Page 102 - TAIWAN TOBACCO CONTROL ANNUAL REPORT 2019

Research and Monitoring Adult Smoking Behavior Survey The HPA regularly implements smoking behavior monitoring surveys for the entire population or targeted age groups required for promoting relevant measures or generate reference for the policies. When compared against interview surveys, telephone surveys allowed the HPA to quickly acquire preliminary and summary referential information within the shortest time possible. Data collected from the telephone interviews could also be used to investigated changes and trends to health-related issues and quickly assess smoking behaviors and awareness of tobacco controls of the general public. In order to understand the current status and trend of smoking behaviors amongst the public throughout Taiwan and in every county and city and acquire data for monitoring and evaluating the performance of tobacco control measures by each health bureau, the HPA began monitoring smoking behaviors of individuals aged 18 years or more via representative sampling in various counties and cities since 2004. To ensure that the collected data could be compared against global standards, the HPA expanded the scope of the survey to include Taiwanese people aged 15 years or above since 2013. The project title was also changed to "Adult Smoking Behavior Survey". This Survey would regularly monitor smoking behaviors of adult population on an annual basis and conduct statistical analysis by counties and cities throughout Taiwan. 16,000 to 26,000 individuals were visited nationwide and the sample size was expanded to over 1,068 for each city or country (excluding Lienchiang County) in 2013. In 2018, a total of 25,992 were called (25,535 of these interviewees were aged 18 or above). The comparison of the counties and cities with past trends, the HPA not only compared trends over the past years but also obtained representative samples in each county and city. Starting in 2017, a two-stage sampling method was adopted to increase the sample coverage rate and improve unequal probability sampling of phone surveys. In the first stage, phone numbers are selected by adopting the following 4 different sampling methods: (1) Selection of numbers from the residential phone directory through systematic internal sampling. (2) Upon selection of numbers, the last two digits are replaced with random numbers. (3) A telephone exchange number database is established in accordance with the Fixed Telecommunication Network Service (FTNS) Local Switching Number Allocation Conditions of the National Communications Commission (NCC) followed by extraction of telephone exchange numbers through Random Digital Dialing (RDD) and addition of 4 digits through random number methods. (4) The same RDD sampling method is adopted for valid phone numbers in the database and the last two digits are replaced with random numbers. This way every landline phone number has a chance of being selected. In the second stage, respondents are selected from chosen households. Since the number of people in each 100