Efnisyfirlit

• Front Matter
• PREFACE TO THE FIRST EDITION
• PREFACE TO THE SECOND EDITION
• ACKNOWLEDGMENTS
• CHAPTER ONE AN INTRODUCTION TO SURVEY METHODOLOGY
• A Note to the Reader
• 1.1 Introduction
• 1.2 A Brief History of Survey Research
• 1.2.1 The Purposes of Surveys
• Schuman (1997) on “Poll” Versus “Survey”
• 1.2.2 The Development of Standardized Questioning
• 1.2.3 The Development of Sampling Methods
• 1.2.4 The Development of Data Collection Methods
• 1.3 Some Examples of Ongoing Surveys
• 1.3.1 The National Crime Victimization Survey
• Table 1.1. Example Survey: National Crime Victimization Survey (NCVS)
• Figure 1.1 Percentage of U.S. households experiencing a crime by type, 1994-2005 National Crime Victimization Survey.
• 1.3.2 The National Survey on Drug Use and Health
• Table 1.2. Example Survey: National Survey of Drug Use and Health (NSDUH)
• Figure 1.2 Percentage of persons reporting illicit drug use in past month, by drug type, 2004-2006
• 1.3.3 The Surveys of Consumers
• Table 1.3. Example Survey: Surveys of Consumers (SOC)
• Figure 1.3 Consumer unemployment expectations and actual change in the U.S. unemployment rate, 1969-2009
• 1.3.4 The National Assessment of Educational Progress
• Table 1.4. Example Survey: National Assessment of Educational Progress (NAEP)
• Figure 1.4 Average scale scores on grade 12 mathematics assessment, by year by type of school.
• 1.3.5 The Behavioral Risk Factor Surveillance System
• Table 1.5. Example Survey: Behavioral Risk Factor Surveillance System (BRFSS)
• Figure 1.5a Percentage of state adults who are obese (body mass index ≥ 30) by state, 1994, BRFSS.
• Figure 1.5b Percentage of state adults who are obese (body mass index ≥ 30) by state, 2001, BRFSS.
• Figure 1.5c Percentage of state adults who are obese (body mass index ≥ 30) by state, 2007, BRFSS.
• 1.3.6 The Current Employment Statistics Program
• Table 1.6. Example Survey: Current Employment Statistics (CES)
• Figure 1.6 Number of employees of all nonfarm employers in thousands, annual estimates 1947-2007, Current Employment Statistics.
• 1.3.7 What Can We Learn From the Six Example Surveys?
• 1.4 What is Survey Methodology?
• 1.5 The Challenge of Survey Methodology
• 1.6 About this Book
• Keywords
• For More In-Depth Reading
• National Crime Victimization Survey
• National Survey of Drug Use and Health
• Surveys of Consumers
• National Assessment of Educational Progress
• Behavioral Risk Factor Surveillance System
• Current Employment Statistics
• Exercises
• CHAPTER TWO INFERENCE AND ERROR IN SURVEYS
• 2.1 Introduction
• Figure 2.1 Two types of survey inference.
• 2.2 The Life Cycle of a Survey From a Design Perspective
• Figure 2.2 Survey lifecycle from a design perspective.
• 2.2.1 Constructs
• 2.2.2 Measurement
• 2.2.3 Response
• 2.2.4 Edited Response
• 2.2.5 The Target Population
• 2.2.6 The Frame Population
• Illustration—Populations of Inference and Target Populations
• 2.2.7 The Sample
• 2.2.8 The Respondents
• Figure 2.3 Unit and item nonresponse in a survey data file.
• 2.2.9 Postsurvey Adjustments
• 2.2.10 How Design Becomes Process
• Figure 2.4 A survey from a process perspective.
• 2.3 The Life Cycle of a Survey from A Quality Perspective
• Figure 2.5 Survey life cycle from a quality perspective.
• 2.3.1 The Observational Gap between Constructs and Measures
• The Notion of Trials
• 2.3.2 Measurement Error: The Observational Gap between the Ideal Measurement and the Response Obtained
• The Notion of Variance or Variable Errors
• 2.3.3 Processing Error: The Observational Gap between the Variable Used in Estimation and that Provided by the Respondent
• 2.3.4 Coverage Error: The Nonobservational Gap between the Target Population and the Sampling Frame
• Figure 2.6 Coverage of a target population by a frame.
• 2.3.5 Sampling Error: The Nonobservational Gap between the Sampling Frame and the Sample
• Figure 2.7 Samples and the sampling distribution of the mean.
• 2.3.6 Nonresponse Error: The Nonobservational Gap between the Sample and the Respondent Pool
• 2.3.7 Adjustment Error
• 2.4 Putting It All Together
• 2.5 Error Notions in Different Kinds of Statistics
• 2.6 Nonstatistical Notions of Survey Quality
• 2.7 Summary
• Keywords
• For More In-Depth Reading
• Exercises
• CHAPTER THREE TARGET POPULATIONS, SAMPLING FRAMES, AND COVERAGE ERROR
• 3.1 Introduction
• 3.2 Populations and Frames
• 3.3 Coverage Properties of Sampling Frames
• 3.3.1 Undercoverage
• Mulry (2007) on U.S. Decennial Census Coverage
• 3.3.2 Ineligible Units
• 3.3.3 Clustering of Target Population Elements Within Frame Elements
• Figure 3.1 Cluster of target population elements associated with one sampling frame element.
• 3.3.4 Duplication of Target Population Elements in Sampling Frames
• Figure 3.2 Duplication of target population elements by more than one sampling frame element.
• 3.3.5 Complicated Mappings between Frame and Target Population Elements
• Figure 3.3 Clustering and duplication of target population elements relative to sampling frame elements.
• 3.4 Alternative Frames for Surveys of the Target Population of Households or Persons
• 3.4.1 Area Frames
• 3.4.2 Telephone Number Frames for Households and Persons
• Figure 3.4 Percentage of U.S. adults with wireless telephone service only and percentage without telephones, January, 2004-June, 2008.
• 3.4.3 Frames for Web Surveys of General Populations
• 3.5 Frame Issues for Other Common Target Populations
• 3.5.1 Customers, Employees, or Members of an Organization
• 3.5.2 Organizations
• 3.5.3 Events
• 3.5.4 Rare Populations
• 3.6 Coverage Error
• 3.7 Reducing Undercoverage
• 3.7.1 The Half-Open Interval
• Figure 3.5 Address list for area household survey block.
• Figure 3.6 Sketch map for area household survey block.
• 3.7.2 Multiplicity Sampling
• 3.7.3 Multiple Frame Designs
• Figure 3.7 Dual frame sample design.
• 3.7.4 Increasing Coverage While Including More Ineligible Elements
• Tourangeau, Shapiro, Kearney, and Ernst (1997) and Martin (1999) on Household Rosters
• 3.8 Summary
• Keywords
• For More in-Depth Reading
• Exercises
• CHAPTER FOUR SAMPLE DESIGN AND SAMPLING ERROR
• 4.1 Introduction
• 4.2 Samples and Estimates
• Figure 4.1 Unknown distribution for variable Y in frame population.
• Figure 4.2 Distributions of y variable from sample realizations samples and the sampling distribution of the mean.
• Figure 4.3 Key notation for sample realization, frame population, and sampling distribution of sample means.
• Warning
• 4.3 Simple Random Sampling
• Comment
• Comment
• 4.4 Cluster Sampling
• Figure 4.4 A bird’s-eye view of a population of 30 “” and 30 “” households clustered into six city blocks, from which two blocks are selected.
• Comment
• 4.4.1 The Design Effect and Within-Cluster Homogeneity
• Table 4.1. Mean roh Values for Area Probability Surveys about Female Fertility Experiences in Five Countries by Type of Variable
• Kish and Frankel (1974) on Design Effects for Regression Coefficients
• 4.4.2 Subsampling within Selected Clusters
• 4.5 Stratification and Stratified Sampling
• Figure 4.5 Frame population of 20 establishments sorted alphabetically, with SRS sample realization of size n = 4.
• Figure 4.6 Frame population of 20 establishments sorted by group, with stratified element sample of size nh = 1 from each stratum.
• 4.5.1 Proportionate Allocation to Strata
• Table 4.2. Proportionate Stratified Random Sample Results from a School Population Divided Into Three Urbanicity Strata
• Cochran (1961) on How Many Strata to Use
• Design Effects for the Stratified Mean
• 4.5.2 Disproportionate Allocation to Strata
• Neyman (1934) on Stratified Random Sampling
• 4.6 Systematic Selection
• Figure 4.7 Frame population of 20 establishments sorted by group with systematic selection; selection interval = 5 and random start = 2.
• Comment
• 4.7 Complications in Practice
• 4.7.1 Two-Stage Cluster Designs with Probabilities Proportionate to Size (PPS)
• Table 4.3. Block Housing Unit Counts and Cumulative Counts for a Population of Nine Blocks
• 4.7.2 Multistage and Other Complex Designs
• 4.7.3 How Complex Sample Designs Are Described: The Sample Design for the NCVS
• 4.8 Sampling U.S. Telephone Households
• Figure 4.8 Number of 100 blocks of number by number listed within the block, 1986 and 2008. (Source: Survey Sampling, Inc.)
• 4.9 Selecting Persons within Households
• 4.10 Summary
• Keywords
• For More In-Depth Reading
• Exercises
• CHAPTER FIVE METHODS OF DATA COLLECTION
• Figure 5.1 A survey from a process perspective.
• 5.1 Alternative Methods of Data Collection
• Figure 5.2 The evolution of survey technology.
• 5.1.1 Degree of Interviewer Involvement
• 5.1.2 Degree of Interaction with the Respondent
• 5.1.3 Degree of Privacy
• 5.1.4 Channels of Communication
• 5.1.5 Technology Use
• 5.1.6 Implications of these Dimensions
• 5.2 Choosing the Appropriate Method
• 5.3 Effects of Different Data Collection Methods on Survey Errors
• 5.3.1 Measuring the Marginal Effect of Mode
• Table 5.1. Design Issues in Research Comparing Face-to-Face and Telephone Surveys
• Hochstim (1967) on Personal Interviews Versus Telephone Interviews Versus Mail Questionnaires
• The de Leeuw and van der Zouwen (1998) Meta-Analysis of Data Quality in Telephone and Face-to-Face Surveys
• 5.3.2 Sampling Frame and Sample Design Implications of Mode Selection
• 5.3.3 Coverage Implications of Mode Selection
• Figure 5.3 Percentage of U.S. adults who ever use the Internet, quarter 1, 2000, to quarter 3, 2008.
• 5.3.4 Nonresponse Implications of Mode Selection
• 5.3.5 Measurement Quality Implications of Mode Selection
• The Tourangeau and Smith (1996) Study of Mode Effects on Answers to Sensitive Questions
• Figure 5.4 Ratio of proportion of respondents reporting illicit drug use in self-administered versus interviewer-administered questionnaires, by time period by drug.
• 5.3.6 Cost Implications
• 5.3.7 Summary on the Choice of Method
• 5.4 Using Multiple Modes of Data Collection
• Figure 5.5 Five different types of mixed mode designs.
• 5.5 Summary
• Keywords
• For More In-Depth Reading
• Exercises
• CHAPTER SIX NONRESPONSE IN SAMPLE SURVEYS
• 6.1 Introduction
• 6.2 Response Rates
• 6.2.1 Computing Response Rates
• Merkle and Edelman (2002) on How Nonresponse Rates Affect Nonresponse Error
• 6.2.2 Trends in Response Rates Over Time
• Figure 6.1 Household nonresponse rate, household refusal rate, and person refusal rate for the National Crime Victimization Survey by year.
• Figure 6.2 Nonresponse and refusal rates for the Current Population Survey by year.
• Figure 6.3 Nonresponse rate and refusal rate for the Survey of Consumers by year.
• Figure 6.4 Median nonresponse rate across states, Behavioral Risk Factor Surveillance System, 1987-2007.
• 6.3 Impact of Nonresponse on the Quality of Survey Estimates
• Figure 6.5 Estimates of the absolute value of the relative nonresponse bias for 959 estimates by nonresponse rate of survey.
• 6.4 Thinking Causally About Survey Nonresponse Error
• Figure 6.6 Alternative models for relationship between response propensity (P) and survey variable (Y), involving auxiliary variables (S, Z).
• 6.5 Dissecting The Nonresponse Phenomenon
• 6.5.1 Unit Nonresponse Due to Failure to Deliver the Survey Request
• Figure 6.7 Causal influences on contact with sample household.
• Figure 6.8 Percentage of eligible sample households by calls to first contact for five surveys.
• Figure 6.9 Percentage household contacted among those previously uncontacted by call number by time of day. (National Survey of Family Growth, Cycle 6.)
• Figure 6.10 Percentage nonresponse bias for estimated proportion of single person households, by number of calls required to reach the house hold, for four surveys.
• 6.5.2 Unit Nonresponse Due to Refusals
• The “I’m Not Selling Anything” Phenomenon
• Figure 6.11 Two sample persons with different leverages for attributes of a survey request.
• What Interviewers Say
• 6.5.3 Unit Nonresponse Due to the Inability to Provide the Requested Data
• 6.6 Design Features to Reduce Unit Nonresponse
• Figure 6.12 Tools for reducing unit nonresponse rates.
• What Interviewers Say about Approaching Sample Households
• Berlin, Mohadjer. Waksberg, Kolstad, Kirsch, Rock, and Yamamoto (1992) on Incentives and Interviewer Productivity
• Morton-Williams (1993) on Tailoring Behavior by Interviewers
• 6.7 Item Nonresponse
• Figure 6.13 Beatty-Herrmann model of response process for item-missing data.
• 6.8 Are Nonresponse Propensities Related To Other Error Sources?
• 6.9 Summary
• Keywords
• For More In-Depth Reading
• Exercises
• CHAPTER SEVEN QUESTIONS AND ANSWERS IN SURVEYS
• 7.1 Alternatives Methods of Survey Measurement
• 7.2 Cognitive Processes in Answering Questions
• Figure 7.1 A simple model of the survey response process.
• 7.2.1 Comprehension
• 7.2.2 Retrieval
• 7.2.3 Estimation and Judgment
• 7.2.4 Reporting
• 7.2.5 Other Models of the Response Process
• Comments on Response Strategies
• 7.3 Problems in Answering Survey Questions
• 7.3.1 Encoding Problems
• Fowler (1992) on Unclear Terms in Questions
• 7.3.2 Misinterpreting the Questions
• 7.3.3 Forgetting and Other Memory Problems
• Figure 7.2 Recall accuracy for types of personal information.
• Table 7.1. Summary of Factors Affecting Recall
• Neter and Waksberg (1964) on Response Errors
• 7.3.4 Estimation Processes for Behavioral Questions
• Overreporting and Underreporting
• Schwarz, Hippler, Deutsch, and Strack (1985) on Response Scale Effects
• 7.3.5 Judgment Processes for Attitude Questions
• 7.3.6 Formatting the Answer
• 7.3.7 Motivated Misreporting
• 7.3.8 Navigational Errors
• Figure 7.3 Example questions from Jenkins and Dillman (1997).
• 7.4 Guidelines for Writing Good Questions
• 7.4.1 Nonsensitive Questions About Behavior
• 7.4.2 Sensitive Questions About Behavior
• 7.4.3 Attitude Questions
• 7.4.4 Self-Administered Questions
• Figure 7.4 Illustration of use of visual contrast to highlight the response box.
• 7.5 Summary
• Keywords
• For More In-Depth Reading
• Exercises
• CHAPTER EIGHT EVALUATING SURVEY QUESTIONS
• 8.1 Introduction
• 8.2 Expert Reviews
• 8.3 Focus Groups
• 8.4 Cognitive Interviews
• Presser and Blair (1994) on Alternative Pretesting Methods
• 8.5 Field Pretests and Behavior Coding
• Table 8.1. Examples of Behavior Codes for Interviewer and Respondent Behaviors
• 8.6 Randomized or Split-Ballot Experiments
• Oksenberg, Cannell, and Kalton (1991) on Probes and Behavior Coding
• Percent of problems per question
• 8.7 Applying Question Standards
• 8.8 Summary of Question Evaluation Tools
• Table 8.2. Studies Comparing Question Evaluation Methods
• 8.9 Linking Concepts of Measurement Quality to Statistical Estimates
• 8.9.1 Validity
• Estimating Validity with Data External to the Survey.
• Estimating Validity with Multiple Indicators of the Same Construct.
• Figure 8.1 Path diagram representing Yαi = λαμi + ɛαi, a measurement model for μi.
• 8.9.2 Response Bias
• Using Data on Individual Target Population Elements.
• Table 8.3. Percentage of Known Hospitalizations Not Reported, by Length of Stay and Time Since Discharge
• Using Population Statistics Not Subject to Survey Response Error.
• 8.9.3 Reliability and Simple Response Variance
• Repeated Interviews with the Same Respondent.
• Table 8.4. Indexes of Inconsistency for Various VictimizationIncident Characteristics, NCVS
• Using Multiple Indicators of the Same Construct.
• Table 8.5. Illustrative Intercorrelations among MHI-5 Items
• O’Muircheartaigh (1991) on Reinterviews to Estimate Simple Response Variance
• 8.10 Summary
• Keywords
• For More In-Depth Reading
• Exercises
• CHAPTER NINE SURVEY INTERVIEWING
• 9.1 The Role of the Interviewer
• 9.2 Interviewer Bias
• 9.2.1 Systematic Interviewer Effects on Reporting of Socially Undesirable Attributes
• 9.2.2 Systematic Interviewer Effects on Topics Related to Observable Interviewer Traits
• Schuman and Converse (1971) on Race of Interviewer Effects in the United States
• Percentage Answering in Given Category by Race of Interviewer
• 9.2.3 Systematic Interviewer Effects Associated with Interviewer Experience
• Table 9.1. Percentage Reporting Lifetime Use of Any Illicit Substance by Interview Order by Interviewer Experience (1998 NSDUH)
• 9.3 Interviewer Variance
• 9.3.1 Randomization Requirements for Estimating Interviewer Variance
• 9.3.2 Estimation of Interviewer Variance
• Kish (1962) on Interviewer Variance
• Statistics Showing High Values of ρint
• 9.4 Strategies for Reducing Interviewer Bias
• 9.4.1 The Role of the Interviewer in Motivating Respondent Behavior
• 9.4.2 Changing Interviewer Behavior
• 9.5 Strategies for Reducing Interviewer-Related Variance
• 9.5.1 Minimizing Questions that Require Nonstandard Interviewer Behavior
• 9.5.2 Professional, Task-Oriented Interviewer Behavior
• 9.5.3 Interviewers Reading Questions as They Are Worded
• 9.5.4 Interviewers Explaining the Survey Process to the Respondent
• 9.5.5 Interviewers Probing Nondirectively
• 9.5.6 Interviewers Recording Answers Exactly as Given
• 9.5.7 Summary on Strategies to Reduce Interviewer Variance
• 9.6 The Controversy About Standardized Interviewing
• 9.7 Interviewer Management
• 9.7.1 Interviewer Selection
• Conrad and Schober (2000) on Standardized versus Conversational Interviewing Techniques
• 9.7.2 Interviewer Training
• Table 9.2. Percentage of Interviewers Rated Excellent or Satisfactory for Six Criteria by Length of Interviewer Training
• 9.7.3 Interviewer Supervision and Monitoring
• 9.7.4 The Size of Interviewer Workloads
• 9.7.5 Interviewers and Computer Use
• 9.8 Validating The Work of Interviewers
• Table 9.3. Percentage of Interviewers Detecting Falsifying for Three Surveys Conducted by the U.S. Bureau of the Census
• 9.9 The Use Of Recorded Voices (And Faces) In Data Collection
• 9.10 Summary
• Keywords
• For More In-Depth Reading
• Exercises
• CHAPTER TEN POSTCOLLECTION PROCESSING OF SURVEY DATA
• 10.1 Introduction
• Figure 10.1 Flow of processing steps in paper surveys.
• Figure 10.2 Flow of processing steps in computer-assisted surveys.
• 10.2 Coding
• 10.2.1 Practical Issues of Coding
• 10.2.2 Theoretical Issues in Coding Activities
• Figure 10.3 Comprehension and judgment task of the coder.
• 10.2.3 “Field Coding”—An Intermediate Design
• Table 10.1. Illustration of Field Coding in the NCVS for the Question “Where Did This Incident Happen?”
• Collins and Courtenay (1985) on Field versus Office Coding
• 10.2.4 Standard Classification Systems
• The Standard Occupational Classification (SOC).
• Table 10.2. 23 Group Standard Occupational Classification
• The North American Industry Classification System (NAICS).
• Table 10.3. Comparison of SIC Divisions and NAICS Sectors
• Table 10.4. NAICS Structure and Nomenclature
• 10.2.5 Other Common Coding Systems
• 10.2.6 Quality Indicators in Coding
• Weaknesses in the Coding Structure.
• Coder Variance.
• Table 10.5. Coder Variance Statistics for Occupation Coding
• 10.2.7 Summary of Coding
• 10.3 Entering Numeric Data into Files
• 10.4 Editing
• Summary of Editing.
• 10.5 Weighting
• 10.5.1 Weighting with a First-Stage Ratio Adjustment
• 10.5.2 Weighting for Differential Selection Probabilities
• 10.5.3 Weighting to Adjust for Unit Nonresponse
• Table 10.6. Hypothetical Equal Allocation for Latinos, with Nonresponse Adjustments
• 10.5.4 Poststratification Weighting
• Table 10.7. Weighted Sample Distribution and Poststratification for Hypothetical NCVS Sample by Gender, Age, and Ethnicity
• 10.5.5 Putting All the Weights Together
• Ekholm and Laaksonen (1991) on Propensity Model Weighting to Adjust for Unit Nonresponse
• 10.6 Imputation for Item-Missing data
• Figure 10.4 Unit and item nonresponse in a survey data file.
• Table 10.8. Illustration of Sequential Hot-Deck Imputation for Family Income, Imputed Data, and Imputation Flag Variable
• 10.7 Sampling Variance Estimation for Complex Samples
• Taylor Series Estimation.
• Balanced Repeated Replication and Jackknife Replication.
• 10.8 Survey Data Documentation and Metadata
• Figure 10.5 Illustration of printable codebook section for the National Crime Victimization Survey.
• 10.9 Summary
• Keywords
• For More In-Depth Reading
• Exercises
• CHAPTER ELEVEN PRINCIPLES AND PRACTICES RELATED TO ETHICAL RESEARCH
• 11.1 Introduction
• 11.2 Standards for the Conduct of Research
• Table 11.1. Key Terminology in Research Misconduct
• Table 11.2. Percentage of Interviewers Detected Falsifying for Three Surveys Conducted by the U.S. Bureau of the Census
• 11.3 Standards for Dealing with Clients
• 11.4 Standards for Dealing with the Public
• Table 11.3. Elements of Minimal Disclosure (AAPOR Code)
• 11.5 Standards for Dealing with Respondents
• 11.5.1 Legal Obligations to Survey Respondents
• The Tuskegee Study of Syphilis
• 11.5.2 Ethical Obligations to Respondents
• 11.5.3 Informed Consent: Respect for Persons
• Table 11.4. Essential Elements of Informed Consent
• Project Metropolitan
• 11.5.4 Beneficence: Protecting Respondents from Harm
• 11.5.5 Efforts at Persuasion
• 11.6 Emerging Ethical Issues
• 11.7 Research About Ethical Issues In Surveys
• 11.7.1 Research on Informed Consent Protocols
• Research on Respondents’ Reactions to Informed Consent Protocols.
• Table 11.5. Self-Reports on Type of Questions Considered Offensive to the Respondent Among Respondents Saying Researchers “Had No Business Asking” Sensitive Questions
• Singer (1978) on Comprehension of Informed Consent
• Research on Informed Consent Complications in Methodological Studies.
• Research on Written versus Oral Informed Consent.
• Summary of Research on Informed Consent in Surveys.
• 11.7.2 Research on Confidentiality Assurances and Survey Participation
• 11.8 Administrative and Technical Procedures for Safeguarding Confidentiality
• 11.8.1 Administrative Procedures
• Figure 11.1 Pledge made by research team members about respondent privacy.
• Table 11.6. Principles and Practices for Protection of Sensitive Data
• 11.8.2 Technical Procedures
• Restricting Access to the Data.
• Restricting the Contents of the Survey Data That May Be Released.
• 11.9 Summary and Conclusions
• Keywords
• For More In-Depth Reading
• Exercises
• CHAPTER TWELVE FAQS ABOUT SURVEY METHODOLOGY
• 12.1 Introduction
• 12.2 The Questions and Their Answers
• Back Matter
• REFERENCES