Psychology
2012. Vol.3, No.2, 217-230
Published Online February 2012 in SciRes (http://www.SciRP.org/journal/psych) http://dx.doi.org/10.4236/psych.2012.32031
Copyright © 2012 SciRes. 217
Personnel Security outside the Box
Semyon Ioffe1, Sergey Yesin2
1Department of the Psycho-Semantics, Northam Psychotechnologies, Ottawa, Canada
2Psychosemantic Clinic, Larnaca, Cyprus
Email: semyon.ioffe@gmail.com
Received September 30th, 2011; revised November 4th, 2011; accepted December 21st, 2011
This study describes the theoretical and methodological bases of research of psycho-semantic spheres of
the psyche of Dari speaking Afghani military personnel. 812 military personnel including 50 intelligence
officers were subjected to two types of tests. The first questioner type investigation was the task where the
individual was to provide some classification about a topic. The response was a judgment of similarity
and/or other association. The second test was a subliminal probing, where the individual had no idea of
topics being tested and was acting in a game like environment. The test measured the subject’s response
time to words or images (semantic stimuli) targeting the unconscious mind. A non-invasive, non-intrusive
unconscious probing provided results which were highly corroborated by eight senior Military Intelligence of-
ficers’ ratings. The questioner type test exposed a number of strategies of evading testing procedures
overwhelmingly used by tested personnel, even by those who were cleared by subliminal probing and ex-
pert opinion. In addition to discussion of the probabilities of false positives and false negative responses;
assessment of one’s mental agility or ability to react and learn under stress; general psychological makeup
of the population tested is discussed. We established that anyone can take a psycho-semantic unconscious
probing test even illiterate people of any race, language or culture. A Semantic Stimuli Response Mea-
surement (SSRM Tek.) test allows for large-scale screening operations to identify unknown threats to na-
tional security while serving as a deterrent to undesirable or illegal behaviours and helps to set high stan-
dards in personal conduct. Key words: security, military, personnel, unconscious, semantic, stress.
Keywords: Security; Military; Personnel; Unconscious; Semantic; Stress
Introduction
Human beings are frequently the “weakest link” in mitigating
threats to national security and commercial industries. People
can be manipulated to conduct espionage against their sworn
nation, conduct sabotage operations against a government or
commercial facility and even inflict mass casualties as a suicide
bomber or carrier of a pathogenic biological agent. People are
frequently influenced by harmful negative factors such as drugs,
alcohol, and traumatic events which radically affect their per-
formance and reliability.
In March of 2008, NATO and the Afghan Government
signed an intelligence sharing agreement designed to facilitate
joint military operations involving the Afghan National Army
(ANA) and Coalition Forces. The Afghan Ministry of Defence
subsequently issued a declaration (i.e., military order) that all
military personnel must submit to a background investigation.
Unlike Western or developed countries, 30+ years of war
have completely destroyed all remnants of the professional
Afghan Army, to include its laws, policies, and procedures.
Illiteracy is approximately 80% across the country and most
new recruits don’t know their actual birthday and many simply
have one name—Abdul, son of Faisal. There are no financial
records to speak of, no electronic connectivity in which to fa-
cilitate efficient and secure investigations and a society plagued
with corruption issues that presently define this nation. Most
Afghan citizens see themselves as being from a tribe as op-
posed to a nation. The Afghan National Army is working hard
to integrate the various tribes into this national structure. Af-
ghanistan is looking for a Personnel Security Program that pro-
vides standards across all organizations and effectively identi-
fies issues such as Anti-Government Activities, Espionage, Cor-
ruption, and Drug/Alcohol Abuse. The psycho-semantic ap-
proach of testing using unconscious response measurement was
chosen as the best fitting test in the Afghani environment and
provides testing in the local Dari language; does not require any
prior information about the subject tested; none or minimal
familiarization with computers is required by the subject tested.
Results of the testing cannot be affected by the subject tested or
operator and immediate results are provided.
The mission of this study was to choose and test technologies
capable of conducting investigation to determine the trust-
worthiness and suitability of literate and illiterate members of
the Afghan Ministry of Defense (MoD) being the basis for
granting Afghan MoD security clearances up to a top secret
level.
Emerging and current technologies attempting to “detect”
performance and reliability issues are as follows:
Polygraph Recordings of Human Vital Signs Assessed
by a Poly graph Exper t. This Approach Is Currently
Widely in Use
The polygraph is commonly referred to as a lie detector, but
it does not recognize lies. A specific physiological lie response
has never been demonstrated and is unlikely to exist (NRC,
2003). The responses it measures are not unique to deception
nor are they always provoked by it. The aim of the polygraph
examiner is to establish a psychological set in the examinee that
S. IOFFE ET AL.
will increase the likelihood that any observed arousal to specific
questions is the result of deceptive responses (Kleiner, 2002;
Grubin, 2010; Senter et al., 2010).
Use of polygraph examination in criminal investigations,
employee screening and security vetting where the emphasis is
on whether or not the individual has passed the test is an area of
much controversy (NRC, 2003). In the event-related investiga-
tion polygraph performed well above chance level, an unac-
ceptable level of false positive results during the screening,
precluding its usage in employment and pre employment
screening procedures (NRC, 2003). Confessions or disclosure
of the individual as the result of either possibility or being a
subject of polygraph testing is considered the highest proof of
polygraph validity and reliability by the polygraph examiner
and the polygraph community at large. Such a situation does
not address the effect of Bogus Pipeline. The effect of Bogus
Pipeline is the expectation that subjects will answer more hon-
estly if they believe that the truth can be tested for accuracy
even in the absence of such a test. Since 1971, when Jones and
Sigall described a new paradigm for measuring affect and atti-
tude and named it “The Bogus Pipeline” attempts to quantify its
effect were undertaken. Damphousse et al. (2007) reported that
in the study, where subjects were informed that their answers
would be analysed by a lie detector, only 14% lied about recent
drug use compared to 40% in the study where no lie detector
was used or mentioned. Both studies participants were the sub-
jects of urinary drug testing. It is important to point out that the
remarkable outcome is the effect of informing the subjects
about the use of a lie detector only. Whether the lie detector
actually does anything or is even physically present is irrelevant.
Telling the subjects that a lie detector will be used, but without
actually using one, will have the same effect as long as the
subjects believe that a lie detector is used. Future scientific
validation of the basic principles of polygraph testing cannot be
considered without factoring in the Bogus Pipeline effects of
such testing.
Voice Stress Analys ers That Are Al so Currentl y in Use
What are the basic principles upon which voice stress ana-
lysers are claimed to be based and have these principals been
verified in scientific studies? Voice stress analysers are based
on the premise that there are tiny frequency modulations in the
human voice called microtremors. When a test subject is lying,
the automatic, or involuntary nervous system causes an inaudi-
ble increase in the microtremor’s frequency. In 1970 Olaf Lip-
pold discovered the muscle microtremor (Lippold, 1970, 1971).
Lippold found that voluntary muscles in the arm generate a
physiological tremor or micro-vibration at about 10 Hz when
the subject is relaxed. When the subject is aroused or stimulated,
the microtremor tends to disappear. There is no evidence in
scientific literature that anyone has uncovered a microtremor in
the throat and larynx muscles involved in voice production.
Eriksson and Lacerda (2007) and Damphousse et al. (2007)
summarised well the validity and reliability in the voice stress
analysers field as follows: review of scientific studies shows
that these machines perform at chance level when tested for
reliability. Given such results and the absence of scientific
support for the underlying principles, it is reasonable to view
the use of these approaches as unjustified either for event re-
lated investigation, employment or pre employment screening.
Eye Movement Tracking to Identify
Familiar/Unfamiliar Faces, Objects and Scenes
It is an emerging technology, though eye movement tracking
by itself is a well technically developed field and is used in
many areas of cognitive science. A pioneer in the field of study
of eye movements, Yarbus (1967) wrote about the relationship
between fixations and interest: The cyclical pattern in the ex-
amination of pictures “is dependent not only on what is shown
on the picture but also on the problem facing the observer and
the information that he hopes to gain from the picture”.
Presently, we still cannot infer specific cognitive processes
directly from a fixation on a particular object in a scene. For
instance, a fixation on a face in a picture may indicate recogni-
tion, liking, dislike, puzzlement etc. Therefore, eye tracking
would require to be coupled with other methodologies to ad-
dress the meaning of the different pattern of fixation for the
different pictures offered to the subject.
Thermal Facial Tracking, an Emerging Technology,
Allows for a Wide Variety of Non Invasive
Measurements: Respiration, Temperature,
Heart Rate, Blinking.
The question is still the same as in the polygraph recording
approach: which measurements, and based on what theory, can
candidate variables be chosen for detecting deception and ex-
plain how their relations for situations of everyday life differ
from the condition of deception?
All the above described methodologies currently used or
emerging are addressing truth verification or credibility assessment
efforts, previously known as lie detection of a different kind. To
understand the premise and frame work of truth verification and
credibility assessment work, we research literature for defini-
tions of truth and credibility. According to the Stanford Ency-
clopaedia of Philosophy the definition of truth is based on nine
different theories: correspondence, coherence,
“http://en.wikipedia.org/wiki/Constructivist_epistemology”,
consensus, pragmatic, deflationary, performative, redundancy,
pluralist. The purpose of this paper does not allow us to go into
details of these theories, but in a survey of professional phi-
losophers and others on their philosophical views which were
carried out in November 2009 (taken by 3226 respondents,
including 1803 philosophy faculty members and/or PhDs and
829 philosophy graduate students) 44.9% of respondents accept
or lean toward correspondence theories, 20.7% accept or lean
toward deflationary theories and 13.8% epistemic theories.
There is also no consensus on the definition of lying and decep-
tion in the literature. Please entertain yourself with the well
known liar paradox: if “This sentence is false” is true, then it is
false, which would in turn mean that it is actually true, but this
would mean that it is false, and so on to infinity. We believe the
whole field of truth verification or credibility assessment efforts
would greatly benefit if a unifying framework allowing charac-
terization across various types of information resources and
diverse information tasks including work problems and per-
sonal interests could be established.
Psycho-Sem antic Methodologies
The name psycho-semantics derives from two words: PSY-
CHE defined in literature as nonphysical attributes of the hu-
man being, faculty for thought, judgment, and emotion; the
Copyright © 2012 SciRes.
218
S. IOFFE ET AL.
mental life, including both conscious and unconscious proc-
esses and SEMANTICS defined as study or science of meaning.
Free association test was the first inroad into the field of psy-
cho-semantics and still used since early 20th century by scien-
tists, psychologists to investigate underlying perception and
meaning, reasoning, and motivation, personality and its pa-
thology. Basically the subject is told to say the first word that
comes to mind in response to a stated word, concept, or other
stimulus. If the word stated has no significant meaning the re-
sponse will slightly fluctuate from other words presented, but
when a stated word provokes an emotionally charged memory
it may produce atypical or revealing associations or, more often,
reaction time becomes either unusually long or unusually short.
The literature describes other approaches for psycho- seman-
tic tests of people, including subjects influenced by harmful
negative factors such as drugs, alcohol, and traumatic events
which radically affect their performance and reliability. In psy-
cho-semantic research, the task for the individual is to provide
some classification about a topic. The response could be a
judgment of similarity, an indication of the extent to which she
or he agrees or disagrees with a statement, or some other asso-
ciation. George Kelly (1955) published a personal construct
theory (PCT) where constructs are described as bipolar catego-
ries that people can use to understand the world. It postulated
that people then behave according to how they construe the
world around themselves. Osgood et al. (1957) described a
technique for the measurement of meaning named semantic
differential (SD). SD was designed to measure the connotative
meaning of concepts. The connotations are used to derive the
attitude towards the given object, event or concept. Since this
time, PCT and SD were widely used in many areas of manage-
ment studies, knowledge modeling in artificial intelligence,
language and attitude studies and a wide range of other disci-
plines (Alford & Strother, 1990; Baker, 1986; Ball, 1983;
Bayard, 1990; Cacioppo & Berntson, 1994; Cheshire, 1982;
El-Dash & Tucker, 1975; Gaies & Beebe, 1991; Gallois & Cal-
lan, 1981; Heise, 2010). In 2010 a single test, PsyExpert, com-
bining PCT (Kelly, 1955, 1970), SD (Osgood, 1959, 1967,
1976), Lentiev’s theory of the emotion (Leontiev, 2002) and
Luscher’s color preferences approach was described (Ioffe &
Yesin, 2010, 2011) and took the whole field of study of human
psyche a step further. Instead of using an expert to uncover
personal constructs as George Kelly described or create adjec-
tival opposites as required for SD, in PsyExpert the subject
being tested reveals his constructs automatically, associates it
with colors from a pallet of 15 colors, using a concept of innate
relations of human emotions and feelings with color. Evaluat-
ing the words and concepts by colors, the subject being tested
does not suspect that he/she is actually “associating” each ele-
ment with an individual specific emotion (the pole of the con-
struct), thus openly revealing their inner world. All of the above
described testing is done in one 30 to 60 minute procedure us-
ing PsyExpert. PCT, SD, PsyExpert and other questionnaire
type tests use conscious mind responses to reveal unconscious
roles, components, influences, actions etc. in human behaviour.
Unconscio us Psyc ho-Sema ntic Methodo logies
There are other researchers who have developed approaches
to bypass consciousness all together (Shevrin & Fritzler, 1968;
Shevrin et al., 1969; Dixon, 1971; Luriya & Vinogradova, 1971;
Reingold & Merikle, 1980; Kunst-Wilson & Zajonc, 1980;
Kostandov, 1983; Kihlstrom, 1987; Greenwald, 1992; Esteves
et al., 1994; Ohman & Soares, 1994; Epstein, 1994; Smirnov et al.,
1995; Dijksterhuis et al., 2005; Dehaene et al., 2006; Ioffe et al.,
2007a) and deal directly with the unconscious mind employing
different masking techniques through use of a tachistoscope
(Renshaw, 1945; Benschop, 1998) or computers (Holender,
1986; Ramachandran & Cobb, 1995; Ansorge et al., 2007).
Tachistoscope can flash images as brief as 1 msec., but inde-
pendently of the length of the exposure, the retina’s afterimage
trace persists approximately 40 msec. Therefore, the subject is
reacting not to the stimuli but to the afterimage of the stimuli.
Exposure time of a visual stimuli (word or picture) using com-
puters is limited by a monitor’s frequency and for LCD moni-
tors a minimal exposure is between 13 and 16 msec. Different
masking techniques are described in the literature (Ansorge et al.,
2007) and though there is no universal agreement on which
masking techniques are the best, forward masking, backward
masking, four dots etc., each of them was validated and are
successfully in use for their own purposes. Over the last four
decades extensive scientific research was directed into uncon-
scious mind mechanisms and its processes (Luriya & Vinogra-
dova, 1971; Shevrin, 1973; Kunst-Wilson & Zajonc, 1980;
Shevrin & Dickman, 1980; Dixon, 1981; Kostandov, 1984;
Wedding & Stalans, 1985; Macmillan, 1986; Kihlstrom, 1987;
Reingold & Merikle, 1988; Lang et al., 1990; Greenwald, 1992;
Shevrin et al., 1992, 1996; Epstein, 1994; Esteves et al., 1994;
Ohman & Soares, 1994; Wong et al., 1994, 1997; Smirnov et al.,
1995; Waller & Mijatovich, 1998; Snodgrass, 2000; Dijksterhuis
et al., 2005; Winkielman et al., 2005; Dehaene et al., 2006;
Ioffe et al., 2007b; Ioffe & Konobeevsky, 2008; Ioffe & Yesin,
2010, 2011; Ioffe, 2011). In 1968 Howard Shevrin published
(Shevrin & Fritzler, 1968) the first report of brain responses to
unconscious visual stimuli, thus providing strong objective
evidence for the existence of the unconscious mind. In the same
study, he showed that unconscious perceptions are processed in
different ways from conscious perceptions. Subsequently,
Shevrin and his colleagues have shown brain markers for un-
conscious factors at work in producing social phobias. The
literature presents many scientists from around the world
working in the unconscious field, but we singled out an Ameri-
can psychoanalyst, Howard Shevrin, as he went beyond just the
theoretical field into practical application and in 1987 applied
for a patent for a system assessing verbal psychobiological
correlates using tachistoscope for presentation of unconscious
stimuli. During the same time, across the world, another scien-
tist, a Russian psychiatrist, Igor Smirnov and his colleagues,
have been studying the unconscious mind using the methodol-
ogy of computer masking techniques of stimuli presentation
and measured evoked potentials or visual motor reaction to
such stimuli (Smirnov et al., 1995). In 1987 and 2002 they ap-
plied and received patents for methodology of psycho-ecology
and a method of psycho-probing. Technology developed by
Igor Smirnov moved studies of human psyche from the scien-
tific laboratory into the medical field of diagnostics using
probing of the unconscious mind.
In 2006 SSRM Tek was born by merging the principals of
free association test with basic principles of masking and vis-
ual-motor response recording as described above (Ioffe et al.,
2007a): we bypass the conscious mind and question the uncon-
scious mind directly using subliminal stimuli and expecting that
if a word shown subliminally provokes subject’s emotionally
charged memory, it will affect his/her reaction time the same
Copyright © 2012 SciRes. 219
S. IOFFE ET AL.
way it was affected in Free association test.
SSRM Tek is developed in the form of a non invasive com-
puter-like game and requires minimum computer knowledge of
the operator or subject being tested. Each and every one of
methodical principals, of which the SSRM Tek. was created,
are validated: Greenwald (1992) after reviewing all available
literature before year 1992 for and against unconscious cogni-
tion, drew the conclusion that “unconscious cognition is now
solidly established in empirical research”. “Unconscious cogni-
tion occurs separately from conscious cognition” (Greenwald,
Klinger, & Schuh, 1995), therefore unconscious mind process-
ing of information is separate, independent phenomena from
conscious mind and our approach to use subliminal stimuli for
testing is valid. “Visually masked prime words were shown to
influence judged meaning of following target words. Uncon-
scious semantic activation is very short-lived (100 msec.).
Unlike supraliminal stimulation, a subliminal leaves no mem-
ory trace that can be observed in response to the next stimuli”
(Greenwald, Draine, & Abrams, 1996). Therefore our usage of
the subliminal stimuli as perfect probes, which are short-lived
and do not contaminate memory—leave no trace by the time
next stimulus arrives, is not simply valid but is very prudent.
“Results show that subliminal priming involves unconscious
categorization of the prime” (Abrams, Klinger, & Greenwald,
2002), based on this study, our approach of describing experi-
ence in words and presenting it subliminally is credible. “Sub-
liminal priming governs by both task context and long-term
semantic memory” (Greenwald et al., 2003). This means that
subliminal stimuli will produce a meaningful response only if a
semantic memory of previous experience exists and subliminal
stimuli can affect time response of pressing a button when a
random row of numbers appears on the screen. “The present
experiment measured an EEG indicator of motor cortex activa-
tion, the lateralized readiness potential (LRP). The LRP data
showed that visually masked words triggered covert motor
activations. These prime-induced motor activations preceded
motor activations by subsequent (to-be-classified) visible target
words” (Zayas, Greenwald, & Osterhout, 2011). During SSRM
Tek test procedure subliminal stimuli inducing motor activa-
tions preceded motor activations by subsequently pressing the
button action when visible row of random numbers appears on
the screen. Therefore able to slow down or speed up the time of
the response or have no change to the time of the response to
row of random numbers. Our methodological approach to de-
velop defence reaction and test it using subliminal stimuli is
supported by Knight, Nguyen, and Bandettini (2003) and Pes-
siglione et al. (2008).
Results of SSRM Tek. applied research were verified using
clinical diagnosis (Ioffe et al., 2007a, 2007b), semantic differ-
ential by Osgood (Ioffe & Konobeevsky, 2008; Ioffe & Yesin,
2010, 2011) and Kelly’s Personal Construct methodology (Ioffe
& Yesin, 2010, 2011), PANSS scores (Tjuvina & Krivtsov,
2006) and intelligence reports (Ioffe, 2011). Additionally, it
was shown that probability of false positive was 0.3% and a
range of 0.08% to 1% with 95% confidence level. In the study
of early diagnostics of relapse of a paranoid schizophrenia
(Tjuvina & Krivtsov, 2006), correct forecasting of an aggrava-
tion of a psychosis using unconscious stimuli testing method-
ology was at the 84.6% accuracy level and overall sensitivity of
the applied method was at the 78.9% accuracy level, specificity
level was 86.9%.
Based on the above discussed we have chosen two method-
ologies: one as the best representative of the psycho-semantic
conscious methodologies (PsyExpert) and another one (SSRM
Tek) as a representative of unconscious psycho-semantic
methodology for our applied research in the Afghan military
personnel.
A current study was performed in two stages:
Stage 1—Development and evaluation of the text-based test-
ing in the literate Afghan Army personnel.
Stage 2—Development and evaluation of the picture-based
testing for the illiterate Afghan Army personnel.
Materials and Methods
Stage 1: Fifty Afghan Military Intelligence officers were
participants of the study, their ages 29 to 54 (43.8 + 6.4). Their
identities were fully protected. No information except age and
their code number between 1 and 50 assigned to them was col-
lected. The Afghanistan MoD’s Ethics Committee approved the
research paradigm. Informed consent was obtained from all of
the officers for the application of psycho-semantic methods for
testing.
Initially the task was to determine if chosen Western-based
technologies could be modified to the Afghan culture and its
main language, Dari, to provide a rapid, automated, non-inva-
sive, non-intrusive screening tool to identify critical issues re-
lating to personnel security.
Using the guidance from senior MoD Intelligence Officers,
the following topics were chosen to be tested: 1) corruption, 2)
espionage, 3) “anti-government activities” and 4) drug/alcohol
abuse.
As the psycho-semantic approaches discussed above were to
be employed in this study, a survey was designed to capture all
semantically relevant terms corresponding to these four topics.
The Dari-language survey was administered to 30 Military
Intelligence Officers from the Afghan MoD. The results were
collected, compiled, and translated from Dari to English, and
then analyzed to identify primary terms corresponding to these
four major topics. Initially 38 primary Dari words were chosen
from the survey conducted for the four topics described above.
Two Technologies Used
The first technology to be tested is called the SSRM Tek and
it is based on the measurement of an individual’s emotional
“reaction” to visual stimuli—words or pictures evoking a
strong semantic connotation and subsequently affecting mo-
tor-skill response (Ioffe et al., 2007ab). A word or picture is
displayed via a computer screen at a fraction of a second called
the “probe”—visual images pass through the retina and are
“seen” only at the unconscious level but not at a conscious level;
then masked by a row of random 15-digit numbers (for text
probes) or a scrambled image (for picture probes) which is the
only image the person taking the test registers at the conscious
level. The person taking the test is trained to press a button
when he sees the random string of numbers or scrambled im-
ages. If the unconscious-based probe evokes a strong-enough
reaction, the person taking the test presses the button either
more slowly or more rapidly as compared to unconscious con-
trol stimuli which are random strings of numbers or scrambled
images. These unconscious-based probes and controls are
shown randomly throughout the testing cycle in word or picture
“groups” and the results are statistically calculated and pre-
Copyright © 2012 SciRes.
220
S. IOFFE ET AL.
sented in an easy-to-read format.
The second technology to be tested is called PsyExpert and it
is an automated color association test—a standardized psycho-
logical exam which was designed to probe the unconscious
mind using computer questionnaire-type probes requiring con-
scious responses. PsyExpert represents a combination of well
known methodologies including Semantic differential by Os-
good 1959, 1967, 1976, Kelly (1955) construct theory and
Lusher color test and theory of the emotions (Kostandov, 1977;
Leontiev, 2002) which indirectly associates and evaluates rela-
tionships of different tested words with emotions.
In this study, PsyExpert was matching the words selected
from the survey with different basic emotions (fear, hope, grief,
pleasure, anger, satisfaction, liking, antipathy, etc.).
PsyExpert was meant also to help validate that the primary
words identified in the survey were in fact words evoking
strong emotional responses.
Each officer was subjected to both of the above described
tests concurrently. On average combine testing would take less
than 90 minutes.
Initially, the plan was to analyze the data of the first PsyEx-
pert and SSRM Tek testing and further make changes to the list
of words initially tested then repeat testing to attempt further
improvement of the results.
Stage 2: The 762 Afghan military personnel tested were be-
tween ages 18 to 69 with the average age 40.64 + 10.54 and
they came from a large cross-section of the Afghan Army.
Testing was done at various Afghan Ministry of Defence sites
and included the MoD Headquarters in Kabul, and the MoD’s
Bala Hissar complex. The instructions on how to take the test
were given to each of the officers in Dari language.
Testing protocol for stage 2 was to compare a series of se-
mantically significant pictures to the semantically significant
Dari words identified in stage 1.
We hypothesized that we might be able to identify a group of
semantically significant images/pictures that have a significant
correlation with the semantically significant group of words
previously shown to be effective during testing and corres-
ponded to anti-government activities. Further, we hypothesized
that the picture based stimuli will evoke a higher number of
significant emotional responses as compared to the text based
stimuli. Once a group of pictures has been identified, a pic-
ture-only based psycho-semantic test could be prepared for
future use with non-literate personnel. Additionally, if our hy-
pothesis will turn out correctly then pictures could be added to
the original four-topic text-based test allowing for a higher
fidelity of screening in literate personnel.
A set of images was initially collected by the two Afghan
technicians and then filtered down after discussions with senior
officers from the Afghan Defence Security Service. A set of 29
pictures were identified for the initial psycho-semantic test,
separated into 4 groups, to cover the “anti-government activities”
topics. The 29 images were separated into the following image
groups: Armed Insurgents; Insurgent Leaders; Sabotage; Sui-
cide Bombing.
We utilized these 29 pictures in 4 separate groups along with
the 20 words matching Armed Insurgents and Suicide Bombing
groups from stage 1 study to form the new (initial) psycho-
semantic test procedure for anti-government activities. To vali-
date the results of such psycho-semantic testing for all topics
described above, we added two more topics: a false positive
“picture” topic consisting of scrambled images; and a false
positive “word” topic consisting of rows of random numbers.
As both false positive “picture” and “word” topics are mean-
ingless stimuli, any statistically significant positive responses to
these topics will represent false positive results of psycho-se-
mantic testing.
Although an initial set of 29 pictures were chosen and priori-
tized for the picture-based test, we designed a survey for the
Afghan Army personnel taking the psycho-semantic testing.
This survey was designed to evaluate a conscious emotional
response to each picture in relation to other pictures in the same
semantic “group”. This test took approximately 2 minutes
(Figure 1).
After a subject took a test of the emotional response to the
pictures, the psycho-semantic test was started as described in
stage 1 above. In addition, the ability to make the right decision
under stress is evaluated based on the four gradations excellent,
good, poor and below poor (unacceptable), calculated as a per-
centile of total number of errors to a permissible number of
errors for the procedure of testing. The results of the defense
reaction teaching procedure are also evaluated in the subjects
being tested.
Statistical analysis, using criteria of Lehmann-Rosenblatt,
one way ANOVA, Wilcoxon non parametric and other meth-
odologies were performed.
Results
Stage 1: Only 8% of the officers tested using PsyExpert
produced justifiable results. These results suggested that the
following words used during testing such as Talib (), Islam
Defenders (), Martyr (), Islam’s Devotee (),
Going to heaven () and Devotee () did not
represent a single meaning in the mind’s semantic space of the
officer tested. These words could not belong to the same testing
topics and were removed from further testing.
92% of the officers tested using PsyExpert developed a
number of different strategies to successfully evade the test
(Figure 2) and therefore produced no meaningful results:
1) Choosing (21 colors were presented to them) manageable
number of colors to remember 2 to 6 colors –33%.
Figu re 1.
Test for emotional response to the 29 pictures randomly presented on the
screen. Each picture appears with five randomly positioned emotions
on the computer screen trestee requires to point and click a mouse on
the one of the emotions corresponding to the picture presented on the
screen.
Copyright © 2012 SciRes. 221
S. IOFFE ET AL.
Figure 2.
Results of the PsyExpert test of one of the Afghani officers. Left part of
the figure from top to bottom shows spheres, colors from the most
likable (white) to the least likable (red), 12 pair of opposite emotions,
numbers with the percent represent percent of the association of the
emotion to the color chosen by the testee. Right part of the figure is the
Runge sphere representing semantic space, white flags represent ele-
ments (words) tested, red flags represent emotions named by the testee
during testing procedure. Take note that most of the emotions are asso-
ciate with two colors white and black (sphere Runge) practically all the
words tested are associated with two colors Black and Red. These sug-
gest that the officer was employing three colors to categorize 24 emo-
tions and 38 words as strategies to control the test outcome.
2) 50% used the same classification values for 6 to 12 colors.
3) 78% used the same color classification for 25 to 40 words
presented to them.
As opposed to the extremely low number of successful tests
using PsyExpert, all 50 Military Intelligence officers successfully
followed SSRM Tek testing procedures. Results of these tests
revealed (Table 1) statistically significant responses of 2 offi-
cers to Espionage topic, 4 officers to anti-government activities
and corruption topics, and 12 officers to drug/alcohol addiction
topics.
To validate these results a “subjective review” was con-
ducted involving 8 senior officers who did not take the SSRM
Tek test; 4 senior officers reviewed two separate groups each
for a total of 8 reviewers. These eight senior Military Intelli-
gence officers rated each of the 50 officers that took the SSRM
Tek test on a scale of 1 - 4 as follows:
I know this officer is involved in this activity;
I’ve heard this officer is involved in this activity;
I don’t know if this officer is involved in this activity; and
I’m sure this officer isn’t involved in this activity.
It did serve to establish a level of “confidence” comparing
the SSRM Tek results to what military intelligence officers had
on file. The comparison between the SSRM Tek testing and the
subjective reviews was overwhelmingly positive (see Figure 3)—
approximately 90% agreement for the anti-government activi-
ties and espionage topics—with less agreement for corruption
and drug/alcohol abuse topics (>83% and 70% respectively).
Table 1.
Show the number of officers for each of topic which had statistically
significant responses. A single officer could have multiple topic re-
sponses.
TOPICS Anti-government
activities Espionage CorruptionDrug/
alcohol abuse
N (number of
officers responded)4 2 4 12
Figure 3.
Comparison of SSRM Tek first set of tests with subjective reviews.
()—did not evoke a statistically significant “range”
among the 50 officers tested. Of the 38 original words chosen
for testing, 14 words were either found to have no statistical
difference or had multiple meanings, either positive or negative,
and therefore were excluded from further testing after con-
sultation with several senior Afghan military officers.
During these consultations, the Afghan military officers sug-
gested several additional words for inclusion in the second sets
of testing to attempt further improvement of the results of test-
ing.
Figure 4 represents a comparison of first and second sets of
SSRM Tek test results, where an exceptionally high degree of
correlation can be observed. Two officers were not available for
the second SSRM Tek test and 42 of the 48 Military Intelli-
gence officers that took both the first set of tests and second set
of tests had statistically identical results –87.5%. Of the re-
maining six individuals, five had changes that were directly
attributable to the change from the first list to the second list of
words tested, and one officer had invalid test results and sug-
gested to be retested after settling back into the local environ-
ment after having made a personal trip outside of Afghanistan.
Stage 2: Average ages of military personnel recruited in all
known countries around the world are between ages of 18 and
22, 62% of the Afghani military personnel tested in our study
were between 40 and 59 (Figure 5).
Evaluation of the Pic tu re Chosen Stim ul i
We verified experimentally the choice for pictures in the
psycho-semantic testing procedure using analysis of the initial
data of 160 officers regarding the emotional content of the 29
pictures. This analysis indicated that several pictures (3, 4, 14
and 29) did not statistically correlate with the other pictures
within each respective group (Figure 6) and were subsequently
removed from further testing. Pictures 3, 4, and 14 were from
the Armed Insurgency Group set, whereas picture 29 was from
the Sabotage set. Criterion for picture exclusion from the topic was
The data from the first set of SSRM Tek testing were also
analyzed in order to determine which words evoked the strong-
est reaction “range”. Some of the Dari language probes had
little-to-no range. For example, the Dari word for bribery
() produced a wide range of responses within the
testing population whereas a slang word for bribery—sweet
Copyright © 2012 SciRes.
222
S. IOFFE ET AL.
Figure 4.
Comparison of SSRM Tek first set of tests with the second set of tests.
Figure 5.
Age Distribution Statistics. Horizontal line represent ages, vertical line—
percentages. Total number of military personnel tested was 762.
the distribution of the emotions allocated to the picture by 160
officers which statistically differed from at least two other pic-
ture distributions in the topic.
We also performed a cluster analysis using Ward’s method
(1963) on the 602 officers’ responses to the emotional content
of the 25 pictures left after four pictures were removed (Figure
7). Ward’s method uses an analysis of variance approach to
evaluate the distances between clusters. In short, this method
attempts to minimize the Sum of Squares (SS) of any two (hy-
pothetical) clusters that can be formed at each step and is re-
garded as very efficient, though tends to create clusters of small
size.
Figure 7 shows that such concepts as Insurgents (pictures 1,
2, 5, 6, 7, 8, 9, 10, 11, 12 and 13) and insurgents’ leaders (pic-
tures 15, 16, 17 and 18) though identified separately are also
associated with each other. The same can be seen for the con-
cepts sabotage (pictures 26, 27, 28) and suicide bombers (pic-
tures 19, 21, 22, 24, 25). It is interesting to note that pictures 20
and 23, which were tested in the suicide bombing context,
clustered separately as belonging to both sabotage and suicide
bombing. These pictures show different suicide vests, where all
other suicide bombing pictures depict a suicide bomber using
explosive vests and sabotage pictures show different bombing
Figure 6.
Two armed insurgent topics and Suicide topic consist of 7 pictures each,
Insurgents leaders and Sabotage topics consist of 4 pictures each. The
numbers from 1 to 29 represent pictures. All the other numbers repre-
sent p values of Wilcoxon signed-rank test, which was used to analyse
officers associative emotions evoked to the pictures they observe on the
computer screen.
Figure 7.
Cluster analysis of 602 officers’ responses to the emotional content of
the 25 pictures. Numbers from 1 to 28 represent different pictures.
Concepts Insurgents (pictures 1, 2, 5, 6, 7, 8, 9, 10, 11, 12 and 13) and
insurgents’ leaders (pictures 15, 16, 17 and 18) though identified sepa-
rately are also associated with each other. The concepts sabotage (Pic-
tures 26, 27, 28) and suicide bombers (pictures 19, 21, 22, 24, 25).
Pictures 20 and 23, which were tested in the suicide bombing context,
clustered separately as belonging to both sabotage and suicide bombing.
Copyright © 2012 SciRes. 223
S. IOFFE ET AL.
Copyright © 2012 SciRes.
224
devices. This suggests that consciously and emotionally the
officers tested distinguished well each of the corresponding
concepts and further showed an emotional separation between
suicide/sabotage and insurgents/insurgents’ leaders concepts.
These provided additional support for the pictures chosen for
the testing and their separation into the topics.
Word Group s—Picture Groups An al y si s
We wanted to investigate how the changes of the stimuli
material (removal of some pictures 3, 4, 14 and 29) affected the
outcome of testing. We used the Wilcoxon Matched-Pairs
Signed-Ranks Test to compare the frequencies of the responses
in word topics to word topics and picture topics to picture top-
ics respectively for both initial set of 160 tests (pictures from 1
to 29) and final set of 602 tests (after we have removed picture
3, 4, 14 and 29). These results showed that there are no differ-
ences between initial and final sets of testing in the word topics
(Table 2) as these topics didn’t change per se. The results also
showed that the picture topics, as a whole, had no statistical
difference between initial and final sets of testing. A paired
preference test comparing picture topics in initial and final sets
of testing showed statistically significant increase for the
Armed Insurgent topics at the 96% confidence level. That is, by
removing irrelevant pictures (pictures 3, 4 and 14) from the
topic within the “semantic group”, we have improved statisti-
cally relevant results.
False Positive Response Analysis
Out of the total 762 tests, 5 Afghan Army personnel showed
statistically significant responses to the false positive word
topic and 4 of the subjects showed statistically significant re-
sponses to the false positive picture topics. Statistical analysis
demonstrated a 99% confidence level that the test has a 0.6%
probability of a false positive response for word topics used in
testing and a 0.5% probability of a false positive response for
picture topics. Most importantly, out of the 762 tests, no one
simultaneously demonstrated a combination of both a word
false positive and a picture false positive. Statistically, this data
means that if the psycho-semantic test involves both pictures
and words for any specific topic, there is a 0% chance of a false
positive response. See Table 3 for a tabular representation of
this data.
Analysis of Word and Picture Topics
The major goal of our study was to discover if the responses
to word topics will have corresponding responses in the picture
topics. We found that 94.6% of the word probes correlated to
84.1% of the picture probes for the armed insurgent group top-
ics (Figure 8). Additionally this graphic shows that 5.4% of the
positive responses to word groups could indicate possible false
negatives if only pictures are used in the psycho-semantic test
and 15.9% of the positive responses to pictures could indicate
possible false negatives if only words are used in the psy-
cho-semantic test. Combining pictures and word probes will
identify a group of officers tested, reducing the potential of
false negative results but increasing the potential false positive
maximum from 0% to 0.6% for unmatched word and picture
topics.
We also found that 91.7% of the word probes correlate to 44%
of the picture probes for the suicide bombing topic (Figure 9).
This graph shows that 8.3% of the positive responses to the word
group could indicate possible false negatives if only pictures are
used in the psycho-semantic test and 56% of the positive re-
sponses to pictures could indicate possible false negatives if
Table 2.
The Wilcoxon Matched-Pairs Signed-Ranks Test of the frequencies of
the responses in word topics and picture topics for initial set of 160
tests and final set of 602 tests—upper part. Preference test/Paired com-
parison for the Armed Insurgent topics in initial set of 160 tests and
final set of 602 tests—lower part.
Word Topics Picture Topics
InitialFinalInitial Final
Al-Qaida 6.5 6.3 25 40
Armed Insurgence
Insurgence
Armed Insurgence9.7 7.7 21 18 Insurgence Leaders
Suicide Bombing6.5 6.3 10 14 Sabotage
Hezb-e-Islami 9.7 7.7 13 15 Suicide Bombing
Taliban 12.98.5
W+ = 15 W– = 0
N = 5 p
0.0625
W+ = 2 W– = 8
N = 4 p0.375
Pictures Initial Final
Armed Insurgence25 Versa40 Size 126
T-Score 2.088 is significant at 96% confidence level.
Sum all positive ranks (W+) and all negative ranks (W), the total number of
pairs (N).
Table 3.
Calculations of False-Positives (FP) for lower and upper limits at 99%
confidence level.
Num.
of testsTopics nLower limit
Probability of
false positivesUpper limit
762FP word topic 50.2% 0.6% 1.9%
762FP picture topic40.15% 0.5% 1.7%
762FP word + picture00 0 0
Figure 8.
Word-picture correlation for insurgent groups (graph not to exact scale). 94.6% of the word probes correlated to 84.1% of the picture
probes. 5.4% of the positive responses to word groups are possible false negatives. 15.9% of the positive responses to pictures are pos-
sible false negatives. 112 and 126 are number of statistically significant responses to words and pictures of insurgent topics respec-
tively.
S. IOFFE ET AL.
Figure 9.
Word-picture correlation for suicide bombings (graph not to exact scale). 91.7% of the word probes correlate to 44% of the picture
probes. 8.3% of the positive responses to the word group are possible false negatives. 56% of the positive responses to pictures are pos-
sible false. 24 and 50 are number of statistically significant responses to words and pictures of suicidal topics respectively.
only words are used in the psycho-semantic test. Combining
pictures and word probes will identify this “total” group, re-
ducing the potential of false negative results but increasing the
potential false positive maximum from 0% to 0.6% for un-
matched word and picture topics.
Stress Factor Analysis
During testing we were able to obtain calculation of the
“stress factor”, a measure of a person’s ability to learn and
make decisions under mental stress. The stress is created by
measuring how quickly the person learns from various audio
and visual “corrections”.
We have rated each testee into the following categories:
Excellent: 42 mistakes out of 1850 control probes (2.3%
errors). 7.5% of the testee scored an “excellent” rating for stress
factor.
Good: 43 - 99 mistakes out of 1850 control probes (2.3% to
5.4% errors). 44.4% of the testees scored a “good” rating for
stress factor.
Poor: 100 - 141 mistakes out of 1850 control probes (5.4% to
7.6% errors). 28.3% of the testees scored a “poor” rating for
stress factor.
Below Poor: 142 mistakes out of 1850 control probes (over
7.6% errors). 19.8% of the testees scored a “below poor” rating
for stress factor.
52% of the 762 tested Afghan Army personnel scored be-
tween “Excellent” and “Good”.
Any testee who makes more than 15 consecutive mistakes
will trigger the test to stop. This may result from extremely
poor cognitive learning abilities or a reluctance to follow direc-
tions. 4 of the 762 officers taking this test needed to be retested
because of this control and all passed on the second attempt.
We conducted one way ANOVA statistical analysis to find
out if and how the stress factors were influenced by age. For
example, if younger officers will cope with the tasks better or
will make less errors than older officers. The results of Analysis
of Variance (ANOVA) showed no significant affect of the tes-
tee’s age on the stress factor (Table 4, Figure 10).
We also divided the 762 testees into two populations: those
who passed the test or had not shown statistically significant
responses to any of the topics tested; and those who had statis-
tically significant responses to one or more of the tested topics.
Each of these two populations was then divided into the four
stress factor groups as above. Results of ANOVA analysis
showed that the stress factor was not affected by age, for either
the officers who passed the test or those who failed it (Table 5).
The probability of this result, assuming the null hypothesis, is
0.70.
Figure 11 represents only Afghan Army personnel exhibiting
an “Excellent” stress factor during psycho-semantic testing. We
wanted to point out a number of the testees between ages 39 and
53 made between 3 and 15 mistakes out of the 1850 test probes.
Table 4.
ANOVA one way analysis of the stress factors versa age. Probability of
this result, assuming the null hypothesis, is equal to 0.65.
Source of
Variation
Sum of
Squares
d.f.
Mean
Squares
F
Between 184.7 3 61.57 0.5456
Error 8.1707E+04 724 112.9
Total 8.1892E+04 727
Between—numerator, error—denominator, d.f.—degrees of freedom, F—ratio.
Figure 10.
Age distribution versa stress factor, mean and standard deviations are
shown. Eexcellent, Ggood, Ppoor, BPbelow poor. Numbers on
the left represent age.
Table 5.
ANOVA one way analysis of the stress factors and age versa officers
who passed or failed the SSRM Tek test. Probability of this result,
assuming the null hypothesis, is equal to 0.70.
Source of
Variation Sum of Squares d.f. Mean Squares F
Between 527.2 7 75.31 0.6665
Error 8.1365E+04 720 113.0
Total 8.1892E+04 727
Between—numerator, error—denominator, d.f.—degrees of freedom, F—ratio.
Copyright © 2012 SciRes. 225
S. IOFFE ET AL.
Figure 11.
Scattergram of age versa number of errors for stress factor “Excellent”
rating.
Analysis of the Results of the Defence Reaction
Development Procedur e
Defence reaction development procedure is words flashing
across the computer screen and someone, by trial and errors, is
required to establish for which words one needs to press the
button and for which not. The general tendency would be that
someone would try to slow down response to such words when
compared to the response to meaningless stimuli. When Afghan
military personnel were tested, 40.2% of the officers showed
statistically significant increase of the time of the response.
54.5% had no changes of the response time when compared to
control/meaningless stimuli response time and 5.3% of the
officers showed a statistically significant decrease of the re-
sponse time. Considering that the age of the personnel tested
was skewed towards 40 to 49 years, we did an ANOVA one
way test in the hope of elucidating the age differences (Table 6,
Figure 12) of the three groups of officers who responded by
increase, decrease or no changes to the defence reaction devel-
opment procedure. Assuming the null hypothesis, the probabil-
ity of the age differences for the above discussed three groups
is very high p < 0.0001.
We also analysed stress factor distribution (number of errors
during this procedure) according to increase, decrease or no
change in the response to the defence reaction development
procedure using a one way ANOVA test (Table 7 and Figure
13). The probability of this result, assuming the null hypothesis,
is less than p < 0001.
Discussion
Though psycho-semantic methods are free from many nega-
tive issues of the traditional investigational methods of observ-
ing the human psyche-suggestive or projective methods, we can
see from this study that they are not without their own limita-
tions:
For example, the direct methods of psycho-diagnostics based
on the self-report as it is presented in PsyExpert, reveals only
consciously realised and not actually operating motives. Even
adequately realised motives can be distorted during testing
because of their various social desirabilities and therefore so-
cially undesirable motives are masked and socially desirable
motives are demonstrated. Administrating the PsyExpert test
presented an array of evading strategies employed by most of
the officers even by those who were cleared by both the uncon-
scious SSRM Tek testing and expert opinion.
Based on PsyExpert data analysis one might conclude that
neither written nor computer based question and answer tests
Table 6.
ANOVA one way analysis of the age differences for three groups of
officers who responded by increase, decrease or no changes to the
defence reaction development procedure. The probability of the age
differences for above three groups is p < 0.0001.
Source of VariationSum of Squares d.f. Mean SquaresF
Between 27.45 2 13.73 17.75
Error 585.5 757 0.7734
Total 612.9 759
Between—numerator, error—denominator, d.f.—degrees of freedom, F—ratio.
Table 7.
One way ANOVA analysis for stress factor distribution of three groups
of officers, those who increased, decreased or had no changes of time
of response to the defence reaction teaching procedure. p < 0.0001.
Source of VariationSum of Squares d.f. Mean SquaresF
Between 9.2324E+04 2 4.6162E+04 14.15
Error 2.4769E+06 759 3263
Total 2.5693E+06 761
Between—numerator, error—denominator, d.f.—degrees of freedom, F—ratio.
Figure 12.
Ages of the officers tested according to the results of their responses to
the defence reaction teaching process. Decrease is statistically signifi-
cant decrease of the time of the response to teaching procedure. In-
crease is statistically significant increase of the time of the response to
teaching procedure. No change is no statistically significant changes of
the time of the response to teaching procedure.
will be successful in Afghan culture, though PsyExpert data in
western culture or with the patient interested in their health
issues showed excellent results (Ioffe, Yesin, 2010, 2011).
The SSRM Tek applied on an unconscious level, as a psy-
cho-semantic method, provided diagnostically significant
structurally quantitative information for the organization of
Copyright © 2012 SciRes.
226
S. IOFFE ET AL.
Figure 13.
Officers with a decreased response to the defence reaction teaching
procedure (mean number of errors 69.7); officers with no change re-
sponse to the defence reaction teaching procedure (mean number of
errors 101.3); officers with an increased response to the defence reac-
tion teaching procedure (mean number of errors 115.6).
individual systems of values and attitudes of the officers tested.
In this unconsciously applied psycho-semantic procedure, the
statistics were collected not within the limits of groups of ex-
aminees but within the limits of repeating probes during testing
procedures in a single examinee. The obtained SSRM Tek re-
ports were well corroborated with expert ratings and repeated
application. These reports can then be used as the basis of rec-
ommending further investigation as to why the person demon-
strated a significant reaction to any of the image groups. There
may be a justly appropriate cause for why a person demon-
strates a strong emotional response to a specific or group of
words/images. For example, a Human Intelligence (HUMINT)
“case officer” might be expected to demonstrate a strong emo-
tional response to the espionage word group if he is hiding his
intelligence profession from family or friends. Additionally, a
person may display a strong emotional response to drug or al-
cohol abuse if they are witness to a significant family member
or friend with a problem.
For the cases of the discrepancies between the expert ratings
and SSRM Tek results for the drug/alcohol dependencies topic
we strongly suspect human emotional error comes more into
play for issues like corruption and drug/alcohol abuse where
moral and religious factors play a more important role than for
issues related to insurgency and espionage topics. SSRM Tek
data and this understanding is well corroborated with a study of
military personnel in a different culture (Costrica, Tkachenko,
Malcev, 1998).
The unconsciously applied psycho-semantic method (SSRM
Tek) appears indirectly, presented to the examinee in the form
of a “verbal game”, appealing, seemingly, only to linguistic
competence, meanwhile actually opening the subjective content
of language symbols that is embodied in the structural forma-
tion work of real motives and goals of the subject tested.
A statistically significant response to any semantic-based
probe does not constitute guilt, but rather, is a manifestation
based on a strong emotional reaction to these stimuli. Thus, the
unconscious testing (SSRM Tek) has been identified in the
Afghan environment as an excellent screening technology that
rapidly “passes” individuals that do not display a measurable
reaction to chosen stimuli and points out the individuals who
require further investigation.
In 762 officers tested, we identified 0.5% false positives for
picture topic, 0.6% false positives for word topic and for si-
multaneously occurring picture and word topics zero false posi-
tives. If we can assume that false positive responses to word
topics and picture topics are two independent variables, than
using the probability of 4 instances of picture topics and 5 in-
stances of word topics in 762 tests, it is possible to estimate the
probability of false positives of simultaneous appearance of
picture and word topic events as 3 out of 100,000 tests. Such a
false positive rate seems to be infinitely low considering 10% to
30% false positive rates of current technologies in use, poly-
graphs and stress voice analysers. In the Afghan military popu-
lation, where 16.5% of tested personnel had word and picture
topics correlation, 16,500 officers, out of 100,000 officers
tested, would not pass a test against 3 possible false positives.
Our data has shown a significant correlation for word and
picture topics, but not a 100% correlation. This implies that in
the human psyche different stimuli modalities, for example
words and pictures, though they overlap each other, never ex-
actly correspond with each other. One likely explanation for
this possible false negative data is no semantic space one-for-
one direct match between every word topic and every picture
topic used. For example, we tested some pictures that represent
insurgent leaders and sabotage in which there was no direct
match with one of the previously chosen words. In this case,
some individuals responded to these picture probes (62 positive
responses for insurgent leaders and 41 positive responses for
sabotage) but there were no correlating word probes. In these
cases, it is possible these “unmatched” pictures resulted in some
positive responses whereas there were no corresponding words
in which to form the same positive semantic response.
We demonstrated that internal controls of psycho-semantic
technology provide a powerful metric in which to assess one’s
mental agility or ability to react and learn under stress. Indi-
viduals who achieved Excellent stress factor status are highly
likely to succeed in most tasks requiring decision making skills
or analytical judgements; achieving Good stress factor status
indicates that these individuals can likely be trained for a vari-
ety of different positions requiring decision making skills; Poor
stress factor indicates that these individuals can possibly be
trained for a variety of different positions not requiring decision
making skills; Below Poor stress factor status indicates that
these personnel will have great difficulties accomplishing tasks
requiring mental agility or learning.
Officers who made less than 15 errors during the test deserve
special attention as they are at the very top of the “Excellent”
Stress Factor group and this represents an extremely high abil-
ity to make decisions and learn under mental stress.
We did not find meaningful correlation between the age and
ability to make decision under stress.
Our testing and analysis has shown high correlation between
previously evaluated word topics and corresponding picture
topics representing anti-government activities considered na-
tional threats to the Afghan Government. Therefore, it is possi-
ble to test illiterate Afghan personnel using picture based test-
Copyright © 2012 SciRes. 227
S. IOFFE ET AL.
ing to identify those with motivations and experiences relating
to these anti-government activities. In addition, because seman-
tic space of the word and pictures topics, though they are
largely overlapping, together provided additional inferences
and therefore, the psycho-semantic test for literate personnel
should include both word and picture topics.
There is no information in scientific literature of the impact
of 30 years of war in Afghanistan on the dynamics of psycho-
logical makeup changes in the public. Our research was not
specifically directed to the area of evaluation of the psycho-
logical makeup of the public, but the Afghani military repre-
sents one of the largest employers for men ages 18 to 60 in the
region. Therefore, we will make an effort to use our data as a
glimpse into the issues describing the psychological state of the
tested population. 40.2% of the officers tested, based on their
responses to conscious and unconscious stimuli, represent a
group of average skilled population with average ability to
learn, generally declaring to conform to social standards, but
looking only for their own needs, easily manipulated into so-
cietal congruency with incentives. 54.5% of the officers tested
can be characterised as having low perceptive accuracy, inabil-
ity to differentiate semantic stimuli, unable to effectively act in
complex situations requiring the need to make fast decisions.
Generally, such subjects pursue not an optimal decision pattern,
based on escapism from the problem. Their basic psychological
defence is denial. 5.3% of the officers can be characterised as
having high perceptive accuracy, high speed of cognitive proc-
esses and high accuracy of semantic analysis. In general, such
subjects have well developed intellectual qualities which are the
necessary requirements for leadership position placements. It
can be assumed that in leadership positions such persons, espe-
cially those who showed the smallest number of errors during
testing, will act in any complicated situation competently, re-
sponsibly and choose optimal decisions. The age differences
between the above groups did not provide valuable insights into
the issue at hand.
This study provides analysis of the operational test con-
ducted during 2009-2010 involving 812 Afghan Army person-
nel. It was discovered that the psycho-semantic approach is a
very unique, relatively quick, easy to learn, highly “fieldable”
tool that can be used in the austere Afghan environment to re-
duce human bias in assessing dangerous and hidden traits, ten-
dencies, and motivations in support of the Afghan Personnel
Security Program. The data from these efforts suggests that
individuals who “pass” this test can be reasonably assessed to
be free of personal experiences, motivations, and predilections
to each of the tested topics. Additionally, if administered as part
of a periodic Personnel Security background check program,
this system allows for large-scale screening operations to iden-
tify unknown threats to national security while serving as a
deterrent to undesirable or illegal behaviours and helps to set
high standards in personal conduct.
Will there ever be a right time to re evaluate, widely used in
security but poorly understood and philosophically underde-
veloped concepts of truth verification and lie detection? We are
using great financial and human resources to plug in newly
developed measurement techniques, such as fMRI or thermal
tracking, into truth verification and lie detection programs;
where unconscious psycho-semantic methodologies are vali-
dated and a new field of unsurpassed enquiries and knowledge
for the organization of individual systems of values and atti-
tudes within peoples’ experiential memories is opened.
Conclusion
The unconscious-based SSRM Tek is an innovative solution
to high-throughput personnel security screening requirements
providing a highly unique, real-time reporting capability indi-
cating potential serious problems which require subsequent
investigation. Although the austere Afghan environment de-
mands such out-of-the-box solutions, this tool is equally valu-
able in solving some of the other security challenges such as
access to sensitive programs (surety programs, restricted areas,
etc.), security clearance programs; screening for post-traumatic
stress disorders; routine inprocessing protocols; border crossing
sites, and visa applications.
REFERENCES
Abrams, R. L., Klinger, M. R., & Greenwald, A. G. (2002). Subliminal
words activate semantic categories (not automated motor responses).
Psychonomic Bulletin and Review, 9, 100-106.
doi:10.3758/BF03196262
Alford, R. L., & Strother, J. B. (1990). Attitudes of native and
non-native speakers toward selected regional accents of US English.
TESOL Quarterly, 3, 479-495. doi:10.2307/3587231
Allen, J. J., Iacono, W. G., & Danielson, K. D. (1992). The identifica-
tion of concealed memories using the event-related brain potential
and implicit behavioral measures: A methodology for prediction in
the face of individual differences. Psychophysiology, 29, 504-522.
doi:10.1111/j.1469-8986.1992.tb02024.x
Ansorge, U., Francis, G., Herzog, M. H., & Ömen, H. (2007). Visual
masking and the dynamics of human perception, cognition, and con-
sciousness. A century of progress, a contemporary synthesis, and fu-
ture directions advances in cognitive psychology, 3, 1-2, 1-8.
Begleiter, H., Porjesz, B., & Garozzo, R. (1979). Visual evoked poten-
tials and affective ratings of semantic stimuli. In H. Begleiter (Ed.),
Evoked brain potentials and behaviour (pp. 127-141). New York:
Plenum Press.
Baker, S. J., Chrzan, G. J., Park, C. N., & Saunders, J. H. (1986). Be-
havioral effects of 0% and 0.05% blood alcohol in male volunteers.
Neurobehavioral Toxicology & Teratology, 8, 77-81.
Ball, P. (1983). Stereotypes of Anglo-Saxon and Non-Anglo-Saxon
accents: Some exploratory Australian studies with the matched guise
technique. Language Sciences, 5, 163-183.
doi:10.1016/S0388-0001(83)80021-7
Bazan, A., Shevrin, H., Brakel, L. A. W., & Snodgrass, M. (2007).
Motivations and emotions contribute to a-rational unconscious dy-
namics: Evidence and conceptual clarification. Cortex; a Journal
Devoted to the Study of the Nervous System and Behavior, 43, 1104.
Bayard, D. (1990). God help us if we all sound like this: Attitudes to
New Zealand and other English accents. In A. Bell, & J. Holmes
(Eds.), New Zealand ways of speaking English (pp. 67-96). Clevedon:
Multilingual Matters Ltd.
Benschop, R. (1998). What is a tachistoscope? Historical explorations
of an instrument. Science in Context, 11, 23-50.
doi:10.1017/S0269889700002908
Beznosjuk, E. V., & Smirnov, I. V. (1990). , -
, //  -
. .1, , 1990. . 21. [Psycho-correction, psycho-
prophylactic, psychotherapy.] Medical Psychology and Psycho- Hy-
giene, M.1, MMI, 21.
Bunce, S. C., Bernat, E., Wong, P. S., & Shevrin, H. (1999). Further
evidence for unconscious learning: Preliminary support for the con-
ditioning of facial EMG to subliminal stimuli. Journal of Psychiatric
Research, 33, 341-347. doi:10.1016/S0022-3956(99)00003-5
Cacioppo, J. T., & Berntson, G. G. (1994). Relationship between atti-
tudes and evaluative space: A critical review, with emphasis on the
separability of positive and negative substrates. Psychological Bulle-
tin, 115, 401-423. doi:10.1037/0033-2909.115.3.401
Cheshire, J. (1982). Linguistic variation and social function. In S. Ro-
Copyright © 2012 SciRes.
228
S. IOFFE ET AL.
maine (Ed.), Sociolinguistic variation in speech communities (pp.
153-166). London: Edward Arnold.
Cobb, S. R., Cole, C., & Rainey, M. T. (1978). The effect of exposure
time upon perceived size. British Jounal Physiology Ophtalmology,
32, 94-96.
Costrica, V. G., Tkachenko, A. E., & Malcev, S. I. (1998). Use of a
computer psycho-semantic analysis allows exposure of persons who
are prone to alcohol and drug abuse and asocial behavior. Proceed-
ings of TRTU, 4, 54-55.
Damphousse, K. R., Pointon, L., Upchurch, D., & Moore, R. K. (2007).
Assessing the validity of voice stress analysis tools in a jail setting.
Report submitted to the US Department of Justice.
http://www.ncjrs.gov/pdffiles1/nij/grants/219031.pdf
Dehaene, S., Changeux, J., Naccache, L., Sackur, J., & Sergent, C.
(2006). Conscious, preconscious and subliminal processing: A test-
able taxonomy. Trends in Cognitive Sciences, 10, 204-211.
Dijksterhuis, A., Aarts, H., & Smith, P. K. (2005). The power of the
subliminal: Subliminal perception and possible applications. In R.
Hassin, J. Uleman, & J. A. Bargh (Eds.), The new unconsciousness
(pp. 77-106). New York, NY: Oxford University Press.
Dixon, N. (1971). Subliminal perception: The nature of controversy.
London: McGraw-Hill.
Dixon, N. (1981). Preconscious processing. New York: John Wiley.
El-Dash, L., & Tucker, R. (1975). Subjective reactions to various
speech styles in Egypt. International Journal of the Sociology of
Language, 6, 33-54. doi:10.1515/ijsl.1975.6.33
Epstein, S. (1994). Integration of the cognitive and the psychodynamic
unconscious. American Psychologist, 49, 709-724.
doi:10.1037/0003-066X.49.8.709
Eriksson, A., & Lacerda, F. (2007). Charlatanry in forensic speech
science: A problem to be taken seriously. The International Journal
of Speech, Language and the Law, 14, 169-193.
Esteves, F., Parra, C., Dimberg, U., & Ohman, A. (1994). Noncon-
scious associative learning: Pavlovian conditioning of skin conduc-
tance responses to masked fear-relevant facial stimuli. Psychophysi-
ology, 31, 375-385.
doi:10.1111/j.1469-8986.1994.tb02446.x
Gaies, S. J., & Beebe, J. D. (1991). The matched-guise technique for
measuring attitudes and their implications for language education: A
critical assessment. In E. Sadtano (Ed.), Language acquisition and
the second/foreign language classroom. Anthology Series 28. Sin-
gapore: SEAMEO Regional Language Centre.
Gallois, C., & Callan, V. J. (1981). Personality impressions elicited by
accented English speech. Journal of Cross-Cultural Psychology, 12,
247-359. doi:10.1177/0022022181123006
Greenwald, A. G. (1992). New look 3: Unconscious cognition re-
claimed. American Psychologist, 47, 766-779.
doi:10.1037/0003-066X.47.6.766
Greenwald, A. G., Klinger, M. R., & Schuh, E. S. (1995). Activation by
marginally perceptible (“subliminal”) stimuli: Dissociation of uncon-
scious from conscious cognition. Journal of Experimental Psychol-
ogy: General, 124, 22-42. doi:10.1037/0096-3445.124.1.22
Greenwald, A. G., Draine, S. C., & Abrams, R. L. (1996). Three cogni-
tive markers of unconscious semantic activation. Science, 273,
1699-1702. doi:10.1126/science.273.5282.1699
Greenwald, A. G., Abrams, R. L., Naccache, L., & Dehaene, S. (2003).
Long-term semantic memory versus contextual memory in unconscious
number processing. Journal of Experimental Psychology: Learning,
Memory, and Cognition, 29, 235-247.
doi:10.1037/0278-7393.29.2.235
Grubin, D. J. (2010) The polygraph and forensic psychiatry. American
Academy of Psychiatry and the Law, 38, 446-451.
Heise, D. R. (2010) Surveying cultures: Discovering shared concep-
tions and sentiments. Hoboken: John Wiley and Sons Ltd.
Holender, D. (1986). Semantic activation without conscious identifica-
tion in dichotic listening, parafoveal vision, and visual masking: A
survey and appraisal. Behavioral and Brain Sciences, 9, 1-23.
doi:10.1017/S0140525X00021269
Ioffe, S., Yesin, S. V., Afanasjev, B. G., & Nezhdanov, I. K. (2007a).
Psychosemantic diagnosis of alcoholic dependencies tested at the
unconscious level in military personnel with Posttraumatic Stress
Disorder (PTSD). Polygraph Journal, 2, 57-69.
Ioffe, S., Yesin, S. V., Afanasjev, B. G., & Nezhdanov, I. K. (2007b).
The role of unconscious effects during the treatment of posttraumatic
stress disorder with alcohol dependencies in military personnel.
Polygraph Journal, 3, 133-148.
Ioffe, S., & Konobeevsky, M. (2008). Psycho-semantic spheres of the
personality among correctional facility employees. Journal of Police
and Criminal Psychology, 23, 23-34.
doi:10.1007/s11896-008-9018-7
Ioffe, S., & Yesin, S. (2010). Psyche from within: Three clinical case
studies. Psychology, 1, 282-294. doi:10.4236/psych.2010.14037
Ioffe, S., & Yesin, S. (2011). Controlled reactivation of psycho- seman-
tic space by unconscious (subthreshold) stimuli. Psychology, 2,
181-186. doi:10.4236/psych.2011.23029
Jones, E. E., & Sigall, H. (1971). The bogus pipeline: A new paradigm
for measuring affect and attitude. Psychological Bulletin, 76,
349-364. doi:10.1037/h0031617
Kelly, G. (1955). The psychology of personal constructs. New York:
Norton.
Kelly, G. A. (1970). Behaviour is an experiment. Perspectives in per-
sonal construct theory. London: Academic Press.
Kihlstrom, J. F. (1987). The cognitive unconscious. Science, 237,
1445-1452. doi:10.1126/science.3629249
Kleiner, M. (2002). Physiological detection of deception in psycho-
logical perspectives: A theoretical proposal, in Handbook of Poly-
graph Testing. London: Academic Press.
Knight, D. C., Nguyen, H. T., & Bandettini, P. A. (2003). Expression of
conditional fear with and without awareness. Proceedings of the Na-
tional Academy of Sciences, 100, 15280-15283.
doi:10.1073/pnas.2535780100
Kostandov, E. A. (1977) Восприятие и эмоции. ., 1977. 316 .
[Perceptions and emotions] M. 716.
Kostandov, E. A. (1983). Функциональная асимметрия полушарий
мозга и неосознаваемое восприятие. ., 1983. 246 . [Functional
asymmetry of cerebral hemispheres and unconscious perception] M.
246.
Kostandov, E. A. (1984).    
     //
  . 1984, .34, N 3, .401-411. [Con-
sciousness and unconscious as a physiology of higher neural activity
in human.] Journal of Higher Neural Activity, 3, 401-411.
Kunst-Wilson, W., & Zajonc, R. (1980). Affective discrimination of
stimuli that cannot be recognized. Science, 207, 557-558.
doi:10.1126/science.7352271
Lang, S. F., Nelson, C. A., & Collins, P. F. (1990). Event-related poten-
tials to emotional and neutral stimuli. Journal of Clinical and Ex-
perimental Neuropsychology, 12, 946-958.
doi:10.1080/01688639008401033
Leontiev, V. O. (2002).  . 2002.: -
 - , [Classification of the emo-
tions] Odessa: Innovation centre.
http://emoatlas.narod.ru/book1.html
Lippold, O. C. (1970). Oscillation in the stretch reflex arc and the ori-
gin of the rhythmical, 8 - 12 c/s component of physiological tremor.
The Journal of Physiology, 206, 359-382.
Lippold, O. C. (1971). Physiological tremor. Scientific American, 224,
65-73. doi:10.1038/scientificamerican0371-65
Luriya, A. R., & Vinogradova, O. S. (1971). 
   //Семантическая
структура слова. .: , 1971, . 27-63. Objective research of
the dynamics of the semantic systems.] In Semantic structure of the
word. M. Science, 27-63.
Macmillan, N. (1986). The psychophysics of subliminal perception.
Behavioral and Brain Sciences, 9, 38-39.
doi:10.1017/S0140525X00021427
National Research Council (2003). The polygraph and lie detection.
Committee to Review the Scientific Evidence on the Polygraph.
Washington DC: The National Academies Press.
Ohman, A., & Soares, J. F. (1994). “Unconscious anxiety”: Phobic
Copyright © 2012 SciRes. 229
S. IOFFE ET AL.
Copyright © 2012 SciRes.
230
responses to masked stimuli. Journal of Abnormal Psychology, 103,
231-240. doi:10.1037/0021-843X.103.2.231
Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. (1957). The meas-
urement of meaning. Urbana, IL: Illinois Press.
Osgood, C. E. (1959). Representational model and relevant research
methods. In I. Pool (Ed.), Trends in content analysis (pp. 33-38).
Urbana, IL: Illinois Press.
Osgood, C. E. (1967). Semantic differential technique in the compara-
tive study of cultures. In L. A. Jakobovite, & M. S. Miron (Eds.),
Readings in the psychology of language (pp. 111-199). Englewood
Cliffs: Prentic-Hall.
Osgood, C. E. (1976). Focus of meaning. Mouton.
Pessiglione, M., Petrovic, P., Daunizeau, J., Palminteri, S., Dolan, R. J.,
& Frith, C. D. (2008) Subliminal instrumental conditioning demon-
strated in the human brain neuron. 59, 561-567.
doi: 10.1016/j.neuron.2008.07.005
Ramachandran, V. S., & Cobb, S. (1995). Visual attention modulates
metacontrast masking. Nature, 373, 21.
Reingold, E., & Merikle, P. (1988). Using direct and indirect measures
to study perception without awareness. Perception and Psychophys-
ics, 44, 563-575. doi:10.3758/BF03207490
Reingold, E., & Merikle, P. (1990). On the inter-relatedness of theory
and measurement in the study of unconscious processes. Mind and
Language, 5, 9-28. doi:10.1111/j.1468-0017.1990.tb00150.x
Renshaw, S. (1945). The visual perception and reproduction of forms
by tachistoscopic methods. Journal of Psychology, 20, 217-232.
doi:10.1080/00223980.1945.9917254
Senter, S., Weatherman, D., Krapohl, D., & Horvath, F. (2010). Psy-
chological set or differential salience: A proposal for reconciling
theory and terminology in polygraph testing. Polygraph, 39,
109-117.
Shevrin, H, & Fritzler, D. (1968). Visual evoked response correlates of
unconscious mental processes. Science, 161, 295-298.
doi:10.1126/science.161.3838.295
Shevrin, H., Smith, W., & Fritzler, D. (1969). Repressiveness as a
factor in the subliminal activation of brain and verbal responses.
Journal of Nervous and Mental Disease, 149, 261-269.
doi:10.1097/00005053-196909000-00003
Shevrin, H. (1973). Brain wave correlates of subliminal stimulation,
unconscious attention, primay- and secondary-process thinking, and
repressiveness. Psychological Issues Monograph 30, Psychoanalytic
Research, 8, 56-87.
Shevrin, H., & Dickman, S. (1980). The psychological unconscious, a
necessary assumption for all psychological theory? American Psy-
chologist, 35, 421-435. doi:10.1037/0003-066X.35.5.421
Shevrin, H., Williams, W. J., Marshall, R. E., Hertel, R. K., Bond, J. A.,
& Brakel, L. A. (1992). Event-related potential indicators of the dy-
namic unconscious. Consciousness and Cognition, 1, 340-366.
doi:10.1016/1053-8100(92)90068-L
Shevrin, H., Bond, J. A., Brakel, L. A. W., Hertel, R. K., & Williams,
W. J. (1996). Conscious and unconscious processes: Psychodynamic,
cognitive, and neurophysioloical nonvergences. New York: Guilford
Press.
Smirnov, I., Beznosjuk, E. & Zhuravlyov, A. (1995). -
:    -
   . ., 1995. . 98-103. [Psychote-
chnologies: Computer psychosemantic analysis and psychocorrection
at unconscious level.] M. “Culture”, 98-103.
Snodgrass, J. M. (2000). Unconscious perception: Theory, method, and
evidence. Amsterdam: John Benjamins.
Tjuvina, N. A., & Krivtsov, K. J. U. (2006)  
    
    
    
.    . . . .
106 (11) 50-53. [Application of a method of the computer psy-
cho-semantic analysis for early diagnostics of possible relapse in the
patients with the paranoid form of a schizophrenia during of
medicamentous remission]. Journal of Neurology and Psychiatry of
S. S. Korsakova.. 106, 50-53.
Waller, G., & Mijatovich, S. (1998) Preconscious processing of threat
cues: Impact on eating among women with unhealthy eating attitudes.
International Journal of Eating Disorders, 24, 83-89.
doi:10.1002/(SICI)1098-108X(199807)24:1<83::AID-EAT7>3.0.CO
;2-X
Ward, J. H. (1963). Hierarchical grouping to optimize an objective
function. Journal of American Statistical Association, 58, 236-244.
doi:10.2307/2282967
Wedding, D., & Stalans, L. (1985) Hemispheric differences in the per-
ception of positive and negative faces. International Journal of Neu-
roscience, 27, 277-281. doi:10.3109/00207458509149773
Winkielman, P., Berridge, K. C., & Wilbarger, J. L. (2005). Uncon-
scious affective reactions to masked happy versus angry faces influ-
ence consumption behavior and judgments of value. Personality &
Social Psychology Bulletin, 31, 121-135.
doi:10.1177/0146167204271309
Wong, P. S., Shevrin, H., & Williams, W. J. (1994). Conscious and
nonconscious processes: An ERP index of an anticipatory response
in a conditioning paradigm using visually masked stimuli. Psychop-
sysiology, 31, 87-101.
doi:10.1111/j.1469-8986.1994.tb01028.x
Wong, P. S., Bernat, E., Bunce, S. C., & Shevrin, H. (1997). Brain
indices of nonconscious associative learning. Consciousness and
Cognition, 6, 519-544. doi:10.1006/ccog.1997.0322
Zayas, V., Greenwald, A. G., & Osterhout, L. (2011). Unintentional
covert motor activations predict behavioral effects: Multilevel mod-
eling of trial-level electrophysiological motor activations. Psycho-
physiology, 48, 208-217.
doi:10.1111/j.1469-8986.2010.01055.x
Yarbus, A. L. (1967). Eye movements and vision. New York: Plenum
Press.
Disclaimer: The views and opinions expressed in the article are those of
the author (or authors) and not the Department of Defense.