面對即時民意－行動電話調查與傳統電話調查的異同

傳統電話調查具有成本較低，能夠快速獲得訪問結果，以及涵蓋率較高的優點，然而傳統電話調查也有難以突破的死角，例如：「先抽戶再抽人」造成樣本扭曲，無法訪問到夜間工作者，年輕人受訪率較低，以及日益增加的沒有安裝家用電話的「行動電話族」等等；行動電話調查正好可以彌補傳統電話調查涵蓋性不足的問題，例如：行動電話調查可直接接觸到受訪者，可以在白天接觸到夜間工作者，以及可以接觸到年輕人與行動電話族等等。然而，行動電話調查也有其尚未突破的盲點存在，例如：「一人多機」，訪問母體不明確以及訪問成本較高等等問題。為了比較行動電話調查與一般傳統電話調查的異同，本研究進行準實驗設計，以同樣的問卷題目同時進行行動電話與傳統電話的訪問。研究結果發現：一、行動電話調查的電話費成本大約是傳統電話調查的2.2倍，成本明顯較高。二、行動電話調查可接觸到較多的年輕人、學生與高教育程度者，但會低估女性、老年人、低教育程度與農林漁牧、勞工、家管或主管人員的意見。三、行動電話調查的地區分佈不比傳統電話調查分層隨機抽樣差。四、在態度性的問題方面，行動電話調查與傳統電話調查的結果差別不大。雖然行動電話調查有著抽樣設計程序較為簡便的優點，並且在地區分佈與態度性問題上的表現和傳統電話民調相比毫不遜色。但是一來由於其花費較高，二來其調查結果推論受到侷限，因此，日後仍應將行動電話調查視為輔助性的工具使用，目前尚無法完全以行動電話調查取代傳統電話調查的功能。

關鍵字

行動電話調查；行動電話涵蓋率；抽樣

並列摘要

Traditional telephone survey has some advantages, such as low cost, quick acquisition of survey outcomes, and high coverage rate. But still, It has some defect, such as: 「Sampling household first then person」will twist the sample structure, unable to reach those who work at nighttime, low interview rate of the young person, and the increasing of the cellular phone users which did not have any household telephone. Cellular phone survey can redeem these defect. For example: Cellular phone survey is able to reach the interviewer directly, it also can reach those who work at night in the daytime, and still, cellular phone survey can reach young person and those who only use cellular phone but household telephone. However, cellular phone survey also has it's blind spot, such as: 「one person has more than one cellular phone」, the sampling population is not clearly defined, and the cost of cellular phone survey is much higher than traditional telephone survey. In order to compare cellular phone survey and traditional telephone survey, we set up a quasi-experimental design, using the same questionnaire to run cellular phone survey and telephone survey at the same period of time. Research finding points out: 1.The survey cost of cellular phone survey is somehow 2.2 times over traditional telephone survey. 2. Cellular phone survey can reach more young person, students and higher educated person, but in the same time, it would underestimate the opinions of female, older ages, lower educational person, and those who work as farmers, fishermen, labors, housewife or supervisors. 3.The district distribution of cellular phone survey seems as good as traditional telephone survey with stratum random sampling. 4. For some attitudinal questions, such as party identification, the outcome of cellular phone survey is similar with traditional telephone survey. Generally speaking, although cellular phone survey has some advantages, such as more simple procedure in sampling design, the district distribution and the outcome of attitudinal questions is no less than traditional telephone survey. But because of it's high survey cost and the outcome of cellular phone survey is hardly inference to the population, thus we should view cellular phone survey as an assistant tool of traditional telephone survey, and till now, we still can't totally replace the functions of traditional telephone survey by cellular phone survey.