Impact of Mailouts on Survey Response Bias in Iraq War Study
Author Information
Author(s): A R Tate, M Jones, L Hull, N T Fear, R Rona, S Wessely, M Hotopf
Primary Institution: King's College London
Hypothesis
Does increasing response rates through multiple mailouts lead to greater misclassification errors in survey results?
Conclusion
Increasing response rates had little effect on the results, suggesting that bias from reporting errors could be more significant than bias from nonresponse.
Supporting Evidence
- The response rate to the survey was 60%.
- 88% of participants responded after one reminder.
- Misclassification errors increased with the number of contact attempts.
Takeaway
The study looked at whether sending more mail to get people to answer a survey made the answers less accurate. It found that trying harder to get responses didn't really help.
Methodology
Data from 17,162 UK military personnel were analyzed to assess response rates and misclassification errors using simulations.
Potential Biases
Potential for misclassification bias due to carelessness in answering questions.
Limitations
The study may not account for all forms of bias, particularly differential misclassification.
Participant Demographics
UK military personnel, including regulars and reservists, with varied age, rank, and service type.
Statistical Information
P-Value
p<0.001
Confidence Interval
95% CI: 1.07, 1.34
Statistical Significance
p<0.001
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website