Detecting Deceptive Responses in Sex Offenders

Detecting Deceptive Responses in Sex Offenders: A Comparison of
Layered Voice Analysis (LVA) and the Polygraph
J. Michael Adler, Ph.D.
Introduction
The polygraph has become a popular tool in sex offender treatment and supervision
planning. Most researchers and clinicians working in the field of sex offender assessment,
treatment, and management recognize the poor credibility of offender self-report (Baker,
Skolnick, Doucette, Levitt, & O’Connor, 2005; Maletzky, 1990; Salter, 1988; Scott, 1997; Steen,
2002). Collateral information, although more reliable than self-report, is often incomplete and
void of important information concerning the offender’s sexual acting out. Even official records
do not accurately report the offender’s history of sexual offending. As a result, the polygraph has
become an increasingly popular tool in attempting to fill in deficits in the offender’s history,
frequency, and scope of their sexual acting out (Abel & Harlowe, 1987; Ahlmeyer & Heil,
Mckee, & English, 2000; Blasingame, 1998; Brake & Shannon, 1997; English & Heil, 2006;
Freeman-Longo & Walls, 1986; Jenson & Jewell, 1988; McGrath, Cumming, Hoke, & BonnMiller, 2007).
The use of the polygraph as a clinical tool in the treatment and supervision of sexual
offenders has been described in the literature. There are three primary types of polygraph
examinations used in sex offender treatment programs. During assessment or early in treatment,
Full Disclosure examinations are used in validating offender self-report regarding sexual deviant
history. A second type of polygraph is used to monitor offender behavior during the course of
treatment and supervision. Maintenance examinations are given periodically, typically every six
months to verify the offender’s compliance with treatment and supervision guidelines in an effort
to increase community safety. The third type of polygraph examination is a Specific or Instant
Offense examination. An instant offense examination (specific issue test) is given when an
offender is in denial regarding their offense of conviction or a specific behavior for which he/she
is denying (Abrams, 1989; Abrams 1992; Baker, et.al., 2005; Blasingame, 1998; Carnes 1995).
Recently layered voice analysis (LVA) has been introduced as an alternative method of
truth verification. Although LVA is currently being used in several capacities, there has been
little independent research on the ability of LVA as a reliable and valid method of truth
verification. To date there has been no independent research on LVA as a reliable and valid
method of truth verification in the field of sexual offender assessment and treatment. Therefore,
the present study was directed toward investigating the effectiveness of LVA in comparison to
the polygraph in detecting deceptive responses from adult and adolescent sex offenders.
Detecting Deceptive Responses in Sex Offenders:
2
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
The polygraph is presently considered the best tool available for detecting deceptive
responses from sexual offenders. The validity of the polygraph has been studied extensively.
Forensic Research, Inc. reviewed 80 research projects published since 1980 involving over 6,380
polygraph examinations for the purpose of assessing the validity and reliability of the polygraph.
Based on this review, polygraph tests on average were 98% accurate (Forensic Research Inc.,
1997).
The use of the polygraph as a method to enhance assessment, treatment, and community
safety with convicted sex offenders has been well documented (Abel & Harlowe, 1987; Abrams,
1989; Abrams, 1991; Abrams, Hoyt, & Jewell, 1991; Abrams & Ogard, 1986; Freeman-Longo &
Walls, 1986; Jenson & Jewell, 1988). It was introduced as a tool in supervising individuals on
probation as early as the 1960’s. Reports of significant success in reducing aberrant behavior
resulted in an expansion of polygraph use to other areas. In 1986, Abrams and Ogard compared
one group of probationers tested periodically using the polygraph with another group of
probationers not tested with the polygraph. They concluded the use of polygraph testing deterred
reoffending. Following the success of the polygraph in supervision, the use of the polygraph
expanded into treatment. Clinicians hoping to break down denial, improve assessment of risk,
and develop more effective treatment began using the polygraph to evaluate the offender’s selfreport. Research looking at the impact polygraph testing had on offender disclosures found
significant increases in the number of offenses and victims disclosed (Ahlmeyer, et.al., 2000;
Emerick & Dutton, 1993). The polygraph measures physiologic changes associated with the sympathetic nervous
system. When the sympathetic nervous system is activated, changes occur in heart rate,
respiration, blood pressure, and skin perspiration. The polygraph measures emotional arousal
associated with the sympathetic nervous system. During a polygraph examination, the person
being tested is connected to sensors which measure respiration (pneumograph), electrodermal
changes due to perspiration (galvanic skin response), blood pressure, and heart rate.
The polygraph test involves four components. The first involves data collection.
Following a review of available information, the examiner conducts a pre-test interview. The
experience and ability of the examiner to develop rapport, collect relevant information regarding
the subject’s present situation, and knowledge of the details of the issue in question are the
essential objectives of the pre-test interview. Following the pre-test interview, the relevant
questions are developed and reviewed with the subject. Relevant questions focus on the issue in
question. In addition to relevant questions, control questions are also developed. Control
questions are developed in which it is assumed the subject is lying. Due to the “accusatory
nature” of the polygraph, it is assumed that the person being tested desires to look good. It is
assumed that the person then will deny bad behavior. Based on these assumptions, acceptable
Detecting Deceptive Responses in Sex Offenders:
3
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
control questions are questions that it is “highly likely” that the person is lying (Abrams, 1992).
Based on the analysis, the test findings are reported as no deception indicated, deception
indicated, or inconclusive (Williams, 1995).
The polygraph provides many benefits to the supervision and treatment of sexual
offenders. The use of the polygraph improves the accuracy of information the offender provides
the treatment provider. English (1998) reported that in a sample of 83 sex offenders who were
polygraphed, 48% had crossed over in age or gender of the victims they offended. Ahlmeyer et.
al. (2000) reported prior to polygraph testing offenders self-reported an average of two victims.
Following the inclusion of polygraph testing, the average number of victims increased to 165.
The use of the polygraph increases the effectiveness of community supervision. Abrams
and Ogard (1986) compared two groups of criminal offenders including sex offenses. One group
was supervised with the polygraph and one group without the polygraph. Seventy-one percent of
the offenders supervised with the polygraph were successful as compared to forty-three percent
of the offenders supervised without the polygraph. Arther (1990) described six benefits of the
use of polygraph including the reduction of crimes committed. The awareness of a future
polygraph has been found to decrease the number of high risk behaviors the offender exhibited in
the community. They also found an increase in disclosures to treatment providers and probation
officers when utilizing the polygraph (Baker, et. al., 2005; Grubin, Madsen, Parsons, Sosnowski,
& Warberg, 2004).
Recently new technology focusing on emotion vocal analysis calculated from a series of
complex algorithms has been developed to provide the user with a fast, accurate access to truth
verification. Layered voice analysis (LVA) uses a series of algorithms to detect and measure
different types of stress and then determines whether the stress is caused from a lie, excitement,
an exaggeration or cognitive conflict. LVA technology is a wide spectrum frequency based
software used in real-time interviews or with pre-recorded material.
According to information supplied from the company that developed the technology,
LVA detects minute, involuntary changes in the voice which reflect various types of brain
activity including emotional activity and information processing states. Using a wide range
spectrum analysis to detect inaudible changes in speech waveform, LVA detects anomalies in
brain activity and classifies them as stress, excitement, deception, and other emotional states. In
clinical terms, changes in cortical perception and interpretation of events manifest as changes in
speech waveform. By comparing the changes to a baseline that has been recorded for a given
individual, one can make sense of the usual and unusual brain activities as manifested in the
voice. Unlike Voice Stress Analysis (VSA or CVSA) which uses micro-tremors that has been
theorized to occur in the frequency range of 8 – 12 Hz, LVA detects micro-changes from normal
voice modulations that are not physiological in nature. LVA monitors changes in the whole
Detecting Deceptive Responses in Sex Offenders:
4
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
range of frequencies and calculates a truthful/emotion free baseline for 129 various proprietary
parameters of voice. Variations from these are measured using several different algorithms to
produce an interpretable finding (Robbins, 2006).
The present study compared the LVA’s interpretation of adult and adolescent sex offenders’ selfreport regarding their sexual offending behavior to the polygraph’s interpretation using a double
blind research method.
METHOD
Participants
Participants in the study were 103 adult (N=80) and adolescent (N=23) sex offenders who
were in treatment or referred for a sex offender risk assessment in a community sex offender
treatment program. 99 of the participants were Anglo/white (97%), two African American (2%),
and one (1%) Hispanic. Eighteen offenders identified to participate in the study were not
included due to not completing a polygraph examination. The adult offenders ranged between 18
and 80 years of age, with a mean age of 39.7. The adolescent offenders ranged between 13 and
17 years of age, with a mean age of 14.9. In this sample, 24 of the participants were in treatment
and had received an average of 20 months of outpatient sex offender treatment. Three of the
adolescents were in treatment and averaged 6 months of sex offender treatment. The outpatient
treatment program utilizes a cognitive/behavioral based treatment program. Adult offenders
receive two hours of group therapy weekly, one hour of family therapy bi-weekly, and one hour
of psycho-educational skills group weekly. Adolescent offenders received 1 ½ hours of group
weekly, one hour of family therapy bi-weekly, and one hour of individual therapy weekly.
The adult offenders in this study (N=80) included 63 convicted of a hands on sexual
offense against a child (79%), ten convicted of child pornography (12%), and seven convicted of
statutory rape (9%). Offender categories were determined by their legal conviction. Of the ten
child pornographers, all reported hands on sexual contact with children. All of the offenses of the
adolescent sexual offenders involved sexual acting out on younger children.
Apparatus
Polygraph testing was conducted by one independent examiner who contracts services
with the treatment center. The examiner is licensed in the State of Tennessee and is an approved
Detecting Deceptive Responses in Sex Offenders:
5
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
examiner by the Tennessee Sex Offender Treatment Board. The polygraph equipment utilized in
this study was the Axciton Computerized Polygraph System and the Lafayette Instrument LX4000, both using the Johns Hopkins Applied Physics laboratory Polyscore computer scoring
algorithm for purposes of score reliability and validity. Polygraph tests consisted of measuring
the offender’s cardiovascular, respiratory, and galvanic skin resistance using a control question
test format.
Layered Voice Analysis testing was conducted by one examiner trained by Voice
Analysis Technologies to conduct LVA examinations. Given the relatively new technology, no
licensure exists. The examiner is an approved treatment provider by the Tennessee Sex Offender
Treatment Board. The LVA equipment utilized in this study was a Dell Inspiron 1520 laptop
computer and LVA software version 6.50. LVA 6.50 technology is a wide spectrum frequency
based software that detects involuntary changes in the voice which reflect varying types of brain
activity, including emotional activity and information processing states. The wide range
spectrum analysis detects inaudible changes in the speech waveform and classifies them in terms
of stress, excitement, and deception. These changes are then compared to a baseline and
analyzed using 8,000 mathematical algorithms applied to 129 voice frequencies. The technology
developed and patented by Nemesysco of Israel measures voice modulations in time and
frequency domain to detect micro-changes from their normal appearance that are not
physiological in nature. All LVA examinations were conducted using off-line method for prerecorded material.
Procedure
Participants in the study were referred for a polygraph as part of a sex offender risk
assessment or completed a maintenance polygraph as part of their normal treatment/supervision
requirements. The clinical interview conducted as part of the risk assessment was audio-taped as
part of normal protocol. The clinical interviews were conducted by a licensed clinician approved
by the Tennessee Sex Offender Treatment Board to complete sex offender evaluations. The
clinician was not part of the study. The type of polygraph (full disclosure, maintenance, and/or
specific) was determined by the information the clinician wanted to verify. The polygraph
examiner in collaboration with the clinician developed relevant questions necessary to verify the
truthfulness of the offender’s self-report concerning disclosures in three areas; (1) the number of
victims and/or offenses (Full disclosure test), (2) offenses that may have occurred since the
offender’s last exam (Maintenance test), (3) a specific instant offense behavior that the offender
was denying (Specific test). Issues related to whether the offender had disclosed all
victims/offenses resulted in a Full Disclosure test. Issues associated with whether the offender
had sexually re-offended over a period of time resulted in a Maintenance test. Finally, issues
Detecting Deceptive Responses in Sex Offenders:
6
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
related to a specific incident/behavior resulted in a Specific test. Once the relevant questions
were developed, the clinician provided the relevant questions to the LVA examiner as well as the
audio-taped clinical interview. The polygraph test and analysis was completed by a separate
individual independent and without knowledge of the LVA analysis. The analysis of the
polygraph resulted in a finding of (1) deception indicated, (2) no deception indicated, or (3)
inconclusive. The LVA analysis was also completed independently and without knowledge of
the polygraph analysis. The LVA analysis resulted in a finding of (1) deception/false statement
or (2) no deception/truthful. Once the clinician had received both analyses, the results were
recorded and provided for comparison.
RESULTS
The criterion for inclusion in the final analysis was completion of both the polygraph and
the LVA examination. Eighteen offenders who were initially identified to participate in the study
were not included in the comparison because they did not complete a polygraph. Of the 121
identified offenders, 103 offenders completed both the polygraph and the LVA examination (80
adult offenders and 23 adolescent offenders), thus were included in the final analysis. Of the 103
polygraph and LVA examinations conducted, 73 Full Disclosure tests were completed, 24
Maintenance tests were completed, and six Specific tests were completed. Regarding the
polygraph analysis, 68 of the polygraph tests were not deceptive (66%), 27 were deceptive
(26%), and eight tests were inconclusive (8%). Regarding the LVA analysis, 68 tests were not
deceptive (66%) and 35 tests were deceptive (34%). No LVA tests were inconclusive.
In order to determine whether the LVA was effective in determining deception with
sexual offenders, 103 adult and adolescent sex offenders were tested by both the polygraph and
the LVA independently. The test results for each offender were described as deceptive, not
deceptive, or inconclusive (polygraph only). The result of the polygraph was then compared to
the result of the LVA using a Chi-square. Results of the Chi-Square support the hypothesis that
there was no significant difference in the outcome results between the polygraph and the LVA
tests. Table I presents the results of the polygraph and LVA tests for the total sample. A one
variable Chi-Square test (χ² = 12.60, p < .0001) determined there were no differences between
the findings of the polygraph and LVA examinations
Detecting Deceptive Responses in Sex Offenders:
7
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
Table I. Results of Polygraph and LVA Tests for total Sample
Total Test completed (N= 103)
Test Outcome
Polygraph
LVA
No Deception
68
68
Deception Indicated
27
35
Inconclusive
8
0
A comparison of the polygraph and LVA results to specific age groups did not demonstrate
significant differences. 80 tests were administered to adult offenders. Regarding the polygraph
analysis, 55 of the polygraph tests were not deceptive (69%), 21 were deceptive (26%), and four
tests were inconclusive (5%). Regarding the LVA analysis, 54 tests were not deceptive (67.5%)
and 26 tests were deceptive (32.5%). No LVA tests were inconclusive. Table II presents the
results of the Polygraph and LVA tests for the Adult Sample.
Table II. Results of Polygraph and LVA Tests for Adult Sample
Total Test completed (N= 80)
Test Outcome
Polygraph
LVA
No Deception
55
54
Deception Indicated
21
26
Inconclusive
4
0
A comparison of the polygraph and the LVA results to the adolescent age group did not
demonstrate significant differences. Twenty-three examinations were administered to adolescent
offenders. Regarding the polygraph analysis, thirteen of the polygraph tests were not deceptive
(57%), six were deceptive (26%), and four tests were inconclusive (17%). Regarding the LVA
analysis, fourteen tests were not deceptive (61%) and nine were deceptive (39%). No LVA tests
were inconclusive. Table III presents the results of the polygraph and LVA tests on the
Adolescent sample.
Detecting Deceptive Responses in Sex Offenders:
8
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
Table III. Results of Polygraph and LVA Tests for Adolescent
Sample
Total Test completed (N= 23)
Test Outcome
Polygraph
LVA
No Deception
13
14
Deception Indicated
6
9
Inconclusive
4
0
A comparison the polygraph and the LVA results by type of examination did not
demonstrate significant differences. 73 examinations were Full Disclosure tests. Regarding the
polygraph analysis, 43 of the polygraph tests were not deceptive (59%), 24 tests were deceptive
(33%), and six tests were inconclusive (8%). Regarding the LVA analysis, 43 tests were not
deceptive (59%) and 30 were deceptive (41%). No LVA tests were inconclusive. Table IV
presents the results of the polygraph and LVA tests for Full Disclosure tests.
Table IV. Results of Polygraph and LVA Tests for Full Disclosure
Tests
Total Test completed (N= 73)
Test Outcome
Polygraph
LVA
No Deception
43
43
Deception Indicated
24
30
Inconclusive
6
0
24 maintenance tests were administered. Regarding the polygraph analysis, 23 of the polygraph
tests were not deceptive (96%) and one test was deceptive (4%). No polygraph tests were
inconclusive. Regarding the LVA analysis, 23 tests were not deceptive (96%) and one test was
Detecting Deceptive Responses in Sex Offenders:
9
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
deceptive (4%). No LVA tests were inconclusive. Table V presents the results of the polygraph
and the LVA tests for maintenance tests.
Table V. Results of Polygraph and LVA Tests for Maintenance Tests
Total Test completed (N= 24)
Test Outcome
Polygraph
LVA
No Deception
23
23
Deception Indicated
1
1
Inconclusive
0
0
Six Specific tests were administered. Regarding the polygraph analysis, two of the polygraph
tests were not deceptive (33%), two tests were deceptive (33%), and two tests were inconclusive
(33%). Regarding the LVA analysis, two tests were not deceptive (33%) and four tests were
deceptive (67%). No LVA tests were inconclusive. Table VI presents the results of the polygraph
and the LVA tests for specific tests.
Table VI. Results of Polygraph and LVA Tests for Specific Tests
Total Test completed (N= 6)
Test Outcome
Polygraph
LVA
No Deception
2
2
Deception Indicated
2
4
Inconclusive
2
0
Detecting Deceptive Responses in Sex Offenders:
10
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
Discussion
The purpose of the present study was to evaluate the ability of LVA technology to
distinguish the truthfulness (and deception) of sex offender self-report by comparing the results
of the LVA test with the results of a polygraph test. The polygraph test is used in the treatment
and supervision of sex offenders to verify the truthfulness of self-report. The reliability and
validity of the polygraph have been studied extensively in the literature. According to the
American Polygraph Association, there have been 80 research projects published since 1980
involving more than 8,000 polygraph examinations resulting in 93.6% to 98% accuracy (Ansley,
1997; Forensic research Inc., 1997). Matte (1996) reported research demonstrating polygraph
testing to be reliable more than 90% of the time.
The LVA technology reports the ability to determine truthful statements from deceptive
statements. They report several field studies demonstrating the validity of the LVA to determine
truth from deceptive statements based on known outcomes (Winscher & Witczak, 2006).
However, this research was done by V solutions, who were also marketing the technology. In
addition, there has been unpublished research regarding the usefulness of the technology as well
as clinical testimonials of the LVA’s ability to identify truth and deception. To date, there is no
known independent research addressing validity and reliability involving LVA technology. Thus,
this study attempted to compare the LVA technology directly to the polygraph in a double blind
research design. In order to accomplish this, the polygraph examiner and the LVA examiner
were independent of each other. Furthermore, neither examiner was aware of or had access to the
test results of the other examiner. Once the offender completed both examinations and a final
report was submitted, the referring agency then documented the outcome results (no deception
indicated, deception indicated, inconclusive) of each exam in the research file. This file was then
provided to this author for analysis.
The present study revealed no significant differences in the findings determined by the
polygraph and the LVA examinations. In addition, there appear to be no significant differences
between the polygraph and the LVA when used with adult or adolescents. Also, there were no
significant differences in the outcomes between the polygraph and LVA when the analysis
controlled for the type of examination (Full Disclosure, Maintenance, Specific). The one
significant difference between the two tests involved the eight polygraph tests that were
inconclusive. An inconclusive result on a polygraph test is due to inadequate physiological data
or physiological data which is insufficient or does not meet a minimum threshold to render a
conclusive opinion (English & Heil, 2000; Williams, 1995). According to the research, rates of
inconclusive tests vary between ten and twenty percent (Reid & Inbau, 1977; Morris, 2009
personal statement). In this particular study, 7.8% of the tests were inconclusive which was
insufficient to cause a significant difference in the group over all.
Detecting Deceptive Responses in Sex Offenders:
11
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
Although not a variable specifically controlled or studied, the LVA technology was a
much less invasive method of evaluating truth from deception than the polygraph. The offender
was not attached to any leads or physiological transducers. It was not necessary to create fear or
anxiety in order to create adequate physiological data to render a conclusive opinion. In addition,
unlike the polygraph, which required a separate appointment, the LVA utilized the audio-taped
clinical interviews which eliminated the need for a separate appointment. As a result, a number
of offenders (N=18; not included in the study) were evaluated by the LVA but did not attend,
keep, or make a polygraph appointment. The LVA also identified deceptive and truthful
information that could not be evaluated by the polygraph without completing a separate
polygraph. An example of this involved an adult offender who acknowledged his offense of his
infant victim and thus was truthful on the Full Disclosure polygraph and LVA test regarding all
sexual contact with children. However, during the clinical interview, the offender had denied any
penetration of the victim which was determined to be deceptive on the LVA. Since the Polygraph
is limited to a single issue that requires a “yes” or “no” response, this issue could not be tested on
the same polygraph.
Finally, inconclusive findings on the polygraph occur in approximately 10 to 20 percent
of the tests. An” inconclusive” finding is of relatively little use to the clinician or probation
officer and usually requires a re-test. In the present study, eight polygraphs were inconclusive.
Of the eight inconclusive polygraph tests, the LVA found seven to be deceptive. The lone
inconclusive involving an adolescent male was found to be truthful on the LVA.
The major limitation to the present study is the limited focus of the work. The present
study attempted to determine if the LVA could differentiate truthful statements from deceptive
statements in an independent, double blind design using the polygraph as the standard. Another
limitation was the low numbers of adolescent participants. Polygraph testing with adolescents is
much less frequent than with adults. In addition, half of the inconclusive (four) polygraph exams
involved adolescent offenders (17% of the adolescent sample).
There was strong interest to address accuracy (validity) and the consistency (reliability)
of the LVA findings. Additional interest was also given to including multiple examiners to
determine whether findings would differ based on examiner.
Further research on the LVA is needed. Research designed to identify criterion-related
validity. Additional research replicating this study will be important. Additional research to
address reliability of the LVA; particularly in the area of having multiple examiners
independently evaluate the recorded material to determine what if any response variability is due
to differences in administration and/or segmenting the recorded material.
Detecting Deceptive Responses in Sex Offenders:
12
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
The present results support the effectiveness of LVA technology to identify truthful and
deceptive statements as compared to the polygraph. Due to the lack of research with LVA, the
focus of this study was limited to whether the findings of the LVA test were consistent with the
findings from the polygraph. The polygraph was used because of the extensive research on its
reliability and validity. At present, the polygraph is considered the best technology for truth
verification. Based on the results of this study, LVA technology appears to be able to provide
clinicians and probation officers with relevant information regarding the accuracy of sex
offender self-report comparable to the polygraph.
The results suggest the LVA and polygraph findings are similar to one another despite
significant differences in their methodology. If these results accurately reflect the LVA’s ability
to discern truth and deception, the addition of this technology in the field could be a major
enhancement to clinical assessment, intervention, and management. The study identified six
strengths of the LVA over the polygraph: (1) The LVA is less invasive. The offender is not
hooked to a number of physiological leads; (2) The LVA process is less investigative. The LVA
uses the clinical interview as the basis for evaluating the offender’s self-report. The offender is
not limited to answering questions using a “yes” or “no” format and the need to experience a
“threat” of consequences does not exist; (3) The LVA can be used in real time. Although this
format was not evaluated as part of this particular study, the clinician can use the feedback not
only whether the statement is deceptive, but can identify inaccuracies and high stress issues that
could be used by the clinician to enhance the usefulness of the clinical interview; (4) The LVA is
much more flexible. It does not require a special room set up or separate appointment. Clinicians
who may be asked to assess in an institution such as a jail or detention center can simply record
the interview. Furthermore, should a situation arise in a group or family session that the clinician
would need to verify for accuracy, the clinician would be able to do so with little disruption or
time lapse; (5) The LVA can evaluate multiple issues at one time. This is a significant issue in
the field. The polygraph is limited to a single issue. Thus if an offender fails a maintenance test,
there is no way to determine what part of the guidelines he/she has been noncompliant without a
separate test or if he acknowledges something to satisfy the results whether that disclosure is
truthful. The LVA would allow follow up without tainting the original issue; (6) The LVA
always makes a determination of the offender’s statements. There are no inconclusive results
using the LVA. The LVA differentiates deceptive from inaccurate, stress/anxiety, and truthful,
thus there is an opinion of the offender’s statements. Combining the above mentioned strengths
with the strengths of the polygraph significantly improve the effectiveness and efficiency of truth
verification in sex offender treatment and management.
Detecting Deceptive Responses in Sex Offenders:
13
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
ACKNOWLEDGEMENTS
The author would like to thank Counseling & Consultation Services, Inc. for their
support, help, and contribution to this research. This research could not have been completed
without the assistance of therapists Tom Herington, M.A. and Amy Mize, M.A. In addition, the
author would like to express his gratitude to Ms. Sandra English for collecting the records and
files as well as providing and maintaining the independence of the polygraph and LVA results.
REFERENCES
Abel, G. G. & Harlowe, N. (1987). The child abuser: how can you spot him?. Redbook, 98-139.
Abrams, S. (1989). Probation polygraph surveillance of child abusers. Prosecutor, 22, 29-36.
Abrams, S. (1991). The use of the polygraph with sex offenders. Annals of Sex Research. 4,
239-263.
Abrams, S. (1992). The polygraph as a tool in the treatment of sexual abusers, Preventing Sexual
Abuse, 2, 10-15.
Abrams, S., Hoyt, D., & Jewell, C. (1991). The effectiveness of the disclosure test with sexual
abusers of children. Polygraph, 20,192-203.
Abrams, S. Ogard, E. (1986). Polygraph surveillance of probationers. Polygraph, 15, 174-182.
Ahlmeyer, S., Heil, P., McKee, B. & English, K. (2000). The impact of polygraphy on
admissions of victims and offenses in adult sexual offenders, Sexual Abuse: A Journal of
Research and Treatment, 12, 123-138.
Ansley, N. (1997). The history and accuracy of guilty knowledge and peak of tension tests,
Polygraph, 21 (3), 174-247.
Arther, R. O. (1990). An old idea whose time has come, The Journal of Polygraph Science, 25,
July-August.
Baker, D., K., Skolnick, J., Doucette, G., Levitt, G., & O’Connor, C. (2005). Intensive parole
supervision of the sex offender-putting the containment approach into practice. In B. K.
Schwartz & H. R. Cellini (Eds.), The Sex Offender: New Insights,Treatment, Innovation,
and Legal Developments, (pp. 5-1 to 5-15). Kingston, NJ: Civic research Institute.
Detecting Deceptive Responses in Sex Offenders:
14
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
Blasingame, G. D. (1998). Suggested clinical uses of polygraphy in community-based sexual
offender treatment programs, Sexual Abuse: A Journal of Research and Treatment, 10,
37-45.
Brake, S. C., & Shannon, D. (1997). Using pre-treatment to increase admission in sex offenders.
In B. K. Schwartz & H. R. Cellini (Eds.), The Sex Offender: Issues in Assessment,
Treatment, and Supervision of Adult and Juvenile Populations (pp. 12-1 to 12-25).
Kingston, NJ: Civic research Institute.
Carnes, P. J. (1995). Polygraph in assessment, treatment, and surveillance of sex offenders,
Sexual Addisction & Compulsivity: The Journal of Treatment and Prevention, 2, 103111.
Dam, J. (1995). Lie detectors are admissible; Should you test your client. Lawyers Weekly USA,
11/20/1995.
Emerick, R. L., & Dutton, W. A. (1993). The effect of polygraphy on the self-report of
adolescent sex offenders: Implications for risk assessment. Annals of Sex Research, 6, 83103.
English, K. (1998). The containment approach: An aggressive strategy for the community
management of adult sex offenders. Psychology, Public Policy, and Law, 4, 218-235.
English, K. & Heil, P. (2006). The need for complete information leads to the polygraph
examination. In B. K. Schwartz & H. R. Cellini (Eds.), The Sexual Predator: Law and
Public Policy (pp. 9-1 to 9-37). Kingston, NJ: Civic research Institute.
Forensic Research, Inc. (1997). The validity and reliability of polygraph testing, Polygraph, 26
(4), 215-223.
Freeman-Longo, R. E. & Walls, R. V. (1986). Changing a lifetime of sexual crime, Psychology
Today, 58-64.
Grubin, D., Madsen, L., Parsons, S., Sosnowski, D., & Warberg, B. (2004). A prospective study
of the impact of polygraphy on high risk behaviors in adult sex offenders. Sexual Abuse:
A Journal of Research and Treatment, 16, 209-222.
Jensen, S.H., & Jewell, C. A. (1988). The sex offender experts. Prosecutor, Fall: 13-20.
Maletzky, B. M. (1991). Treating the Sexual Offender. Newbury Park: Sage.
Detecting Deceptive Responses in Sex Offenders:
15
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
Matte, J. A. (1996). Examination and cross examination of experts. Forensic psychophysiology
using the polygraph: Scientific truth verification-lie detection (pp.29-34). Williamsville,
NY: J.A.M. Publications.
McGrath, R. J., Cumming, G. F., Hoke, S. E., & Bonn-Miller, M. O. (2007). Outcomes in a
community sex offender treatment program: A comparison between polygraphed and
matched non-polygraphed offenders. Sexual Abuse: A Journal of Research and
Treatment, 19, 381-393.
Reid, J. E. & Inbau, F. E. (1977). Truth and Deception: The Polygraph (“lie detection”)
technique (2nd ed.). Baltimore: Williams and Wilkins.
Robbins, L. (2007). Layered voice analysis 6.50; white paper, (unpublished Technical Report).
Salter, A. (1988). Treating child sexual offenders and victims, a practical guide, CA: Sage
Publications.
Scott, L. K. (1997). Community management of sex offenders. In B. K. Schwartz & H. R. Cellini
(Eds.), The Sex Offender: Issues in Assessment, Treatment, and Supervision of Adult and
Juvenile Populations (pp. 16-1 to 16-12). Kingston, NJ: Civic research Institute.
Steen, C. (2002). The expert witness in the sex offender case-A practical guide. In B. K.
Schwartz & H. R. Cellini (Eds.), The Sex Offender: Issues in Assessment, Treatment, and
Supervision of Adult and Juvenile Populations (pp. 12-1 to 12-25). Kingston, NJ: Civic
research Institute.
Williams, V. L. (1995). Response to Cross and Saxe’s “acritique of the validity of polygraph
testing in child sexual abuse cases, Journal of Child Sexual Abuse, 4, 55-71.
Winscher, T. & Witczak, T. (2006) Layered voice analysis training, LVA Training Manual
(unpublished).
Detecting Deceptive Responses in Sex Offenders:
A Comparison of Layered Voice Analysis (LVA) and the Polygraph
16