本研究比較兩種電話抽樣方法，一種是以住宅電話簿等距抽樣後再將中選號碼的最後一碼隨機處理，另一種是隨機撥號法（RDD），先以PPS方式抽出局碼，再加上後四碼隨機。經過五年實驗結果顯示RDD抽樣雖然接通率平均低了大約九個百分點，每一個典型調查案大約多支出一萬元費用，但具備了理論上涵蓋率完整的優點，訪問結果和地區代表性也都不錯，而控制組後一碼隨機的接通率雖然較高，但必須依靠電話簿抽樣，且在理論和實務上涵蓋率仍有缺失，比較起來，RDD應是較好的電話抽樣方法。

The study compared two telephone-sampling methods, one is the typical directory-based systematic sampling with random number re-placement of the last digit, and another is random digit dialing (RDD) using PPS method. Results showed that the RDD method performed well with its advantage on perfect coverage. The cost of lower contact rate due to broader coverage is not high, and there is no difference between the two methods in terms of sample's geo-graphical representation.