SEARCH

SEARCH BY CITATION

Keywords:

  • Motivation;
  • Respiration;
  • Electrodermal;
  • Social factors;
  • Normal volunteers

Abstract

  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. References

A mock crime experiment was designed in which 100 participants were randomly assigned to five conditions: four experimental conditions in a 2 × 2 factorial design (two guilt conditions—guilty or informed innocents, crossed with two incentive level conditions—with or without a promised reward for proving innocence), and one control condition of uninformed innocents. Along with the common dependent polygraph measures, attitudes towards cooperating with the test were assessed. Results indicated that the informed innocents preferred to cooperate whereas guilty participants preferred to try and obstruct the test. These tendencies were amplified among participants who were promised a reward. The cooperative choice attenuated electrodermal responses to the critical items. Respiration measures were sensitive to the incentive level manipulation. Implications of the results for future research and for actual detection of information tests were discussed.

Criminal justice experts are eager to know whether a given person is withholding information regarding a criminal event. Therefore, scientists and forensic experts are in a continuous search for appropriate approaches and methods that can help the legal system in this respect. A psychophysiological detection of information method that serves this purpose was proposed by Lykken (1959) and since then has gained considerable scientific attention (see Ben-Shakhar & Elaad, 2003, for a review). The method is known as the Concealed Information Test (CIT), or the Guilty Knowledge Test (GKT). The CIT utilizes a series of multiple-choice questions, each having one critical or correct alternative pertaining to a feature of the crime under investigation, and several incorrect alternatives serving as controls. The critical and control items are chosen so that an innocent suspect who has no crime-related knowledge cannot discriminate among them (Lykken, 1998). Typically, if the suspect's physiological responses to the correct alternatives are consistently larger than to the controls, possession of concealed information about the event in question is inferred. This is the classic cognitive approach to the CIT (Lykken, 1974).

Research on the CIT in laboratory settings has revealed highly accurate decisions for both guilty and innocent participants (Ben-Shakhar & Elaad, 2003). However, in most of these studies, detection rates were computed for fully informed guilty participants and for fully ignorant innocents. From a practical point of view, such comparisons may raise questions. For example, innocent examinees may gain access to critical information through the mass media description of the crime. In addition, suspects could be exposed to critical information during the interrogation itself. For example, Krapohl (2011) cited a case study where an innocent suspect whom the police wanted to fail the CIT was placed in a jail cell with an agent who implanted the critical information of the investigated crime in the innocent suspect by repeating an invented story in which the critical items appeared. Critical information may also leak through contact with witnesses and other people who may have some knowledge of the crime. However, as long as innocent suspects can explain how they became aware of the critical information, the problem is not severe. But, innocent suspects who were exposed to critical details of information are frequently unaware of this exposure and are unable to account for the sources of their knowledge. Furthermore, some knowledgeable, but innocent, suspects might refuse to admit their knowledge because they might fear reprisal by the culprit.

Studying informed innocent participants is also important from a theoretical perspective. Ben-Shakhar and Elaad (2003) reported that highly motivated deceptive guilty participants were more often detected in the CIT than less motivated deceptive guilty participants. It seems that whenever motivation to prove innocence is involved, informed guilty participants are encouraged to invest more effort to obstruct the test and avoid detection (motivation to cope with the test) while ignorant innocent participants are not affected by the motivation manipulation at all. Therefore, whenever guilty and innocent participants are compared when accounting for factors other than mere knowledge (e.g., emotions, act of deception, motivation to prove innocence), one should consider informed innocents in addition to ignorant innocents.

Results of mock-crime studies where innocent participants were exposed to critical information in an innocent context (e.g., Ben-Shakhar, Gronau, & Elaad, 1999; Bradley & Rettinger, 1992; Bradley & Warfield, 1984) showed that the detection rates of guilty participants were higher than those of innocent participants who shared the same critical information (but see also Gamer, Kosiol, & Vossel, 2010; Nahari & Ben-Shakhar, 2011, for different reports). The following question is whether informed innocents can be discriminated from noninformed innocents. Elaad and Ben-Shakhar (1989) reanalyzed Bradley and Warfield's (1984) data and demonstrated that the mean detection score based on the pooled data of all three informed innocent groups tested in this study was greater than chance. Similar results were reported by Ben-Shakhar et al. (1999) and more recently by Zvi, Nachson, and Elaad, (2012) and by Elaad (2013). To conclude, the mere knowledge of the relevant information in itself enhances physiological responses of informed innocent examinees.

In all these studies, the Guilty Action Test (GAT) questioning format (Bradley & Warfield, 1984) was used. The GAT format changed the question wording from passive knowledge (e.g., “What was the color of the stolen envelope?”) to active participation (e.g., “Was the color of the envelope you stole …?”). This guarantees that informed innocents are truthful when denying involvement with the crime-related items whereas guilty participants' denials are lies. However, when both deception and knowledge were controlled, and the effect of guilt was examined alone, using the standard CIT (“Do you know …?”) format, very high rates of false positives (90% and 70%) were reported for informed innocent participants (Bradley, MacLaren, & Carle, 1996). Results suggested that, when informed innocent examinees were put into a situation where they deny information they know is true (as is the case with the CIT), they are especially reactive on the critical items. When informed innocent participants were truthful on the GAT, 50% were still judged as guilty. This rate is higher than typically found with innocent participants who have no knowledge. It seems that, in experimental settings, guilt alone is insufficient to differentiate between informed guilty and informed innocent participants (Elaad, 2009, 2011).

Hartwig, Granhag, and Stromwall (2007) found that a majority of guilty participants in a mock study reported to have a plan before facing the interrogation whereas only a few innocent participants reported to having prepared for the interview. Innocent suspects' behavior was explained by their belief that telling the truth “as it is” is sufficient to exonerate them. Applying these results to real-life interrogations may suggest that guilty suspects are threatened by the police polygraph test because they know that their odds of “beating” the test are low. Actually, they expect to be punished after the completion of the test. Those who ultimately agree to take the test despite the poor odds have some hope of success based on rumors and popular literature indicating that the polygraph makes mistakes. Such guilty suspects understand that their interest (to hide the truth) conflicts with the polygraph examiners' interest (to reveal the truth) and in order to succeed they have to obstruct the polygraph operation. They mistakenly believe that this can be achieved by increased attention to the critical items. However, attention enhances physiological responding to the identified information.

Innocent suspects who agree to take the test expect to be released from further questioning after its completion. They feel confident that they are able to prove their innocence in the test and that the polygraph operators' interest (to arrive at a correct decision) coincides with their own interests (to be found truthful). Furthermore, innocent suspects tend to have a naïve faith in the power of their own innocence to set them free, and seem to trust the criminal justice system (Kassin, 2005).

This attitude of innocent suspects was observed in earlier accounts on guilty and innocent people's verbal behavior that showed that liars and truth tellers have different strategies in response to the interrogation (e.g., Colwell, Hiscock-Anisman, Memon, Woods, & Michlik, 2006; Granhag & Hartwig, 2008). Specifically, the psychology of guilt predicts that guilty suspects will use uncooperative strategies with respect to possibly self-incriminating information (Stromwall, Hartwig, & Granhag, 2006), whereas innocent suspects will use much more straightforward strategies (Kassin, 2005). For example, innocent people may volunteer being at the crime scene, while guilty people tend to avoid mentioning such information (Hartwig, Granhag, Stromwall, & Vrij, 2005). Overall there is support for the notion that truth tellers are more cooperative and forthcoming than liars (DePaulo et al., 2003).

Cooperation instructions were previously included in an altered polygraph Comparison Question Test (Bradley, MacLaren, & Black, 1996; Cullen & Bradley, 2004). It was reported that, although informed innocent participants recognized the relevant questions as threatening, their cooperation (truthful answers to the relevant questions) was associated with lower physiological responses than those of the less cooperative guilty participants.

The two different states of mind (that of guilty examinees and that of innocents) were the focus of Zvi et al.'s (2012) study. Zvi and colleagues instructed guilty and informed innocent participants to either cope (exert effort to avoid detection) with the polygraph or be cooperative. Results indicated that guilt was associated with stronger physiological responses to the critical items than informed innocence and that coping instructions were associated with stronger physiological responses to the critical items than cooperative instructions. Note that Zvi et al. (2012) used the GAT questioning format and therefore compared deceptive guilty participants with truthful informed innocents. In another recent study, Elaad (2013) used only informed innocent participants in the GAT format and showed that cooperation instructions, supported by incentive to cooperate, attenuated the responses to identified critical details.

Both Zvi et al. (2012) and Elaad (2013) reported that informed innocent participants who were motivated to pass the test could have deeply encoded crime-related memories but still attenuate their responses to the critical items during the GAT. This stands in contrast to the results of other recent studies that reported no different responses to the critical items between informed innocents and guilty participants (Gamer, 2010; Gamer et al., 2010; Nahari & Ben-Shakhar, 2011). Gamer et al., (2010) concluded that, when informed innocents are able to remember critical details of the crime during the GAT, they have a high risk of failing the test. These contrasting results need more research, and the present study will provide further answers to the deeply encoded information question.

Most important, the participants in both previous studies (Elaad, 2013; Zvi et al., 2012) were not free to decide for themselves whether they were willing to cope or cooperate on the test. It may be argued that such a procedure lacks ecological validity. The present study seeks to examine whether guilty and informed innocent participants demonstrate different levels of willingness to cooperate on the test when they are informed of the advantages that cooperation has for innocent but not for guilty examinees. Results should shed light on the potential benefits of incorporating cooperation instructions in real-life detection of information test settings. It is suggested that a proper warning that cooperation is beneficial to the interests of innocent examinees and operates against the basic interests of deceptive participants encourages innocent but not guilty participants to adopt a cooperative attitude toward the test, and behave accordingly.

Unfortunately, guilt and innocence in low-stakes experimental settings fail to simulate real guilt and real innocence. Participants in an experiment know perfectly well that they are participating in a simulation and that no harm will be inflicted upon them as a consequence of the polygraph test results. One way to compensate for this shortcoming is to manipulate the level of motivation to prove innocence, on the assumption that higher motivation levels simulate better than lower motivation levels in actual testing conditions. The present study will therefore manipulate the participants' motivation to prove their innocence. Hence, the option to either cooperate or not on the test will be presented to motivated and less motivated guilty and informed innocent participants tested with the GAT questioning format. It is hypothesized that results will show a clear preference of informed innocents to cooperate. In contrast, guilty examinees will prefer to cope. From a theoretical point of view, results may provide further support for the distinctive state of mind that guilty and informed innocent examinees possess.

Method

  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. References

Participants

The participants were 100 undergraduate psychology students (23 males and 77 females) who were recruited through ads placed on bulletin boards throughout the campus. Their age ranged between 19–34 years with a mean age of 23.5 (SD = 2.71). Participants received course credit for participation in the experiment. Some received extra course credit for truthful classification on the test.

Design

Two factors, guilt (guilty and informed innocent participants) and incentive to prove innocence (promised reward for success and no promised reward), were put to a test in a 2 × 2 between-subjects design. The combination of both factors forms four experimental conditions: (1) guilty participants who actually committed a mock theft and were promised a bonus for proving their innocence in the test, (2) guilty participants who actually committed a mock theft and were instructed to prove their innocence in the test but were promised no reward for success, (3) informed innocent participants who were promised a reward for a truthful outcome in the test, and (4) informed innocent participants who were instructed to yield a truthful outcome in the test but were promised no reward.

In addition, a fifth control group of uninformed innocent participants was included. All five groups were offered the opportunity to cooperate through receiving a detailed description of how to cooperate, and what interests such a cooperation would serve. The dependent variables were (a) detection of concealed information with three physiological indices (skin conductance, respiration, and finger pulse changes), and (b) adoption of cooperative attitude and behavior during the test as indicated by self-reports after the test. Participants were randomly allocated into the five conditions. Each group was comprised of 20 participants.

Procedure

The procedure of the present experiment was similar in many aspects to the procedure described by Zvi et al. (2012). Therefore, only differences and main points will be outlined.

The experiment was conducted by two female experimenters who performed two different roles. They changed roles for about one half of the participants. Experimenter A greeted the participants individually and informed them that they were going to participate in a polygraph experiment. Participants were then asked for their name, gender, and age, and were handed a consent form, which by signing it they indicated their consent to undergo a polygraph test. All participants agreed to sign the form.

The participants were assigned to one of the five conditions according to a predetermined random order. Note that the ratio of male and female participants was preserved in the different conditions (7 males were assigned to the guilty–high incentive condition, and 4 males were assigned to each of the other four conditions). Participants were randomly assigned to one of four different crime profiles according to their gender and month of birth (see Zvi et al., 2012, for details).

Each crime profile specified six features of an envelope theft: The envelope's color and addressee, as well as a piece of jewelry, a sum of money, an object, and a photograph of an animal—all were contained in the stolen envelope. The profiles were (1) a yellow envelope addressed to the research authority, containing 57 New Israeli Shekels (NIS), a necklace, a calculator, and a photograph of a bear; (2) a green envelope addressed to the computer department, containing 24 NIS, a ring, a pen, and a photograph of a rhinoceros; (3) a red envelope addressed to the security officer, containing 49 NIS, earrings, spectacles, and a photograph of a zebra; and (4) a blue envelope addressed to the academic secretary, containing 63 NIS, a bracelet, a key, and a photograph of a lion.

The assignment of profiles to the examinees was random as demonstrated by applying a chi-square test for independence on the profiles' frequencies in the four experimental conditions, χ2(9) = 12.01, ns.

A fifth profile of neutral control items (a gray envelope addressed to the students' dean, containing 31 NIS, a watch, a camera, and a photograph of a vulture) was used only in the interrogation phase of the experiment. Finally, a sixth buffer profile of neutral items (a black envelope addressed to the finance section, containing 16 NIS, a pendant, cell phone, and a photograph of an elephant) followed the presentation of each question in order to absorb the initial orienting response (OR).

Guilt and informed innocence manipulations

The participants simulating guilt received an explanation of the nature of the experiment from Experimenter A, who told them to take one of six instruction sheets placed on a small table, read it carefully, and follow the instructions. Experimenter A warned the participants that any failure to complete the indicated task would result in losing the promised course credit.

Participants then picked up an instruction sheet that instructed them to commit a mock theft in an office adjacent to the polygraph laboratory where they were later tested. Four alternative crime profiles were used: (1) removal of a book from a shelf in order to get a yellow envelope, (2) removal of a briefcase in order to get a green envelope, (3) removal of a coat in order to get a red envelope, and (4) pushing aside a telephone in order to get a blue envelope (contents of the envelopes are described above). The participants were instructed to take the contents of the envelope and put them in their pocket or bag, and then return to the experimenter.

The informed innocent participants were told by Experimenter A that the purpose of the study was to find out whether they could prove their innocence in a polygraph test. The participants were asked to take an instruction sheet respective to their gender and month of birth, and leave the room. Once outside, they were directed to a trash bin. They looked inside for a specific package (out of five) that contained the envelope with a piece of jewelry, a sum of money, an object, and a photograph of an animal. The participants were then instructed to look inside the envelope, make sure that all the articles listed on the instruction sheet were in place, and that the envelope's listed color and addressee were correct. Then, they were instructed to take the envelope with all of its contents and deliver it to Experimenter A. Such a manipulation guarantees that both the guilty and the innocent participants obtained the critical information by handling the critical items (Zvi et al., 2012).

Upon returning, the experimenter told the guilty and the informed innocent participants that they were suspected as thieves because they were seen near the office from which the envelope was stolen and that the following polygraph test would determine their involvement in the mock crime.

Cooperation and incentive manipulations

All participants received written and oral instructions that informed them that the polygraph is designed to help people prove their innocence. Therefore, if they are innocent and did not steal the envelope, they are advised to cooperate with the polygraph by adopting a relaxed and calm attitude and by following the examiner's instructions so as to allow the polygraph to perform under the best possible conditions. They were further told that if they adopt a coping attitude by staying alert, ready to act, and attentive to the questions asked, the polygraph would operate under less than optimal conditions and its effectiveness may be impaired. All the participants were further instructed not to move during the entire examination so as to avoid the impression that they were trying to obstruct the test.

Half of the innocent and half of the guilty participants were informed that they would have an extra course hour to their credit that they could keep if the polygraph indicated that they were telling the truth. The other half of the participants were promised no reward for a truthful outcome.

Uninformed innocence manipulation

The uninformed innocent participants were asked to indicate their gender and month of birth and then leave the room and wait outside for 3 min. Upon returning, they were told that they were suspected of stealing money from an envelope in an adjacent office. The participants were told that they had to prove their innocence in a polygraph test and were informed that the polygraph is designed to help people prove their innocence. They were promised no reward for a truthful outcome on the test.

The inclusion of another control group with a promised reward could have balanced the design. However, there is no theoretical or other reason to believe that the promised reward would elicit different responses than those of any other uninformed group. It was therefore decided not to include a second control group.

The polygraph test

After performing their role in the experiment, the participants met again with the experimenter who directed them to the examination room where they met the other experimenter who conducted the polygraph test. Experimenter B was aware of the participants' guilt condition and the instructions they were given, but was unaware of their randomly assigned critical profile. Experimenter B told the participants that they would undergo a polygraph test regarding a theft of a missing envelope. She reminded them that their task was to prove their innocence in the polygraph test.

Experimenter B accompanied the participants to the examination room, and asked them to sit down in the examination chair, lean on the back support, place their hands on the arms support, look at the monitor in front of them, and refrain from moving during the entire test. The experimenter then attached the polygraph devices to the examinees, while explaining the function of each device (e.g., “The device that I am attaching to your finger will record your blood flow and heart rate”).

The experimenter returned to the adjacent control room and, through a speaker, explained that the polygraph would determine whether the participants were involved in the theft on the basis of their physiological responses. She reminded the participants that the polygraph works in accord with the best interest of innocent examinees to yield truthful outcomes, and made it clear that the purpose of the polygraph was to determine whether the participants are guilty or innocent.

The questions were presented to the participants after an initial rest period of 2 min during which skin conductance baseline was recorded. Six questions were presented, each focusing on a different feature of the mock crime (the color of the stolen envelope, the name of the addressee, the stolen sum of money, the stolen jewelry, the stolen object, and the stolen photograph). The questions were presented on the computer monitor and were read aloud to the participants through prerecorded sound files. Items were recorded by a male assistant who was unaware of the critical items. The visual presentation of each item lasted for 5 s and the acoustic presentation for about 1 s. The acoustic presentation appeared about 1 s after the onset of the visual presentation.

Each question contained 11 items: one critical (e.g., the color of the stolen envelope) and four neutral control items (other colors of envelopes), which were presented in a random order and then repeated in a different order. A neutral buffer item (a different color) presented at the beginning of each question served to absorb the initial orienting response. Interstimulus intervals ranged from 16 to 24 s, with a mean interval of 20 s. The questions were presented in a random sequence with a break after the first three questions, designed to give the participants the opportunity to take a deep breath and move their limbs. Subsequently, the remaining three questions were presented. Whenever necessary, the experimenter commented on the behavior of the participants in the first part of the experiment (e.g., do not yawn, do not cross your legs), and reminded them to sit straight and refrain from moving while being tested. All questions were formulated according to the GAT questioning format. For example, with relation to the stolen sum of money, participants were asked: “Was the amount of money that you stole …?” The question was followed by a sequence of different alternative answers (75 NIS, 63 NIS, etc.). Participants replied “No” indicating that they denied stealing the respected sum of money. The GAT format guaranteed that guilty participants lied in response to the critical items, and told the truth in response to the controls. The innocent participants (both informed and uninformed) told the truth in response to both the critical and the neutral items.

Following the polygraph test, the participants in the four experimental groups were asked to name the six critical items of their profile, and to indicate their level of excitement in the test. Excitement level was indicated on scales ranging from 0 (not at all excited) through 100 (highly excited). The next question referred to cooperation. Participants were asked to indicate how much they cooperated on the test in order to help it arrive at a correct decision. Answers ranged from 0 (not at all) to 100 (all the time) with a midpoint of 50 (sometimes yes and sometimes no). Finally, participants were asked to estimate how successful they were in the test, with success defined as a polygraph's indication that they were not involved in the theft raised by a particular question. Their answers ranged from perfect success (7; proved their innocence on all 6 questions) to absolute failure (1; couldn't prove their innocence on even one question). Finally, participants were debriefed about the purpose of the study and, if deserved, they received the promised reward. They were asked not to discuss the experiment with their fellow students during the semester.

Response scoring and analysis

Participants' physiological responses were transmitted in real time to the computer. A sampling rate of 20 per second was always used.

Electrodermal responses

Skin conductance response (SCR) was computed as the maximal increase in conductance obtained from the participant, from 1 to 5 s after stimulus onset (Ben-Shakhar et al., 1999).

Finger pulse

Finger pulse waveform length (FPWL) responses were defined as the measured line length of the pattern that describes the peripheral blood vessel activity within 15 s from stimulus onset. The FPWL response combines pulse rate deceleration and pulse amplitude constriction. This is reflected by a shorter line length (the shorter the line, the stronger is the response) (see Elaad & Ben-Shakhar, 2006, for details).

Respiration

Respiration responses were defined on the basis of the total respiration line length (RLL) during the 15-s interval following stimulus onset, where shorter lines correspond to stronger responses (see Elaad, Ginton, & Jungman, 1992, for details).

Results

  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. References

Recall of Crime-Related Items

At the end of the experiment, the participants were asked to recall the items that appeared on their instruction sheet. The mean recall scores were 5.70 (SD = 0.57) and 5.65 (SD = 0.55) for the high- and low-incentive guilty groups, respectively, and 5.40 (SD = 0.60) and 5.65 (SD = 0.93) for the respective high- and low-incentive informed innocents.

A 2 (Guilt; guilty vs. innocent) × 2 (Incentive Level; high vs. low) analysis of variance (ANOVA) was conducted on the recalled items. Neither guilt, F(1,76) = 0.43, nor incentive level, F(1,76) = 0.97, nor their interaction, F(1,76) = 1.72, yielded a significant effect. This indicates that guilt and incentive level were not affected by memory. The uninformed innocent participants (control group) were asked to guess the critical items and correctly guessed on average 0.75 (SD = 0.96) of the items.

Cooperation

Following the completion of the test, the participants were asked to “indicate your level of cooperation with the polygraph during the test.” The answers to the question were given on a scale ranging from 0 (not at all) through 100 (all the time). The mean cooperation scores are displayed in Table 1.

Table 1. Means (SDs) of Cooperation Ratings on the Test
 High-incentiveLow-incentiveAcross
Informed participants
Guilty65.0 (40.8)69.5 (40.1)67.3 (40.0)
Innocent97.5 (7.2)83.0 (30.1)90.3 (22.8)
Across81.3 (33.3)76.3 (35.6)78.8 (34.4)
Uninformed innocents  96.5 (7.5)

The present control group does not fit into a balanced design. Therefore, an approach that permits us to consider the entire variance (Winer, 1962) was used. First, a one-way ANOVA on the cooperation scores with condition (the five experimental conditions) as the between-subjects factor was performed. The analysis revealed a significant condition effect, F(4,95) = 5.23, p = .001, inline image, indicating that the conditions differ in the reported cooperation on the test.

To assess the effect of knowledge on the differences in the cooperation scores, a planned orthogonal Helmert contrast was used to compare the mean cooperation score of the four knowledge conditions against the mean of the uninformed innocent control condition. The mean difference (MD) was significant (MD = 17.75, SE = 7.32, p = .017), indicating that participants in the control group reported more cooperation than informed participant.

Following Winer (1962), we conducted a multifactor analysis on the experimental treatments to assess whether guilty and informed innocents differ in their reported cooperation on the test, and whether the incentive level affected these accounts as well. The cooperation scores of the experimental groups were analyzed using a 2 (Guilt) × 2 (Incentive Level) ANOVA. The data revealed a significant, F(1,76) = 10.00, p = .002, inline image, main effect for guilt, indicating that informed innocent participants were more ready than guilty participants to cooperate on the test. Neither the incentive level effect, F(1,76) = 0.47, nor the interaction effect, F(1,76) = 1.71, were significant.

Excitement

Participants were asked at the completion of the test to indicate their level of excitement during the test. They gave their answers on a scale ranging from 0 (not at all excited) through 100 (highly excited). The mean excitement scores were 68.75 (SD = 16.04) and 52.00 (SD = 26.47) for the high- and low-incentive guilty groups, respectively, and 60.25 (SD = 28.67) and 40.50 (SD = 24.16) for the high- and low-incentive informed innocent groups, respectively. The mean excitement score computed for the control group was 42.50 (25.52). A one-way ANOVA on the excitement scores, with the five levels of condition serving as the between-subjects factor, was performed. The analysis showed a significant condition effect, F(4,95) = 4.71, p = .002, inline image, indicating that the conditions differ in the excitement participants experienced during the test.

The effect of knowledge on the differences in the excitement scores was assessed using a planned orthogonal Helmert contrast. The mean excitement score of the four knowledge conditions was compared against the mean of the uninformed innocent control condition. Results revealed a significant mean difference (MD = −12.88, SE = 6.14, p = .039), indicating that the uninformed innocent participants reported less excitement during the test than the informed innocent and guilty participants.

A 2 (Guilt) × 2 (Incentive Level) ANOVA on the excitement reports yielded a significant, F(1,76) = 11.27, p = .001, inline image, incentive level effect, which indicates that a promised reward for successful outcomes was associated with higher levels of reported excitement (M = 64.50, SD = 23.33) than no promised reward (M = 46.25, SD = 25.69). Guilt effects on excitement reports were marginal, F(1,76) = 3.38, p = .07, inline image. There was a tendency of guilty participants to report more excitement (M = 60.38, SD = 23.22) than their fellow informed innocents (M = 50.38, SD = 28.02). The Guilt × Motivation interaction effect, F(1,76) = 0.08, was not significant.

Perceived Success on the Test

A major difference between the states of mind of guilty and innocent suspects concerns the perceived prospect of success, which was defined as a polygraph indication that they were not involved in the theft. The mean success scores were: 3.50 (SD = 1.60) and 3.60 (SD = 1.23) for guilty participants in the high- and low-incentive groups, respectively, and 4.15 (SD = 1.84) and 5.25 (SD = 1.29) for the informed innocent participants in the respective high- and low-incentive groups. The mean success score of the control group was 5.50 (1.39). A one-way ANOVA on the success scores where condition (5 levels) served as the between-subjects factor, exhibited a significant condition effect, F(4,95) = 4.71, p = .002, inline image, indicating that the conditions differ in the perceived success scores.

The effect of knowledge on the success scores was assessed through a planned orthogonal Helmert contrast. It was found that the mean difference was significant (MD = 2.63, SE = 0.37, p < .001). Results indicated that the control group predicted more success in the test than the informed participants.

A 2 (Guilt) × 2 (Incentive Level) ANOVA was conducted on the participants' predicted success scores. A significant guilt effect, F(1,76) = 11.55, p < .001, inline image, showed that the informed innocent participants predicted more success than the guilty participants. This implies that the former felt more secure about the outcomes of the test than the latter. The incentive level effect was marginally significant, F(1,76) = 3.14, p = .08, inline image. The interaction effect was insignificant, F(1,76) = 2.18, p = .14, inline image.

Tonic Skin Conductance Level

The subjective nature of questionnaires and the sensitivity of questionnaires to demand characteristics would suggest the addition of a more objective method of manipulation check. This can be achieved by looking at the tonic responses of the skin conductance level (SCL) in various stages of the test (Elaad, 2013). To this end, the minimal and maximal SCL raw scores (in image units) from 1 to 5 s after stimulus onset, which were elicited in response to critical details, were used. First, the mean of the two raw scores was computed. Then, mean SCL scores to the critical items were computed across the six test series, for each participant. Mean scores across participants within each experimental condition served to indicate the SCL of that condition. The SCL means are presented in Table 2.

Table 2. Tonic SCL Means (SDs) Elicited for Critical Items
 High-incentiveLow-incentiveAcross
Informed participants
Guilty7.80 (3.32)6.13 (2.82)6.96 (3.15)
Innocent7.06 (3.82)7.83 (3.38)7.44 (3.58)
Across7.42 (3.55)6.98 (3.19)7.21 (3.36)
Uninformed innocents  7.44 (3.55)

Note that for guilty participants the incentive for success raised the SCL. For informed innocents, the opposite is true and the incentive attenuated SCL. Furthermore, the high-incentive guilty group showed higher SCL scores than the control group whereas the high-incentive informed innocent group showed lower SCL rates than the control group. In practice, SCL scores supported the expected guilt and incentive-level effects.

Psychophysiological Z Score Analysis

In order to eliminate individual differences in responsivity and to enable a meaningful summation of different measures, all responses, including the buffer item, to each GAT question were transformed into within-subjects standard scores relative to the respective means and standard deviations. These standard scores were computed for each of the three physiological indices (SCR, FPWL, and RLL). Standardization transformation was similar to all three measures, but because concealed knowledge is indicated by smaller rather than by larger RLLs and FPWLs, all of these Z scores were multiplied by −1.

For each participant, mean standardized responses to the critical alternatives were computed across GAT series. Mean scores across participants within each experimental condition served as the response score of that condition. To avoid inflating the response scores, the physiological data was gathered and analyzed irrespective of the participants' memory to the crime-related items. The mean of each individual physiological score, as well as the mean of a combined measure defined as (SCR + FPWL + RLL)/3, are presented in Table 3.

Table 3. Means (SDs) of Physiological Responses
 High-incentiveLow-incentiveAcross
  1. Note. RLLb = respiration responses derived from the covert back measure.

Guilty participants
SCR.50 (.27).34 (.42).42 (.36)
RLLb.41 (.45).12 (.32).27 (.41)
FPWL.38 (.36).38 (.31).38 (.33)
Combined.43 (.27).28 (.22).36 (.26)
Informed innocents
SCR.23 (.41).17 (.31).20 (.37)
RLLb.26 (.41).20 (.28).23 (.34)
FPWL.33 (.34).14 (.31).23 (.33)
Combined.27 (.26).17 (.20).22 (.24)
Across guilt
SCR.37 (.37).26 (.37).31 (.38)
RLLb.33 (.43).16 (.30).25 (.38)
FPWL.35 (.34).26 (.35).31 (.34)
Combined.35 (.28).23 (.22).29 (.26)
Uninformed innocents
SCR  −.02 (.28)
RLLb  .06 (.27)
FPWL  −.01 (.16)
Combined  .01 (.12)

As this is a 2 × 2 factorial design of experimental treatments plus one control group, we followed Winer's (1962) suggestions and first used a one-way multivariate analysis of variance (MANOVA) to assess the differences between all five conditions. Hence, Z scores of the three physiological measures were used in a MANOVA with condition (the five experimental conditions) as the between-subjects factor, and SCR, FPWL, and RLL as the dependent factors.

The MANOVA showed a significant overall condition effect, λ = .63, F(3,12) = 3.90, p < .001, inline image. Turning to the individual measures, a significant condition effect was obtained for all three measures: SCR, F(4,95) = 6.32, p < .001, inline image; FPWL, F(4,195) = 4.37, p = .003, inline image; RLL, F(4,95) = 4.44, p = .002, inline image. Results indicate that the various conditions differ in the responses participants exhibited to the critical items.

To weigh the effect of knowledge on the observed differences, a planned orthogonal Helmert contrast was performed on the mean response scores. The contrast compared the average of the four knowledge conditions against the mean of the uninformed innocent control condition. A separate comparison was made for each physiological measure. It was found that the MD was significant for all three measures (SCR: MD = −.328, SE = .086, p < .001; FPWL: MD = −.248, SE = .079, p = .002; RLL: MD = −.261, SE = .085, p = .003). Results indicate that mere knowledge of the critical details contributed to enhanced responsivity to these items.

Following Winer (1962), we proceeded with a multifactor analysis on the experimental treatments to assess whether the guilty participants responded to the critical items more than the informed innocents, whether the incentive level affected these responses too, and whether the reported cooperation is another independent variable that affects physiological responding. For this purpose, Z scores data on the physiological measures were analyzed using multiple analysis of covariance (MANCOVA) with guilt (guilty vs. innocent), and incentive level (high vs. low) serving as independent factors, and SCR, FPWL, and RLL as the dependent factors. To unravel the possible cooperation effects, the participants' cooperating ratings were added to the analysis as covariance. The MANCOVA showed a significant, F(1,75) = 4.44, p = .038, inline image, SCR effect for guilt, indicating that guilty participants' SCRs to the critical items were stronger than those of the informed innocents (Table 3). The FPWL effect was marginally significant, F(1,75) = 2.90, p = .093, inline image, whereas the RLL effect was not significant, F(1,75) = 0.21. A significant, F(1,75) = 4.39, p = .04, inline image, RLL effect for incentive level was also obtained, showing that the opportunity to be rewarded for successful outcomes yielded stronger RLLs to the critical items than when no reward for success was promised. The effects for SCR, F(1,75) = 2.26, and for FPWL, F(1,75) = 1.92, were not significant. No significant interactive effects were obtained. The cooperation effects reported for SCR, FPWL, and RLL, F(1,75) = 2.35, F(1,75) = 0.18, and F(1,75) = 0.01, respectively, were not significant.

Finally, a 2 (Guilt) × 2 (Incentive Level) analysis of covariance (ANCOVA) with cooperation as covariance, was conducted on the combined measure. Results revealed a significant, F(1,75) = 5.73, p = .019, inline image, incentive-level effect and a significant guilt effect, F(1,75) = 4.14, p = .045, inline image. No significant interaction effect, F(1,75) = 0.11, was obtained. Cooperation was not significant either, F(1,75) = 0.80.

Another way to examine the contribution of cooperation to the physiological responses in the GAT procedure is by using linear regressions. The actual cooperation ratings were entered as the independent variable, predicting the responses to the critical items of the three physiological measures (SCR, RLL, FPWL) and the combined measure. Results revealed a significant effect of cooperation on SCRs (β = −.25, p = .025). No significant effects were found for RLLs (β = −.01, p = ns), FPWLs (β = −.09, p = ns), and for the combined measure (β = −.17, p = ns). Note that cooperation affected all measures in the same direction (negative relation between cooperation and the magnitude of the physiological responses), indicating that choosing to cooperate predicts lower physiological responding.

As indicated earlier, the two informed innocence conditions provided high cooperation ratings that are significantly higher than the cooperation ratings provided by guilty participants. However, the high-incentive innocent participants rated their cooperation higher than the low-incentive innocent participants. To find out if the difference in cooperation affected the magnitude of SCRs to critical items, linear regressions were conducted where cooperation ratings were entered as the independent variable. The regressions were performed for each individual condition. Results revealed a significant effect of cooperation on SCRs (β = −.61, p = .004) for the high-incentive innocence condition. For the low-incentive innocence condition, the effect was not significant (β = −.26, p = ns). The two guilt conditions have not provided significant results, either.

ROC Analysis

The receiver operating characteristic (ROC) analysis has been frequently employed in CIT studies (e.g., Ben-Shakhar, 1977; Elaad, 2010; Elaad & Ben-Shakhar, 1989, 1997; Gamer & Berti, 2010; Verschuere, Crombez, Koster, & De Clercq, 2007; Vossel, Gamer, Godert, & Rill, 2003), and was recommended by the National Research Council (2003) report as highly relevant for descriptions of the diagnostic value of polygraph tests. This method is derived from the signal detection theory where detection efficacy is defined in terms of the degree of separation between the distributions of the responses to the critical items produced by experimental and control (uninformed innocents) participants. For this purpose, the distributions of the mean Z scores, computed for each informed participant across all the items of the relevant profile, and the distribution of the mean Z scores, computed for each of the uninformed control participants, were calculated for each physiological measure. On the basis of these distributions, ROC curves were generated, and the areas under these ROC curves, along with the corresponding 95% confidence intervals, were computed (Bamber, 1975). ROC statistics computed for each physiological measure (SCR, RLL, FPWL, and the combined measure) of the four experimental conditions are presented in Table 4. Significantly better than chance ROC areas for the three measures across experimental conditions are displayed (note that the lower bounds of the ROC areas are no less than 0.5, which indicates that the two distributions cannot be differentiated). This shows that across experimental conditions the informed participants responded to the critical information at a higher level than the control group of uninformed innocent participants.

Table 4. Areas Under the ROC Curves and 95% Confidence Intervals
 High-incentiveLow-incentiveAcross
 Area95% CIArea95% CIArea95% CI
  1. Note. RLLb = respiration responses derived from the unobtrusive back measure.

Guilty participants
SCR.910.824–.996.760.611–.909.835.732–.938
RLLb.793.639–.946.632.456–.809.713.585–.840
FPWL.767.620–.915.780.637–.923.774.649–.898
Combined.922.835–1.00.845.718–.972.884.800–.968
Informed innocents
SCR.651.480–.823.660.488–.832.656.510–.801
RLLb.700.525–.875.730.570–.890.715.588–.842
FPWL.745.889–.901.592.409–.776.669.524–.814
Combined.823.694–.951.728.566–.889.775.657–.893
Across guilt
SCR.781.665–.897.710.572–.848.745.633–.857
RLLb.746.624–.869.681.549–.814.714.613–.814
FPWL.756.629–.884.686.542–.830.721.600–.843
Combined.872.785–.960.786.672–.901.829.745–.914

The ROC analysis (Table 4) revealed that high-incentive guilty participants disclosed the critical information at a significant rate for all individual measures (note that the lower bounds of the areas computed for the three measures is greater than .5). Low-incentive guilty participants produced significant ROC areas for SCR and for FPWL but not for RLL. High-incentive innocents showed significant areas for RLL and FPWL but not for SCR. Finally, low-incentive innocents produced a significant ROC area only for RLL. The combined score presented significant areas for all four experimental conditions. Across conditions, all areas fall above chance level (Table 4). It may be concluded that informed participants were more responsive to the critical information than uninformed innocents.

The most sensitive measure to the guilt manipulation was SCR. Using a method proposed by Hanley and McNeil (1983) for comparing two ROC areas, the area of guilty participants was significantly (Z = 1.97) greater than the corresponding area of informed innocent participants. Table 4 reveals that the difference is due to the large area that was presented by high-incentive guilty participants, which is significantly larger than that of high-incentive innocents (Z = 2.66) and low-incentive innocents (Z = 2.54).

Discussion

  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. References

The purpose of the current study was to further explore the different states of mind of guilty and informed innocent examinees undergoing detection of information tests. These states of mind trigger different responses to cooperative instructions from guilty and innocent examinees. Two earlier studies already examined this matter (Elaad, 2013; Zvi et al., 2012). Zvi et al. (2012) manipulated the guilt of the examinees (guilty or informed innocent) and the instructions they received (to cope or to cooperate on the test), and controlled the incentive level (all participants were promised a reward for success). Elaad (2013) manipulated the instructions (whether or not cooperating instructions were delivered) and the incentive level (with and without an external reward for success), where guilt was controlled (only informed innocent participants were examined). The present study manipulated guilt and incentive level while controlling the instructions to cooperate. Specifically, the present study offered both guilty and informed innocents two options, to cope or to cooperate, enabling them to make their own choice which strategy to adopt.

The present cooperating ratings show that guilty participants were less inclined to cooperate on the test than informed innocents who tended to be quite cooperative. The cooperation tendency was enhanced when a reward for successful outcomes was promised. It was further determined that choosing to cooperate predicts attenuated SCR responses to the critical items. Furthermore, although not significant, cooperation reports predicted the tendency to reduce FPWL and RLL responses as well.

A rather speculative explanation for the association between the willingness to cooperate and attenuated SCRs rests on the point that SCRs are affected by whether a probe in the CIT comprises a central or a peripheral aspect of the respective crime (e.g., Gamer & Berti, 2012). Thus, for participants in a coping state of mind, the critical items may become “central” whereas for those in a cooperative state of mind the critical item may become peripheral. Future research is encouraged to further examine and develop this idea.

Overall, guilt or innocence and the incentive levels were not associated with recall of crime-related items. This enables a better comparison between the various experimental conditions without a possible recall artifact.

Results showed that sheer possession of concealed information contributed to enhanced responses to the critical items when compared with uninformed participants. Cooperation attenuated SCRs and to some extent also FPWL responses to the critical items of informed innocents. On the other hand, RLL was the only measure that was sensitive to the incentive level manipulation. ROC analysis further supported these results. SCR was the most effective measure in detecting the critical information and the most sensitive measure to the guilt manipulation. These results are of no surprise as the distinctive sensitivity to critical information of electrodermal responses in general (e.g., Bradley, Malik, & Cullen, 2011), and SCR in particular (e.g., Cutrow, Parks, Lucas, & Thomas, 1972; Verschuere et al., 2007), are well documented in the literature of CIT and GAT questioning formats.

SCR provided a clear distinction between high-incentive guilty participants, who reacted to the critical information much more than either high-incentive or low-incentive innocents. These results are straightforward but important. They suggest that providing cooperating instructions and clear guidance on how to implement such cooperation may enhance correct classification of guilty and innocent examinees in the GAT.

The present results support those of Zvi et al. (2012) who reported that both innocence and cooperating instructions attenuated SCR responses to the critical items. Specifically, the informed innocent participants were less responsive to the critical items than the guilty examinees, and when the participants were instructed to cooperate on the test, SCR responses decreased as well. Results also coincide with those of Elaad (2013). Elaad examined the association between a promised reward for either success in proving innocence (goal-oriented instructions) or success in being cooperative (task-oriented instructions) and the willingness of informed innocent examinees to adopt the goal- or task-oriented attitude. He showed that high-incentive cooperation instructions attenuated informed innocents' responses to the critical items.

The fact that high-incentive informed innocent participants were more inclined to cooperate on the test than the low-incentive informed innocents may suggest that in real life, where stakes are higher (Verschuere, Meijer, & De Clercq, 2011), and where motivation to prove innocence is much stronger, the tendency of innocent examinees to cooperate may be further enhanced. The reverse effect may be true for guilty examinees. Although not significant, high-incentive guilty participants were less inclined than their low-incentive counterparts to cooperate on the test. As a result, their physiological responses to the critical items were enhanced.

The present results may provide further support for previous suggestions (e.g., Zvi et al., 2012) to advise examinees in real life to cooperate on the test and guide them how to do so. A proper warning should be followed that cooperation serves the interests of innocent examinees and operates against the basic interests of guilty examinees. This may encourage innocents to adopt a cooperative attitude and behave accordingly. They are expected to calm down and try to assist the polygraph examiner reach the expected (correct) decision. Such cooperation instructions would convince guilty examinees to avoid cooperation, adopt a coping attitude, and double their efforts to obstruct the test.

Limitations of the Present Study

The present study entails a laboratory experiment where stakes are low and participants were well aware that no harm would be inflicted upon them as a result of their performance on the test. Specifically, the threat is not the same as the threats that criminal suspects usually experience during actual testing. To compensate for this shortcoming, we manipulated two incentive levels to prove innocence with the expectation that the higher incentive level will simulate better than the lower incentive level real-life conditions. Still, the higher incentive level does not simulate real-life testing. This dictates caution when applying the present results to actual testing.

The present sample of participants consisted of undergraduate students enrolled in a psychology course who participated in the study for course credit. One argument against using psychology students is the possible resentment they may feel toward the requirement to participate in a departmental study. However, in actual testing, innocent suspects may often manifest resentment when they are asked to take a polygraph test.

The present sample is largely a female sample. Although females were no more or less inclined than males to either cope or cooperate on the test, caution is still recommended when generalizing the results to largely male suspects.

Final Remarks

To sum up, the present study extends previous accounts on the distinctive motivational processes or states of mind of informed innocent and guilty examinees to polygraph tests in general and to detection of information tests in particular. It has demonstrated that innocent examinees prefer to cooperate on the test whereas guilty examinees prefer not to cooperate. These distinctive states of mind add another theoretical and practical level to the cognitive approach that emphasized the recognition of the critical information in detection of information tests. This addition is important because the cognitive rationale fails to explain why informed innocent examinees respond less to critical items than informed guilty examinees. The present motivational approach deserves more research attention than is currently provided.

References

  1. Top of page
  2. Abstract
  3. Method
  4. Results
  5. Discussion
  6. References
  • Bamber, D. (1975). The area above the ordinal dominance graph and the area below the receiver operating characteristic graph. Journal of Mathematical Psychology, 12, 387415.
  • Ben-Shakhar, G. (1977). A further study of the dichotomization theory in detection of information. Psychophysiology 14, 408413.
  • Ben-Shakhar, G., & Elaad, E. (2003). The validity of psychophysiological detection of information with the guilty knowledge test: A meta-analytic review. Journal of Applied Psychology, 88, 131151.
  • Ben-Shakhar, G., Gronau, N., & Elaad, E. (1999). Leakage of relevant information to innocent examinees in the GKT: An attempt to reduce false-positive outcomes by introducing target stimuli. Journal of Applied Psychology, 84, 651660.
  • Bradley, M. T., MacLaren, V. V., & Black, M. E. (1996). The control question test in polygraphic examinations with actual controls for truth. Perceptual and Motor Skills, 83, 755762.
  • Bradley, M. T., MacLaren, V. V., & Carle, S. B. (1996). Deception and nondeception in guilty knowledge and guilty actions polygraph tests. Journal of Applied Psychology, 81, 153160.
  • Bradley, M. T., Malik, F. J., & Cullen, M. C. (2011). Memory instructions, vocalization, mock crimes, and concealed information tests with a polygraph. Perceptual and Motor Skills, 113, 840858.
  • Bradley, M. T., & Rettinger, J. (1992). Awareness of crime-relevant information and the guilty knowledge test. Journal of Applied Psychology, 77, 5559.
  • Bradley, M. T., & Warfield, J. F. (1984). Innocence, information, and the guilty knowledge test in the detection of deception. Psychophysiology, 21, 683689.
  • Colwell, K., Hiscock-Anisman, C., Memon, A., Woods, D., & Michlik, P. M. (2006). Strategies of impression management among deceivers and truth-tellers: How liars attempt to convince. American Journal of Forensic Psychology, 24, 19.
  • Cullen, M. C., & Bradley, M. T. (2004). Positions of truthfully answered controls on control question tests with the polygraph. Canadian Journal of Behavioral Science, 36, 167176.
  • Cutrow, R. J., Parks, A., Lucas, N., & Thomas, K. (1972). The objective use of multiple physiological indices in the detection of deception. Psychophysiology, 9, 578588.
  • DePaulo, B. M., Lindsay, J. J., Malone, B. E., Muhlenbruck, L., Charlton, K., & Cooper, H. (2003). Cues to deception. Psychological Bulletin, 129, 74118.
  • Elaad, E. (2009). Effects of context and state of guilt on the detection of concealed crime information. International Journal of Psychophysiology, 71, 225234.
  • Elaad, E. (2010). Effects of perceived reliability and generalization of crime-related information on detection in the concealed information test. International Journal of Psychophysiology, 75, 295303.
  • Elaad, E. (2011). Effects of incomplete information on the detection of concealed crime details. Applied Psychophysiology and Biofeedback, 36, 159171. doi: 10.1007/s10484-011-9153-2.
  • Elaad, E. (2013). Effects of goal- and task-oriented motivation in the guilty action test. International Journal of Psychophysiology, 88, 8290.
  • Elaad, E., & Ben-Shakhar, G. (1989). Effects of motivation and verbal response type on psychophysiological detection of information. Psychophysiology, 26, 442451.
  • Elaad, E., & Ben-Shakhar, G. (1997). Effects of item repetitions and variations on the efficiency of the guilty knowledge test. Psychophysiology, 34, 587596.
  • Elaad, E., & Ben-Shakhar G. (2006). Finger pulse waveform length in the detection of concealed information. International Journal of Psychophysiology, 61, 226234.
  • Elaad, E., Ginton, A., & Jungman, N. (1992). Detection measures in real-life criminal guilty knowledge tests. Journal of Applied Psychology, 77, 757767.
  • Gamer, M. (2010). Does the guilty action test allow for differentiating guilty participants from informed innocents? A re-examination. International Journal of Psychophysiology, 76, 1924.
  • Gamer, M. & Berti, S. (2010). Task relevance and recognition of concealed information have different influences on electrodermal activity and event-related brain potentials. Psychophysiology, 47, 355364.
  • Gamer, M., & Berti, S. (2012). P300 amplitudes in the concealed information test are less affected by depth of processing than electrodermal responses. Frontiers in Human Neuroscience, 6, 110. doi: 10.3389/fnhum.2012.00308.
  • Gamer, M., Kosiol, D., & Vossel, G. (2010). Strength of memory encoding affects physiological responses in the guilty action test. Biological Psychology, 83, 101107.
  • Granhag, P. A., & Hartwig, M. (2008). A new theoretical perspective on deception detection: On the psychology of instrumental mind reading. Psychology, Crime and Law, 14, 189200.
  • Hanley, J. A., & McNeil, B. J. (1983). A method of comparing the areas under receiver operating characteristic (ROC) curves derived from the same cases. Radiology, 148, 839843.
  • Hartwig, M., Granhag, P. A., & Stromwall, L. A. (2007). Guilty and innocent suspects' strategies during police interrogations. Psychology, Crime & Law, 13, 213227.
  • Hartwig, M., Granhag, P. A., Stromwall, L. A., & Vrij, A. (2005). Detecting deception via strategic disclosure of evidence. Law and Human Behavior, 29, 469484.
  • Kassin, S. M. (2005). On the psychology of confessions: Does innocence put innocents at risk? American Psychologist, 60, 215228.
  • Krapohl, D. J. (2011). Limitations of the concealed information test in criminal cases. In B. Verschuere , G. Ben-Shakhar , & E. Meijer (Eds.), Memory detection: Theory and application of the concealed information test (pp. 151170). Cambridge, UK: Cambridge University Press.
  • Lykken, D. T. (1959). The GSR in the detection of guilt. Journal of Applied Psychology, 43, 385388.
  • Lykken, D. T. (1974). Psychology and the lie detection industry. American Psychologists, 29, 725739.
  • Lykken, D. T. (1998). A tremor in the blood: Uses and abuses of the lie detector (2nd ed.). New York, NY: Plenum Press.
  • Nahari, G., & Ben-Shakhar, G. (2011). Psychophysiological and behavioral measures for detecting concealed information: The role of memory for crime details. Psychophysiology, 48, 733875.
  • National Research Council. (2003). The polygraph and lie detection: Committee to review the scientific evidence on the polygraph, division of behavioral and social sciences and education. Washington, DC: The National Academic Press.
  • Stromwall, L. A., Hartwig, M., & Granhag, P. A. (2006). To act truthfully: Nonverbal behavior and strategies during a police interrogation. Psychology, Crime and Law, 12, 207219.
  • Verschuere, B., Crombez, G., Koster, E. H. W., & De Clercq, A. (2007). Antisociality, underarousal and the validity of the concealed information polygraph test. Biological Psychology, 74, 309318.
  • Verschuere, B., Meijer, E., & De Clercq, A. (2011). Concealed information under stress: A test of the orienting theory in real-life police interrogations. Legal and Criminological Psychology, 16, 348356.
  • Vossel, G., Gamer, M., Godert, H. W., & Rill, H. G. (2003). The efficiency of detecting concealed information with the guilty knowledge test: A comparison of different physiological variables and non-physiological methods of credibility assessment using signal detection theory [Abstract]. Journal of Psychophysiology, 17, 184.
  • Winer, B. J. (1962). Statistical principles in experimental design. New York, NY: McGraw-Hill.
  • Zvi, L., Nachson, I., & Elaad, E. (2012). Effects of coping and cooperative instructions on guilty and informed innocents' physiological responses to concealed information. International Journal of Psychophysiology, 84, 140148. doi: 10.1016/j.ijpsycho.2012.01.022.