research article the impact of presentation modality on...

Hindawi Publishing CorporationJournal of CriminologyVolume 2013 Article ID 164546 10 pageshttpdxdoiorg1011552013164546

Research ArticleThe Impact of Presentation Modality on Perceptions ofTruthful and Deceptive Confessions

Deborah Bradford1 Jane Goodman-Delahunty2 and Kevin R Brooks3

1 National Drug and Alcohol Research Centre University of New South Wales Sydney NSW 2052 Australia2 Australian Graduate School of Policing and School of Psychology Charles Sturt University Sydney NSW 1655 Australia3 Department of Psychology Macquarie University Sydney NSW 2109 Australia

Correspondence should be addressed to Deborah Bradford deborahbradfordunsweduau

Received 24 January 2013 Revised 26 March 2013 Accepted 26 March 2013

Academic Editor Augustine Joseph Kposowa

Copyright copy 2013 Deborah Bradford et al This is an open access article distributed under the Creative Commons AttributionLicense which permits unrestricted use distribution and reproduction in any medium provided the original work is properlycited

This study examined the impact of presentation modality and the effectiveness of direct and indirect measures of deception todistinguish truthful from deceptive confessions Confession statements were presented in one of three formats audiovisual audio-only or written text Forty-six observers classified each statement as true or false and provided ratings of confidence informationsufficiency perceived cognitive load and suspiciousness Compared to audio and written confessions exposure to audiovisualrecordings yielded significantly lower accuracy rates for direct veracity judgements with below chance level performance Therewas no evidence that indirect measures assisted observers in discriminating truthful from deceptive confessions Overall observersshowed a strong bias to believe confessions with poor detection rates for false statements Reliance on video recordings to assessthe veracity of confession evidence is unlikely to reduce wrongful convictions arising from false confessions

1 Introduction

Due to the pervasive nature of deception in society individ-uals are faced with the difficult task of distinguishing truthsfrom lies on a regular basis These judgments are especiallyimportant in the context of the criminal justice system wherelaw enforcement and legal professionals are routinely taskedwith assessing the credibility of witnesses and suspects inpolice investigations and other criminal proceedings Theability to accurately discriminate truths from lies in legalsettings is crucial as judgment errors can have grave con-sequences This is especially relevant to confession evidencegiven the weight accorded this type of evidence in decisionsto prosecute and trial verdicts Evidence of a confession isconsidered the most incriminating form of trial evidenceand can have a powerful and persuasive effect on convictionrates [1ndash3] Given the ever-growing body of research detailingnumerous examples of individuals wrongly convicted on thebasis of false confession evidence [1 4ndash8] it is vital that theveracity of confessions be accurately evaluated

Ensuring that wrongful convictions on the basis of falseconfessions are avoided involves both the use of appropriateand noncoercive police interviewing techniques and the im-plementation of effective safeguards to assess the credibilityof suspect admissions Historically police interviews havebeen conducted in an accusatorial style aimed at inducinga suspect to confess [9 10] This objective was supported bythe use of psychologically coercive and manipulative tacticsthat can exert strong pressure on suspects to confess [11 12]While this approach has largely been abandoned in somejurisdictions such as the United Kingdom and Australia itremains the dominant approach to police interviewing in theUnited States [12 13] and has been identified as a key factorin many documented cases of false confessions [1 5 8 10 14]

In recent years these practices have been strongly criti-cised A number of scholars [6] have advocated for the fullaudiovisual recording of police interviews both to discouragethe use of coercive methods and to ensure that an accuraterecord of suspect statements is available for subsequentevaluations of veracity [15 16] This led to the introduction

2 Journal of Criminology

of formal policy changes in many jurisdictions requiringaudiovisual recording [9 13] However these changes do notprovide an absolute safeguard against wrongful convictionson the basis of false confessions [9 13 17] Indeed thepossibility remains that policemay engage in preliminary un-recorded questioning with a suspect which may compromisethe credibility of any subsequent recorded admissionsThis isclearly illustrated in the Australian case of Andrew Mallardwho was wrongly convicted and imprisoned in 1995 formurder based in part on a 20-minute false videotaped con-fession recorded following 11 hours of police questioning[18] Perhaps of greater concern is the possibility that factfinders may use a video recording as a tool to assess theveracity of suspect admissions in the event of a disputed con-fession [13] Decades of deception research showing that factfinders are generally nomore accurate than chance at discern-ing between videotaped truthful and untruthful statements[19] calls into question the reliability of video recordingsto prevent erroneous prosecution on the basis of a falseconfession

The use of video recordings to assess the veracity of con-fession evidence is further complicated by the inherentlybelievable nature of confessions and the possibility that inthis context truths and lies may be especially difficult todiscern This bias was confirmed by experimental researchshowing that confessions were overwhelmingly believed(gt80 of the time) irrespective of actual message veracityresulting in very low detection rates for false confessions[20] Kassin [10 21] noted two possible reasons why falseconfessions may be particularly difficult to detect first dueto a truth-biasing attribution error (fundamental attributionerror) [22] which ldquoleads people to expect self-serving behav-iour in othersmdashand hence to trust confessionsrdquo [10 page223] Secondly and perhaps more appreciably false con-fessions typically contain highly detailed persuasive andemotional features that are commonly associated with truthtelling (eg details of how and why a crime was committedand expressions of remorse) [10] The presence of thesepersuasive features combined with a predisposition to believeself-incriminating accounts may make false confessions par-ticularly difficult to detect

Research directly investigating observersrsquo performanceat distinguishing truths from lies in the specific context ofconfessions is limited as is the number of studies comparingmode of communication in this context Perhaps the mostpertinent study on this topic examined observer assessmentsof the credibility of five truthful andfive deceptive confessionsprovided by incarcerated prisoners [23] Results revealedan overall accuracy rate of 539 a relatively unimpres-sive performance level not significantly better than chancealone Similarly an analysis of 20 brief (writtenaudiotaped)excerpts (mean length 10 seconds) of actual criminal con-fessions [24] yielded comparable levels of accuracy (55 to62) Even lower accuracy rates (33) were reported forobservers making veracity assessments on the basis of video-recorded confessions to mock crimes in the absence ofsupporting contextual information [25]While these findingsafford little optimism for success in discerning true fromfalse confessions performance accuracy improved when the

modality of a confession was varied For instance partici-pants achieved higher accuracy rates when confessions werepresented in audio-only compared to audiovisual format[23] These findings were congruent with most outcomes inthe general deception literature (see [24 26] for exceptions)indicating that observers tend to achieve higher accuracyrates when they attend to speech content (written transcripts)or vocal cues (audio recordings) compared to exposure tothe full audiovisual presentation of the target (audiovisualrecordings) [27ndash34] As noted by Vrij [35] these findingscontroverted the widely-held belief that ldquolie detection iseasiest when the lie detector has access to the full picture ofthe potential liar and that just reading a textual version ofa statement or just listening to someonersquos voice hampers liedetectionrdquo (p 303)

Differences in performance accuracy between presenta-tion modalities may be attributed to the fact that when visualinformation is available observers rely most heavily on non-verbal behaviours and are particularly drawn to stereotypicalcues to deception (eg gaze aversion and fidgeting) to guidetheir deception judgments [26 29 36ndash39] Given the body ofevidence showing that these cues are not indicative of deceit[40] it is unsurprising that judgments made on this basis areless accurate than those in response to vignettes where thesepotentially distracting cues are unavailableThis contention issupported by research suggesting that reliance on demeanourcues (eg physical movements eye gaze and interactionstyle) alone can result in poor accuracy at discriminatingtruths from lies [41] Furthermore research has shown thatthere are generally more valid verbal and speech-relatedcues to deception than nonverbal cues [19 40 42] andthat observersrsquo notions about the validity of verbal cues todeception are generally more accurate than those concerningnonverbal cues [43] Hence observers may show increaseddetection accuracy when visual information is ignored orunavailable perhaps allowing them to direct their attentionto the more diagnostic speech content andor vocal aspectsof the message

Furthermore to the extent that truthful confessors expe-rience feelings of guilt or shame associated with acknowledg-ing responsibility for a transgression [40] they may exhibitbehaviours consistent with the stereotypical image of decep-tionThese behaviours are more apparent in some modalitiesthan others Since stereotypical beliefs about deception aremost commonly nonverbal [26 44] observers who viewvideotaped confessions may be particularly impaired in theirability to discern truth fromdeception compared to observerswho are not exposed to nonverbal behaviours of confessorsHowever these findings have not been used to guide policydevelopments regarding the uses of police video recordingsto assess the credibility of confession evidence indicatinga need for further clarity on the impact of presentationmodality on perceptions of confessions The primary aim ofthe current research was to address this gap in the literatureto help inform the development of guidelines on how policeinterview recordings should be used to better evaluate theveracity of confessions

While restricting message modality is one procedure thatmay improve veracity discrimination other strategies may

Journal of Criminology 3

further assist in focusing observersrsquo attention on valid cuesto deception and in turn improve detection accuracy Onesuch strategy is to ask observers to make indirect judgmentsof veracity (eg ldquoDoes the person have to think hardrdquo)instead of explicit truth versus lie classifications (eg ldquoIsthe person lyingrdquo) Indeed some studies have shown thatindirect measures of deception may discriminate truthfulfrom deceptive messages more accurately than explicit truthversus lie classifications [45ndash47] possibly because observershave implicit or intuitive notions about deception that aremore accurate than their explicit ideas [48] For examplein a videotaped lie detection task Anderson et al [45]found that explicit detection accuracy was no better thanchance yet observers were able to distinguishmessages usingindirect measures such as observer confidence (higher fortruthful statements) suspiciousness of the message (greaterfor deceptive statements) and the extent to which observersfelt that they had enough information to make a veracityrating (higher for truthful statements) Similarly when usinga theoretically based indirect measure of deception to testthe proposition that lying draws on more cognitive resourcesthan does truth telling (ldquoDoes the person have to thinkhardrdquo) police officers were able to distinguish betweenvideotaped truths and lies whereas those officers asked tomake explicit veracity ratings of the videotaped statementswere unable to make this distinction [47] Findings fromthis study further suggested that using the indirect methodfocused observersrsquo attention on behavioural cues that wereindicative of deceit while those who made explicit truthliejudgments focused on stereotypical cues that were not reli-ably associated with deception As these strategies have yet tobe tested in the context of confessions the secondary aim ofthe current research was to address this gap in the literatureand to provide further information to assist practitionersin developing guidelines to effectively assess the veracity ofconfession evidence

11 Aims of the Research The objectives of the current studywere threefold

(1) to further investigate the influence of presentationmodality on deception detection accuracy in the con-text of confessions using a larger set of statementslonger in duration than those used in previous stud-ies This was achieved by comparing the deceptiondetection performance of lay observers when assess-ing autobiographical confessions to personal trans-gressions presented in one of three different modal-ities (audiovisual recordings audio recordings andverbatim written transcripts)

(2) to shed light on behavioural cues that may be associ-ated with deception in the context of confessions bycoding selected verbal and nonverbal cues exhibitedby truthful and deceptive confessors and

(3) to extend the literature on the utility of indirect mea-sures of deception by examining whether measuresthat benefited detection accuracy in earlier studies areeffective in the context of confessions This approachhas not been explored previously with confessions

12 Research Hypotheses We predicted that

(1) participants exposed to audiovisual recordings wouldattend to potentially uninformative and biasing visualinformation and would therefore be less accurate onthe explicit truthlie discrimination task than partici-pants exposed to the same information in audio-onlyrecordings or written transcripts

(2) to the extent that stereotypical nonverbal cues todeception are displayed more frequently by truthfulthan deceptive confessors audiovisual observersrsquo per-formance would be significantly impaired relative tochance

(3) confidence would be higher and observers would feelthey had more information when assessing truthfulthan deceptive confessions

(4) deceptive confessors would be perceived as having toldquothink harderrdquo than truthful confessors

(5) deceptive messages would be perceived as more sus-picious than truthful messages

2 Method

21 Research Design The study was conducted in two stagesFirst 60 university students gave video recorded statementsof truthful and deceptive confessions Second a subset ofthe 60 confession statements was presented to observers inaudiovisual audio-only or written format All confessionstatements comprised narratives up to several minutes inlength

22 Stage One Elicitation of StimulusMaterials To elicit con-vincing truthful and deceptive confessions we adapted therealsimulating paradigm [49ndash51] a procedure traditionallyused in hypnosis research to elicit real and simulated hypnoticnarratives A number of modifications were made as outlinedby Barnier et al [52] First a group of participants gavegenuine autobiographical confessions describing social trans-gressions to an experimenter who video recorded their state-ments Next a group of ldquosimulatorsrdquo adopted one of the gen-uine confessional events already described (ensuring it wasone that they had not themselves experienced) and concoctedand presented a convincing deceptive confession based onthat event as if it were a real autobiographical experienceIn this yoked design genuine confessions were the basis formatched simulated deceptive confessions

221 Participants Participants were 60 students (41 females19 males) enrolled at the University of New South Wales inSydney Australia with an average age of 2305 years (SD =558 years)

222 Procedure Eachparticipant gave one truthful or decep-tive confession The initial 4ndash6 participants gave genuineaccounts as the experimental paradigm required that at leastthree genuine accounts be available for adoption by partici-pants in the deceptive condition prior to recording statementsby simulators Once the initial genuine accounts were elicited


Table 1 Behavioural coding of truthful and deceptive confessions means and standard deviations

Behaviour Truthful confession Deceptive confession 119875-valuePauses 689 (413) 679 (629) 0970Speech hesitations 1946 (926) 1746 (619) 0540Speech errors 783 (404) 1025 (494) 0203Illustrators 4304 (2671) 4525 (2187) 0827Eye blinks 11079 (6258) 9429 (4816) 0477Smiling 1200 (938) 817 (702) 0269Gaze aversion 12408 (5599) 10158 (4856) 0304Note Frequency counts were employed for all coded behaviours except gaze aversion which was coded as the number of seconds the speaker looked awayStandard deviations are displayed in parentheses

assignment to truthful versus deceptive conditions was alter-nated for all subsequent participants

Participants in the genuine condition (truth tellers) wereasked to recall and briefly describe three autobiographicalevents where they did something that they ldquofelt guilty forand thought a lot about afterwardsrdquo Participants stated howold they were at the time of the event and how certain theywere of the experience using a ten-point scale (10 = absolutecertainty) The experimenter selected one confessional eventfor the participant to describe based primarily on partici-pantsrsquo certainty of having experienced the event (ie ratingscloser to 10) and on recency (the most recent event waschosen) Participants in the genuine condition were givenfive minutes alone to think about the selected event detailsbefore describing it ldquoin asmuch detail as possiblerdquo to a secondexperimenter who video recorded their statement

Participants in the deceptive condition were presentedwith three brief summaries (3ndash5 sentences in length) ofconfessions elicited from participants in the genuine condi-tion and were asked if they had ever experienced the event(yesno) and their certainty using a ten-point scale (10 =definitely experienced the event) The experimenter selectedone event (whichever the participant was most certain theyhad not experienced) Participants had five minutes alone toprepare a detailed and convincing confession of the selectedevent as if they had truly experienced Following the prepa-ration period participants were informed that the secondexperimenter was unaware whether they were telling thetruth or a lie and that the participantrsquos task was to convincethe second experimenter that the statement was truthful byldquodescribing the event in as much detail as you can and byappearing as genuine as you canrdquo The participantrsquos narrativewas video recorded without interruption

Selection and Preparation of Experimental Accounts A ran-dom subset of 24 statements was selected subject to thefollowing constraints an equal number of males and females(age range 18ndash41 119872 = 2400 years SD= 646 years) an equalnumber of truthful and deceptive confessions and adequateframing of the participant in the video The confessionsdescribed experiences such as stealing property damagelying to a loved one and cheating on a partner Statementsranged in length from 334 to 988 words (119872 = 57225 wordsSD = 18981 words) and their recorded duration ranged fromtwo to five minutes (119872 = 18913 seconds SD = 5754

seconds) The 24 experimental confession statements werenot significantly different in length (number of words) as afunction of truth status (119905(22) = 027 119875 = 079) or gender ofthe storyteller (119905(22) = 049 119875 = 063)

All experimental statements were transcribed verbatimand verified by an observer blind to the truth status of eachaccount Nonfluencies such as ldquoumrdquo ldquoahrdquo and ldquoerrdquo repetitionsand pauses (greater than two seconds) were retained (andnoted in written transcripts) to preserve the quality ofaccounts The 24 experimental video recorded statementswere randomly organized for presentation in audiovisualaudio and written modalities An interval of approximately20ndash25 seconds separated the audio and audiovisual presenta-tions

Behavioural Coding To examine whether there were differ-ences between the behaviours of truth tellers versus liarsvisual and vocal cues were coded independently by twotrained raters who were blind to the experimental conditionsand study objectives A series of behaviours drawn from acoding scheme employed in a number of earlier deceptiondetection studies (see [53] for a description) were analysedand intercoder reliability was assessed using Krippendorff rsquosalpha (120572) [54] As the two ratersrsquo estimates were sufficientlyreliable [55] for all behavioural cues (120572 = 073 to 099)mean values were used in all subsequent analyses Vocal cues(coded from transcripts) included frequency counts of thenumber of pauses (120572 = 099) speech hesitations (eg ldquoahrdquoldquoerrdquo 120572 = 098) and speech errors (eg sentence repetitionstutters false starts 120572 = 077) Visual cues (coded fromvideo recordings) included frequency counts of illustrators(armhand movements that accompany speech 120572 = 095)eye blinks (120572 = 085) smiling (grinsmile or laugh continu-ous smiles scored anew every two seconds 120572 = 073) andgaze aversion (the number of seconds that the participantlooked away from the interviewer 120572 = 086) No statisti-cally significant differences emerged for any of the codedbehaviours across truthful versus deceptive confessions (all119875 values gt 020) A similar pattern of nonsignificant findingswas observed after controlling for the duration of eachstatement Means and standard deviations are presented inTable 1

23 Stage Two Observer Classifications In the second stageof the experiment observers none of whom were involved


Table 2 Signal detection measures for presentation modality groups

Signal detection measures Written (119899 = 19) Audiovisual (119899 = 15) Audio (119899 = 12)Hits 412 278 368Misses 588 722 632Correct rejections 632 550 681False alarms 368 450 319Total correct 522 414 524Discrimination accuracy (1198891015840) 014 minus055 014Response bias (c) 030 040 043

in the first stage of the study were randomly assigned to oneof three conditions in which they (1) watched audiovisualrecordings (2) listened to audio recordings or (3) read writ-ten transcripts of confession statements Participants indicat-ed (a) whether each statement was true or false and ona 7-point Likert-type scale rated (b) confidence in theirveracity judgment (c) the extent to which they had adequateinformation to judge veracity (d) the extent to which theyperceived the speaker to be ldquothinking hardrdquo (experiencinghigh cognitive load) and (e) how suspicious they were ofthe statement The primary dependent variable was theoverall accuracy of participant veracity ratings analysedusing signal detection theory [56] This framework yieldedtwo independent parameters of performance discriminationaccuracy (1198891015840) and response bias (119888) measured by com-bining judgments for truthful and deceptive statementsDiscrimination accuracy referred to participantsrsquo sensitivityin correctly detecting a signal when it was present (ie deceit)and correctly rejecting the absence of the signal when notpresent (ie truth) Higher positive values of 1198891015840 indicatedhigher discrimination accuracy a value of zero indicatedchance level performance and negative values indicated thataccuracy was below chance [56] Response bias estimated theextent to which each participant was predisposed to reportldquodeceptionrdquo versus ldquotruthrdquo with positive values indicating atruth bias and negative values indicating a deception bias

231 Participants Participants were 48 community volun-teers and university students recruited via study email adver-tisements Two participants failed to complete the experi-ment leaving a total of 46 participants (38 females 8 males)ranging in age from 18 to 75 years (119872 = 3996 years SD =1322 years)

232 Procedure Participants in groups of 5ndash10 read listenedto or watched 24 randomly organized confessions some ofwhich were true and some of which were false They wereadvised that the percentage of truthful statements in thesample fell between 25 and 75 [47 57]

3 Results

31 Accuracy Ratings for Explicit Credibility Assessments Theoverall accuracy rate in assessing confession veracity (globaljudgment accuracy) across conditions participants andstatements was 487 353 for deceptive confessions and

621 for truthful confessions Overall accuracy was not sig-nificantly different fromchance (two-tailed binomial test119875 =040) There was no relationship between overall accuracyand confidence ratings across modality groups (rho = minus018119873 = 46 119875 = 023 two-tailed)

To examine differences in classification accuracy as afunction of presentation modality and account truth statusa 2 (truth status) times 3 (mode of presentation) mixed modelANOVA was conducted This analysis yielded a significantmain effect of truth status 119865(1 43) = 6683 119875 lt 00011205782= 060 showing that across all presentation modalities

observers were significantly better at identifying true (119872 =062 SD = 013) than deceptive confessions (119872 = 036 SD =016)Themain effect of presentationmodality was significant119865(2 43) = 764 119875 lt 0005 1205782 = 026 Planned contrastsusing a Bonferroni correction revealed that participantsassessing written confessions (119905(43) = 353 119875 lt 0005119872 = 052 SD = 008) and those assessing audio recordedconfessions (119905(43) = 322 119875 lt 0005 119872 = 052 SD =007) classified significantly more statements correctly thanparticipants who assessed audiovisual confessions (119872 = 041SD = 011) There were no significant differences in overallaccuracy between audio and written modalities and two-tailed binomial testing revealed that neither of these twogroups performed at levels significantly greater than chanceHowever in the audiovisual condition overall accuracy(414) was significantly worse than chance (two tailed bino-mial test 119875 lt 00001) The interaction between truth statusand presentation modality was not significant 119865(2 43) =070 119875 = 050

32 Signal Detection Measures For each individual partici-pant measures of discrimination accuracy (1198891015840) and responsebias (119888) were calculated across the 24 experimental state-ments These values were averaged to obtain means for eachmodality group Signal detection and performance accuracymeasures are displayed in Table 2

Separate one-way ANOVAs assessed differences in dis-crimination accuracy (1198891015840) and response bias (119888) across themodes of presentation For discrimination accuracy resultsindicated a significant main effect of mode of presenta-tion 119865(2 43) = 837 119875 lt 001 1205782 = 028 Consistent withthe findings above thosewho assessed audiovisual recordingsof confessions showed significantly lower discriminationaccuracy relative to participants who assessed written tran-scripts (119905(43) = 373 119875 lt 0005) and to those who assessed


Table 3 Means and standard deviations for ratings of indirect measures of deception as a function of truth status and modality group

Audiovisual Audio Written TotalT L T L T L T L

Confidence 487 503 463 460 469 477 474 481(052) (073) (082) (094) (075) (072) (069) (079)

Information sufficiency 467 474 385 390 409 436 422 436(070) (089) (110) (116) (097) (092) (097) (101)

Cognitive loadlowast 350 304 340 335 373 361 357lowast 336lowast

(066) (050) (063) (059) (053) (072) (060) (066)

Suspiciousnesslowast 388lowast 334lowast 372lowast 376lowast 378lowast 383lowast 380 365(057) (071) (041) (049) (065) (070) (056) (068)

Note lowastSignificant differences (119875 lt 005) standard deviations in parentheses

audio recorded confessions (119905(43) = 332 119875 lt 0005) Nostatistically significant differences emerged in the degree ofresponse bias across groups (119865(2 43) = 079 119875 = 046)However a series of one-sample 119905-tests comparing bias scoresto zero revealed that respondents in all groups exhibited atruth bias (all 119875 values lt 0002)

33 Indirect Measures of Deception To examine observersrsquoperformance on the indirect measures of deception a seriesof 2 (truth status) times 3 (presentation mode) mixed modelANOVAs were conducted on ratings of confidence informa-tion sufficiency cognitive load and suspiciousness Resultsshowing means and standard deviations for all indirectmeasures as a function truth status and presentationmodalityare presented in Table 3

For perceived cognitive load there was a significant maineffect of truth status 119865(1 43) = 511 119875 lt 005 WilksrsquoLambda = 089 1205782 = 010 showing that overall participantsperceived a higher level of cognitive load among actual truth-ful confessors compared to deceptive confessors Neither themain effect of presentation modality 119865(2 43) = 263 119875 =008 nor the interaction between truth status and presenta-tion modality 119865(2 43) = 191 119875 = 016 reached statisticalsignificance For perceived suspiciousness results showed nosignificant main effects for truth status 119865(1 43) = 303 119875 =009 or presentation modality 119865(2 43) = 054 119875 = 059However the interaction between truth status and presen-tation modality was significant 119865(2 43) = 545 119875 lt 001Wilkrsquos Lambda = 080 1205782 = 019 Closer examination ofgroup means suggested that participants who assessed writ-ten transcripts and audio recordings were more suspicious ofdeceptive confessions (written 119872 = 383 SD = 070 Audio119872 = 376 SD = 049) compared to truthful confessions(written 119872 = 378 SD = 065 Audio 119872 = 372 SD = 041)By contrast participants who assessed audiovisual recordingsof confessions showed the opposite pattern of results demon-strating more suspicion in response to objectively truthfulconfessions (119872 = 388 SD = 057) relative to deceptiveconfessions (119872 = 334 SD = 071)

Analysis of confidence and information sufficiency rat-ings revealed no statistically significant results for eitherthe main effects of truth status presentation modalityor the interactions Results and associated 119865 statistics for

confidence were truth status 119865(1 43) = 095 119875 = 034 pre-sentationmodality 119865(2 43) = 078 119875 = 047 and interaction119865(2 43) = 056 119875 = 058 Results and associated 119865 statisticsfor information sufficiency were truth status 119865(1 43) =311 119875 = 009 presentation modality 119865(2 43) = 276 119875 =007 and interaction 119865(2 43) = 092 119875 = 041

4 Discussion

41 The Impact of Presentation Modality on Perceptions ofConfessions The first objective of this study was to test theinfluence of presentation modality (audio recordings audio-visual recordings and written transcripts) on observersrsquoability to accurately discriminate true from false confessionsAs predicted and consistent with the findings of Kassinand associates [23] participants exposed to audiovisualrecordings were significantly less accurate at discriminatingbetween truthful and deceptive confessions relative to thoseexposed to audio recordings or written transcripts Althoughexposure to audio recordings and transcripts did not yieldaccuracy rates significantly greater than would be expectedby chance alone as predicted performance below chancelevel by audiovisual observers demonstrated that exposureto visual images in this study significantly impaired detec-tion accuracy In fact while observers assessing audiovisualrecordings were poor at detecting both truthful and deceptiveconfessions they were particularly deficient in their abilityto identify deceptive confessions barely more than a quarter(278) of deceptive confessions were correctly identified

One possible explanation for audiovisual observersrsquoexceptionally poor performance is that they perceivedbehaviours that weremore prevalent in truthful statements tobe indicative of deception systematically resulting inmisclas-sifications of deceptive statements Often when there is littlecorroborating information available to verify a statementobservers who have access to visual cues focus their attentionon nonverbal behaviours including stereotypical cues todeception [19 37 44 58ndash60] Since the typical stereotype of aliar is visual in nature such stereotypes are more likely to beactivated by testimony presented visually [26] Although inthis study there were no statistically significant visible differ-ences between the behavioural cues displayed by truth tellersversus liars a nonsignificant trend indicated that truthful


confessors displayed more gaze aversionmdasha stereotypicalnonverbal cue commonly associated with deceptionmdashthandid deceptive confessors Veracity judgments by observerswho relied on this behaviour would have erred systemati-cally accounting for the finding that audiovisual observerswere more likely to misclassify a true confession than tocorrectly identify a deceptive confession Since the cueswhichobservers relied on in making credibility judgments were notexamined in detail this explanation requires further researchHowever this contention is supported by prior findingsshowing that observers assessing audiovisual recordings ofconfessions attended most closely to nonverbal behavioursand that reliance on stereotypical visual cues (eg gaze aver-sion and fidgeting) impaired detection accuracy [36] If com-mon stereotypical visual cues believed by many laypersons toindicate to deception [40] are in fact associatedwith true con-fessions thismay account for the finding that video observersperformed at levels significantly worse than chancemdasha mostunusual phenomenon in psychological research

An interesting aspect of our results was the observersrsquostrong tendency to believe both true and false confessionsWhile there were no differences across modality groups interms of response bias all observers were biased to perceivethe confessions as true resulting in greater accuracy ratesin classifying true versus false confessions These outcomesare well documented in the deception literature [19 34 6162] and are consistent with the notion that confessions areinherently believable due both to the persuasive features theyoften contain and to the presence of a common observerbias to expect ldquoself-serving behaviour in othersrdquo [10 20] Thefact that confessions in the current study described plausiblesocial transgressions may have increased observer suscepti-bility to believe both true and false confessions

42 Indirect Measures of Deception and Confessions Thesecond objective of this study was to examine the utility ofindirect measures of deception to distinguish true from falseconfessions Overall contrary to our predictions none ofthe indirect measures discriminated truthful from decep-tive messages in the expected direction in any modalityPrior research examining indirect measures of deceptiondemonstrated that these measures can outperform explicitlie detection decisions showing that observers tend to bemore confident and feel that they have more informationin their assessments of truthful messages while being moresuspicious and likely to perceive deceivers as having toldquothink hardrdquo [45 47] Interestingly the current findings werecontrary to those observed in prior studies in that observersperceived true confessors as having to ldquothink harderrdquo thandeceptive confessors Moreover unlike observers who lis-tened to or read confessions observers who watched audio-visual recordings of confessions were more suspicious oftrue than false confessions This was perhaps due to the factthat the indirect assessments were made by observers afterthey reported explicit veracity judgements thereby reducingthe force of the indirect ratings However previous researchhas demonstrated that indirect measures helped observersto discriminate true from deceptive statements even whenboth explicit and indirect assessments were made [45]

Alternatively it may be the case that some indirect measuresof deceptionwere ineffective when applied to confessions dueto the uniquely persuasive features of confessions that makethe task of detecting deception more difficult compared toother types of witness statements

43 Cognitive Load and Confessions The finding that obser-vers perceived truthful confessors as experiencing greatercognitive load than deceptive confessors was potentiallyimportant in highlighting that individuals making a genuineconfession may experience an increased cognitive load Inthis study true confessions weremade about social transgres-sions (eg cheating on a partner stealing and committingproperty damage) that were accompanied by feelings ofembarrassment guilt and remorse In certain situations suchas whenmaking a confession giving truthful accounts can bemore cognitively complex and demanding than communicat-ing deceptive messages [63] In situations where ldquopotentiallydamagingrdquo content is presented for instance individualsconveying true messages must ldquodecipher a way in which thetruth can be palatably packagedrdquo [63 page 110] whereas thosechoosing to deceive are required perhaps more simply tofabricate information suited to the demands of the situationMcCornack [63] contended that generating truthfulmessagesin a ldquopotentially damagingrdquo context was more constrainedand hence more cognitively complex than generating deceitTruth tellers in the current study may have conveyed animpression of elevated cognitive load via cues that werenot captured in the current behavioural coding protocolConversely if deceivers adopted the strategy of insertingautobiographical truths into their deceptive accounts [64]they might not have to ldquothink very hardrdquo about concoctingtheir statements and thusmaynot have exhibited signs of highcognitive load

The foregoing findings raise the possibility that assess-ing the credibility of a confession wherein an individualacknowledges culpability for a transgression or crime may bean especially difficult undertaking since genuine confessionsmay contain features or cues typically judged to be indicativeof deceit DePaulo et al [40] noted that truthful communi-cations associated with feelings of guilt or shame may elicitbehavioural responses or cues which are traditionally linkedto deception Individuals who have engaged in questionablebehaviour or committed a transgression often experiencefeelings of guilt [65] thus truthful confessors may experiencestronger guilt in confessing than individuals who are know-ingly making a false or deceptive confession As a result ldquoifthe behaviour of truthful transgressors was compared withthat of deceptive transgressors cues to these self-consciousemotions would be more in evidence for the truth tellersif they distinguished them from the liars at allrdquo [40 page81] In addition DePaulo et al [40] argued that deceiverstypically experience a higher degree of ldquodeliberatenessrdquo andmay appear less forthcoming and more tense than truthtellers However in contexts involving self-incriminatingstatements truthful communicators may exhibit a greatersense of deliberateness and consequently may appear lessforthcoming and provide less compelling statements thandeceivers [40] Taking these results together it is plausible


that truthful confessions could contain certain characteristicsconventionally assumed to occur more often in deceptivecommunication This potential was emphasised by Kassin[10 21] who noted that false confessions elicited in real-life investigative settings typically contain several featurescommonly associated with truth telling such as rich detailshighly accurate statements that often included explanationsor justifications for commission of a crime and statements ofremorse The inclusion of these features combined with thebehavioural nuances described above may produce persua-sive and credible false accounts that are difficult to discount

44 Limitations of the Study While the foregoing findingshave important implications for assessing the credibility ofconfessions some limitations must be acknowledged Firstthe statement sample size of 24may not have yielded adequatestatistical power to detect differences in behaviours displayedby truthful versus deceptive confessors Further researchexamining the cues displayed by truthful and deceptiveconfessors in larger samples is necessary to shed light on thebehavioural features that may differentiate true from falseconfessions Perhaps the most significant limitation of thisstudy was that the confession statements were confessionsto social transgressions by students not high-stakes criminaladmissions While significant efforts were made to motivateparticipants to discuss salient life events and to provideconvincing accounts we cannot rule out the possibility thatour deception manipulation may not have elicited strongdeception cues [40 66] which may have increased thedifficulty of the deception task for observers across allpresentation modalities However results of the study on theimpact of presentation modality were consistent with thoseconducted with relatively higher stakes [23]

45 Conclusions and Implications for Policy and PracticeThe current findings have important implications for theimplementation of policies that require a full audiovisualrecord of police interviews In particular the poor detectionrate for false confessions (278) shown by audiovisualobservers suggests that reliance on video recorded statementsto assess the veracity of confessions is unlikely to be aneffective method of reducing wrongful convictions arisingfrom false confessions While an audiovisual record maygo some way to ensuring adherence to police interviewingprotocols [15] the current body of evidence indicates thatthe veracity of confessions is best determined via either audiorecordings or written text-based statements as these appearless susceptible to the biases that can impair visually basedcredibility assessments

References

[1] S A Drizin and R A Leo ldquoThe problem of false confessions inthe post-DNA worldrdquo North Carolina Law Review 82 vol 82pp 891ndash1007 2004

[2] SM Kassin andK Neumann ldquoOn the power of confession evi-dence an experimental test of the fundamental difference hypo-thesisrdquo Law and Human Behavior vol 21 no 5 pp 469ndash4841997

[3] R A Leo and R J Ofshe ldquoThe consequences of false confes-sions deprivations of liberty and miscarriages of justice in theage of psychological interrogationrdquo Journal of Criminal Law andCriminology vol 88 no 2 p 429 1998

[4] G H GudjonssonThe Psychology of Interrogations and Confes-sions A Handbook Wiley West Sussex UK 2003

[5] Innocence Project ldquoFalse Confessionsrdquo 2011 httpwwwinno-cenceprojectorgunderstandFalse-Confessionsphp

[6] S M Kassin S A Drizin T Grisso G H Gudjonsson R ALeo andAD Redlich ldquoPolice-induced confessions risk factorsand recommendationsrdquo Law and Human Behavior vol 34 no1 pp 3ndash38 2010

[7] S M Kassin and G H Gudjonsson ldquoThe psychology ofconfessions a review of the literature and issuesrdquo PsychologicalScience in the Public Interest vol 5 no 2 pp 33ndash67 2004

[8] B Scheck P Neufeld and J Dwyer Actual Innocence Double-day New York NY USA 2000

[9] DDixon ldquolsquoAwindow into the interviewing processrsquoThe audio-visual recording of police interrogation in New South WalesAustraliardquo Policing amp Society vol 16 no 4 pp 323ndash348 2006

[10] S M Kassin ldquoOn the psychology of confessions does inno-cence put innocents at riskrdquo American Psychologist vol 60 no3 pp 215ndash228 2005

[11] F E Inbau J E Reid J P Buckley and B C Jayne CriminalInterrogations and Confessions Aspen Montgomery Md USA4th edition 2001

[12] R Leo Police Interrogation and American Justice Harvard Uni-versity Cambridge UK 2008

[13] D Dixon ldquoQuestioning suspects a comparative perspectiverdquoJournal of Contemporary Criminal Justice vol 26 no 4 pp 426ndash440 2010

[14] D Davis andW T OrsquoDonohue ldquoThe road to perdition extremeinfluence tactics in the interrogation roomrdquo in Handbook ofForensic PsychologyW TOrsquoDonohue P R Laws andCHollinEds pp 897ndash996 Basic Books New York NY USA 2003

[15] G D Lassiter and M J Lindberg ldquoVideo recording custodialinterrogations the devilrsquos in the detailsrdquo Journal of ForensicPsychology vol 1 pp E3ndashE10 2009

[16] T P Sullivan ldquoElectronic recording of custodial interrogationseverybody winsrdquoThe Journal of Criminal Law and Criminologyvol 95 pp 1127ndash1140 2005

[17] G D Lassiter ldquoVideotaped interrogations and confessionswhatrsquos obvious in hindsight may not be in foresightrdquo Law andHuman Behavior vol 34 no 1 pp 41ndash42 2010

[18] Innocence Project WA ldquoThe Faces of Exoneration AndrewMallardrdquo 2013 httpwwwinnocenceprojectwaorgauand-rew mallardhtml

[19] A Vrij Detecting Lies and Deceit Pitfalls and OpportunitiesWiley Chichester UK 2nd edition 2008

[20] T R Levine R K Kim and J P Blair ldquo(In)accuracy at de-tecting true and false confessions and denials an initial testof a projected motive model of veracity judgmentsrdquo HumanCommunication Research vol 36 no 1 pp 82ndash102 2010

[21] S M Kassin ldquoTrue or false lsquoirsquod know a false confession if I sawonersquordquo inDeception Detection in Forensic Contexts P A Granhagand L A Stromwall Eds pp 172ndash194 Cambridge UniversityPress Cambridge UK 2004

[22] L Ross ldquoThe intuitive psychologist and his shortcomings dis-tortions in the attribution processrdquo Advances in ExperimentalSocial Psychology vol 10 no C pp 173ndash220 1977


[23] SM Kassin C AMeissner andR J Norwick ldquolsquoIrsquod know a falseconfession if i saw onersquo a comparative study of college studentsand police investigatorsrdquo Law and Human Behavior vol 29 no2 pp 211ndash227 2005

[24] M Davis K A Markus and S B Walters ldquoJudging the credi-bility of criminal suspect statements doesmode of presentationmatterrdquo Journal of Nonverbal Behavior vol 30 no 4 pp 181ndash198 2006

[25] J P Blair T R Levine and A S Shaw ldquoContent in context im-proves deception detection accuracyrdquo Human CommunicationResearch vol 36 no 3 pp 423ndash442 2010

[26] C F Bond Jr and B M DePaulo ldquoAccuracy of deceptionjudgmentsrdquo Personality and Social Psychology Review vol 10no 3 pp 214ndash234 2006

[27] D E Anderson BMDePauloM EAnsfield J J Tickle andEGreen ldquoBeliefs about cues to deceptionmindless stereotypes oruntapped wisdomrdquo Journal of Nonverbal Behavior vol 23 no1 pp 67ndash89 1999

[28] K Brooks I Watkins and D Bradford ldquoUsing eye contactinstructions to facilitate lie detection through increased cogni-tive load sensitivity and bias for scripted and unscripted liesrdquosubmitted to Journal of Criminology

[29] J K Burgoon J P Blair and R E Strom ldquoCognitive biasesand nonverbal cue availability in detecting deceptionrdquo HumanCommunication Research vol 34 no 4 pp 572ndash599 2008

[30] B M DePaulo G D Lassiter and J I Stone ldquoAttentional deter-minants of success at detecting deception and truthrdquoPersonalityand Social Psychology Bulletin vol 8 pp 273ndash279 1982

[31] J E Hocking J Bauchner E P Kaminski and G R MillerldquoDetecting deceptive communication from verbal visual andparalinguistic cuesrdquo Human Communication Research vol 6pp 33ndash46 1979

[32] N RMaier and J AThurber ldquoAccuracy of judgments of decep-tion when an interview is watched heard and readrdquo PersonnelPsychology vol 21 pp 23ndash30 1968

[33] RWiseman ldquoThemegalab truth testrdquoNature vol 373 no 6513p 391 1995

[34] M Zuckerman B M DePaulo and R Rosenthal ldquoVerbal andnonverbal communication of deceptionrdquo in Advances in Exper-imental Social Psychology L Berkowitz Ed vol 14 pp 1ndash57Academic Press New York NY USA 1981

[35] A Vrij ldquoGuidelines to catch a liarrdquo in Deception Detection inForensic Contexts P A Granhag and L A Stromwall Eds pp287ndash314 Cambridge University Press Cambridge UK 2004

[36] M Davis and K A Markus ldquoMisleading cues misplaced con-fidence an analysis of deception detection patternsrdquo AmericanJournal of Dance Therapy vol 28 no 2 pp 107ndash126 2006

[37] J L Hale and J B Stiff ldquoNonverbal primacy in veracity judg-mentsrdquo Communication Reports vol 3 pp 75ndash83 1990

[38] A Vrij ldquoWhy professionals fail to catch liars and how they canimproverdquo Legal and Criminological Psychology vol 9 no 2 pp159ndash181 2004

[39] A Vrij P A Granhag and S Porter ldquoPitfalls and opportunitiesin nonverbal and verbal lie detectionrdquo Psychological Science inthe Public Interest vol 11 no 3 pp 89ndash121 2010

[40] B M DePaulo B E Malone J J Lindsay L MuhlenbruckK Charlton and H Cooper ldquoCues to deceptionrdquo PsychologicalBulletin vol 129 no 1 pp 74ndash118 2003

[41] T R Levine K B Serota H Shulman et al ldquoSender demeanorindividual differences in sender believability have a powerfulimpact on deception detection judgmentsrdquo Human Communi-cation Research vol 37 no 3 pp 377ndash403 2011

[42] S L Sporer and B Schwandt ldquoParaverbal indicators of decep-tion a meta-analytic synthesisrdquo Applied Cognitive Psychologyvol 20 no 4 pp 421ndash446 2006

[43] L A Stromwall P AGranhag andMHartwig inPractitionersrsquoBeliefs about Deception Detection in Forensic Contexts P AGranhag and L A Stromwall Eds pp 229ndash250 CambridgeUniversity Press Cambridge UK 2004

[44] Global Deception Research Team ldquoA world of liesrdquo Journal ofCross-Cultural Psychology vol 37 pp 60ndash74 2006

[45] D E Anderson B M DePaulo andM E Ansfield ldquoThe devel-opment of deception detection skill a longitudinal study ofsame-sex friendsrdquo Personality and Social Psychology Bulletinvol 28 no 4 pp 536ndash545 2002

[46] B M DePaulo R Rosenthal C R Green and J RosenkrantzldquoDiagnosing deceptive and mixed messages from verbal andnonverbal cuesrdquo Journal of Experimental Social Psychology vol18 no 5 pp 433ndash446 1982

[47] A Vrij K Edward and R Bull ldquoPolice officersrsquo ability to detectdeceit the benefit of indirect deception detection measuresrdquoLegal and Criminological Psychology vol 6 no 2 pp 185ndash1962001

[48] M Hartwig and C F Bond ldquoWhy do lie-catchers fail A lensmodel meta-analysis of human lie judgmentsrdquo PsychologicalBulletin vol 137 no 4 pp 643ndash659 2011

[49] R A Bryant A J Barnier D Mallard and R Tibbits ldquoPost-hypnotic amnesia for material learned before hypnosisrdquo Inter-national Journal of Clinical and Experimental Hypnosis vol 47no 1 pp 46ndash64 1999

[50] M T Orne ldquoThe nature of hypnosis artifact and essencerdquo Jour-nal of Abnormal and Social Psychology vol 58 no 3 pp 277ndash299 1959

[51] M T Orne ldquoThe simulation of hypnosis why how and whatit meansrdquo International Journal of Clinical and ExperimentalHypnosis vol 19 no 4 pp 183ndash210 1971

[52] A J Barnier S J Sharman P Ashkar J D Leland A Marshand K McConkey Account Qualities and Interpersonal SourceMonitoring of Autobiographical Memories The Impact of Decep-tion Emotion and Instruction School of Psychology Universityof New South Wales Sydney Australia 2006

[53] A Vrij S A Mann R P Fisher S Leal R Milne and RBull ldquoIncreasing cognitive load to facilitate lie detection thebenefit of recalling an event in reverse orderrdquo Law and HumanBehavior vol 32 no 3 pp 253ndash265 2008

[54] A F Hayes and K Krippendorff ldquoAnswering the call for astandard reliability measure for coding datardquo CommunicationMethods and Measures vol 1 pp 77ndash89 2007

[55] K KrippendorffContent Analysis An Introduction to ItsMetho-dology Sage Thousand Oaks Calif USA 2nd edition 2004

[56] D M Green and J A Swets Signal Detection Theory andPsychophysics Wiley New York NY USA 1966

[57] C A Meissner and S M Kassin ldquolsquoHersquos guiltyrsquo investigatorbias in judgments of truth and deceptionrdquo Law and HumanBehavior vol 26 no 5 pp 469ndash480 2002

[58] SMann A Vrij and R Bull ldquoDetecting true lies police officersrsquoability to detect suspectsrsquo liesrdquo Journal of Applied Psychology vol89 no 1 pp 137ndash149 2004

[59] J B Stiff G R Miller C Sleight P Mongeau R Garlick andR Rogan ldquoExplanations for visual cue primacy in judgments ofhonesty and deceitrdquo Journal of Personality and Social Psychologyvol 56 no 4 pp 555ndash564 1989


[60] A Vrij and G R Semin ldquoLie expertsrsquo beliefs about nonverbalindicators of deceptionrdquo Journal of Nonverbal Behavior vol 20no 1 pp 65ndash80 1996

[61] T R Levine H S Park and S A McCornack ldquoAccuracy indetecting truths and lies documenting the lsquoveracity effectrsquordquoCommunication Monographs vol 66 no 2 pp 125ndash144 1999

[62] S A McCornack and M R Parks ldquoDeception detection andrelationship development the other side of trustrdquo in Commu-nication Yearbook 9 M McLaughlin Ed pp 377ndash389 SageBeverly Hills Calif USA 1986

[63] S AMcCornack ldquoThe generation of deceptivemessages layingthe groundwork for a viable theory of interpersonal deceptionrdquoin Message Production Advances in Communication Theory JO Greene Ed pp 91ndash126 Lawrence Erlbaum AssociatesMahwah NJ USA 1997

[64] B E Malone R B Adams D E Anderson M E Ansfield andBMDePaulo ldquoStrategies of deception and their correlates overthe course of friendshiprdquo inProceedings of the AnnualMeeting ofthe American Psychological Society Washington DC USAMay1997

[65] R F Baumeister A M Stillwell and T F Heatherton ldquoGuiltan interpersonal approachrdquo Psychological Bulletin vol 115 no2 pp 243ndash267 1994

[66] M G Frank and T H Feeley ldquoTo catch a liar challenges forresearch in lie detection trainingrdquo Journal of Applied Communi-cation Research vol 31 no 1 pp 58ndash75 2003

Submit your manuscripts athttpwwwhindawicom

Child Development Research

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

Education Research International


Biomedical EducationJournal of



Psychiatry Journal

ArchaeologyJournal of



AnthropologyJournal of


Research and TreatmentSchizophrenia


Urban Studies Research

Population ResearchInternational Journal of


CriminologyJournal of


Aging ResearchJournal of



NursingResearch and Practice

Current Gerontologyamp Geriatrics Research

Hindawi Publishing Corporationhttpwwwhindawicom

Volume 2014

Sleep DisordersHindawi Publishing Corporationhttpwwwhindawicom Volume 2014

AddictionJournal of


Depression Research and TreatmentHindawi Publishing Corporationhttpwwwhindawicom Volume 2014


Geography Journal


Research and TreatmentAutism


Economics Research International


of formal policy changes in many jurisdictions requiringaudiovisual recording [9 13] However these changes do notprovide an absolute safeguard against wrongful convictionson the basis of false confessions [9 13 17] Indeed thepossibility remains that policemay engage in preliminary un-recorded questioning with a suspect which may compromisethe credibility of any subsequent recorded admissionsThis isclearly illustrated in the Australian case of Andrew Mallardwho was wrongly convicted and imprisoned in 1995 formurder based in part on a 20-minute false videotaped con-fession recorded following 11 hours of police questioning[18] Perhaps of greater concern is the possibility that factfinders may use a video recording as a tool to assess theveracity of suspect admissions in the event of a disputed con-fession [13] Decades of deception research showing that factfinders are generally nomore accurate than chance at discern-ing between videotaped truthful and untruthful statements[19] calls into question the reliability of video recordingsto prevent erroneous prosecution on the basis of a falseconfession

The use of video recordings to assess the veracity of con-fession evidence is further complicated by the inherentlybelievable nature of confessions and the possibility that inthis context truths and lies may be especially difficult todiscern This bias was confirmed by experimental researchshowing that confessions were overwhelmingly believed(gt80 of the time) irrespective of actual message veracityresulting in very low detection rates for false confessions[20] Kassin [10 21] noted two possible reasons why falseconfessions may be particularly difficult to detect first dueto a truth-biasing attribution error (fundamental attributionerror) [22] which ldquoleads people to expect self-serving behav-iour in othersmdashand hence to trust confessionsrdquo [10 page223] Secondly and perhaps more appreciably false con-fessions typically contain highly detailed persuasive andemotional features that are commonly associated with truthtelling (eg details of how and why a crime was committedand expressions of remorse) [10] The presence of thesepersuasive features combined with a predisposition to believeself-incriminating accounts may make false confessions par-ticularly difficult to detect

Research directly investigating observersrsquo performanceat distinguishing truths from lies in the specific context ofconfessions is limited as is the number of studies comparingmode of communication in this context Perhaps the mostpertinent study on this topic examined observer assessmentsof the credibility of five truthful andfive deceptive confessionsprovided by incarcerated prisoners [23] Results revealedan overall accuracy rate of 539 a relatively unimpres-sive performance level not significantly better than chancealone Similarly an analysis of 20 brief (writtenaudiotaped)excerpts (mean length 10 seconds) of actual criminal con-fessions [24] yielded comparable levels of accuracy (55 to62) Even lower accuracy rates (33) were reported forobservers making veracity assessments on the basis of video-recorded confessions to mock crimes in the absence ofsupporting contextual information [25]While these findingsafford little optimism for success in discerning true fromfalse confessions performance accuracy improved when the

modality of a confession was varied For instance partici-pants achieved higher accuracy rates when confessions werepresented in audio-only compared to audiovisual format[23] These findings were congruent with most outcomes inthe general deception literature (see [24 26] for exceptions)indicating that observers tend to achieve higher accuracyrates when they attend to speech content (written transcripts)or vocal cues (audio recordings) compared to exposure tothe full audiovisual presentation of the target (audiovisualrecordings) [27ndash34] As noted by Vrij [35] these findingscontroverted the widely-held belief that ldquolie detection iseasiest when the lie detector has access to the full picture ofthe potential liar and that just reading a textual version ofa statement or just listening to someonersquos voice hampers liedetectionrdquo (p 303)

Differences in performance accuracy between presenta-tion modalities may be attributed to the fact that when visualinformation is available observers rely most heavily on non-verbal behaviours and are particularly drawn to stereotypicalcues to deception (eg gaze aversion and fidgeting) to guidetheir deception judgments [26 29 36ndash39] Given the body ofevidence showing that these cues are not indicative of deceit[40] it is unsurprising that judgments made on this basis areless accurate than those in response to vignettes where thesepotentially distracting cues are unavailableThis contention issupported by research suggesting that reliance on demeanourcues (eg physical movements eye gaze and interactionstyle) alone can result in poor accuracy at discriminatingtruths from lies [41] Furthermore research has shown thatthere are generally more valid verbal and speech-relatedcues to deception than nonverbal cues [19 40 42] andthat observersrsquo notions about the validity of verbal cues todeception are generally more accurate than those concerningnonverbal cues [43] Hence observers may show increaseddetection accuracy when visual information is ignored orunavailable perhaps allowing them to direct their attentionto the more diagnostic speech content andor vocal aspectsof the message

Furthermore to the extent that truthful confessors expe-rience feelings of guilt or shame associated with acknowledg-ing responsibility for a transgression [40] they may exhibitbehaviours consistent with the stereotypical image of decep-tionThese behaviours are more apparent in some modalitiesthan others Since stereotypical beliefs about deception aremost commonly nonverbal [26 44] observers who viewvideotaped confessions may be particularly impaired in theirability to discern truth fromdeception compared to observerswho are not exposed to nonverbal behaviours of confessorsHowever these findings have not been used to guide policydevelopments regarding the uses of police video recordingsto assess the credibility of confession evidence indicatinga need for further clarity on the impact of presentationmodality on perceptions of confessions The primary aim ofthe current research was to address this gap in the literatureto help inform the development of guidelines on how policeinterview recordings should be used to better evaluate theveracity of confessions

While restricting message modality is one procedure thatmay improve veracity discrimination other strategies may













2 Method






















3 Results










Confidence 487 503 463 460 469 477 474 481(052) (073) (082) (094) (075) (072) (069) (079)



(066) (050) (063) (059) (053) (072) (060) (066)








4 Discussion














References













































































Psychiatry Journal



















Volume 2014


AddictionJournal of




Geography Journal

















2 Method






















3 Results










Confidence 487 503 463 460 469 477 474 481(052) (073) (082) (094) (075) (072) (069) (079)



(066) (050) (063) (059) (053) (072) (060) (066)








4 Discussion














References













































































Psychiatry Journal



















Volume 2014


AddictionJournal of




Geography Journal






















3 Results










Confidence 487 503 463 460 469 477 474 481(052) (073) (082) (094) (075) (072) (069) (079)



(066) (050) (063) (059) (053) (072) (060) (066)








4 Discussion














References













































































Psychiatry Journal



















Volume 2014


AddictionJournal of




Geography Journal








Confidence 487 503 463 460 469 477 474 481(052) (073) (082) (094) (075) (072) (069) (079)



(066) (050) (063) (059) (053) (072) (060) (066)








4 Discussion














References













































































Psychiatry Journal



















Volume 2014


AddictionJournal of




Geography Journal
















References













































































Psychiatry Journal



















Volume 2014


AddictionJournal of




Geography Journal



























































Psychiatry Journal



















Volume 2014


AddictionJournal of




Geography Journal





research article the impact of presentation modality on...

Documents