EROC) entry by device.
5. Exhibit 12.3 Visualization of GIS data to detect falsification.
6. Exhibit 12.4 Sample of dashboard alerts.
7. Exhibit 12.5 Annotated example of field supervisor dashboard.
8. Exhibit 12.6 Example of a CARI portlet on the dashboard.
Chapter 13
1. Figure 13.1 Proportion of “don’t know” and “refusal” answers to personal gross income question in last five waves before attrition in BHPS.
Chapter 14
1. Figure 14.1 Texting rates among U.S. adults based on Pew telephone surveys from December 2007 through April 2013.
2. Figure 14.2 Data quality in text versus voice and human versus automated interviews; (a) rounding, (b) straightlining, and (c) disclosure.
3. Figure 14.3 Interview duration (median duration of question–answer sequences in minutes) and median number of turns per survey question in text and voice interviews.
Chapter 15
1. Figure 15.1 Cumulative distribution of score g_k, normalized to between 0 and 1. The graph is limited to the 10,000 highest scores monthly in 2014. Some 500 highest scores are excluded for scalability. Each isarithm line corresponds to 50 units.
2. Figure 15.2 Plot of log of sum of absolute value of errors against log of sum of scores. Each sum of 100 units classified by order of the score. Crosses, sample exports; dots, SE imports; pyramids, sample imports; stars, SE exports.
Chapter 16
1. Figure 16.1 Neighborhoods of residence (“stadsdeel”) in Amsterdam, the Netherlands.
2. Figure 16.2 Formulation of the 2013 Sportmonitor survey questions on neighborhood of residence. The postcode (ZIP code) is asked at the most detailed level possible.
3. Figure 16.3 Latent class measurement model for three measurements of neighborhood of residence. Latent variables shown in shaded ovals, observed variables as rectangles.
4. Figure 16.4 Model‐based estimates of the proportion of residents within each (true) neighborhood who received one of the survey modes (relationship between x and S).
5. Figure 16.5 Estimated correct classification probabilities (excluding missing) by survey mode for the two survey measures.
6. Figure 16.6 Estimated conditional probabilities of choosing each category within method classes (profile plot). Method class 1 is estimated to contain 9%, and Method class 2 is estimated to contain 91% of respondents.
7. Figure 16.7 Scatterplot of a constructed neighborhood SES score (see Table 16.2) and the log odds ratio (logistic regression coefficient) of choosing each neighborhood when the Method class equals 1. A, Centrum; E, West; F, Nieuw–West; K, Zuid; M, Oost; N, Noord; T, Zuidoost.
Chapter 17
1. Figure 17.1 Process for estimating GDP by current and constant price approaches.
Chapter 18
1. Figure 18.1 Illustration of relationship between NCVS interview waves and analysis groups.
Chapter 19
1. Figure 19.1 Screenshot of CHI used for the Consumer Expenditure Interview Survey.
Chapter 20
1. Figure 20.1 Distribution of total survey error (TSE) for the 4 : 3 : 1 : 3 : 3 : 1^a vaccination coverage rate (2012 National Immunization Survey). (Note: ^a4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT); 3 or more doses of polio vaccine; 1 or more doses of measles‐containing vaccine, 3 or more doses of Haemophilus influenzae Type b (Hib)‐containing vaccine, 3 or more doses of Hepatitis B vaccine, and 1 or more doses of varicella vaccine (at or after 12 months of age).)
2. Figure 20.2 Mean TSE and 95% credible intervals using the two‐stage model (incorporating sample‐frame coverage and nonresponse error only) for national‐level vaccination coverage rates corresponding to 4 : 3 : 1 : 3 : 3 : 1^a, 4+DTaP^b, and 1+MMR^c (2009–2012 National Immunization Survey). (Note: ^a4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT); 3 or more doses of polio vaccine; 1 or more doses of measles‐containing vaccine, 3 or more doses of Haemophilus influenzae Type b (Hib)‐containing vaccine, 3 or more doses of Hepatitis B vaccine, and 1 or more doses of varicella vaccine (at or after 12 months of age). ^b4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT). ^c1 or more doses of measles/mumps/rubella vaccine (MMR).)
3. Figure 20.3 Mean TSE using the two‐ and three‐stage models for national‐level vaccination coverage rates corresponding to 4 : 3 : 1 : 3 : 3 : 1^a, 4+DTaP^b and 1+MMR^c (2012 National Immunization Survey). (Note: Mean TSE estimate for the overall group area based solely on 2012 data. ^a4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT); 3 or more doses of polio vaccine; 1 or more doses of measles‐containing vaccine, 3 or more doses of Haemophilus influenzae Type b (Hib)‐containing vaccine, 3 or more doses of Hepatitis B vaccine, and 1 or more doses of varicella vaccine (at or after 12 months of age). ^b4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT). ^c1 or more doses of measles/mumps/rubella vaccine (MMR).)
4. Figure 20.4 Mean TSE and 95% credible intervals using the three‐stage model for vaccination coverage rates corresponding to 4 : 3 : 1 : 3 : 3 : 1^a, 4+DTaP^b, and 1+MMR^c by child’s race/ethnicity (2011–2012 National Immunization Survey). (Note: Mean TSE estimate for the overall group area based solely on 2012 data. ^a4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT); 3 or more doses of polio vaccine; 1 or more doses of measles‐containing vaccine, 3 or more doses of Haemophilus influenzae Type b (Hib)‐containing vaccine, 3 or more doses of Hepatitis B vaccine, and 1 or more doses of varicella vaccine (at or after 12 months of age). ^b4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT). ^c1 or more doses of measles/mumps/rubella vaccine (MMR).)
5. Figure 20.5 Mean TSE and 95% credible intervals using the three‐stage model for vaccination coverage rates corresponding to 4 : 3 : 1 : 3 : 3 : 1^a, 4+DTaP^b and 1+MMR^c by child’s mother’s education (2011–2012 National Immunization Survey). (Note: Mean TSE estimate for the overall group area based solely on 2012 data. ^a4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT); 3 or more doses of polio vaccine; 1 or more doses of measles‐containing vaccine, 3 or more doses of Haemophilus influenzae Type b (Hib)‐containing vaccine, 3 or more doses of Hepatitis B vaccine, and 1 or more doses of varicella vaccine (at or after 12 months of age). ^b4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT). ^c1 or more doses of measles/mumps/rubella vaccine (MMR).)
6. Figure 20.6 Mean TSE and 95% credible interval using the three‐stage model for vaccination coverage rates corresponding to 4 : 3 : 1 : 3 : 3 : 1^a, 4+DTaP^b and 1+MMR^c by child’s sex (2011–2012 National Immunization Survey). (Note: Mean TSE estimate for the overall group area based solely on 2012 data. ^a4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT); 3 or more doses of polio vaccine; 1 or more doses of measles‐containing vaccine, 3 or more doses of Haemophilus influenzae Type b (Hib)‐containing vaccine, 3 or more doses of Hepatitis B vaccine, and 1 or more doses of varicella vaccine (at or after 12 months of age). ^b4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT). ^c1 or more doses of measles/mumps/rubella vaccine (MMR).)
7. Figure 20.7 Mean TSE and 95% credible intervals using the three‐stage model for vaccination coverage rates corresponding to 4 : 3 : 1 : 3 : 3 : 1^a, 4+DTaP^b and 1+MMR^c by child’s MSA status (2011–2012 National Immunization Survey). (Note: Mean TSE estimate for the overall group area based solely on 2012 data. ^a4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT); 3 or more doses of polio vaccine; 1 or more doses of measles‐containing vaccine, 3 or more doses of Haemophilus influenzae Type b (Hib)‐containing vaccine, 3 or more doses of Hepatitis B vaccine, and 1 or more doses of varicella vaccine (at or after 12 months of age). ^b4 or more doses of diphtheria and tetanus toxoids and acellular pertussis vaccine (DTaP/DTP/DT). ^c1 or more doses of measles/mumps/rubella vaccine (MMR).)
Chapter 21
1. Figure 21.1.1 Comparison of different wage records.
Chapter 22
1. Figure 22.1 Percentages of all potential errors identified in the 100 articles, by type of error.
2. Figure 22.2 Pearson correlations between the number of errors of a particular type and the impact factor of the journal (^* p < 0.10).
3. Figure 22.3 Pearson correlations between indicators of particular types of errors and journal impact factors (see Table 22.1 for indicator names; ^** p < 0.05, ^* p < 0.10).
4. Figure 22.4 Proportions of articles with particular types of estimation errors over time.
5. Figure 22.5 Proportions of articles with particular types of quality errors over time.
6. Figure 22.6 Proportions of articles with particular types of software errors over time.
7. Figure 22.7 Proportions of articles with particular types of language/presentation errors over time.
8. Figure 22.8 Percentages of all possible errors identified in the articles in each time period, for each type of error.
Chapter 23
1. Figure 23.1 Covariate adjustment in a CFA. Note that b1–b4 indicate equality constraints on regression coefficients across groups, and c indicates an equality constraint on the covariance.
2. Figure 23.2 The mixture model for the extended mixed‐mode survey.
3. Figure 23.3 Path diagram for adjustment model, eta is the latent Parenting factor.
Chapter 24
1. Figure 24.1 Wage regression coefficients (web sample).
2. Figure 24.2 Wage regression coefficients (telephone sample).
3. Figure 24.3 Weighted wage regression coefficients (web sample).
4. Figure 24.4 Weighted wage regression coefficients (telephone sample).
5. Figure 24.5 Weighted wage regression coefficients (pooled sample, unweighted).
Chapter 25
1. Figure 25.1 Conceptual framework of the record linkage process.

Notes on Contributors

Manfred Antoni
Research Data Centre (FDZ)
Institute for Employment Research (IAB)
Nuremberg
Germany

Christopher Antoun
Center for Survey Measurement
U.S. Census Bureau
Suitland, MD
USA

Reg Baker
Marketing Research Institute International
Ann Arbor, MI
USA

Stefan Bender
Research Data and Service Centre
Deutsche Bundesbank
Frankfurt am Main
Germany

Grant Benson
Survey Research Center
University of Michigan
Ann Arbor, MI
USA

Heather Bergdahl
Process Department
Statistics Sweden
Stockholm
Sweden

Marcus E. Berzofsky
Division for Statistics and Data Science
RTI International
Research Triangle Park, NC
USA

Paul P. Biemer
Social, Statistical, and Environmental Sciences
RTI International
Research Triangle Park, NC
Odum Institute for Research in Social Science
University of North Carolina
Chapel Hill, NC
USA

Paul Burton
Survey Research Center
University of Michigan
Ann Arbor, MI
USA

Christine Bycroft
Statistics New Zealand
Wellington
New Zealand

Jennifer Hunter Childs
Research and Methodology Directorate
U.S. Census Bureau
Washington, DC
USA

Sue Connor
Westat
Rockville, MD
USA

Frederick G. Conrad
Survey Research Center
University of Michigan
Ann Arbor, MI
Joint Program in Survey Methodology
University of Maryland
College Park, MD
USA

Mick P. Couper
Survey Research Center
University of Michigan
Ann Arbor, MI
Joint Program in Survey Methodology
University of Maryland
College Park, MD
USA

Edith de Leeuw
Department of Methodology and Statistics
Utrecht University
Utrecht
The Netherlands

Stephanie Eckman
Survey Research Division
RTI International
Washington, DC
USA

Brad Edwards
Westat
Rockville, MD
USA

Barbara Felderer
Collaborative Research Center SBF 884
“Political Economy of Reforms”
University of Mannheim
Mannheim
Germany

Jamie Griffin
Survey Research Center
University of Michigan
Ann Arbor, MI
USA

Heidi Guyer
Survey Research Center
University of Michigan
Ann Arbor, MI
USA

Kristen Cibelli Hibben
Survey Research Center
University of Michigan
Ann Arbor, MI
USA

Daniela Hochfellner
Center for Urban Science and Progress
New York University
New York, NY
USA

Anders Holmberg
Statistics Norway
Oslo
Norway

Joop Hox
Department of Methodology and Statistics
Utrecht University
Utrecht
The Netherlands

Yuli Patrick Hsieh
Survey Research Division
RTI International
Chicago, IL
USA

Frost Hubbard
Survey Solutions Division
IMPAQ International
Columbia, MD
USA

Andrew L. Hupp
Survey Research Center
University of Michigan
Ann Arbor, MI
USA

Joost Kappelhof
Department of Education, Minorities, and Methodology
Institute for Social Research/SCP
The Hague
The Netherlands

Alan F. Karr
Center of Excellence for Complex Data Analysis
RTI International
Research Triangle Park, NC
USA

Jennifer Kelley
Institute for Social and Economic Research
University of Essex
Colchester
UK

Meena Khare
National Center for Health Statistics
Centers for Disease Control and Prevention
Hyattsville, MD
USA

Yumi Kim
Department of Research Methods
Market Strategies International
Livonia, MI
USA

Antje Kirchner
Department of Sociology
University of Nebraska‐Lincoln
Lincoln, NE
Survey Research Division
RTI International
Research Triangle Park, NC
USA

Thomas Klausch
Department for Epidemiology and Biostatistics
VU University Medical Center
Amsterdam
The Netherlands

Frauke Kreuter
Joint Program in Survey Methodology
University of Maryland
College Park, MD
USA
Department of Sociology
University of Mannheim
Mannheim
Statistical Methods Group
Institute for Employment Research (IAB)
Nuremberg
Germany

John LaFrance
Market Strategies International
Livonia, MI
USA

Thomas Laitila
Department of Research and Development
Statistics Sweden
Department of Statistics
Örebro University School of Business
Örebro
Sweden

JiaoJiao Li
Market Strategies International
Livonia, MI
USA

Karin Lindgren
Process Department
Statistics Sweden
Stockholm
Sweden

Peter J. Lugtig
Institute for Social and Economic Research
University of Essex
Colchester
UK
Department of Methodology and Statistics
Utrecht University
Utrecht
The Netherlands

Lars E. Lyberg
Inizio
Stockholm
Sweden

Peter Lynn
Institute for Social and Economic Research
University of Essex
Colchester
UK

Aaron Maitland
Westat
Rockville, MD
USA

Aigul Mavletova
Department of Sociology
National Research University Higher School of Economics
Moscow
Russia

Peter Ph. Mohler
University of Mannheim
Mannheim
Germany

William D. Mosher
Bloomberg School of Public Health
Johns Hopkins University
Baltimore, MD
USA

Mary H. Mulry
Research and Methodology Directorate
U.S. Census Bureau
Washington, DC
USA

Joe Murphy
Survey Research Division
RTI International
Chicago, IL
USA

Elizabeth M. Nichols
Research and Methodology Directorate
U.S. Census Bureau
Washington, DC
USA

Anders Norberg
Process Department
Statistics Sweden
Stockholm
Sweden

Daniel L. Oberski
Department of Methodology and Statistics
Utrecht University
Utrecht
The Netherlands

Beth‐Ellen Pennell
Survey Research Center
University of Michigan
Ann Arbor, MI
USA

Gregg Peterson
Survey Research Center
University of Michigan
Ann Arbor, MI
USA

Vicki J. Pineau
NORC at the University of Chicago
Chicago, IL
USA

Joseph W. Sakshaug
Cathie Marsh Institute for Social Research
University of Manchester
Manchester
UK
Department of Statistical Methods
Institute for Employment Research (IAB)
Nuremberg
Germany

Michael F. Schober
Department of Psychology
New School for Social Research
New York, NY
USA

James A. Singleton
National Center for Immunization and Respiratory Diseases
Centers for Disease Control and Prevention
Atlanta, GA
USA

Benjamin Skalland
NORC at the University of Chicago
Chicago, IL
USA

Philip J. Smith
National Center for Immunization and Respiratory Diseases
Centers for Disease Control and Prevention
Atlanta, GA
USA

Diana Maria Stukel
FHI 360
Washington, DC
USA

Can Tongur
Process Department
Statistics Sweden
Stockholm
Sweden

Roger Tourangeau
Westat
Rockville, MD
USA

Dennis Trewin
Former Australian Statistician
Australian Bureau of Statistics
Canberra
Australia

James Wagner
Survey Research Center
University of Michigan
Ann Arbor, MI
Joint Program in Survey Methodology
University of Maryland
College Park, MD
USA

Brady T. West
Survey Research Center
University of Michigan
Ann Arbor, MI
Joint Program in Survey Methodology
University of Maryland
College Park, MD
USA

Kirk M. Wolter
NORC at the University of Chicago
Chicago, IL
USA

Gelaye Worku
Department of Statistics
Stockholm University
Stockholm
Sweden

Yingfu Xie
Process Department
Statistics Sweden
Stockholm
Sweden

H. Yanna Yan
Survey Research Center
University of Michigan
Ann Arbor, MI
USA

Ting Yan
Methodology Unit
Westat
Rockville, MD
USA

David Yankey
National Center for Immunization and Respiratory Diseases
Centers for Disease Control and Prevention
Atlanta, GA
USA

Wei Zeng
NORC at the University of Chicago
Chicago, IL
USA

Zhen Zhao
National Center for Immunization and Respiratory Diseases
Centers for Disease Control and Prevention
Atlanta, GA
USA

WILEY SERIES IN SURVEY METHODOLOGY

Total Survey Error in Practice

Notes on Contributors