Model-Based Estimation
The primary goal of survey sampling is the accurate estimation of totals, means, and ratios for characteristics of interest within a finite population. Rather than assuming that sample observations are realizations of random variables satisfying some model, it is standard to treat only the sample selection process itself as random. This is called randomization or design-based inference. Because they rely on averages taken across all possible samples and not on the sample actually drawn, design-based methods can sometimes produce misleading results. Model-based estimation, by contrast, is conditioned on the realized sample but requires more assumptions about the behavior of the characteristics of interest. Model-based methods can be used along with or as a substitute for design-based inference.
Let U denote the population of N elements. ...
Looks like you do not have access to this content.
Reader's Guide
Ethical Issues In Survey Research
Measurement - Interviewer
Measurement - Mode
Measurement - Questionnaire
Measurement - Respondent
Measurement - Miscellaneous
Nonresponse - Item-Level
Nonresponse - Outcome Codes And Rates
Nonresponse - Unit-Level
Operations - General
Operations - In-Person Surveys
Operations - Interviewer-Administered Surveys
Operations - Mall Surveys
Operations - Telephone Surveys
Political And Election Polling
Public Opinion
Sampling, Coverage, And Weighting
Survey Industry
Survey Statistics
- All
- A
- B
- C
- D
- E
- F
- G
- H
- I
- J
- K
- L
- M
- N
- O
- P
- Q
- R
- S
- T
- U
- V
- W
- X
- Y
- Z