|
Size: 1685
Comment: Killing SurveyStatistics page
|
← Revision 4 as of 2025-01-10 15:52:51 ⇥
Size: 1739
Comment: Killing SurveyStatistics page
|
| Deletions are marked like this. | Additions are marked like this. |
| Line 1: | Line 1: |
| ## page was renamed from SurveyStatistics/ResponseRates = Response Rates = |
= Response Rate = The quality of a survey instrument, and resulting survey measurements, can be partially described by the '''response rate'''. |
| Line 18: | Line 19: |
| The predominant formulas for response rates come from AAPOR. | Response rate calculations are standardized AAPOR. |
| Line 32: | Line 33: |
| A response rate's calculation largely flows from the determination of [[SurveyStatistics/Dispositions|case dispositions]]. In any survey, there is an unclear distinction between a partial interview and a refusal. Many surveys ''furthermore'' allow questions to be implicitly skipped, explicitly refused, or include a 'Prefer not to answer' option. It becomes unclear what distinction functionally exists between complete interviews, partial interviews, and refusals. | A response rate's calculation largely flows from the determination of [[Statistics/Disposition|case dispositions]]. In any survey, there is an unclear distinction between a partial interview and a refusal. Many surveys ''furthermore'' allow questions to be implicitly skipped, explicitly refused, or include a 'Prefer not to answer' option. It becomes unclear what distinction functionally exists between complete interviews, partial interviews, and refusals. |
Response Rate
The quality of a survey instrument, and resulting survey measurements, can be partially described by the response rate.
Contents
Description
Calculation
Response rate calculations are standardized AAPOR.
Response Rate 1 (RR1) is a simple rate of complete interviews (I) over all other types of presumed eligible cases: partial interviews (P), refusals (R), non-contacts (NC), others (O, a catch-all category for types of non-response that aren't necessarily a 'refusal'), and cases with unknown eligibility (UNK).
RR1 is functionally an underestimate, because we expect some number of cases with unknown eligibility to actually be ineligible. This sets up Response Rate 3 (RR3), which incorporates an estimate of e, the eligibility rate among those unknown cases.
A common method for estimating e comes from CASRO: proportional allocation of cases with unknown eligibility according to those with known eligiblity. Adapting the above notation, and noting that ineligible cases are represented by IN, this is calculated as:
A response rate's calculation largely flows from the determination of case dispositions. In any survey, there is an unclear distinction between a partial interview and a refusal. Many surveys furthermore allow questions to be implicitly skipped, explicitly refused, or include a 'Prefer not to answer' option. It becomes unclear what distinction functionally exists between complete interviews, partial interviews, and refusals.