電話訪問調查雖然是今日台灣最常見的調查方法，但是國內過去對此一方法的研究卻異常地不足。有鑑於此，本研究採取「過關」的角度對電話訪問全程進行檢驗，目的在瞭解電話撥號樣本在各個「關卡」中的流失的情形，以及在過關過程中，訪員與撥號時段等因素對於調查完訪之影響。此一研究發現電話樣本資料庫存在著地區的差異，從而形成某些地區空號過多的情形。此外，樣本發放的策略存在著偏差，以致造成樣本發放的早晚次序影響該樣本被嘗試撥接的次數。在訪員重撥策略方面，本研究也發現訪員雖然對電腦銀幕上出現的電話號碼，無法進行策略上的選擇，卻偏向於選擇特定時段，以利於重撥時能夠成功地完成訪問。本文最後的分析發現，日期與時段因素對於是否成能成功地完成訪問存在著相當的影響。

While Computer-Assisted Telephone Interviewing is a popular mode of data collection in Taiwan, little work has been done in the past with regard to the examination of the quality of telephone survey in Taiwan. This paper takes a procedural approach to examine the steps of conducting a telephone survey. The sampling pools that were drawn from a telephone database are first examined against the retention of samples. It is found that there are huge regional differences with regard to the invalid telephone numbers. The article then reviews recalling strategy and finds that samples from earlier pools were over used and that interviewers tend to choose 8 to 9 o'clock in the evening as the prime hour for recalling. In the end, the factors affecting the completion of a successful interview are examined. It is concluded that interview effects are minimal, but completion of a successful interview is greatly affected by calling schedule.