Related Experiment Videos
Measurement errors stemming from nonrespondents present at in-person interviews
S L Edwards1, M L Slattery, K N Ma
1Department of Oncological Sciences, University of Utah Medical School, Salt Lake City, USA.
Annals of Epidemiology
|May 20, 1998
Summary
Third-party presence during interviews in epidemiological studies can affect data accuracy. Researchers should consider privacy measures to ensure reliable exposure data collection for studies like colon cancer research.
Area of Science:
- Epidemiology
- Biostatistics
- Public Health
Background:
- In-person interviews are common in epidemiological research.
- Third-party presence during interviews may introduce measurement error, particularly if related to case status.
Purpose of the Study:
- To assess the frequency of third-party presence during in-person interviews.
- To evaluate the impact of third-party presence on the reporting of exposure data in a colon cancer case-control study.
Main Methods:
- Case-control study design.
- Analysis of interview data including third-party presence and exposure variables.
- Statistical comparison of exposure reporting between interviews with and without third-party presence.
Main Results:
- Third parties were present in 28% of case interviews and 22% of control interviews.
- Men reported significantly lower alcohol consumption when a third party was present.
- Associations between colon cancer and alcohol/energy intake were not significantly altered by third-party presence, though energy intake showed a stronger association without third-party presence.
Conclusions:
- The presence of a third party can influence reported exposure data.
- Emphasis on privacy during data collection is crucial.
- Interviewers may require enhanced training to ensure privacy and data integrity.