Presentation on theme: "Countermeasures to P300- based Guilty Knowledge Tests of Deception J.Peter Rosenfeld, Matt Soskins,Joanna Blackburn, & Ann Mary Robertson Northwestern."— Presentation transcript:
Countermeasures to P300- based Guilty Knowledge Tests of Deception J.Peter Rosenfeld, Matt Soskins,Joanna Blackburn, & Ann Mary Robertson Northwestern University. Supported by DoDPI
Some History (earliest publications) zRosenfeld et al., 1987,1988,1991 zFarwell and Donchin, 1991 zAllen, Iacono, & Danielson, 1992 zJohnson and Rosenfeld, 1992 zSince we were there at beginning, why do we challenge now with countermeasures? (1) It’s about time….
Stimuli: zProbes (P or R in figures): Items which subject is suspected of knowing (e.g., murder weapons). Subject denies(lies). zTargets (TR) Items: Items to which subject presses ‘YES’. (Benchmark P300). zIrrelevants (I or W in figures): Items of which subject has no knowledge and denies, honestly, by pressing ‘NO’.
How P300 amplitude is supposed to catch Liars: 1)P>I (‘BAD’) 2)P-TR corr >P-I corr(‘BC-AD’) 1)P=I 2)P-I corr >P-TR corr
Whither R-TR correlation if there are latency differences? Probe P3 Target P3 Nothing should happen to bootstrapped amplitude difference test (BAD) but bootstrapped cross-correlation test (BC-AD) should fail.
Experiment 1, based on Farwell & Donchin (1991): z--6 Different Probes z--Innocent, Guilty, and Countermeasure(CM) Groups z--Countermeasure: Associate various latent responses to different categories (jewelry type, drawer color, operation name, etc.), all irrelevant members of the category. z--”Off the Street “ subjects (Psych 101).
General Instructions…. zMock crime scenario zPress “Yes” to Targets (on list) zPress “No” to all other stimuli (Possibly guilty probes and Irrelevants).
More simply…. zProbeTargetI1I2 I3 I4 ring bracelet necklace watch broach tiara zpink brown yellow purple red blue zdonkey tiger lion cow pig horse z etc., etc…... (only half the matrix here.) zAll these are shuffled, presented in random order, involving 4 repetitions of each item.
What are the covert countermeasures for the 6 categories of 6 probes? z1) Jewelry category……….micro right finger wiggle z2) drawer lining category…. “ left “ “ z3) owner’s name category…. “ “ toe “ z4) operation name category…” right “ “ z5) location of item category…… Imagine professor slaps you z6) desktop category……………Do Nothing zI.e., make irrelevants into relevant targets.
Results, Exp. 1: CM works, and analysis method matters: Diagnoses of Guilty Guilty Group Innocent Group CM Group 9/11(82%)1/11(9%) 2/11(18%) Amplitude Difference (BAD) method,p=.1 Cross-Correlation(BC-AD) Method, p=.1 6/11(54%)0/11(0%)6/11(54%) --with “Off the street” subjects; see next 2 slides.
Guilty Subject, probe(R) vs irrelevant(W) R >> W: subject clearly guilty
…but clear latency shift in TR and R P300s..so BC-AD fails, BAD catches the S.
Do Reaction Times help? Guilty Group CM Group 9/10 90% BAD 3/10* 30% BART * These are guilty calls based on bootstrapped (Probe-Irrelevant) RTs, at p =.05. * Bootstrapped Analysis of RT difference, R-W.
Absolute Irrel. RTs for CM and Inno. Groups overlap*...and 3 fastest CM-ers beat tes, 3 slowest did not *I.e., they don’t help either.
Experiment 2: The one probe protocol (Rosenfeld et al., 1991). z100% simpler to conceptualize: zThere are 6 items, repeated 30-40 times randomly, consisting of.. zA probe, subject maybe guilty, zA target to force attention (not really necessary but allows us to test Farwell cross-correlation method vs. ours.) z4 irrelevants.
Standard autobiographical oddball paradigm. z6 dates are presented, one is the birth date. z5 other dates with no personal significance, one is a target, four are irrelevant. zSubject is feigning head injury, the inability to recognize birth date 50% of the time, so he presses yes and no to all dates on 50-50 basis.
Design: zOne group of advanced Ss run in 3 successive weeks. (Also Control group.) zWeek 1: All naïve(about CM) and guilty (of having birth date and recognizing it despite overt lies. zWeek 2: Use explicit countermeasure (to be explained). zWeek 3: Repeat first week, stop the CM.
What’s the CM? zWhen you see the first Irrelevant, do micro left finger wiggle. zWhen you see second Irrelevant, do micro right toe wiggle. zWhen you see third irrelevant, imagine me slapping your face. zWhen you see fourth Irrelevant, do nothing. zI.e., make irrels. into rel. targets.
Results, week 1(guilty): Probe(R) vs.Irrelevant (W) As usual, R>>W
Week 1: Probe(R) and Target(TR): Both have nice P3
Week 2: Explicit CM, R v W Not so different anymore...
Week 3: 5/12 test beaters(effortless), R v W: Surprise! R = W
Week 3: 5/12 test beaters, R v TR--”classic defeats”:..and TR >> R as with innocents.
Exp. 2; quantitative. yWK BAD BC-AD zno CM 12/13(.92) 9/13(.69) zCM 6/12(.50) 3/12(.25) zno CM 7/12(.58) 3/12(.25) z(Control group: nothing much happened over 3 weeks of repeating week 1.)
RTs for 3 weeks; week 1= week 3, proving CM not used in week 3.
Irrelevant RTs, with and without CM. No overlap!
Conclusions, bottom lines.. z6-probe protocol beat-able, RT is no help, and the 6 probe combination lacks a real rationale anyway. (Lykken wouldn’t like?) z1-probe protocol may be explicitly beat- able, but the very slow Irrelevant RT distribution will raise suspicions. 1 probe per run is more Lykkenable. zBUT---1-probe paradigm after CM practice is beat-able, period.
What to do? zWe have found(submitted) that within individuals, the scaled scalp distribution method detects 73% (not great) using statistical criteria yielding 0 % false positives. zThis method should be worked on, because there is no obvious CM as there is with simple amplitude.