意在言外？－－ 以口語化行為徵候進行『測謊鑑定前篩試驗』之可行性

Beyond Words？ Evaluating Verbal Behavior to Assess the Feasibility of「Pre-Polygraph Examination Screening Instrument」

徐國超(Kuo-Chao Hsu)

指導教授 ： 周愫嫻

國立臺北大學/社會科學學院/犯罪學研究所/碩士(2011年)

摘要

本研究根據害怕偵測理論、心向理論及犯罪感制約理論，設計一套前篩試驗工具，藉由受測人針對特定議題說謊、隱瞞與說真話時，回答內容上的差異，評估此工具辨識個體說謊的可能性。主要研究目的是找出工具分數之最佳判定閾值（optimal cut-off value）、分析工具之準確度、工具中各題鑑別度，及相關基本變項對本工具之影響。
本研究從2010年11月份至2011年2月份間，以刑事警察局所有受理委託機關申請所為之測謊鑑定受測人為對象，由專業測謊人員根據本工具施測，再由研究者對受測人之回答內容分類、給分。
研究結果發現測謊結果中「不實反應組」（0.13）的「平均單題得分」平均值明顯低於「無不實反應組」（0.52），代表本前篩工具分數可有效的區辨個體對特定議題說謊與誠實兩種情況。再經由「接收者操作特徵」曲線 (Receiver Operating Characteristic Curve)分析，找出判定上的最佳閾值（optimal cut-off value）為0.32456，曲線下面積為0.901，可知本前篩工具準確度高；若與測謊結果之後端效標比較，本研究也發現總體命中率為83.65%，陽性預測值與陰性預測值分別為82%、85.2%，是一個適合做為前篩試驗之工具。另從項目分析結果發現研究工具中之各題具有不同的鑑別力，此一發現將有利於開啟後續優化工具之相關研究；同時本研究也證實樣本之性別、年齡、教育程度及受測人角色對研究工具之區辨能力，並無影響。
整體而言，本研究發現受測人的確會因說謊與否，觸發出不同的口語化行為，筆者認為自行設計之工具堪可為初步判定受測人是否說謊、隱瞞與說真話之用。
本文建議在良好的軟硬體及嚴謹的品管下，本前篩工具應可用於排定偵查的優先順序，紓解目前測謊實務量過大之困境；再者，以本前篩工具作為一項測謊鑑定之外的輔助工具，能有效幫助施測者觀察、評估受測人之口語化行為徵候，讓施測者在解讀測謊結果時注意到更多面向，減少不必要的失誤。

關鍵字

口語化行為 ； 前篩試驗 ； 說謊 ； 測謊

並列摘要

Based on the concept of 「Fear of detection」,「Psychological Set」and 「Conditioning」, the research is trying to design a 「Screening Question Set」. By evaluating the Verbal Behavior, the author attempted to access the feasibility of「Pre-Polygraph Examination Screening Instrument（P.E.S.I.）」. So the purpose of the research is to define the optimal cut-off value of this question set, to determine the degree of accuracy, the discrimination between every single question and the affection of different population variables.
From November 2010 to February 2011, the research sample had collected all examinees who applied by the judiciary to the Criminal Investigation Bureau. After questioning the 「Screening Instrument」by the polygraph examiners, the author should evaluate the responded content obtained from the examinees and score it.
The result shown that the score of the 「Screening Instrument」in the 「deception Indicated group」（0.13）is significantly lower than that in the「Non-Deception indicated group」（0.52）. It means the 「Screening Instrument」really can distinguish the truth from lies. After that, by using the 「Receiver Operating Characteristic curve」, it is shown that the optimal cut-off value of the score is 0.32456, and the predictability expressed as the area under the curve (AUC) of the receiver operator characteristic curve (ROC) was 0.901. Supposed comparing with the polygraph conclusion, the total consistency rate was 83.65% and the positive and negative predictive value was 82% and 85.2%. In addition, every single question in「Screening Instrument」contributes a different degree of discrimination. It could be implied that the「Screening Instrument」can be optimized by choosing 「Good」questions. Finally, the research also showed that the differences between gender, age, education and status did not affect the results.
The author believed that by using「Pre-Polygraph Examination Screening Instrument（P.E.S.I.）」, the examiner can get an initial conclusion about whether the examinee was lying before Polygraph Examination. The research confirmed this opinion.
The findings also considered that the 「Pre-Polygraph Examination Screening Instrument（P.E.S.I.）」, in conjunction with a carefully controlled procedure, could be used to prioritize suspects. It assists to relieve the heavy workload of polygraph examination. Furthermore, using the「Pre-Polygraph Examination Screening Instrument（P.E.S.I.）」can help the examiner notice wider aspects of the Polygraph Examination itself, avoiding unnecessary errors.

並列關鍵字

Verbal Behavior ； Screening Instrument ； Lie ； Polygraph

參考文獻

羅時強，2002，《測謊「區域比對法」之本土化研究初探》，高雄醫學院行為科學研究所碩士論文。

Akobeng, Anthony K. 2007. Understanding diagnostic test 3 : Receiver operating characteristic curves. Acta Paediatrica, 96：644-647.

Biggerstaff, B.J. 2000. Comparison diagnostic tests: a simple graphic using likelihood ratio. Stat. Med. 19：649 - 663.

Crichton, N. 2002. Information point: receiver operating characteristic (ROC) curves. Journal of Clinical Nursing, 11(1)：136.

Department of Defense Polygraph Institute【DoDPI】 1994. Polygraph program, United States Department of Defense. Polygraph, 23：61-84.