Coverage error (undercoverage), When members of the target population have zero chance of being selected because they are not on the sampling frame or cannot be reached by the chosen mode., Selection Bias, Systematic difference between who is sampled and the target population, often due to non-probability sampling, differential access to the survey mode, non-response error, etc.., Nonresponse Bias, When people who do not respond to the survey are systematically different from those who do on key demographic variables., Measurement Error, Difference between a respondent’s true value and their recorded answer, caused by question wording, mode, or interviewer., Social Desirability Bias, Tendency to give answers that are socially acceptable rather than truthful, stronger in interviewer-administered modes., Interviewer Effects, Systematic influence of the interviewer’s presence, identity, or behaviour on respondent answers., Satisficing, When respondents do not carefully consider the answers—e.g., straight-lining, choosing the first acceptable option., Penetration rate, Mobile phone subscriptions per 100 inhabitants., Coverage Rate, Percentage of individuals who own at least one mobile phone. The survey-relevant measure., Proxy Coverage, People who do not own a phone but live with someone who does, making them potentially reachable through shared devices , Sequential Mixed-Mode, Modes offered one after another (e.g., web first, then mail follow-up for nonrespondents)., Concurrent Mixed-Mode, Different modes offered simultaneously, or different modes used for different subpopulations, Total Survey Error , A framework considering all sources of error in surveys: measurement, sampling, nonresponse errors, etc, Validity, Whether the survey measures what it claims to measure., Mode Effect, A difference in results caused by the survey collection method, Mode selection effect, Different modes attract different types of respondents , Mode measurement effect, Same person answers differently depending on the mode (undesired), Push-to-web, Paper letter directing respondents to complete an online survey, Reliability, Consistency of results when the survey is repeated, Sampling Frame, A list or database from which the sample is drawn, Recall bias, Errors from asking respondents to remember distant past events, Double-barrelled questions, A single question that asks about two or more different things, CATI, Computer-Assisted Telephone Interviewing, Construct validity, Whether the survey measures the concept it claims to measure.

Tutorial 6. Concepts

by
Print
Embed

Leaderboard

Visual style

Options

Switch template

)
Continue editing: ?