Is the sample used by Taiwan's Manpower Survey representative? Are its statistical results accurate?
1. The Manpower Survey's monthly sample consists of 20,000 randomly-selected households (nearly 60,000 persons aged 15 & over). The sampling rate is 0.25%, which is higher than the rate of 0.23% in Korea, 0.09% in Japan, and 0.05% in the United States; the sample is therefore quite representative.
2. The statistical results of the Manpower Survey are based on the detailed, objective data obtained from the monthly surveys, and overall survey design and statistical methods are very rigorous. Furthermore, in order to enhance the survey's accuracy, apart from improving sampling methods and regularly updating the sample population files in order to control sampling error, DGBAS is also continuing to improve its question design methods and training for survey personnel, and unifies survey question definitions when needed, ensuring that survey personnel will not allow their subjective judgments affect their actual form completion. DGBAS is also actively implementing a follow-up control mechanism to reduce untruthful and inaccurate answers and improve the quality of survey data.