Reasons for Non-Response to Mailed Questionnaires – Experience of a Large Cohort Study in Thailand

Kawin Thinkhamrop, Sam-ang Seubsman, Adrian Sleigh, Bandit Thinkhamrop, Thai Cohort Study Team

Abstract


Non-response is an important problem in survey research and should be minimized. Researchers always aim to maximize response rates to maintain representativeness of the sample with regard to the corresponding source population. In particular, it is an issue for health surveys that are conducted longitudinally. Objective: To determine the factors that are associated with non-response to mailed questionnaires in health research in Thailand. Methods: This study used the data collected for the Thai Health-Risk Transition: a National Cohort Study, Sukhothai Thammathirat Open University. It is a prospective cohort study for a period of 5 years where the baseline survey was conducted in 2005 and the follow-up survey was in 2009. Data were collected by mailed questionnaires. Factors determining non-response were investigated among a total of 85,217 respondents of the 2005 baseline survey who were then followed-up in 2009 using a 12 page questionnaire. Multiple logistic regression was used to determine factors that were a risk for or protected against non-response on follow-up in 2009. Results: Among a total of 85,217, there were 24,440 non-responses. The rate of non-response and its 95% confidence interval (CI) was 28.7 (28.4 - 29.0). Factors associated with protection against non-response are presented as odds ratio (OR) < 1 and 95%CI; they included being female (OR = 0.88; 95%CI: 0.85 - 0.91; p-value < 0.001), older age (compared to age ≤ 25, OR for age 26-39 = 0.56; 95%CI: 0.53 - 0.58, OR for age ≥ 40 = 0.27; 95%CI: 0.26 - 0.29; p-value < 0.001), higher educational level (compared to secondary school, OR of diploma or equivalent = 0.91; 95%CI: 0.87 - 0.94, OR of bachelor degree or higher = 0.68; 95%CI: 0.65 - 0.71; p-value < 0.001), and having children (OR = 0.82; 95%CI: 0.78 - 0.85; p-value < 0.001). Living in urban areas was a risk for non-response as well (OR = 1.16; 95%CI: 1.12 - 1.20; p-value < 0.001). Conclusion: To maximize the rate of response to mailed questionnaire, researchers should focus efforts on the target sample who were male, young of age (≤ 25), with low education, having no children, and living in urban areas.

Keyword: non-response, mailed questionnaire, health survey


Full Text:

Full Text in Thai

Refbacks

  • There are currently no refbacks.


ISSN : 1906-1137