Test of Espionage and Sabotage (TES) Expansion and Validation Study
Sheila D. Reed, Ph.D., Department of Defense Polygraph Institute, Ft McClellan, AL 36205
Two previous studies conducted by the Department of Defense Polygraph Institute (DoDPI) indicated that programmed guilty (PG) and innocent examinees were correctly identified at levels significantly greater than chance when the new psychophysiological detection of deception (PDD) format-the test for espionage and sabotage (TES)-was administered. However, in both of those studies, the relevant questions contained the caveat "during this project." One objective of this study was to assess the impact of the caveat on the accuracy of the examiners' decisions.
A second objective of this study was to increase the generalizability of the TES format to the field. Several modifications were made to the format to make it more applicable to field situations. The modifications included: (a) the relevant and sacrifice relevant questions were reworded; (b) new lists of acceptable directed lie control (DLC) questions and irrelevant questions were developed; (c) the order in which relevant questions were asked on a sub-test and the order in which the sub-tests were administered was changed; and (e) a new method to handle artifacts was developed.
Many government agencies currently are using computerized polygraph systems to collect the data during the administration of PDD examinations. A computer algorithm to analyze the data would be useful. Therefore, a third objective of this study was to collect and store digitized data that could be used to develop a computer algorithm.
The 304 examinees recruited from the communities surrounding Ft. McClellan, AL., were paid $30.00 for their participation. One hundred five of the examinees were programmed guilty (PG), by enacting one of six possible mock crimes: (a) two espionage, (b) two sabotage, (c) one unauthorized contact, or (d) one unauthorized disclosure. All PG examinees were given $100.00 which they kept in a pocket during the examination. Twenty-seven certified federal examiners conducted the examinations. The examiners did not know which examinees were PG. The examiners used either standard field polygraphs or Axciton computerized polygraphs to conduct the examinations. The examiners received two hours of training to familiarize them with the format, pretest, scoring, and the use of directed lie controls. Due to the uniqueness of one of the experimental conditions, each examiner conducted six practice examinations before conducting an examination for the study. The examinations were conducted individually, in large rooms in a building at Fort McClellan, AL. All examinations were audio taped.
There were two experimental groups. Examiners in group one (individual) each conducted the complete examination, including the pretest, intest, and the data analysis. Later, a different group-one examiner blindly evaluated the test data. Examiners in group two (teams) worked in teams. One examiner conducted the pretest, and the second examiner conducted the intest. Both examiners, independently, evaluated the data. After each pretest was administered, the examiner wrote down an opinion regarding the examinee's guilt. After each intest, the administering examiner wrote down an opinion regarding the examinee's guilt. The data from the experimental groups and the examiners' opinions are reported elsewhere.
In order to test the affect of the question caveat, approximately one half of the examiners in each of the experimental groups included the caveat "during this project" as part of each relevant question. The other half of the examiners in each group asked the relevant questions without the caveat. In order to increase the generalizability of the results to the field, the format of the TES examination was slightly different from the format utilized in the previous studies. The differences included: (a) The relevant questions were worded differently; (b) a different list of ten DLCs was developed; (c) a different list of irrelevant questions was developed; (d) sub-test B was administered prior to sub-test A; (e) during the administration of sub-test B, the order of the two relevant questions was reversed; (f) the sacrifice relevant question was reworded; and (g) artifacts were handled in a different manner.
The examinations were administered according to the TES administration manual (including the modifications). Similar to the earlier studies, an innocent examinee who responded significantly to the relevant questions--a false positive (FP) decision--was questioned by the examiner to determine if there was a legitimate, "real-world" explanation for the examinee's physiological responses to the relevant questions. If there was a real-world explanation for an innocent examinee's physiological responses to the relevant questions, the examinee was not included in the initial analyses. These examinees were considered false positives with justification (FPWJ).
The data from 264 examinees were included in the original analyses. The remaining 40 examinees were excluded for the following reasons: five PG examinees confessed their guilt to the examiner prior to the examination; ten examinees were not suitable, medically, to be tested; two examinations were incomplete; thirteen FPWJ examinees were excluded, and ten examinees did not meet the eligibility requirements (e.g. educational level).
The major finding was that even with the numerous format changes, when a conclusive decision was made (i. e. inconclusive decisions were excluded), PG and innocent examinees were identified at levels significantly greater than chance. In addition, the decisions of the examiners who asked the relevant questions with the caveat "during this project" were not significantly more accurate identifying the PG or innocent examinees than were the decisions of the examiners who asked the relevant questions without the caveat.
Due to the small number of final INC decisions, INC decisions were excluded from all analyses. All of the percentages of correct decisions were significantly greater than chance. When the TES was administered without the question caveat, 83.7% (z = 3.54, p < .001) of the PG examinees and 84.0% (z = 4.6, p < .001) of the innocent examinees were correctly identified. When the TES was administered using the question caveat, 70.3% (z = 1.78 , p < .04) of the PG examinees and 85.5% (z = 4.89, p < .001) of the innocent examinees were correctly identified. The percentages of correct decisions obtained when the caveat was included are not significantly different from the percentages of correct decisions obtained when the caveat was not included. When the results of all the tests are combined, 77.9% (z = 3.81, p < .001) of the PG examinees and 84.8% (z = 6.72, p < .001) of the innocent examinees were correctly identified.
In conclusion, the TES format appears to be fairly robust. Minor alterations in the format, such as the specific questions utilized, the order of the sub-tests, and how distortions are handled do not appear to impact, adversely, on the accuracies of the decisions. In addition, greater weight can be given to the results of the two previous studies, because, during this study, the accuracies of the decisions were not significantly different when the TES was administered with the caveat "during this project" and when the TES was administered without the caveat. Future studies should begin to evaluate what aspects of the TES format contribute to the greater detection of PG examinees.
Detection of Deception, PDD, TES, Polygraph, Security screening