29 Examiners Available Now — Contact Them DirectlyNo Booking Fee — You Pay Your Examiner Directly270+ Testing Locations Across the U.S., Canada & UKFree Polygraph Research Database — 44,000+ StudiesWe Promote Clear Standards & Best Practice in Private Polygraph TestingLearning Center — Independent Guides to Every Type of Polygraph TestSame-Day Verbal Result · Written Report Within 48 HoursVetted Examiners Trained at Accredited Polygraph Schools100% Confidential — Private, Appointment-Only TestingPolygraph Glossary (600+ Terms), Book Library & School DirectoryTrusted by 10,000+ Clients, Attorneys & Organizations
LieDetectorTest.com Stop guessing. Get the truth.
Research ledger
How to accurately detect autobiographical events

Home › Polygraph Research › How to accurately detect autobiographical events

Catalogue entry · Reaction Time & Cognitive Load Methods

How to accurately detect autobiographical events

Sartori, G., Agosta, S., Zogmaister, C., Ferrara, S.D., Castiello, U. — Psychological Science,

2008Published
209 subjects across eight validation experiments; laboratory replication samples and real-life forensic casesSample size
Key findings

The aIAT achieved 91% correct classification rates at both group and individual levels, with test-retest reliability exceeding 90% and split-half reliability showing 88% agreement. The method successfully identified true autobiographical memories in laboratory settings and real-life forensic cases involving violent crime confessions.

Abstract

We describe a new method, based on indirect measures of implicit autobiographical memory, that allows evaluation of which of two contrasting autobiographical events (e.g., crimes) is true for a given individual. Participants were requested to classify sentences describing possible autobiographical events by pressing one of two response keys. Responses were faster when sentences related to truly autobiographical events shared the same response key with other sentences reporting true events and slower when sentences related to truly autobiographical events shared the same response key with sentences reporting false events. This method has possible application in forensic settings and as a lie-detection technique.

Methodology

Six experiments employed a five-block computerized IAT variant in which participants categorized logical truth statements and alternative versions of autobiographical events, with scoring based on the IAT-D Index measuring reaction time associations between true statements and correct memory versions. Validation involved 209 subjects across eight experiments with standardized stimulus presentation and motor-response mapping.

Detailed summary

Sartori and colleagues developed the autobiographical Implicit Association Test (aIAT) to distinguish between two competing autobiographical narratives by measuring automatic, unmeasurable reaction-time associations between events and truth values. The method adapts the classic IAT paradigm to measure which autobiographical event is genuinely encoded in episodic memory, moving beyond traditional physiological polygraph approaches. Across six core experiments and eight validation studies, the aIAT achieved 91% accuracy with strong reliability metrics (test-retest >90%, split-half 88%). The test required only 10-15 minutes of administration on a standard PC with no specialized training for interpretation, and demonstrated efficacy in both controlled laboratory conditions and real forensic cases involving serious crimes. The approach offers theoretical advantages by grounding deception detection in cognitive psychology rather than autonomic arousal patterns.

Implications for polygraph practice

The aIAT provides a practical, low-cost alternative to neuroimaging and traditional polygraph methods for forensic memory detection, with potential for remote web-based administration for screening purposes. Findings suggest that reaction-time cognitive paradigms may offer superior accuracy and accessibility compared to physiological approaches, with implications for future development of automated, psychologically-grounded deception detection systems.

Comprehensive study analysis

An in-depth, original analysis of this research study's methodology, findings, and significance for the polygraph profession.

Background & Context

The detection of autobiographical memories has long been a critical challenge in forensic settings, where establishing what individuals truly remember about crimes can determine guilt or innocence. Traditional memory-detection methods like the Concealed Information Test rely on recognizing crime details unknown to innocent persons, but these approaches have limitations when attempting to distinguish between two competing narratives of events.

The 2008 study by Sartori and colleagues introduced the autobiographical Implicit Association Test (aIAT), adapting the classic IAT to determine which of two contrasting autobiographical events is encoded as true in an individual's memory. This represented a paradigm shift from measuring simple recognition to assessing the implicit associations between autobiographical events and logical truth values.

The method emerged from cognitive psychology's understanding that reaction times can assess automatic associations between concepts that individuals cannot willfully override, offering potential advantages over traditional polygraph techniques that rely on physiological arousal patterns.

Research Design & Methodology

Sartori and colleagues conducted six experiments demonstrating extremely high accuracy (overall 91%) in determining which of two autobiographical events was true. The paradigm employed a computerized classification task using four stimulus categories.

The test structure included:

  • Logical categories with sentences objectively true ("I am in front of a computer") or false ("I am climbing a mountain") for respondents at testing time
  • Alternative versions of an autobiographical event (e.g., "I went to Paris for Christmas" vs. "I went to New York for Christmas"), only one being true
  • Five-block design with initial logical discrimination (20 trials) and event discrimination blocks, followed by combined classification tasks
  • Scoring using the IAT-D Index to measure association strength between true statements and autobiographical versions, inferring which matches episodic memory

Experiment 1 used a playing card paradigm where participants chose between two cards (AUC = .88; Hit rate = 81% in subsequent replications). The procedure followed a standardized five-block format with combined blocks mapping both tasks onto the same response keys.

Results & Key Findings

The aIAT demonstrated correct classification rates of 91% at both group and individual levels. The true autobiographical event was identified because it produced faster reaction times when sharing the same motor response with true sentences, with validation experiments finding more than 90% accuracy.

Key statistical outcomes included:

  • Data from eight validation experiments with 209 subjects calculated classification accuracy based on D-IAT values
  • Weighted average D-IAT of 0.59 for first administration and 0.70 for second administration
  • Split-half reliability showed 88% agreement in identifying true autobiographical memory using odd versus even stimuli
  • Test-retest reliability exceeding 90% when the same aIAT was administered twice

The test performed well not only in laboratory settings but also in real-life forensic cases involving individuals who had confessed to murder or attempted murder.

Discussion & Significance

The aIAT offered several practical advantages: administration time of 10-15 minutes, unmanned analysis requiring no specialized training, low-tech equipment needs (standard PC sufficient), and potential for remote web-based administration for screening purposes. These features made it substantially more accessible than neuroimaging approaches like fMRI.

The theoretical significance lay in demonstrating that differences in categorization latency between compatible and incompatible blocks indicated association strengths between autobiographical events and logical truth categories. During incongruent blocks, the cognitive system activates control processes to inhibit prepotent responses (associating true events with "true") and emit alternative responses, reflected in reaction time patterns and ERP components like LPC and N200.

Subsequent research applied the aIAT across diverse contexts including past intentions, white lies, mock crimes, holidays, drug consumption, driver's licenses, flashbulb memories, whiplash malingering, and real cases of suspected crime-related amnesia.

Limitations & Considerations

Verschuere and De Houwer (2009) demonstrated that coached individuals could fake aIAT responses similar to other lie-detection tests, though fakers could often be identified. Participants easily became familiar with IAT procedures through websites like implicit.harvard.edu, where millions had already completed similar tests and received feedback, enabling self-training to alter performance.

Additional methodological constraints included:

  • Subsequent investigations revealed that negative sentences or reminder labels generated unreliable and inaccurate results, requiring exclusion from validation analyses
  • The aIAT detects what is stored and believed true in an individual's mind rather than objective truth—if an event is strongly believed true and encoded as such, the aIAT identifies it as true regardless of actual occurrence
  • Wide variability in classification accuracy emerged across contexts, with particularly low performance when testing non-specific events like drug abuse
  • The emotionally neutral autobiographical memories tested in laboratory settings differed from more emotional-laden forensic memories, requiring replication in ecological settings

Practical Applications

For forensic examiners, confidence in single-case results could be enhanced by building aIATs on known autobiographical data (birthdate, children's names) to verify correct identification, then testing central facts and peripheral details in separate aIATs administered at least one week apart. Individual D-values could be interpreted using accuracy estimates—for example, a D of 0.43 falls in the 0.4-0.49 range with average 81% accuracy.

For consumers and legal settings, the aIAT's 91% accuracy must be weighed against countermeasure vulnerabilities and the distinction between subjective memory encoding versus objective truth. The technique shows promise as an investigative tool but requires careful interpretation within the broader evidentiary context, particularly given that the aIAT is based on associations between response labels rather than pure information recognition, offering potential resilience against information contamination.

Read the original study

The analysis above is original editorial content based on our review of this research. For the complete study including full data, methodology details, and author discussion, access the original publication below.

Related research

Other studies in this category that may be of interest.

Join Our Examiner Network

APA-trained examiners using validated techniques can apply to join the LieDetectorTest.com network.

Apply now →

Keep reading the ledger.

Every peer-reviewed study on polygraph and deception detection we track — catalogued, searchable and citable.

Need to book now? Our online booking system is open 24/7. Speak directly with our team about your test or booking.