Week 3 Flashcards
Chapters 6, 7 & 14
Coverage error
Sampling frame does not cover all population
Sampling error
Estimates are based on samples, not the intended population
Non-response error
Respondents differ from non-respondents in a systematic way
Adjustment error
The difference between the population and the adjusted data (weighting, scaling)
Validity
The accuracy of the measure
Reliability
The consistency of the measure
Measurement error
The used instrument does not measure what was intended to measure
Processing error
Error connected to the processing of survey data, such as data entry, handling of questionnaires
The Tourangeau model
The process of reading the question and answering the question, Four stages: Interpretation, retrieval, judgement, reporting
The Tourangeau model, stage 1
Interpretation, respondents interprets the question. Affected by: the clarity of the question, previous questions and the respondent characteristics
The Tourangeau model, stage 2
Retrieval, retrieving relevant information from memory. Affected by: ordering, priming, framing
The Tourangeau model, stage 3
Judgement, how to use the retrieved information for this question
The Tourangeau model, stage 4
Reporting, a compromise between the internal answer and the response categories
Types of measurement errors:
Type I: Misreporting
Type II: Loss in precision
Type III: Cognitive shortcuts (don’t know, answer not there, straight lining)
Potential interviewer effects
Response rates - rates increase with more experience
Social desirability - interviewers gender, age and race can influence the responses
Longitudinal surveys
Surveys where the same survey is distributed at multiple times. Either with the same respondents or new respondents (cross-sectional)
Cross-sectional surveys
Collecting data from a lot of individuals for a single point in time
Consequences of longitudinal surveys
Attrition bias, panel conditioning
Attrition bias
Individuals who fall out of the panel who significantly differ from the rest of the panel
Panel conditioning
Learning effects because respondents fill out the survey multiple times
Mixed-mode designs
Using different research methods, concurrent mixed-mode designs and sequential mixed-mode designs
Concurrent mixed-mode designs
Different modes for data collection used simultaneously
Sequential mixed-mode design
Different modes for data collection are used at different stages or one main mode
Heuristic rule for compliance
Participation in the survey is based on personal considerations.
Social exchange theory
Individuals feel a sense of obligation to fill out the survey because of the money they received.
Economic exchange theory
Respondents choose to participate in the survey based on a rational cost-benefit calculation.
Prepaid/unconditional incentives
Given to all individuals invited to fill out the questionnaire, regardless if they choose to participate. More powerful than postpaid (see theories)
Postpaid/conditional incentives
Given to only the participants who fulfilled the survey
Consequences of the use of mixed-mode designs
Social desirability
Acquiescence - tendency to agree
Norm of even-handedness - take order of questions into account
Primacy/recency effect - tendency to choose the first or the last answer
Priming
Previous questions/information influence the answer of the respondent
Framing
The way a question is asked, information given on the subject
