The Vasomotor Response In The Comparison Question Test

The Vasomotor Response In The Comparison Question Testsigned a waiver consenting to placement ofthe components and testing was thencontinued. Each subject was then given apractice examination to acquaint them withthe testing procedures and all subjects weregiven positive feedback on their practice testperformance (Kircher et al., 2001).Thepractice examination was a known solutionexamination where the subject wrote thenumber 3 on a piece of paper with preprinted numbers 1, 2, 4 and 5. Then theexaminer asked them “Regarding the numberyou wrote on that paper, was it the number1?”and continued through the number 5.base rate but unaware of any subjects’assignment. The experienced examiner haddisplayed proficiency in numerical scoring inseveral prior experiments, including onewhere FPA was evaluated (Honts & Reavy,2015). The experienced examiner analyzedthe data using the ESS for the respiration,EDA and cardio components and used thevasomotor assessment tool the FPA score for36 of the 40 examinations. One student usedan LX-4000 and did not identify the auxiliaryport to which the plethysmograph wasconnected.This resulted in a loss ofvasomotor assessment tool data for fourexams.In those cases, the t visually.FPA scores wereassigned with the standard ESS three-pointscale, -1, 0, 1. The experienced examinerentered all scores into an Excel (Microsoft,2013) spreadsheet and sent that spreadsheetto the school director (third author). Theschool director then sent the experiencedexaminer’s scores and the individual testingand criterion assignment to the first authorfor analysis.Data were collected using a singleissue three relevant question examinationfollowing the Utah protocol (Raskin & Honts,2002).The examinations were conductedwith Directed-Lie comparison questions(DLC).All examinations consisted of fiverepetitions of the questions with datarecording (see Appendix C for questions andsubjects’ answers).After each datacollection, the examiner released the pressurein the cuff and gave the subject about oneminute of rest before adding air andcollecting the next chart. After data for fiverepetitions were collected, the examinerremoved the components and dismissed thetest subject.Each student took fourexaminations (with one exception for astudent with an excused medical absence oneday who was replaced with another studentprogrammed with the same criterion state).Half the examinations were randomlyprogrammed guilty and a total of 40examinations were collected.The schooldirector or a senior instructor reviewed alldata for correct collection procedures. Onestudent examiner was asked to repeat anexamination because the pretest wasinsufficient and did not conform to standardpractices. The inadequate examination wasnot included in our study data analyses.ResultsSince the vasomotor responses in fourexaminations were scored using the Utahnumerical scoring criteria rather than thenewly developed vasomotor scoring tool, weran an initial analysis to see if there weresignificantdifferencesbetweentotalvasomotor scores generated by the tool andthe Utah scoring. A Guilt (Guilty, Innocent) XMethod (Utah, Tool) ANOVA of the totalvasomotor scores produced no significanteffects involving the Method variable. Themain effect of Guilt was significant, F(1, 36) 9.88, p .003, ηp2 0.215.Numerical ScoresWith 40 examinations, five questionrepetitions, and three relevant questions,there were 600 possible occasions to assign anumerical score in each of the components.The rate of assigning scores is shown in Table1. Vasomotor response were scored the mostandrespirationresponsestheleastAfter all the data were collected thestudent examiners scored the tests theyconducted.Their recorded results werescanned and kept as part of their schoolrecord. The individual examinations wereanonymized and emailed to an experiencedexaminer who was aware of the approximate.Polygraph, 2015, Volume 44 (1)65

Honts, Handler, Shaw, GouglerTable 1 Frequency and Percentage of Scoring of the Various Components From the 600 OpportunitiesComponentFrequencyPercentage Scored518.5%Electrodermal37061.7%Blood orthephysiologicalcomponents and total scores are shown inTable 2 for Innocent and Guilty subjects.Differences between guilty and innocentmeans, assessed with independent sample ttests (df 38) were significant for all of thecomponents and the total score, p .01.Component and total scores were correlatedwith each other and with the Guilt criterion.The correlation values are presented in Table3. Component scores and the total scorewere all significantly correlated with the Guiltcriterions and were ranked in predictivepower as follows: electrodermal vasomotor blood pressure respiration.Table 2 Means and Standard Deviations for Component and Total Scores for Innocent andGuilty SubjectsMeasureGuiltyInnocentRespiration-1.85 (2.25)0.00 (0.72)Electrodermal Activity-7.10 (6.54)6.30 (7.52)Mean Blood Pressure-1.05 (3.28)3.55 (3.71)Vasomotor Activity-1.80 (3.43)3.65 (3.17)-11.80 (7.69)13.50 (9.30)Total ESS Score66Polygraph, 2015, Volume 44 (1)

The Vasomotor Response In The Comparison Question TestTable 3 Correlation of Component and Total Scores with GuiltMeasureRespirationElectrodermalBlood PressureVasomotorElectrodermal0.41bBlood 4bp .05p .01The component numerical scores werealso submitted to a stepwise discriminantanalysis. The discriminant analysis used theWilk’s method with the probability of entryset to 0.05 and the probability for removal setto 0.10.Three component scores wereretained in the following order of stepwiseentry, electrodermal, vasomotor, and bloodpressure.The standardized canonicaldiscriminant function coefficients were 0.685,0.591, and 0.485, respectively.Thecanonical correlation was 0.831, and theWilk’s Lambda was 0.309, p 0.001. The.discriminant analysis also produced adecision classification table that is presenthere as Table 4. The initial discriminantmodel classified 95% of the cases correctly. Across-validationoftheclassificationperformance was conducted using the N - 1process whereby each case is classified by thefunctions derived from all cases other thanthat case. The classifications from the crossvalidation analysis are also presented inTable 4.The cross-validation correctlyclassified 92.5% of the casesTable 4 Original and Cross-Validated Discriminant Analysis ClassificationsAnalysis and StatusOriginalGuiltyInnocentPredicted Group atedGuiltyInnocentPolygraph, 2015, Volume 44 (1)67

Honts, Handler, Shaw, GouglerESS DecisionsDecisions with ESS were made withthe standard -4 and 2 cutting scores. Noadjustment in the cutting scores was madefor the addition of the vasomotor component.ESS decisions made with and without thevasomotor component are shown in Table 5.Detection Efficiency Coefficients (Kircher,Horowitz, & Raskin, 1988) were calculated forESS decisions with and without thevasomotor component.The ionalvaluethatassessestheclassification power of deception detection.Without the vasomotor component theDetection Efficiency Coefficient was 0.86.With the vasomotor component the DetectionEfficiency Coefficient was 0.87.Thedifference between these Detection EfficiencyCoefficients was not significant.Table 5 ESS Outcomes With and Without the Vasomotor ComponentOutcomesWithout 1118GuiltyInnocentWith VasomotorGuiltyInnocentDiscussionrespiration at all. It is notable that theseresults are very similar to those recentlyreported by Honts and Reavy (2015). Hontsand Reavy found that their Utah ScoringSystemvasomotorcomponentscoresoutperformed the blood pressure andrespiration.Similarly, their discriminantanalysis solution loaded the vasomotorcomponent second and did not ssment tool tested is more reliable thanthe human evaluators, and may be moreprecise, as are most computer scoringalgorithms (Nelson, et al., 2008). Bell et al.(1999) reported the vasomotor response isscored about 30% of the time when using theUtah Scoring System. In this study, theexperienced examiner using the assessmenttool entered scores for the vasomotorresponse in 74.5% of the cases. In this studythe increased reliability and frequency ofscoring produced a vasomotor measure thatwas strongly correlated with the Guiltcriterion and one that outperformed thetraditional respiration and blood pressuremeasures.Discriminant analyses of thecomponent ESS scores produced a solutionthat included the vasomotor component asthe second predictor and did not includeA possible flaw and limitation to thisstudy was presented in the fact that four ofthe student examinations were collectedusing an LX 4000 instrument and thestudent examiner erred in not assigning theplethysmograph an auxiliary port status,resulting in a loss of the assessment plot/toolfor those examinations. In place of the tool,an experienced examiner used the Utahnumerical scoring criteria for FPA listed byBell et al. (1999). However, our analysis68Polygraph, 2015, Volume 44 (1)

The Vasomotor Response In The Comparison Question Testfailed to find any significant differencesbetween the vasomotor scores generated withthe Utah Scoring system and those scoresgenerated with that assessment tool.Moreover, our results are very similar tothose obtained by Honts and Reavy (2015)who used only the Utah Scoring System.One additional finding in this study isworth noting.Honts and Handler (2014)addressed the issue of numerically evaluatingrespiration in directed-lie tests.Theyreviewed the literature and concluded thattheavailabledatasupportedscoringrespiration as normal when directed-liecomparison questions were used. This studyemployed directed lies and significantcorrelations with the guilt criterion wereobtained with the ESS scores from therespiration component that was scored withthe standard ESS rules. While the stepwisediscriminantmodeldidnotincluderespiration, this should not be taken as anindication that respiration should not beevaluated.The lack of inclusion ofrespiration in the stepwise discriminantanalysis indicates that after the other threevariables were entered, respiration did notsignificantly improve the accuracy of themodel.The original model was alreadyclassifying 95% of the subjects correctly.Thus the ability of the statistical model toimprove more was limited by approaching theceiling on accuracy. Respiration produced asignificant and substantial correlation withthe guilt criterion of 0.49. That correlation isa strong indication that respiration is avaluable measure for scoring. Thus, theseresults provide additional evidence to ttherespirationcomponent be scored as normal with theUtah or the ESS criteria and rules.Inthisstudytheexperiencedexaminer was familiar with the hardware andsoftware of the assessment tool andrecognized how artifacts might adverselyaffect the reaction ratio.When theexperienced examiner noted artifacts in thepre or post-stimulus visual waveform, thosedata were excluded from consideration. If theunusable data were from a comparisonquestion, the experienced examiner used thenearest comparison question that waswithout artifact. If the artifact was in arelevant question the experienced examinerscored it by assigning the value of zero. It isuncertain how well the assessment toolwould work without training on datacollection and artifacts. Additionally, it is notknown how the repeated testing ofindividuals within this study affected the testdata.Little work in test-retest reliabilityexists, though perhaps this scenario offersopportunity to fill in the knowledge gap. Inthis study serving as a subject in multipletests seems to have had little impact on theaccuracy of the examinations, as thataccuracy was very high.Moreover, theresults of this study were very similar toHonts and Reavy (2015) who used lusionThe results of this study providegeneral support for the effectiveness of thealgorithm-based assessment tool we tested.The vasomotor scores derived using theassessment tool strongly correlated with theGuiltcriterionandoutperformedthetraditional blood pressure and respirationcomponents. These results are consistentwith the large body of research supportingthe use of the vasomotor response fordeception detection. In total this body ofscientific studies provides strong support forthe use of the vasomotor response in fieldapplication.Finally, it is unknown how theinclusion of the vasomotor scores might affectthe published reference distributions used todevelop the ESS cutting scores. The ESS haspublished p-values for subtotal and grandtotal scores that were derived without thevasomotor component score. Since the ESSis being taught at many polygraph seminarsand schools it would be important to know if,and how, these plethysmograph scores affectthe distributions. We made no adjustmentshere and achieved high performance, but thegeneralizability of our findings should beestablished through replication with othersubject populations.Polygraph, 2015, Volume 44 (1)69

Honts, Handler, Shaw, GouglerReferencesBarland, G. H., & Raskin, D. C. (1975). An evaluation of field techniques in detection of deception.Psychophysiology, 12, 321-330.Bell, B. G., Raskin, D. C., Honts, C. R., & Kircher, J. C. (1999).system. Polygraph, 28, 1-9.The Utah numerical scoringDepartment of Defense (2006). Federal Psychophysiological Detection of Deception ExaminerHandbook. Retrieved from ygraphhandbook-02-10-2006.pdf on 3-31-2007. Reprinted in Polygraph, 40(1), 2-66.Handler, M.D. (2006). Utah Probable Lie Comparison Test. Polygraph 35 (3), 139-148.Honts, C. R., Amato, S., & Gordon, A. (2000). Validity of outside-issue questions in the control question test: Final report on grant no. N00014-98-1-0725. Submitted to the Office of Naval Research and the Department of Defense Polygraph Institute. Applied Cognition Research Institute, Boise State University. DTIC# ADA376666.Honts, C.R. & Handler, M., (2014) Scoring Respiration When Using Directed Lie ComparisonQuestions, Polygraph, 43 (3).Honts, C. R., & Reavy, R., (2015). The comparison question polygraph test: A contrast of methodsand scoring. Manuscript revised and resubmitted.Horowitz, S. W., Kircher, J. C., Honts, C. R., & Raskin, D. C. (1997). The role of comparison questions in physiological detection of deception. Psychophysiology, 34, 108-115.Kircher, J.C. (1983). Computerized decision-making and patterns of activation in the detection of deception. Unpublished doctoral dissertation, University of Utah, Salt Lake City.Kircher, J. C., Horowitz, S. W. & Raskin, D. C. (1988). Meta-analysis of mock crime studies of thecontrol question polygraph technique. Law and Human Behavior, 12, 79-90.Kircher, J. C., Packard, T., Bell, B. G., & Bernhardt, P. C. (2001). Effects of prior demonstrations ofpolygraph accuracy on outcomes of probable-lie and directed-lie polygraph tests. Polygraph39, 22–67.Kircher, J. C., & Raskin, D. C. (1988). Human versus computerized evaluationsof polygraph data in a laboratory setting. Journal of Applied Psychology, 73, 291-302.Kubis, J.F. 1962 Studies in Lie Detection: Computer Feasibility Considerations. RADC-TR 62-205,Contract AF 30(602)-2270. Air Force Systems Command, U.S. Air Force, Griffiss AirForce Base. New York: Rome Air Development Center.Nelson, R., Krapohl, D. & Handler, M. (2008). Brute force comparison: A Monte Carlo study of theObjective Scoring System version 3 (OSS-3) and human polygraph scorers. Polygraph, 37,185-215.Nelson, R. Handler, M. Shaw, P., Gougler, M., Blalock, B., Russell, C., Cushman, B., & Oelrich, M.(2011). Using the Empirical Scoring System, Polygraph, 40 (2).Podlesny, J. A., & Raskin, D. C. (1978). Effectiveness of techniques and physiological measures inthe detection of deception. Psychophysiology, 15, 344-358.70Polygraph, 2015, Volume 44 (1)

The Vasomotor Response In The Comparison Question TestPodlesny, J. A., & Truslow, C. M. (1993). Validity of an expanded-issue (modified general question)polygraph technique in a simulated distributed-crime-roles context. Journal of AppliedPsychology, 78, 788-797.Raskin, D. C. & Hare, R.D. (1978). Psychopathy and detection of deception in a prison population.Psychophysiology, 15, 126-136.Raskin, D.C. & Honts, C.R. (2002). The comparison question test. In M. Kleiner (Ed.), Handbook ofpolygraph testing. London: Academic (1-49).Raskin, D. C. & Kircher, J. C. (1990). Development of a computerized polygraph system andphysiological measures for detection of deception and countermeasures: A pilot study(Contract 88-L55300-000). Salt Lake City: Scientific Assessment Technologies.Rovner, L. I. (1986). The accuracy of physiological detection of deception for subjects with priorknowledge. Polygraph, 15, 1-39.Polygraph, 2015, Volume 44 (1)71

Honts, Handler, Shaw, GouglerAppendix A.Instructions, Innocent ConditionBy now you have agreed to be a participant in this study.There are different conditions in this study. These instructions will tell you about your role inthe experiment.Your condition assignment was made on a random basis, and you actually chose it with theenvelope you selected.None of the polygraph examiners know whether you will be telling the truth during the polygraph examinations. Do not tell them!You have been selected to be in the innocent condition.Please read these instructions carefully and make sure that you understand exactly what youare to do. Reread these instructions if necessary. You may make a few notes to help you remember what to do as you carry out these instructions - there are writing materials on the desk foryou.This is a polygraph, or psychophysiological detection of deception experiment. About half ofthe subjects in the experiment are instructed to commit a theft. They are to go to a room and stealsome money from an envelope. Then they report back for a polygraph examination. If they arefound innocent on the test, they are placed in a pool for a considerable bonus on their final examination grade - in addition to e

the FPA criteria into numerical scoring evaluations. One reason may be a concern for the reliability of evaluating the FPA criteria due to possible subjectivity in using the scoring criteria. Trying to separate similar amplitude changes and durations might be difficult for some examiners,