Muutke küpsiste eelistusi

Mixed-Mode Official Surveys: Design and Analysis [Kõva köide]

5.00/5 (2 hinnangut Goodreads-ist)
(Statistics Netherlands, The Hague, The Netherlands), , , , ,
Teised raamatud teemal:
Mixed-Mode Official Surveys: Design and AnalysisSuurem pilt
Teised raamatud teemal:
Püsilink: https://www.kriso.ee/db/9781138618459.html
Märksõnad:

This book covers mixed-mode data collection methods and its consequence for the survey process from design, data collection, and estimation. It is an invaluable resource for professionals involved in survey research and analysis, graduate students of survey methodology and academics working in the area.



Mixed-mode surveys have become a standard at many statistical institutes. However, the introduction of multiple modes in one design goes with challenges to both methodology and logistics. Mode-specific representation and measurement differences become explicit and demand for solutions in data collection design, questionnaire design and estimation. This is especially true when surveys are repeated and are input to long statistics’ time series. So how can statistical institutes deal with such changes? What are the origins of mode-specific error? And how can they be dealt with? In this book, the authors provide answers to these questions, and much more.

Features

  • Concise introduction to all the key elements of mixed-mode survey design and analysis

  • Realistic official statistics examples from three general population surveys

  • Suitable for survey managers and survey statisticians alike

  • An overview of mode-specific representation and measurement errors and how to avoid, reduce and adjust them

Preface xi
Acknowledgements xiii
Authors xv
Part I Introduction
1 Foreword to Mixed-Mode Official Surveys: Design and Analysis
3(6)
1.1 Why a Book on Mixed-Mode Survey Design and Analysis?
3(1)
1.2 Modes and Devices
4(2)
1.3 Outline of the Book
6(3)
2 Designing Mixed-Mode Surveys
9(24)
2.1 Introduction
9(1)
2.2 Why Mixed-Mode Survey Designs?
9(4)
2.3 How Can Modes Be Combined?
13(3)
2.4 How Do Modes Affect Quality?
16(6)
2.5 How to Reduce Mode Impact on Quality?
22(4)
2.6 Case Studies
26(4)
2.6.1 Dutch Crime Victimization Survey
27(1)
2.6.2 Dutch Labour Force Survey
28(1)
2.6.3 European Health Interview Survey
29(1)
2.7 Summary
30(3)
Part II Mode Effects
3 Mode-Specific Measurement Effects
33(24)
3.1 Introduction
33(2)
3.2 Measurement Features of Modes
35(7)
3.3 Mode-Specific Answering Behaviors and Response Styles
42(7)
3.4 Detection of Mode-Specific Measurement Effects
49(6)
3.5 Summary
55(2)
4 Mode-Specific Selection Effects
57(26)
4.1 Introduction
57(1)
4.2 Coverage Issues in CAPI, CATI, Mail, and Web Surveys
58(10)
4.2.1 CAPI Coverage
58(1)
4.2.2 CATI Coverage
59(4)
4.2.3 Mail Coverage
63(1)
4.2.4 Web Coverage
63(5)
4.3 Response and Subgroup Response Rates in Mixed-Mode Designs
68(11)
4.3.1 Mixed-Mode Designs at Statistics Netherlands
70(9)
4.4 Summary
79(1)
Notes
79(4)
Part III Design
5 Mixed-Mode Data Collection Design
83(40)
5.1 Introduction
83(2)
5.2 Mode and Mode Combinations
85(5)
5.3 Concurrent and Sequential Designs
90(4)
5.3.1 Concurrent or Sequential Order?
90(3)
5.3.2 Sequential, but Which Sequence?
93(1)
5.4 Costs
94(1)
5.5 Sample Composition and Data Quality
95(2)
5.6 Communication Strategies to Increase Web Response
97(23)
5.6.1 Contacting the Sample Person or Household
99(1)
5.6.2 Reminders
100(1)
5.6.3 Invitation Letters, Flyers, and Envelopes: Statistics Netherlands' Experiments
101(9)
5.6.3.1 Experimental Manipulations
102(4)
5.6.3.2 Results
106(4)
5.6.4 Incentives in Web and Mixed-Mode Surveys
110(13)
5.6.4.1 Effect on Sample Composition and Data Quality
112(1)
5.6.4.2 Incentives in European NSIs
113(1)
5.6.4.3 Incentive Experiments by Statistics Netherlands
114(4)
5.6.4.4 An Incentive Experiment in the Longitudinal Labor Force Survey
118(2)
5.7 Summary
120(1)
Notes
121(2)
6 Mixed-Mode Questionnaire Design
123(26)
6.1 Introduction
123(1)
6.2 General Goals and Challenges in Questionnaire Design
123(4)
6.2.1 Why Questionnaire Matters
123(1)
6.2.2 Striking a Balance between Conflicting Goals and Stakeholders
124(2)
Clients (Partly Overlap with Data Users)
125(1)
Respondents
125(1)
Interviewers
125(1)
Programmers
125(1)
Data Users (Partly Overlap with Clients)
125(1)
Project Managers
126(1)
Questionnaire Designers
126(1)
6.2.3 Questionnaire Design as an Iterative Process
126(1)
6.3 Questionnaire Design and Mode-Specific Measurement Errors
127(6)
6.3.1 The Cognitive Response Process Model
127(1)
6.3.2 Main Sources of Mode-Specific Measurement Error
128(1)
6.3.3 Even Small Design Differences May Matter
129(1)
6.3.4 Device-Specific Measurement Errors
130(3)
6.4 Questionnaire Design for Mixed-Mode Surveys
133(5)
6.4.1 Minimizing Differences or Minimizing Error?
133(1)
6.4.2 Mixed-Mode Requirements in the Questionnaire Design Process
134(2)
6.4.3 Mixed-Device Questionnaire Design
136(2)
6.5 Testing and Evaluating Mixed-Mode Questionnaires
138(5)
6.5.1 Testing and Evaluation as Part of the Questionnaire Design Process
138(3)
6.5.2 Mixed-Mode Questionnaire Testing and Evaluation
141(2)
Step 1 Make an Explicit Mode Risk Assessment
141(1)
Step 2 Decide if the Test Should Compare Modes
141(1)
Step 3 Decide Which Modes to Test, and When in the Development Process
142(1)
Step 4 Test Relevant Modes as Realistically as Practically Possible
142(1)
Step 5 In Analyses and Reporting, Distinguish between Usability and Content Findings, and Reflect on Mode and Device Specificity Regarding Findings
143(1)
6.6 Guidelines for Mixed-Mode Questionnaire Design
143(3)
6.6.1 Keep the Stimulus and Response Task as Similar as Possible in All Modes
143(1)
6.6.2 Organize the Questionnaire Design Process to Prevent and Detect Mode-Specific Measurement Errors
144(1)
6.6.3 Prevent Social Desirability Bias
145(1)
6.6.4 Prevent Satisficing
145(1)
6.7 Summary
146(3)
Part IV Analysis
7 Field Tests and Implementation of Mixed-Mode Surveys
149(26)
7.1 Introduction
149(3)
7.2 Design of Field Experiments
152(2)
7.3 Design-Based Inference for Field Experiments
154(7)
7.3.1 Measurement Error Model and Hypotheses Testing
154(2)
7.3.2 Parameter and Variance Estimation
156(2)
7.3.3 Wald Test and Related Tests
158(1)
7.3.4 Extensions
159(1)
7.3.5 Software
160(1)
7.4 Time Series Methods
161(3)
7.5 Small Area Estimation Methods
164(2)
7.6 The Introduction of a Hybrid Mixed-Mode Design in the Dutch Crime Victimization Survey
166(6)
7.6.1 Design
166(2)
7.6.2 Results
168(4)
7.7 Discussion
172(2)
7.8 Summary
174(1)
8 Re-interview Designs to Disentangle and Adjust for Mode Effects
175(26)
8.1 Introduction
175(2)
8.2 Decomposition of Relative Mode Effects
177(2)
8.3 Estimating Components of the Relative Mode Effects
179(5)
8.3.1 Experimental Design
179(2)
8.3.2 Estimation Strategies
181(2)
8.3.3 Assumptions
183(1)
8.4 Application
184(6)
8.5 Adjusting Measurement Bias Using Re-interview Designs
190(5)
8.6 Simulation Study
195(2)
8.7 Extension of Re-interview Design to Multiple Modes
197(1)
8.8 Conclusions
198(1)
8.9 Summary
199(2)
9 Mixed-Mode Data Analysis
201(22)
9.1 Introduction
201(1)
9.2 Analyzing Mixed-Mode Survey Response Data
202(3)
9.2.1 Literature
202(2)
9.2.2 Problem Statement, Definitions, and Notation
204(1)
9.3 Correcting Differential Measurement Bias
205(3)
9.3.1 Counterfactuals
205(2)
9.3.2 Mode-Specific Estimators
207(1)
9.4 Balancing Differential Measurement Bias
208(4)
9.4.1 Mode Calibration
208(3)
9.4.2 Choosing Calibration Levels
211(1)
9.5 Handling Measurement Bias in Practice
212(2)
9.5.1 Testing Assumptions
212(1)
9.5.2 Calibration Levels and Mixing Coefficients
212(2)
9.5.3 Comparison with Single-Mode Designs
214(1)
9.6 Applications
214(5)
9.6.1 Health Survey
214(1)
9.6.2 Crime Victimization Survey
215(3)
9.6.3 Labor Force Survey
218(1)
9.7 Summary
219(4)
Part V The Future of Mixed-Mode Surveys
10 Multi-Device Surveys
223(28)
10.1 Introduction
223(2)
10.2 Smart Surveys
225(8)
10.2.1 A Taxonomy of Smart Surveys
225(3)
10.2.2 Sensor Data
228(3)
10.2.3 Other Types of External Data
231(2)
10.3 Total Survey Error
233(4)
10.3.1 Representation and Measurement of Smart Surveys
233(2)
10.3.2 Criteria to Include New Types of Data
235(2)
10.4 Methodology for Hybrid Data Collection
237(6)
10.4.1 Active Versus Passive Data Collection
237(1)
10.4.2 Data Collection Strategies
238(2)
10.4.3 Measurement Strategies
240(1)
10.4.4 Estimation Strategies
241(1)
10.4.5 Logistics and Operations
241(2)
10.5 Case Studies
243(5)
10.5.1 Smart Survey Criteria
243(3)
10.5.2 A Case Study Elaborated: Physical Activity
246(2)
10.6 Summary
248(1)
Note
249(2)
11 Adaptive Mixed-Mode Survey Designs
251(22)
11.1 Introduction
251(2)
11.2 Elements of Adaptive Multi-Mode Survey Designs
253(13)
11.2.1 Survey Mode as a Design Feature
254(2)
11.2.2 Population Strata
256(5)
11.2.3 Quality and Cost Objectives in Multi-Mode Surveys
261(4)
11.2.4 Optimization Strategies
265(1)
11.3 Case Studies
266(6)
11.3.1 The Dutch LFS
267(2)
11.3.2 The Dutch HS
269(3)
11.4 Summary
272(1)
12 The Future of Mixed-Mode Surveys
273(8)
12.1 Introduction
273(1)
12.2 What Are the Current Developments?
273(2)
12.3 What Are the Open Areas in the Methodology of Mixed-Mode Surveys
275(2)
12.4 Open Areas in Logistics and Implementation of Mixed-Mode Surveys
277(1)
12.5 Will There Be Mixed-Mode Surveys in the Future?
278(3)
References 281(20)
Index 301
Bart Buelens Bart Buelens has worked in data analytics after graduating in mathematics and obtaining a PhD in computer science. As statistician and data scientist at Statistics Netherlands he conducted research on inference in mixed-mode surveys, model-based estimation and machine learning. In 2018 he moved to VITO, a Belgian research and technology organization, where he contributes to data science research in the area of sustainability with emphasis on applied artificial intelligence.

Jan van den Brakel After finishing a Master in Biometrics, Jan van den Brakel started as a junior methodologist at the Methodology Department of Statistics Netherlands in 1994. Based on his research work at Statistics Netherlands on design-based inference methods for randomized experiments embedded in probability samples he finalized his PhD in Statistics in 2001. In 2005 he became senior methodologist, responsible for coordinating research into model based inference methods. Since 2010 he is an Extended professor of Survey Methodology at Maastricht University. His research interest are sampling, design and analysis of experiments, small area estimation, time series analysis and statistical methods for measuring the effects of redesigns of repeated surveys.

Deirdre Giesen Deirdre Giesen holds a master in health sciences and in sociology. She has worked as a survey methodologist at Statistics Netherlands since 2000. She is a senior methodologist and responsible for coordination of cognitive lab/userlab tests at Statistics Netherlands. Part of her work is pre-testing and evaluating questionnaires, both for businesses and household surveys. Her research interests include questionnaire testing methodology, mixed-mode mixed-device questionnaire development and the measurement and reduction of response burden.

Annemieke Luiten - Annemieke Luiten holds a Master in psychology and a PhD in survey methodology. She is a senior methodologist at Statistics Netherlands, with a specialty in data collection methodology. Her research areas comprise nonresponse reduction, fieldwork, interviewer behavior, mixed mode-surveys and the role of sensor measurement in surveys.

Barry Schouten - After a Master and PhD in Mathematics, Barry Schouten started as junior methodologist at the Methodology Department of Statistics Netherlands in 2002. In 2009, he became a senior methodologist and coordinator for research into primary data collection. His research interests gradually widened from nonresponse reduction and adjustment to multi-mode surveys, measurement error and adaptive survey design. In 2017, he became a professor at Utrecht University, holding a special chair on mixed-mode survey designs. He is one of the coordinators of a joint data collection innovation network (WIN in Dutch) between Statistics Netherlands and Utrecht University that was established in 2016.

Vivian Meertens - Vivian Meertens holds a Master and PhD in Sociology at the University Medical Centre Nijmegen and Interuniversity Centre for Social Theory and Methodology (ICS). After some years working as an associate professor at the department of Social Medical Science, she started as a survey methodologist at Statistics Netherlands in 2007. She has worked on pre-testing and evaluating of mixed mode and mixed device questionnaires for household surveys. Her research interest focuses on developments and pre-testing of several European model questionnaires measuring social phenomena to produce (inter) national social statistics.