Congress logo
Back to Conference page

The 13th World Conference on Tobacco OR Health

Building capacity for a tobacco-free world

July 12-15, 2006, Washington, DC, USA

Thursday, July 13, 2006 - 12:00 PM

The extent of unreliability of smoking data on adolescents obtained from personal interview

Ting Yuan David Cheng, MS, Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, 615 N. Wolfe Street, Baltimore, MD 21215, Chi Pang Wen, MD, DrPH, Health Policy, National Health Research Institutes, Taiwan, 35 Keyan Rd., Zhunan Town, Maioli County, Taiwan, and Hui-Ting Chan, MS, Division of Health Policy Research, National Health Research Institutes, Taiwan, 35 Keyan Rd., Zhunan Town, Maioli County, Taiwan.

Objective: Smoking by adolescent youth is a major issue in tobacco control, and validity of collected data is essential, and yet, data collected by different methods were used interchangeably in different studies, as if they had the same validity. The smoking status by personal interview was compared with those from written questionnaire.

Methods: Data from National Health Interview Survey of Taiwan in 2001 came from nationally representative samples with 94% response rate. A total of 2,540 adolescents aged 1217 were interviewed and they also simultaneously completed questionnaires in writing, with confidentiality assured. Smoking definitions came from those by WHO.

Results: Data from personal interviews on girls were far less reliable than on boys: While there were 3 times more smokers identified from written survey than from personal interview for girls (6.6% versus 2.2%), only 2 times higher for boys (25.9% versus 13.3%). The magnitude of the discrepancy by ex-smokers and occasional smokers was excessively large, between 17-47 folds (13.5% versus 0.8% in boys; 9.3% versus 0.2% in girls) and 5-7 folds (9.0% versus 1.7% in boys; 2.6% versus 0.4% in girls) difference, respectively.

Conclusion: Smoking data obtained from personal interview was seriously biased and unreliable. The values on ex-smokers were by the least reliable, followed by data on occasional smokers. Girls were less reliable than boys. Only one half to one third of smokers could be identified through interview process. Adolescents covered up smoking history, an act initiated out of the concern that smoking by youth is socially unacceptable.