Background
    According to the estimation made by the World Health Organization (WHO), tobacco kills more than 8 million people each year worldwide. More than 7 million of those deaths result from direct tobacco use, while around 1.2 million are the result of non-smokers being exposed to second-hand smoke. About 25,000 smokers died attributed to of smoking, and nearly 3,000 non-smokers died related to secondhand smoke each year in Taiwan. On average, one person lost their life due to 
smoke in less than 20 minutes. Based upon the increasing problems of tobacco use and related hazards, the Ministry of Health and Welfare and its former organization (the Department of Health) implemented the Tobacco Hazards Prevention Act in 1997 and has been promoted tobacco control policies since then.

     To monitor the smoking behaviors and provide references for assessing the effectiveness of tobacco control policies and strategies in Taiwan, the Health Promotion Administration (HPA) launched the annual Adult Smoking Behavior Survey (ASBS) in 2004. The telephone survey was conducted yearly from July to the end of September. In 2018, the survey frequency was modified to every other year. Due to the increase of mobile phone users in the population, HPA has adapted the 
dual-frame design that combined landline and mobile phones survey since 2020.

Aim of the Survey
  1. To understand smoking behaviors, secondhand smoke exposure among people  over the age of         15 in Taiwan.

  2. To establish the nationwide and county/city-wide database of the smoking   prevalence of people over the age of 15, and as references for policy planning and  evaluation of tobacco control programs and interventions.

 The sampling frame in the landline survey is based on the Chunghwa Telecom residential telephone directory and the National Communications Commission's Code verification and allocation of local exchanges for Fixed Communication Network services as the sampling frame. It was applied the Probability Proportional to Size (PPS) as the sampling method in landline survey. The sampling is conducted with the Random Digit Dialing principle, which selects the prefix firstly, and then randomly selects the last two or four digits of the telephone numbers with those selected area 
codes. After phone calls are answered, within-household sampling is conducted to pick up the qualified respondent.  The sampling frame in the mobile phone survey is based on the National Communications Commission's Mobile communication network service user number allocation data. Mobile phone numbers are selected by Random Digit Dialing (RDD). The county's population size selected entries of the last five digits as the sample mobile numbers.

Survey Methods
 ASBS used the CATI (Computer Assisted Telephone Interviewing) system which integrates the computer (both hardware and software) with the telephone system and is equipped with “Within-household Sampling” function. The specialist pre-programs the survey contents and the skipping rules of questions into the CATI system. The interviewers input and record respondents’ answers. The supervisors can monitor phone calls, and tackle possible errors to ensure each call complies with the standard operating procedure of the survey.

Data Processing and Analysis
  The survey data were checked and cleaned before analysis. Post-stratification weighting is applied, and goodness of fit test is used to determine whether the distribution of the respondents in gender, age, and residence matched the target population's characteristics to ensure a good representative of the target population.

Results of the survey are published in the format of survey reports, news release, and conference or journal papers. Descriptive statistics are accessible from the Statistical yearbook of Health Promotion.