Confirmation bias and suspect projection in investigations

Confirmation bias and suspect projection in investigations is the tendency to make a favored suspect fit the evidence, rather than let the evidence determine whether the suspect fits. In the book, the danger is not limited to amateurs or to careless police work: a plausible clue, a dense file, or the emotional force of a decades-old case can make unrelated facts seem to converge. DNA eventually disproves several promising theories, while methodical indexing and file review offer a counterweight to the impulse to turn resemblance into certainty.

How a Suspect Becomes the Answer

The book describes suspect fixation as a narrowing of attention. Once a particular face occupies the investigator’s mind, ordinary details can be recruited to support the imagined case. The narrator compares this process to the early intensity of falling in love: curiosity about the suspect keeps expanding, and confirmation bias encourages the investigator to project the desired story onto ambiguous information. A photograph of a smiling man beside a cake, for example, can be misread as an image of someone holding a knife. The distortion lies not in the photograph, but in the interpretation imposed on it.

That pattern also makes the abundance of modern information hazardous. The narrator reflects that returning to the past with more data and newer tools does not automatically produce better judgment. More facts create more opportunities to connect details that may have no meaningful relationship. Pattern-seeking can make a rough resemblance feel like a discovery, and the investigator may remain attached to it even when the evidence offers a reason to move on. The book treats this not as proof that investigation is futile, but as a recurring risk that must be recognized and checked.

Larry Pool’s False Leads

Larry Pool recalls an early episode in his work on the Original Night Stalker case. In 2001, investigators told him that a fingerprint found on a lamp at an East Area Rapist scene was believed to belong to the killer. The circumstances made the print seem compelling: the victim had heard the intruder turn on the light, and the lamp had recently been unpacked, reducing the likelihood that someone else had touched it. A retired Contra Costa investigator had recovered an old copy of the print and sent it to Orange County.

The suspect identified through the print had died five years earlier. Pool was skeptical from the beginning, in part because the man’s age and criminal history did not seem right. Further inquiry supplied an alternative explanation for the print. The dead man had been a friend of the victim’s boyfriend; the two had fallen out shortly before the attack, and a stereo had been stolen from the home around the same time. Pool theorized that the suspect might have entered the house to steal the stereo and left the print, without being the murderer. His superiors nevertheless wanted a conclusive test. After the suspect’s remains were exhumed and DNA was examined, the result confirmed Pool’s suspicion: the man was not the killer.

The episode shows both the appeal and the limits of a striking forensic clue. The print was real, and its location seemed significant, but its presence did not establish who committed the attack. The surrounding circumstances supported another account of how it could have been left. DNA testing supplied a decisive check where interpretation alone could not. Pool’s disappointment at the investigative excitement gathering around the suspect also reflects the emotional cycle of the work: a long-imagined breakthrough can collapse when the next piece of evidence arrives.

The Narrator’s Projection of a Suspect

The narrator experiences a similar rush while searching for the killer. A tip about library holdings of Larry Crompton’s book Sudden Terror leads to a search of registered sex offenders in relevant areas. The narrator finds Doug Fiedler, whose height and weight seem compatible with the profile and whose California convictions included sexual offenses. Additional details appear to reinforce the theory: Fiedler’s mother came from a large Sacramento County family, and his former wife had lived at an address in Dana Point, near the site of Keith and Patty Harrington’s murders.

The accumulating details produce a sense of momentum, but they do not establish that Fiedler committed the crimes. The narrator sends Pool the information, believing there is a strong chance it has identified the killer. Pool replies that Fiedler has already been eliminated through DNA in CODIS. The apparent trail of green lights ends with a result that does not fit the narrator’s emerging story. In retrospect, the narrator recognizes how readily a plentiful collection of addresses, biographical details, and geographic overlaps can be arranged into a persuasive villain. The case of Fiedler makes the book’s warning personal: the researcher who understands projection in principle can still be caught by it in practice.

Pool’s attitude has changed through repetition. Early in his work, he admitted to feeling more excitement about suspects. Fifteen years later, he described a more restrained response: he would review a report and set it aside unless the evidence warranted further attention. That caution does not mean he has stopped caring about the case. The book also recalls his vivid response when a room of colleagues brought him news of a possible fingerprint match. Rather, experience has taught him that excitement is not verification, and that a suspect who initially seems promising may still be wrong.

Collective Research and Checks on Interpretation

The book extends the problem beyond individual investigators to online communities following the East Area Rapist / Original Night Stalker case. Members of the A&E Cold Case Files message board analyzed newly uncovered incidents and exchanged theories; the board had more than a thousand topics and twenty thousand posts. Its members could assemble detailed dossiers, including maps and yearbook photographs, but that diligence did not guarantee sound conclusions. A computer-science graduate student, for instance, sent the narrator a seven-page dossier on a person of interest and worried over whether the suspect’s shoe size matched the offender’s. Attention to detail could sharpen an inquiry, but it could also deepen commitment to a theory before the central claim had been established.

The community’s information was also partial and unevenly shared. The Social Worker, one of the board’s de facto leaders, sometimes acted as a gatekeeper between investigators and forum members. On July 2, 2011, she posted a drawing of a decal reportedly seen on a suspicious vehicle near a Sacramento rape scene, asking whether anyone could identify it while noting that a possible Naval Air Station North Island connection was unconfirmed. The detail was a lead to examine, not proof of the vehicle’s connection to the crimes. The account further reveals that investigators were subtly steering the online sleuths, so apparent grassroots research operated within a larger exchange between public contributors and law enforcement.

A more systematic approach appears in Michelle McNamara’s work with official case materials. She indexed people, places, and objects in the reports to make investigative details easier to find, distinguish people who might be of interest because of later geographic movement, and locate possible overlaps among victims. She shared findings and the Master List with active investigators. In January 2016, she and Paul Holes were allowed to review sixty-five Bankers Boxes of case files held at the Orange County Sheriff’s Department. Among the most valued material were supplemental reports about overlooked suspects, witnesses, vehicles, and prowlers—items that had drifted away from the main case folders.

Michelle bought two high-volume scanners so the material could be digitized, made searchable, and shared with investigators, including Paul Holes, Ken Clark, and Erika Hutchcraft. The method did not guarantee that the offender’s name would be found, although Michelle estimated that the likelihood was about 80 percent. Its significance for this subject is methodological: rather than build a suspect from whichever details most vividly attract attention, the work organized a large, dispersed record so that connections and omissions could be examined. The book thus sets projection against procedures that preserve uncertainty, invite verification, and make neglected evidence available for renewed scrutiny.