Content uploaded by Charles Honts
Author content
All content in this area was uploaded by Charles Honts on Aug 30, 2019
Content may be subject to copyright.
ANALYZING IACONO’S THOUGHT EXPERIMENT 1
Analyzing Iacono’s Thought Experiment About Polygraph Field Studies:
Reason or Fantasy?
Charles R. Honts
Boise State University
and
Steven Thurber
Minnesota Department of Human Services
Author Note
Correspondence should be addressed to Charles R. Honts, Ph. D., Department of Psychological
Science, Boise State University, 1910 University Drive MS-1715, Boise ID 83725-1715. The
authors would like to thank Adela Stephanescu for her help in editing the completed manuscript.
Prepublication Manuscript:
Honts, C. R., & Thurber, S. (in press). Analyzing Iacono’s Thought Experiment
About Polygraph Field Studies: Reason or Fantasy? Polygraph & Forensic
Credibility Assessment: A Journal of Science and Field Practice, 48(2).!
ANALYZING IACONO’S THOUGHT EXPERIMENT 2
Abstract
We review and analyze a thought experiment first published in Iacono (1991) and reintroduced in
Iacono and Ben-Shakhar (2019). The Iacono Thought Experiment (ITE) appears to have used
backtracking methods to generate a series of assumptions and preconditions which would make
it possible to have a polygraph test with chance accuracy that produces a confession-criterion
field study with high accuracy. From this thought experiment, Iacono promulgated a hypothesis
that all polygraph confession criterion studies produce exaggeratedly high estimates of accuracy.
Our analysis of the assumptions and preconditions of the ITE found them to be unrepresentative
and highly unlikely to be met in real world settings. We used a converging evidence approach
that applied meta-analytic results, field studies that did not use a confession criterion, and data
from wrongful conviction cases that involved polygraph examinations to test the Iacono
hypothesis. We found strong falsification evidence to the Iacono hypothesis and conclude that it
should be abandoned as a meaning description of field polygraph research.
Keywords: deception detection, polygraph, Comparison Question Test
ANALYZING IACONO’S THOUGHT EXPERIMENT 3
Analyzing Iacono’s Thought Experiment About Polygraph Field Studies:
Reason or Fantasy?
Polygraph tests represent an important and widespread application of a psychological test in
law enforcement, national security, and employment around the world. Internationally, the
American Polygraph Association shows members from 62 countries. Zhang (2011) estimated
that there were as many as 8000 polygraph examiners operating in China alone. Despite the
ubiquitous nature of polygraph testing, it has received relatively little attention in academic
psychology and often, that attention has been in the form of negative commentary.
The most commonly used - and criticized - polygraph test is the Comparison Question Test
(CQT). The CQT comes in several variants, but in all cases, it monitors the subject’s autonomic
physiology (usually, respiration, electrodermal activity, relative blood pressure, and often
peripheral vasomotor activity) while the subject answers a series of questions. The question
series contains two categories of critical questions (usually two or three of each). Relevant
questions directly address the matter under investigation. Comparison questions are designed and
presented in such a way that every subject lies, or is assumed to lie, in their response to them
during the test. The subject’s physiological responses are expected to show an interaction, so
that subjects who are deceptive to the relevant questions show larger physiological responses to
relevant questions as compared to their comparison questions. Subjects who are being truthful to
the relevant questions are expected to show the opposite pattern, with physiological responses to
comparison questions being larger than those to relevant questions.
There are a number of reviews of CQT research, typified by but not limited to the following
examples: Raskin, Honts, and Kircher (1997), Iacono & Lykken (1997), The National Research
ANALYZING IACONO’S THOUGHT EXPERIMENT 4
Council (NRC, 2003), Honts (2004), Vrij (2008); American Polygraph Association (2011);
Raskin, Honts, and Kircher (2014), and Iacono and Ben-Shakhar (2018). There is variation
across the reviews, but nevertheless they generally produced overall accuracy estimates over
80%.
However, all of those reviews can be criticized for selective study choices and a lack of meta-
analytic scrutiny. None attempted to test moderator variables, although they sometimes reached
conclusions that hypothesized or even assumed powerful moderator effects. The NRC (2003)
report was particularly egregious in that regard. NRC found objectively high discrimination
estimating the area under the Receiver Operating Characteristic (ROC) curve (AUC) at 0.89. The
use of ROC analysis and AUC as an effect size has been criticized as an inappropriate
application of a technology developed to examine the performance of signal detection with
technology (specifically an operator's ability to view RADAR screens and distinguish enemy
ships, friendly ships, and noise; Tape, 2019) to psychology in general (Balakrishnan, 1999), and
to polygraph testing in particular (Honts & Schweinle, 2009).
Nevertheless, the NRC used AUC as their index of effect size, but what does AUC actually
mean and what does an AUC value of 0.89 imply about polygraph performance? The value of
AUC can range from 0.50, which represents chance performance, to 1.00 which represents
perfect performance (100% accuracy; Tape, 2019). Tape (2019) qualitatively characterizes AUC
values between 0.80 and 0.90 as indicating a good discriminator and AUC values above 0.90 as
excellent. Tables (Rice & Harris, 2005) and software (DeCoster, 2012) to convert between AUC
and other measures of effect size are readily available. Reference to those tables and software
show that an AUC value of 0.89 corresponded to a Cohen’s d value of 1.74 and an rpb of 0.66.
ANALYZING IACONO’S THOUGHT EXPERIMENT 5
Cohen (1969, 1988, 1992) described large effects in psychology as those with d values above
0.80 (corresponding rpb > 0.49). Cohen famously said that, in applied psychology, effect sizes of
d = 0.8 are “about as high as they come” (Cohen, 1988, p. 81). Thus, the AUC effect size
reported by NRC (2003) indicates extremely high performance for the CQT as compared to other
psychological tests and measures.
In spite of powerful empirical evidence of the usefulness of the CQT as a discriminator of
truth and deception, the NRC discounted those findings, saying the research methods were
substandard. To the present authors this seems to be an arrogant conclusion as the NRC
substituted their judgment about the qualities of research published in first tier peer-reviewed
journals of psychological science. Such a position is insulting to the editors of those first-tier
journals and the working scientists who peer-review for them. The NRC’s opinion is all the
worse for the fact that none of the members of the NRC committee who wrote the report had
ever published a study on deception detection.
Additionally, the NRC and others were notably critical of the use of experimental
(laboratory) studies for assessing CQT validity. Iacono & Lykken (1997) completely dismiss the
experimental research on the CQT, arguing that the real-world motivational contexts could not be
modeled experimentally and therefore laboratory results were qualitatively different from those
in real cases. However, the conclusions of the NRC and Iacono & Lykken (1997) about the
generalizability of the research methods in archival peer-reviewed journals of psychological
science and the generalizability of experimental CQT research should all be viewed as opinion
and not as fact, as none of those opinions were or are data-based.
ANALYZING IACONO’S THOUGHT EXPERIMENT 6
The history of academic disagreement over the accuracy (criterion validity) of the CQT is
long and has at times been polemic. Those disagreements are typified by, but are not limited to,
published exchanges between researchers from the University of Utah and the University of
Minnesota, beginning in 1978 in the journals Psychophysiology (Raskin and Hare, 1978; Lykken,
1978; Raskin, 1978) and Psychological Bulletin (Lykken, 1979; Raskin and Podlesny, 1979).
Direct exchanges in the literature between these groups continued until 2002 (Honts, Raskin, and
Kircher, 2002; Iacono and Lykken, 2002). Those disagreements were argued at a number of
different levels on various topics. Throughout the disagreements, the Minnesota group radically
rejected the notion that deception detection could be validly modeled in the laboratory and held
that the results of laboratory studies were not useful for estimating criterion validity in field
applications because they lacked external validity (generalizability). The Minnesota group holds
that position until today despite the general rejection of such criticisms across the entirety of
Psychological Science and specifically for deception detection. Hartwig and Bond (2014)
provided a general discussion about generalizability of laboratory studies and provided a specific
empirical rejection of differences between experimental and field settings, within a meta-analysis
of the interpersonal deception detection research literature.
The scientific issues surrounding the contrast of experimental and field settings for research
in interpersonal deception detection are nearly identical to those with the CQT. Recently, Honts
and Thurber (2019) reported a comprehensive meta-analysis of the CQT that followed the
analytic approach of Hartwig and Bond (2019). Honts and Thurber (2019) reported no
statistically detectable effects for moderators of motivation, subject population or setting
(experiment vs. field) in their comprehensive meta-analysis of the CQT.
ANALYZING IACONO’S THOUGHT EXPERIMENT 7
The Minnesota group was initially supportive of field studies that fit their criterion for useful
field studies. However, starting in the 1980s, field studies were published that produced high
levels of accuracy with the CQT (Honts & Raskin, 1988; Raskin, Kircher, Honts, & Horowitz,
1988). Those studies were specifically designed to meet the Minnesota group's criteria.
Subsequently, the Minnesota group rejected all field studies and took the radical position that
valid research on the CQT could not be conducted. One keystone of that position was a thought
experiment first reported by Iacono (1991) and then with some modification reintroduced in
Iacono and Ben-Shakhar (2018) . The Iacono (1991) thought experiment was originally
1
presented as follows:
Suppose that 800 crimes are being investigated using a polygraph technique that
operates with exactly chance accuracy; i.e. half of both the guilty and innocent
suspects will fail and half will pass. Because the polygraph is often used in crimes
for which there are multiple suspects, let us assume, without loss of generality,
that we are dealing with 800 two-suspect crimes, and that for each, one suspect is
guilty and the other innocent. Let us assume further that (1) the guilty suspect is
tested first 50% of the time, (2) the second suspect will not be tested if the result
of the first test indicates deception, (3) neither innocent suspects nor those guilty
suspects who pass the test will confess, and (4) 20% of the guilty who fail the test
and are subsequently interrogated confess. (Iacono, 1991, pp. 202-203).
In Iacono and Ben-Shakhar provided two simplified versions of the Iacono Thought Experiment with single
1
subjects and with paired subjects. Assumptions 3, 4, and 5 do not apply to either the single subject or the paired
tests as all subjects are tested regardless of the outcomes. Confession rates are not specified for either analysis thus
Assumption 7 is not specific. The other assumptions are either explicit or implied in the latter version of the Iacono
Thought Experiment.
ANALYZING IACONO’S THOUGHT EXPERIMENT 8
Thought experiments are well known in philosophy and science. Thought experiments can be
defined as “devices of the imagination used to investigate the nature of things” (Brown, 2014).
One of the most famous scientific thought experiments was Galileo’s reasoning that two objects
of different weight must fall at the same speed. Galileo’s thought experiment is easily validated
from observations, such as the conclusive demonstration by Neil Armstrong on the moon when
he dropped a feather and a hammer simultaneously and they landed on the surface at the same
time (Pigllucci, 2006). Pigllucci further notes that thought experiments can also be wrong and be
falsified by data. Had Galileo’s thought experiment been invalidated by data, it would have been
lost to history and forgotten.
Thought experiments can take on a number of forms or types. While a discussion of the
multiple types of thought experiments is beyond the scope of this paper, it is worth noting that
Iacono’s thought experiment appears to be a type known as Backcasting (Robinson, 1982). In
Backcasting one imagines a desired or possible state of the world and then reasons backward
from that end-state to the necessary precursors. By definition, such logic necessarily does not
provide a description of reality, it only provides a chain of precursors that might produce the
desired end-state. Such thought experiments, like all thought experiments, are useful in the real
world only to the extent that they can be tested and validated or falsified with data. We begin our
analysis of the Iacono Thought Experiment (ITE) by defining the hypothetical precursors that he
either invented or selected to reach the desired end-state where polygraph tests with chance
accuracy could produce a field study with high accuracy rates.
ANALYZING IACONO’S THOUGHT EXPERIMENT 9
Elucidation and Analysis of the Hypothetical Preconditions and Assumption of the Iacono
Thought Experiment
Explicit Assumptions of the Iacono (1991) Thought Experiment.
Iacono (1991) makes a number of explicit assumptions that were used to create a possible path to
the desired end state.
1. Eight hundred subjects are tested where 400 are Innocent and 400 are Guilty.
2. The polygraph preforms exactly at chance accuracy of 50% correct, 50% incorrect, and
no inconclusive outcomes. This assumption is part of the overall desired end-state where a
chance polygraph test could produce high accuracy outcomes. All of the other assumptions
also serve the establishment of that end-state.
3. Each crime has only two suspects. (Iacono makes this assumption and states that it is
made “without loss of generality” (p. 203).
4. The Guilty suspect is tested first in half of the cases
5. If the first suspect fails the polygraph test, the second suspect will not be tested.
6. Neither innocent nor guilty suspects who pass the test will confess.
7. Only 20% of the Guilty suspects who fail and are interrogated will confess.
Implicit Assumptions of the Iacono Thought Experiment. The following implicit
assumptions are also necessary for the mathematics and logic of the Iacono Thought Experiment
to reach the desired end-state.
8. The polygraph is the only source of information about who is guilty in a criminal case.
9. Guilty people only confess after polygraph examinations.
ANALYZING IACONO’S THOUGHT EXPERIMENT 10
Analysis of the Explicit Assumptions of the Iacono Thought Experiment
Assumption 1 is that the base rate of guilty to innocent subjects is equal. The base rate of
guilt in a criminal case will vary greatly depending upon when the polygraph is used. If it is used
early in an investigation, there are likely to be far more innocent than guilty subjects; if it is used
very late in an investigation, there may be many more guilty than innocent subjects. The
assumption of equal base rates is acceptable for a thought experiment as long as one recognizes
that variations in the base rate could dramatically alter the end state results and that a base rate of
50% will be unusual in actual practice.
Assumption 2 is that the polygraph performs exactly as a coin flip. This assumption is made
as a premise of the thought experiment and it is a necessary component of the desired end state.
However, this premise is without empirical support in the real world. To our knowledge, there
are no studies that show any version of the CQT to perform at chance levels.
Assumption 3 is that each case has only two suspects. This premise simplifies the
mathematics necessary to achieve the desired end sate of the Iacono thought experiment, but it is
a premise that is rarely met in the real world, and is not at all representative of the field at the
time Iacono (1991) was written (for example, Honts & Raskin, 1988, Raskin, Kircher, Honts, &
Horowitz, 1988 all contain many single and multiple suspect cases as do the more recent field
studies). Iacono’s assertion that this assumption is made without a loss of generality (for the
Backcasting thought experiment) is clearly not supported by data.
Assumption 4 is that a guilty suspect is tested first in half of the cases. This assumption is
tenable only if there are only two suspects and that the examiner has no reason to test one or the
ANALYZING IACONO’S THOUGHT EXPERIMENT 11
other suspect first. It is a convenient assumption for the thought experiment, but it is unlikely to
be widely representative of field polygraph testing.
Assumption 5 states that if the first person is tested and fails the second suspect will not be
tested. However, this is not the case in real investigations. If the first subject is tested and fails
but does not confess, then the remaining suspect or suspects will likely be tested to assess their
involvement in the crime. The likely logic of investigators would be that the additional suspect(s)
would not be suspects unless there was a reason to suspect them, and they may well be involved.
In our experience it is, in fact, common practice to test all suspects in a case during an
investigation.
Assumption 6 states that neither innocent suspects nor guilty suspects who pass the test will
confess. This is manifestly not true. Under certain circumstances, such as a wrongfully failed or
deliberately misrepresented polygraph test result, innocent suspects will confess to crimes they
did not commit. The White Paper of the American Psychology Law Society (Kassin, Drizin,
Grisso, Gudjonsson, Leo, & Redlich, 2010) specifically notes that wrongfully failed or willfully
misrepresented polygraph outcomes are a powerful false evidence ploy that puts the actually
innocent at increased risk of false confession. Moreover, Bonpasse (2013), provides examples
and discussion of actual cases where incorrect or misrepresented polygraph outcomes have
contributed to miscarriages of justice through their role in eliciting false confessions.
Assumption 6 is also incorrect for guilty suspects who pass polygraphs, as it ignores the fact that
investigations rarely stop just because a polygraph has been passed. If the subsequent
investigation continues and additional information is obtained, then the suspect will likely be
interviewed a second time, despite the passed polygraph, and may provide a confession then or
ANALYZING IACONO’S THOUGHT EXPERIMENT 12
confess later as part of a plea bargain. At least one such case was included in Honts & Raskin
(1988).
Assumption 7. This assumption states that only 20% of the guilty suspects who are
interrogated will confess. The choice of 20% is arbitrary and has no empirical basis. The actual
confession rate will depend upon the situation in which the tests were conducted. Polygraph
tests conducted for defense attorneys, or by the police on subjects who have defense counsel, are
unlikely to be followed up with interrogations regardless of the polygraph outcome. On the other
hand, the U. S. Department of Defense (2002) has reported data indicating that in one polygraph
program, more than 90% of the failed polygraph examinations resulted in relevant admissions.
Clearly, the rate chosen for this assumption will have a major impact on the resultant outcomes
of the ITE. Moreover, since all confession rate values are situationally specific, it is non-sensical
to provide a single value for central tendency as such a value would be meaningless to any
specific applied setting.
Analysis of the Implicit Assumptions of the Iacono Thought Experiment
Assumption 8 asserts that the polygraph is the only source of information about who is
guilty in a criminal investigation. This ignores that fact that individuals can confess in contexts
other than polygraph examinations, or that other incontrovertible evidence of guilt or innocence
may be obtained independent of the polygraph test. This was examined explicitly in Honts
(1996) and no differences in numerical scores or accuracy rates were found between confession
confirmed and evidence confirmed cases for guilty or innocent subjects. Interestingly, in Honts
(1996) none of the innocent subjects were confirmed by a confession obtained in the context of a
ANALYZING IACONO’S THOUGHT EXPERIMENT 13
polygraph examination. Assumption 8 is preposterous on its face, but the ITE cannot work
without it.
Assumption 9 is that guilty people only confess following polygraph tests. As covered in
our discussion of Assumption 6, we noted that inaccurate and misrepresented polygraph tests can
result in false confessions from innocent subjects. We also noted that even guilty subjects who
pass polygraph tests will sometimes confess later, when faced with new or overwhelming
evidence. Moreover, guilty suspects, and occasionally innocent subjects, will confess as part of a
plea bargain. Thus, it is obviously true that guilty people who fail a polygraph test but either are
not interrogated, or initially resist an interrogation, may confess later. At least three field studies
have explicitly taken this into account and looked for confirming information in an exhaustive
sample of cases within a particular period of time, and used all of the information available not
only in the polygraph examination file, but in the complete police record of the case (Honts,
1996; Patrick & Iacono, 1991; Raskin et al., 2019).
Summary of the Analysis of Preconditions and Assumptions
Our analysis shows that the assumptions of the Iacono thought experiment were generally
chosen without reference to data or professional practice, in the service of developing what
became a highly improbable set of preconditions and assumptions leading to a specific solution
showing that a polygraph test with chance accuracy could produce a field study with high
accuracy rates. The Iacono thought experiment was then transmogrified into a normative
statement that all field studies of the CQT were, are, and forever will be unreliable and
overestimate actually accuracy. We do not believe that this normative conclusion is justified
unless it can withstand empirical examination and falsification. For the remainder of this paper
ANALYZING IACONO’S THOUGHT EXPERIMENT 14
we will refer to the hypothesis derived from the Iacono thought experiment, that the CQT is no
more accurate than chance and that all confession criterion studies are biased to dramatically
overestimate the accuracy of the CQT as the Iacono Thought Experiment Hypothesis (ITEH).
Data That Could Falsify the Iacono Thought Experiment Hypothesis (ITEH)
Just like Galileo’s thought experiment concerning falling objects, the ITEH survives the test
of science based upon a lack of falsification data in the scientific research. This leads to the
question of what data would falsify the ITEH? The remainder of this paper addresses several
sources of converging data that do, in fact, lead to the conclusion that the ITEH is false.
Convergence of Experimental and Field Data Without Detectable Moderator Effects.
Recent studies summarized by Hartwig and Bond (2014) have indicated strong convergence
between experimental and field studies in psychological science and interpersonal deception
detection. Hartwig and Bond explicitly rejected the notion that experiments and field research
on interpersonal deception detection produced significantly different results. If the ITEH were
true, then polygraph testing would have to be qualitatively different underlying mechanisms from
interpersonal deception. Under such circumstances we would expect that laboratory studies of
the CQT would produce dramatically lower accuracies than the (according to the ITEH)
exaggeratedly high accuracies produced by the supposedly unavoidable effects of the ITEH on
field studies of the CQT. Existing reviews simply do not reveal dramatically more accurate
results in field than in the laboratory (NRC, 2003; Honts & Thurber, 2019).
ANALYZING IACONO’S THOUGHT EXPERIMENT 15
Lack of Differences in Accuracy Between Field Studies that Rely on the Confession
Criteria and Those That Do Not.
Since the ITEH is critically bound to the use of confessions as a criterion of confirmation of
Guilt and Innocence, and the ITEH predicts that the confession criterion critically biases field
studies to show high accuracy, we should expect that field studies that included or used other
methods of confirmation would produce accuracies that approach chance levels of accuracy.
Empirically, this is simply not the case. Honts (1996) directly tested this hypothesis, rating
strength of confirmation on a scale that ranged from confessions with the generation of new
evidence at one end of the scale to no confirmation at the other. Honts (1996) tested that scale
against decision accuracy and against numerical scores. In direct opposition to the predictions of
the ITEH, Honts found no effects for the level of confirmation. That is, confession confirmed
cases did not have higher accuracy levels than cases that were confirmed by methods other than
confession (physical evidence and/or witness statements).
Similarly, there are two field studies that use paired testing and mathematics to estimate
accuracy (Ginton, 2013; Mao, Liang, & Hu, 2015). This paired testing approach, while not
without problems (Iacono & Ben-Shakhar, 2018), is not dependent upon confessions and so is
outside the scope of the ITEH. Estimated accuracy rates from the paired subjects studies
converge with data from both laboratory and field studies and thus provide support for both.
Lack of Concurrence Between Wrongful Convictions and Failed CQT Polygraph Tests.
If the ITEH is correct that CQT polygraphs are no more accurate than chance, we would
expect that, on average, half of the innocent subjects tested in criminal justice settings would
produce false positive errors. In the criminal justice settings, innocent subjects who failed the
ANALYZING IACONO’S THOUGHT EXPERIMENT 16
polygraph would be exposed to interrogation and thus put at risk of false confession. Under such
circumstances, we would expect there to be a relatively large number of false positive outcomes
among the ranks of the wrongfully convicted. Bonpasse (2013) reviewed the case files of the
National Registry of Exonerations which was founded in 1989 as a joint project of the
2
University of Michigan and Northwestern University Law School. Bonpasse reported finding
215 exoneration cases where polygraph tests were involved. Of those 215 cases only 23 (10.7%)
contained information that an Innocent subject had been tested before trial and had failed the
polygraph. However, there were 44 (20.5%) Innocent subjects who had been tested with the
polygraph before trial, produced truthful outcomes, but those favorable outcomes did not help
them avoid wrongful conviction. Although the ITEH predicts that false positive errors should be
common among the wrongfully convicted, they occurred at only half the rate of true negative
outcomes. Bonpasse also reported that across all testing, before and after trial and including tests
of the immediate suspect and others (co-defendants and witnesses), 135 (62.9%) of the
polygraph test outcome were favorable to the wrongfully convicted person while only 31
(14.4%) produced unfavorable outcomes. Data from the wrongfully convicted strongly
contradicts and thus falsifies ITEH.
Discussion
To our knowledge, there is not a single study of the CQT, either laboratory or field, that
produced chance accuracy rates. While there is a substantial amount of variability between
studies of the CQT, no review has found that laboratory studies are dramatically less accurate
than field studies (NRC, 2003; Honts & Thurber, 2019). Thus, the ITEH completely lacks
http://www.law.umich.edu/special/exoneration/Pages/about.aspx
2
ANALYZING IACONO’S THOUGHT EXPERIMENT 17
empirical substantiation. Moreover, data from the Honts & Thurber (2019) meta-analysis, field
studies that do not use the confession criterion, and the wrongfully convicted all provide
evidence that the ITEH is false. The results of the ITE are therefore seen as a failed thought
experiment that is completely without empirical support, and which should be relegated to the
trash heap of history’s failed ideas.
ANALYZING IACONO’S THOUGHT EXPERIMENT 18
References
American Polygraph Association (2011). Meta-analytic survey of criterion accuracy of validated
polygraph techniques. Polygraph, 40, 194-305.
Balakrishnan, J. D. (1999). Decision processes in discrimination: Fundamental
misrepresentations of signal detection theory. Journal of Experimental Psychology:
Human Perception and Performance, 25, 1189–1206.
Bonpasse, M. (2013). Polygraph and 215 wrongful conviction exonerations. Polygraph, 42,
112-127.
Brown, J. R. (2014). Thought experiments. Stanford Encyclopedia of Philosophy. Retrieved from
https://plato.stanford.edu/entries/thought-experiment/
Cohen, J. (1969). Statistical power analysis for the behavioral sciences. New York: Academic
Press.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd Edition). Hillsdale
NJ:Erlbaum.
Cohen, J., (1992). A power primer. Psychological Bulletin, 122, 155-159.
DeCoster, J. (2012). Converting effect sizes 2012-06-19-4.xls Retrieved from http://www.stat-
help.com/spreadsheets.html
Ginton, A. (2013). A non-standard method for estimating accuracy of lie detection techniques
demonstrated on a self-validating set of field polygraph examinations. Psychology,
Crime, & Law, 19, 577-594. DOI: http://dx.doi.org/10.1080/1068316X.2012.656118
Hartwig, M., & Bond, C. F. (2014). Lie detection from multiple cues: A meta-analysis. Applied
Cognitive Psychology, 28, 661-676.
ANALYZING IACONO’S THOUGHT EXPERIMENT 19
Honts, C. R. (1996). Criterion development and validity of the control question test in field
application. The Journal of General Psychology, 123, 309-324.
Honts, C. R. (2004). The psychophysiological detection of deception, in P. Granhag and L.
Strömwall (Eds.) Detection of deception in forensic contexts. London: Cambridge
University Press 103-123.
Honts, C. R., & Raskin, D. C. (1988). A field study of the validity of the directed lie control
question. Journal of Police Science and Administration, 16, 56-61.
Honts, C. R., & Schweinle, W., (2009). Information gain of psychophysiological detection of
deception in forensic and screening settings. Applied Psychophysiology and Biofeedback,
34, 161-172. (Available online July 2009)
Honts, C. R., & Thurber, S. (2019, March). A Comprehensive Meta-Analysis of the Comparison
Question Polygraph Test. Paper presented at the annual meeting of the American
Psychology Law Society, Portland, Oregon.
Honts, C. R., Raskin, D. C., & Kircher, J. C. (2002). The scientific status of research on
polygraph techniques: The case for polygraph tests. In, D. L. Faigman, D. Kaye, M. J.
Saks, & J. Sanders (Eds.) Science in the Law: Social and Behavioral Sciences Issue,
American Casebook Series (pp. 598-634). West Group: St. Paul, Minnesota.
Iacono, W. G. (1991). Can we determine the accuracy of polygraph tests? In J. R. Jennings, P.K.
Ackles & M. G. H. Coles (Eds.) Advances in psychophysiology (pp. 201-207). London,
UK: Jessica Kingsley Publishers.
ANALYZING IACONO’S THOUGHT EXPERIMENT 20
Iacono, W. G., & Ben-Shakhar, G. (2019). Current Status of forensic lie detection with the
comparison question test: An Update of the 2003 National Academy of Sciences report
on polygraph testing. Law and Human Behavior, 43, 86-98.
Iacono, W. G., & Lykken, D. T. (1997). The scientific status of research on polygraph techniques:
The case against polygraph tests. In, D. L. Faigman, D. Kaye, M. J. Saks, & J. Sanders
(Eds.) Science in the Law: Social and Behavioral Sciences Issue, American Casebook
Series (pp. 582-618-). West Group: St. Paul, Minnesota.
Iacono, W. G., & Lykken, D. T. (2002). The scientific status of research on polygraph techniques:
The case against polygraph tests. In, D. L. Faigman, D. Kaye, M. J. Saks, & J. Sanders
(Eds.) Science in the Law: Social and Behavioral Sciences Issue, American Casebook
Series (pp. 634-688). West Group: St. Paul, Minnesota.
Kassin, S. M., Drizin, S. A., Grisso, T., Gudjonsson, G. H., Leo, R. A., & Redlich, A. D. (2010).
Police-induced confessions: Risk factors and recommendations. Law and Human
Behavior, 34, 3-38.
Lykken, D. T. (1978). The psychopath and the lie detector. Psychophysiology, 15, 137–142.
http://dx.doi.org/10.1111/j.1469-8986.1978.tb01349.x
Lykken, D. T. (1979). The detection of deception. Psychological Bulletin, 86, 47–53. http://
dx.doi.org/10.1037/0033-2909.86.1.47
Mao, Y., Liang, Y., & Hu, Z. (2015). Accuracy rate of lie-detection in China: Estimate the
validity of CQT on field cases. Physiology & Behavior, 140, 104-110.
National Research Council (2003). The Polygraph and Lie Detection. Washington, DC: The
National Academies Press.
ANALYZING IACONO’S THOUGHT EXPERIMENT 21
Patrick, C. J., & Iacono, W. G. (1991). Validity of the control question polygraph test: The
problem of sampling bias. Journal of Applied Psychology, 76, 229-238.
Piglucci, M. (2006). What is a thought experiment, anyhow? Philosophy Now: A magazine of
Ideas, 58, Retrieved from
https://philosophynow.org/issues/58/What_is_a_Thought_Experiment_Anyhow
Raskin, D. C. (1978). Scientific assessment of the accuracy of deception of detection: A reply to
Lykken. Psychophysiology, 15, 143-147.
Raskin, D. C., & Hare, R. D. (1978). Psychopathy and detection of deception in a prison
population. Psychophysiology, 15, 126-136.
Raskin, D. C., & Podlesny, J. A. (1979). Truth and deception: A reply to Lykken. Psychological
Bulletin, 86, 54–59. http://dx.doi.org/10.1037/0033-2909.86.1.54
Raskin, D. C., Honts, C. R., & Kircher, J. C. (1997). The scientific status of research on
polygraph techniques: The case for polygraph tests. Chapter in, D. L. Faigman, D. Kaye,
M. J. Saks, & J. Sanders (Eds.) Modern scientific evidence: The law and science of
expert testimony (pp. 565-582).
Raskin, D. C., Honts, C. R., & Kircher, J. C. (2014). Credibility assessment: Scientific research
and applications. Oxford, UK: Academic Press. ISBN: 978-0-12-394433-7 (ebook
version available online 17 December 2013).
Raskin, D. C., Kircher, J. C., Honts, C. R., & Horowitz, S. W. (2019). A study of the validity of
polygraph examinations in criminal investigations. Final report to the National Institute
of Justice, Grant Number 85-IJ-CX-0400. Polygraph & Forensic Credibility Assessment:
A Journal of Science and Field Practice, 48, 10-39.
ANALYZING IACONO’S THOUGHT EXPERIMENT 22
Rice, M. E. and Harris, G. T. (2005). Comparing effect sizes in follow-up studies: ROC Area,
Cohen’s d, and r. Law and Human Behavior, 29, 615-620.
Robinson, J. B. (1982) Energy backcasting: A proposed method of policy analysis. Energy
Policy, December.
Tape, T. G. (2019). Interpreting Diagnostic Tests. University of Nebraska Medical Center.
Retrieved from, http://gim.unmc.edu/dxtests/Default.htm also referencing subordinate
web pages on the same topic at this URL.
U. S. Department of Defense (2002). Department of Defense Polygraph Program Annual Report
to Congress, Fiscal Year 2002. Office of the Assistant Secretary of Defense (Command,
Control, Communications, and Intelligence.
Vrij, A. (2008). Detecting Lies and Deceit: Pitfalls and Opportunities, Second Edition.
Chichester, UK: Wiley.
Zhang, X. (2011). The evolution of polygraph testing in the People’s Republic of China.
Polygraph, 40, 181-193.
