The Journal of Credibility Assessment and Witness Psychology
1997, Vol. 1, No. 1, 4-8
Published by the Department of Psychology of Boise State University
On Furedy's (1993) Call for Abolition of Research and Practice With the Control Question Test
J. Peter Rosenfeld, Department
of Psychology, Northwestern University
Copyright 1997 by the Department of Psychology of Boise State University and the Author. Permission for non-profit electronic dissemination of this article is granted. Reproduction in hardcopy/print format for educational purposes or by non-profit organizations such as libraries and schools is permitted. Any modification of this document is expressly forbidden. For all other uses of the this article, prior advance written notice is required. Send inquiries by hardcopy to: Charles R. Honts, Ph. D., Editor, The Journal of Credibility Assessment and Witness Psychology, Department of Psychology, Boise State University, 1910 University Drive, Boise, Idaho 83725, USA.
The two paradigms for detection of deception--the Guilty Knowledge Test (GKT) and Control Question Test (CQT) -- have been the source of acrimonious disagreement in the psychophysiology community for many years. Furedy (1993) and Honts, Kircher & Raskin (1995) are the most recent examples of the two sides of this dispute and provide its full bibliographic details. In Furedy (1993) the attack on the CQT lead to an extraordinary, proposed final solution to what he calls the "polygrapher's dilemma;" a call "for all practitioners and researchers to abandon the CQT." A much more fully developed set of related anti-CQT arguments had been presented in Ben-Shakhar and Furedy (1990).
Honts et al. (1995) responded to Furedy's (1993) attack by attempting to point out logical and factual flaws in his critique of the CQT. Their arguments may be judged on their merits. One of the points they made, however, appears on reflection to bear more emphasis and illustration than actually given by Honts et al. (1995). This point is that Furedy's (1993) attack, though not always represented as such, is mostly on the faulty practices of some polygraph professionals, rather than on the basic principles of control question testing. Indeed a careful reading of Furedy's own co-authored book (Ben-Shakhar & Furedy, 1990) provides much evidence that the CQT could be given in such a way that even more theoretically-based objections to the CQT (e.g., the lack of true "control" items) are remediable. (Of course, much of the book by Ben-Shakhar and Furedy (1990) is a critique of the principles of an early form of CQT utilized in North America.)
*Note: This paper was originally submitted to the International Journal of Psychophysiology, in response to the Furedy (1993) paper which appeared there. The editor, however, felt that he did not want to devote more space to the controversy. The paper was also submitted to Polygraph, whose editor also declined to devote that journal's pages to an academic controversy.
Moreover, Honts et al. (1995) did not deal with the positioning by Furedy and others of the GKT as being (unlike the CQT) a standardizable, true psychological test. This omission is probably because the GKT as exemplar was not emphasized in Furedy (1993), although it is in Ben-Shakhar & Furedy (1990) and elsewhere. Moreover all parties interested in the CQT-GKT controversy are well-aware that an attack on the CQT implies the superiority of the GKT. The present comments are addressed to the issues listed in this and in the preceding paragraph.
As an aside, it should be noted that the most frequent criticism of the GKT by CQT partisans is likewise on the basis of practice versus principles issues: it is typically pointed out that the GKT is not practical to broadly administer because of a) the demanding investigatory and methodological challenge of developing an adequate number of key items of guilty knowledge (Podlesny, 1993; Elaad, 1990), a problem actually articulated by a leading proponent and developer of the GKT, Lykken (1981), and b) the GKT requirement that crime details not be publicly leaked, thereby providing guilty knowledge to innocent persons. In Ben-Shakhar & Furedy (1990, p.120), it is pointed out that unlike the case in North America, in Japan, cooperation on information leaks between police and news media is such that the GKT is used 50% of the time. If problems with the GKT are avoidable through appropriate modifications of practice, the same remediation should be allowed for the CQT, as illustrated below.
One of the major points of contrast of GKT vs CQT (according to
Furedy, 1993, and Ben-Shakhar & Furedy, 1990, p.10) is that
although the former test can be made standardizable, the CQT cannot,
since the development of its control questions depends on pretest
interviews which must vary from case to case. As stated by Ben-Shakhar
& Furedy, (1990, p.28):
In the GKT...the questions are developed by the
features of the investigation, and they do not depend [as in the
CQT] upon the examiner-examinee interaction. Since the set of
all features of a given event may be fairly large, different examiners
may construct different questions for a GKT procedure. However,
because the set of questions chosen by a GKT examiner is sampled
from the features of the event, two GKT procedures may be viewed
as equivalent forms of a vocabulary test, each containing different
items sampled from the same set of words.
This assertion appears to idealize GKT construction procedures, and glosses over some potentially serious shortcomings in the GKT's alleged superior standardization. One might first wonder if there is any evidence in support of the analogy of different GKTs for a given situation resembling different forms of a vocabulary test. Such could be the case, but I am not aware of any data showing that different "forms" of a GKT would produce results that correlate in the way that results of different forms of a vocabulary test do. Indeed, I have had more than a dozen teaching assistants in my Psychobiology course over the years whose job it is to construct tests on a body of material of consistent difficulty and coverage from year to year. Some of these people are brighter and harder-working than others and therefore construct better tests than their less-gifted counterparts. A similar situation must exist in GKT construction. Indeed, will the depth of criminal investigation be constant from examiner to examiner or from crime to crime? Will the details of a crime be equally rich or complex from situation to situation? Standardization is typically achieved in testing by repeated administration of the same items to sample after sample of a large population, as is the case with college entrance exams. It is difficult to imagine how this more familiar standard of standardization might be met with GKT procedures. Obviously, much of the above can be said about the CQT. The point here is not to assert the CQT over the GKT, but simply to note that the latter procedure may not be as flawless as is suggested by the idealizing selection quoted above.
Furedy (1993) also faults the CQT for requiring a psychologically
stressful pre-test interview, implying that in contrast, the GKT
is free of such undesirable interrogation. Honts et al. (1995)
respond that merely being under suspicion of having committed
a crime is sufficiently aversive whether or not one is also to
take a polygraph test of either variety. It might be added that
the police investigation following any crime would more than likely
also involve an aversive interrogation of suspects, no matter
which form of polygraph test (if either) was anticipated. Also,
the pre-test interview associated with the CQT need not be as
aversive as Furedy (1993) asserts; this is another practice
issue, and indeed, Leonarde Keeler, who developed the first polygraph,
It is most important in this type of test that
no methods shall be resorted to which will excite the suspect.
All exciting factors must be eliminated, so that the responses
will be due only to the case in question and not to physical or
other psychological disturbances. This subject must be treated
kindly and with respect at all times, in order to induce relaxation
and as far as possible to eliminate emotional tension. Best results
are obtained when the operator works on the theory that the subject
is innocent and attempts to obtain as regular a curve as possible.
The guilty side of the situation will be well handled by the suspect
himself. (Keeler, 1930.)
The most frequent and difficult-to-defend critique of the early, North-American form of the CQT is that its "control" questions are not real scientific controls since they may be discriminated by examinees from relevant questions; this is often related to the additional problem that they may differ from relevant questions in arousal value (Furedy, 1993). Honts et al. (1995) responded to this criticism, but did not make explicit the important additional point that derivative versions of the CQT are possible in which these problems can be really attenuated, if not eliminated. In support of this point, one need not look much further than Ben-Shakhar and Furedy's own book, where they remind the reader of Lykken's (1981) "truth control test," in which questions about equivalently arousing but fictitious crimes are used as (true, scientific) control items (Ben-Shakhar & Furedy, 1990, p. 24). The authors point out that such a procedure may have a problem in requiring suspects to believe they are under suspicion for fictitious as well as real crimes, which imposes an additional burden on investigators. This is, again, a practice issue. Interestingly, Ben-Shakhar & Furedy, (1990) speculated that this additional burden probably "accounts for the infrequent use of [these special control] questions." Of course, additional burdens on investigators also accounts for the infrequent use of the GKT. Indeed, we are further informed (in the Ben-Shakhar & Furedy 1990 volume, p.119) that in Japan, the CQT is given in what to the authors is a wholly acceptable format: control questions involve other crimes, and additionally, no pre-test interrogation is involved: "So in principle, this version of the CQT does have a control in the normal, scientific sense of that term." (I might add that we have used such a modified CQT format successfully in two research studies in a laboratory situation; Rosenfeld, Angell, Johnson, & Quien, 1991; Johnson & Rosenfeld, 1992, although one earlier paper by Podesny & Raskin, 1978, found that a "truth-control" CQT was not as effective as a regular CQT.)
This quotation contains a remarkable statement from a writer who frequently illustrates the impossibility of equilibrating the arousal value of relevant and control questions with a particularly egregious example of a relevant question: "'Did you lick X's vagina?' (where X was a four-year old girl)...In such an extreme case...there probably is no way to bring the C [control] questions up to the emotional level of the R [elevant] questions..." (Furedy, 1993, p.265). Actually, the "truth-control" question approach certainly exists here; one could ask control questions about oral sex with Y (i.e., a different 4-year old), bestiality, and so on, however one could also utilize a less arousing relevant question. Again, this is a practice vs. principle issue.
The question of whether control and relevant questions can be equilibrated in terms of arousal value may be best considered an empirical one. Lykken (1981, p.275) emphasizes the importance of testing prospective guilty knowledge items for a GKT on persons known to be ignorant of a crime's details. If the various multiple choices differ consistently in response production in known innocent, crime-ignorant subjects, (e.g., other police personnel, not involved in the case being investigated), the items are discarded. It would seem possible to accomplish the same thing with prospective control and relevant items in a CQT of the "truth control" variety. Granted, this is not now done in the field, but it would seem that it could be done. If the recommended GKT practice of equilibrating items for intrinsic arousal value can become widely accepted, then so can an analogous CQT practice, both based on sound scientific principle.
Finally, I think all would agree with Furedy (1993) that CQT practices which are patently damaging and/or unethical should be abandoned in lab as well as field. Indeed, the older form of the CQT, which is Furedy's (1993) typical CQT model, does have the problems he has repeatedly cited, and, unfortunately, it is probably still the "test" utilized by many federal and private polygraphers in the U.S. and Canada. However, his call for abandonment of an ethically-conducted research area containing intriguing theoretical questions (e.g., what are the cognitive and emotional concomitants of lying?), an area also holding a rare genuine example of applied psychophysiology, seems overstated. This is also a research area dealing with a universal psychological phenomenon, a fact which was emphasized by Lykken: "Deception is so pervasive in human intercourse that it is difficult to imagine a world in which dissembling had become impossible." (Lykken, 1981, p.51). CQT researchers include some of the few people who think seriously about deception. If we abolish CQT research, no new principles will be found to inform the ongoing practice of deception detection, which may then be free to perpetuate forensic psychomythology. In the Ben-Shakhar & Furedy (1990, p.27) volume, the authors decry a shortage of information about the CQT "...because little good appropriate research is available..." Why then would they want the research--in an ethically acceptable format--abandoned?
In their response to Furedy (1993), Honts et al. describe a novel derivative version of the CQT, the directed-lie test, and cogently demonstrate (with appropriate research citations) how this procedure overcomes most if not all of Furedy's (1993) objections to the original form of the CQT. My final point here, in noting the directed-lie procedure (which Furedy, 1993, did not mention) is that its development was the logical result of a long program of CQT research. This very promising new procedure may not have evolved had CQT research been previously "outlawed."
Elaad, E. (1990). Detection of guilty knowledge in real life criminal investigations. Journal of Applied Psychology, 75, 521-530.
Furedy, J.J. (1993). The "control" question "test" (CQT) polygrapher's dilemma: Logico-ethical considerations for psychophysiological practitioners and researchers. International Journal of Psychophysiology, 15, 263-267.
Honts, C.R., Kircher, J.C., & Raskin, D.C. (1995). Polygrapher's dilemma or Psychologist's chimaera: A reply to Furedy's logico-ethical considerations for psychophysiological practitioners and researchers. International Journal of Psychophysiology, 20, 199-207.
Johnson, M.M., & Rosenfeld, J.P. (1992). Oddball-evoked P300 -based method of deception detection in the laboratory. II: Utilization of non-selective activatiaon of relevant knowledge. International Journal of Psychophysiology, 12, 289-306.
Keeler (1930). Deception tests and the lie detectors. Proc. Int. Assoc. for Identification, 16, 186-193. (Reprinted in Polygraph (1994) vol. 23, pp. 134-144.)
Lykken, D.T. (1981). A tremor in the blood. New York: McGraw-Hill.
Podlesny, J.A. and Raskin, D.C. (1978). Effectiveness of techniques and physiological measures in the detection of deception. Psychophysiology, 15, 344-359.
Podlesny, J.A. (1993). Technical Paper. Is the guilty knowledge polygraph technique applicable in criminal investigations? A review of FBI case records. Crime Laboratory Digest, 20, 57-61.
Rosenfeld, J.P., Angell, A., Johnson, M., Qian, J. (1991). An
ERP-based, control-question lie detector analog: Algorithms for
discriminating effects within individual's average waveforms.
Psychophysiology, 18, 319-335.
Article Submitted 4 March 1996
Accepted for Publication 5 April 1996
Published 5 February 1997
Number of page accesses since 5 February 1997:
Return to the JCAAWP Home Page