Scientific Content Analysis

This encyclopedia entry examines Scientific Content Analysis (SCAN), a verbal deception detection technique developed by former Israeli polygraph examiner Avinoam Sapir that analyzes written statements for linguistic cues to deception. Despite worldwide use by law enforcement and federal agencies, empirical research consistently demonstrates that SCAN fails to discriminate between truthful and deceptive statements above chance levels.

Review of SCAN as a verbal credibility assessment technique, including examination of its claimed criteria, procedural methods, and empirical validation studies involving laboratory experiments and field studies of sexual abuse cases.

Background & Context

Scientific Content Analysis (SCAN) was developed by former Israeli police lieutenant and polygraph examiner Avinoam Sapir, emerging from his experience conducting polygraph examinations. Based on his experience with polygraph examinees, Sapir argues that people who tell the truth differ from liars in the type of language they use, and developed criteria that he claims can assist in differentiating between true and fabricated statements, though without reporting a theoretical foundation as to why these specific criteria should differ.

SCAN is very popular and used worldwide, with practitioners from many different countries reporting it as the most frequently used lie detection tool at investigative interviewing seminars. The method is used worldwide in countries including Australia, Belgium, Canada, Israel, Mexico, UK, US, the Netherlands, Qatar, Singapore, and South Africa, and is also used by federal agencies, military law enforcement, private corporations, and social services. This widespread adoption occurred despite limited scientific validation, representing a significant gap between forensic practice and empirical evidence.

Research Design & Methodology

In a typical SCAN procedure, the examinee is asked to write down "everything that happened" in a particular period of time to get a "pure version" of the facts, typically obtained without the interviewer interrupting or influencing the examinee, then a SCAN trained analyst investigates a copy of the handwritten statement using several criteria described throughout the SCAN manual. Criteria that are present within the written statements are highlighted according to a specific color scheme, circled or underlined, with the presence of a specific criterion indicating either truthfulness or deception, depending on the criterion itself.

SCAN lacks a well-defined list of criteria as well as a standardized scoring system. Research has shown that 12 criteria primarily drove SCAN in sexual abuse cases, and only six published studies examined the validity of SCAN, of which only four were published in peer reviewed journals. Key research examining SCAN includes:

• A study with 61 participants who wrote truthful, outright lies, or concealment lies about activities they had just completed, with statements coded using both SCAN and Reality Monitoring (RM)
• Analysis of 82 sexual abuse cases from Dutch police in which SCAN had been applied, with two independent coders scoring various SCAN criteria in written statements from victims, suspects, and witnesses
• A study where participants wrote one truthful and one fabricated autobiographical statement about a negative event, with two raters indicating the presence of 12 SCAN criteria

Results & Key Findings

SCAN failed to demonstrate validity above chance levels. Reality Monitoring discriminated significantly between truth tellers and outright liars and between truth tellers and concealment liars, whereas SCAN did not discriminate between truth tellers and either kind of liar.

Specific empirical findings from validation studies:

• Two raters indicated the presence of 12 SCAN criteria, but no significant differences emerged between truth tellers and liars
• SCAN trained police officers classified four statements with an average accuracy of 68%, police officers without SCAN achieved 72%, and students 65%, with the SCAN group not differing significantly from police officers who did not use SCAN, leading to the conclusion that SCAN did not have an incremental value in detecting deceit
• Cross-validated classification showed that 49.60% of liars and 53% of truth tellers were correctly classified, demonstrating that SCAN performed around chance level
• SCAN is primarily driven by 12 criteria, with results indicating low inter-rater agreement for most SCAN criteria, suggesting SCAN is insufficiently developed as a forensic tool

Discussion & Significance

In contrast to CBCA and RM, SCAN presents no theoretical rationale, and there is no evidence that these criteria are actually diagnostic. Leading deception detection authority Aldert Vrij points out that most studies did not rely on ground truth being established, there is no standardization among different methods of analysis implying much depends on subjective interpretation and skill of the individual, and attributes this to an absence of theoretical underpinning behind SCAN.

Subsequent empirical studies found that SCAN techniques are applied inconsistently and are not reliable at detecting deceptive statements, and the use of SCAN techniques has been found to be vulnerable to contextual bias on the part of investigators. A 2016 U.S. government review concluded that SCAN "did not distinguish truth-tellers from liars above the level of chance" and that some of its "indicators of deception" were in fact signs that a suspect was telling the truth. The scientific consensus characterizes SCAN as lacking empirical support despite its widespread use in forensic practice.

Limitations & Considerations

The fundamental limitations of SCAN are severe and multiple. While findings about verbal cues are less variable and more strongly related to deception than non-verbal cues, and verbal cues are found in the content and meaning of statements, SCAN's specific criteria lack theoretical foundation. Ground truth was not established in field studies referenced by proponents, and for CBCA and SCAN criteria that overlap, predictions contradict each other with CBCA experts claiming some criteria are more prevalent in truthful statements while SCAN experts claim the same criteria are more prevalent in deceptive statements.

Critics argue that the technique encourages investigators to prejudge a suspect as deceptive and affirm a presumption of guilt before interrogation has begun, and SCAN has been criticized as "theoretically vague" with little or no empirical evidence in its favor and characterized as "junk science" with classification as pseudoscience.

Practical Applications

Despite its widespread use, the scientific evidence suggests SCAN should not be employed as a credibility assessment tool in investigative or forensic contexts. Research shows a lack of validity of SCAN, with little agreement between raters in identifying SCAN criteria, and overall results indicating that the psychometric qualities of SCAN as an investigative tool are insufficient for use in police practice.

For practitioners seeking evidence-based verbal credibility assessment tools, scientifically validated alternatives such as Criteria-Based Content Analysis (CBCA) when properly applied, Reality Monitoring, or newer approaches like the Verifiability Approach offer superior empirical support. The disconnect between SCAN's popularity and its lack of scientific validation underscores the critical need for law enforcement and investigative agencies to adopt only empirically validated deception detection methods.