Chapter 28: Forensic Psychology and Criminal Profiling: What Works (Not Much) and What Doesn't

"There is something about the unknown that draws the eye — and the imagination fills it." — a paraphrase of a sentiment common in the psychology of perception [constructed teaching line, after a widely repeated observation about how the mind completes a gap].

Overview

On television, a profiler walks into a crime scene, narrows their eyes, and delivers a portrait of the killer with the precision of a passport: a white male, twenty-five to thirty-five, who lives alone, was abused as a child, works a menial job below his intelligence, drives an old American car, and is — the camera lingers here — returning to the scene to watch. The detectives nod, the net tightens, and the profile is vindicated in the final act. It is one of the most seductive images forensic science has ever produced, and it is the one this chapter is going to take apart.

Here is the uncomfortable order of things. The discipline that can deliver real value to courts and investigations — forensic psychology — is narrow, careful, and almost entirely unglamorous: it assesses whether a defendant is mentally able to stand trial, whether they were legally responsible at the time of an offense, and how likely a particular person is to commit violence in the future. None of that involves divining a stranger's childhood from the position of a body. The discipline that television sells as the crown jewel — criminal profiling, the construction of a psychological and behavioral portrait of an unknown offender — is the single most over-sold forensic product in popular culture, and the evidence that it actually works, in the way it claims to work, is thin to the point of embarrassment.

This is the chapter where the book's fourth theme — the CSI effect, the over-selling of forensics — meets its purest case. We will be precise about a distinction the public never hears: the gap between forensic psychology, which has standards, error rates that people at least try to measure, and a genuine place in court, and profiling, which mostly does not. And we will advance the validity spectrum into territory it has not yet visited. Profiling, judged honestly against the yardstick of the 2009 NAS and 2016 PCAST reports, sits low — not because psychology is unscientific, but because the specific claim profiling makes (that scene behavior reliably predicts an unknown offender's personal characteristics) has almost no validated foundation.

In this chapter, you will learn to:

• Tell forensic psychology — the working discipline of competency, responsibility, and risk — apart from the criminal profiler of fiction.
• Explain how competency to stand trial and the insanity defense are actually evaluated, and what they do and do not decide.
• Describe the FBI profiling model, the organized/disorganized typology, and the difference between modus operandi and signature.
• Weigh the real evidence on whether profiling works — and understand why "it helped catch him" is not the same as "it works."
• Define the Barnum effect and use it to explain why a profile feels accurate even when it predicts nothing.
• Locate profiling near the low end of the validity spectrum, and name the few places psychology genuinely helps a case.

Learning Paths

🔎 Investigator/CSI: Your temptation is the profile — the tidy portrait that tells you whom to look for. Sections 28.3 and 28.4 are your inoculation: they show how a profile can narrow your search in the wrong direction and harden into tunnel vision. The most useful thing here is §28.6's honest list of what psychology can do for an investigation, so you ask it for that and not for magic. 🧪 Lab analyst: You will rarely produce profiling evidence, and you should be glad. Note §28.2 — the assessments (competency, risk) that are the field's real, testable work — and §28.4–28.5, where you will recognize the same validity logic (error rates, falsifiability) that governs every method in this book. ⚖️ Law/courtroom: Sections 28.2 and 28.4 are yours. Competency and sanity testimony is admitted routinely and governed by real standards; profiling testimony as proof of identity is generally not admitted, and §28.4 explains why. Learn to hear the difference between an opinion a court will take and one it will (rightly) refuse. 👥 General reader/juror: This entire chapter is the antidote to the most beloved forensic fiction there is. §28.1 and §28.5 are the core: the profiler is mostly a character, and the reason the character feels so right — the Barnum effect — is a fact about your mind, not about the killer's.

28.1 Forensic psychology vs. the TV profiler

Begin with the word, because almost everyone who uses it means the wrong thing. Forensic psychology is the application of psychological science and clinical practice to questions that a legal system must answer. Like every discipline in this book, it takes an established field — here, clinical and research psychology — and turns it toward the law: Is this defendant mentally able to participate in their own trial? Were they legally responsible when the offense occurred? How dangerous is this person if released? Is this child's testimony reliable? Does this person meet the legal standard for civil commitment? These are real questions, asked thousands of times a day in courts across the country, and answering them is the actual job of the actual profession.

Notice what is not on that list: figuring out who an unknown killer is from the way a crime was committed. That activity — criminal profiling, also called criminal investigative analysis or offender profiling — is the inference of an unknown offender's likely characteristics (demographics, personality, background, behavior) from the evidence of the crime scene and the offense. It is a small, specialized, and — as we will see — empirically shaky corner of the field, practiced by relatively few people, and it is almost the only part of forensic psychology that television depicts. The result is a public image that is nearly the inverse of the reality: the profession's least validated activity stands in for the whole, while its genuine, careful, everyday work is invisible.

The two activities differ in a way worth stating sharply, because the difference is the chapter:

• The working discipline (competency, responsibility, risk) evaluates a known person who is present — you interview them, test them, review their records, and reach an opinion about a defined legal question, an opinion that can be checked against standards and, sometimes, against outcomes.
• Profiling describes an unknown person who is absent — you reason backward from a scene to a portrait of someone you have never met and cannot test, and the portrait can almost never be checked until (and unless) someone is caught, at which point the checking is hopelessly contaminated by hindsight.

That asymmetry — present-and-testable versus absent-and-unfalsifiable — is why the two sit at opposite ends of the validity spectrum despite sharing a profession. Hold it in mind; it explains nearly everything that follows.

🔬 At the Bench What does a forensic psychologist actually do on a Tuesday? Far more often than anything resembling profiling: a court-ordered competency evaluation. The psychologist (or psychiatrist) receives a referral, reviews the charging documents and the defendant's mental-health and medical history, conducts a structured clinical interview, often administers standardized instruments designed for the specific legal question, and writes a report answering that question for the judge — for example, whether the defendant can understand the proceedings and assist their attorney. The work is documented, structured, governed by professional guidelines, and subject to a second opinion from an evaluator retained by the other side. It looks like medicine, not like mind-reading. The drama is in the legal stakes, not in a flash of intuition.

There is one more reason the distinction matters here, in a forensic-science book specifically. The other chapters have taught you to ask, of any method, "what scientific discipline does it borrow from, and do that discipline's standards travel intact into court?" (Chapter 1). Forensic psychology borrows from clinical and research psychology, and for its core tasks — competency, risk — a good deal of that rigor does travel: there are validated instruments, peer-reviewed reliability studies, and explicit standards. Profiling borrows the prestige of psychology — the lab coat, the credential, the language of "personality" — without bringing a validated method for the specific claim it makes. It is the bite-mark problem (Chapter 16) in a different costume: a borrowed authority standing in for a borrowed science. Learning to see that substitution is, again, the whole point.

28.2 Competency, sanity, and risk assessment: the real bread and butter

If you want to know what forensic psychology genuinely contributes to justice, look here, not at the profilers. Three tasks make up the bulk of the working field, and all three are admitted in court routinely, governed by real legal standards, and supported — to varying but genuine degrees — by validated methods. They are unglamorous, which is exactly why television ignores them and why this book foregrounds them.

Competency to stand trial

The most common forensic-psychology referral is the evaluation of competency to stand trial — whether a defendant, right now, has sufficient present ability to understand the proceedings against them and to assist in their own defense. This is not a question about guilt, and it is not a question about the defendant's state of mind at the time of the crime. It is a present-tense question about the fairness of trying them at all: a legal system cannot justly prosecute someone who cannot comprehend the charges or participate in their defense. The governing standard in U.S. federal courts was set by the Supreme Court in Dusky v. United States (1960), which articulated the test in terms of whether the defendant has a rational and factual understanding of the proceedings and can consult with counsel with a reasonable degree of rational understanding. Most states adopted substantially the same standard.

Competency is the field's strongest, most defensible product, and it is worth understanding why. The question is concrete and present-tense (can this person, today, do these specific things?), the person is available to be examined and re-examined, structured assessment instruments exist specifically for it, and — crucially — a wrong answer can be corrected, because competency can be restored (often through treatment) and re-evaluated. Inter-rater reliability for competency determinations, when structured instruments are used, is generally good. This is forensic psychology at its most science-like: a defined question, a testable subject, validated tools, and a feedback loop.

⚖️ In the Courtroom Competency and sanity are routinely confused by the public and must never be confused by you. Competency to stand trial is present-tense: can the defendant participate in the trial happening now? A defendant found incompetent is not acquitted; the proceedings are typically paused while competency restoration is attempted. The insanity defense is past-tense: it concerns the defendant's mental state at the moment of the offense, and it goes to criminal responsibility (guilt), not to the fairness of the trial. A person can be perfectly competent to stand trial for an act committed while, the defense argues, legally insane — and the two questions are decided separately, by different evidence, sometimes years apart. Collapsing them is one of the most common errors a juror (or a screenwriter) makes.

Sanity and criminal responsibility

The second task is the evaluation of criminal responsibility — the defendant's mental state at the time of the offense, the question at the heart of the insanity defense. Here the law, not psychology, defines the standard, and the standards vary by jurisdiction. The historical anchor is the M'Naghten rule (from an 1843 English case), under which a defendant may be excused if, due to a disease of the mind, they did not know the nature and quality of the act, or did not know that it was wrong. Some U.S. jurisdictions use M'Naghten; others have used a broader "irresistible impulse" addition or the American Law Institute's Model Penal Code formulation; a few have abolished the defense in its traditional form. After the 1981 attempted assassination of President Reagan, the acquittal of John Hinckley Jr. by reason of insanity provoked a national backlash, and Congress passed the Insanity Defense Reform Act of 1984, which narrowed the federal standard and shifted the burden of proof to the defense.

Two facts about the insanity defense cut directly against its television image. First, it is rare and rarely successful — it is raised in a small fraction of felony cases and succeeds in a small fraction of those. The image of defendants routinely "getting off" by faking madness is a myth; the defense is hard to win, and a successful insanity acquittal generally leads to commitment in a secure psychiatric facility, frequently for longer than a prison sentence would have run. Second, the evaluation is genuinely difficult and more contestable than competency, because it requires reconstructing a past mental state from interview, records, and the circumstances of the offense — and a past internal state cannot be examined directly the way present competency can. This is why "battles of the experts" cluster around sanity more than around competency: the question is intrinsically harder to pin down, and reasonable, qualified evaluators can disagree.

Risk assessment

The third task is violence risk assessment — estimating the likelihood that a particular individual will commit violence in the future. It arises everywhere in the legal system: bail and sentencing, civil commitment, parole, release decisions, threat assessment. And it has a real, instructive history that maps onto this book's central argument about validity.

For most of the twentieth century, risk was assessed by unstructured clinical judgment — an experienced clinician forming an impression. The research verdict on that approach was damning: unstructured clinical predictions of violence were, in landmark reviews, only modestly better than chance and prone to large errors, especially over-prediction of danger. In response, the field developed actuarial and structured professional judgment tools — instruments that combine empirically validated risk factors according to explicit rules, rather than relying on a clinician's gut. These structured approaches measurably outperform unstructured judgment. The lesson is precisely the one Chapter 31 will make about bias and Chapter 1 made about method: a structured, validated, error-rate-bearing procedure beats expert intuition, even very experienced intuition — and a field becomes more scientific exactly when it admits this and adopts the structure.

🔬 At the Bench Even the best risk instruments share an honesty constraint you must understand: they estimate a probability for a group, not a certainty for an individual. A tool may place a person in a category whose members reoffend at, say, a certain rate over a certain period — but it cannot tell you whether this person will be among those who do. Communicating that distinction is exactly the prosecutor's-fallacy discipline from Chapter 9, in a new setting: a group base rate is not an individual verdict. An honest risk assessment reports a probabilistic category with its uncertainty, never "this person will be violent." The good news is that this is real, validated science with measured predictive accuracy. The constraint is that its accuracy is modest and group-based, and the honest practitioner says so.

Three tasks, then — competency, responsibility, risk — and notice what unites them. Each concerns a real person who can be examined; each is governed by an explicit legal standard; each has structured, validated tools that beat unaided intuition; and for each, the honest output is bounded and checkable. This is the genuine science of forensic psychology, and it is admitted in court every day. Keep it firmly in view, because the rest of the chapter is about an activity that has almost none of these properties — and yet is the only one anyone has heard of.

🔍 Check Your Understanding 1. A defendant is found incompetent to stand trial. Does that mean they have been acquitted, or found "not guilty by reason of insanity"? Explain the difference between competency and the insanity defense. 2. Why does structured/actuarial risk assessment outperform unstructured clinical judgment — and what is the book-wide principle (from Chapter 1 and previewing Chapter 31) that this illustrates?

28.3 Criminal profiling: the FBI model and its claims

Now we turn to the activity television actually depicts, and we will describe it fairly before we test it. Criminal profiling is the attempt to infer the characteristics of an unknown offender — age range, sex, race, employment, relationship status, personality, residence relative to the crime, likely prior record, and post-offense behavior — from the evidence of the crime and the crime scene. The premise is that behavior reflects personality, and that the way an offender behaves during a crime therefore leaves a readable trace of who they are.

The most famous version is the model developed by the FBI's Behavioral Science Unit (later the Behavioral Analysis Unit) beginning in the 1970s. Its agents interviewed a sample of incarcerated offenders — including a number of convicted murderers — about their crimes, their methods, and their histories, and from those interviews built a typology and a method. The signature product of that early work was the organized/disorganized dichotomy: the claim that violent offenders, particularly killers, fall broadly into two types whose crime scenes reveal which they are.

THE FBI "ORGANIZED / DISORGANIZED" TYPOLOGY  (as classically presented — illustrative)

   ORGANIZED offender                         DISORGANIZED offender
   ─────────────────────                      ──────────────────────
   crime scene:  controlled, planned          crime scene:  chaotic, spontaneous
   victim:       targeted, a stranger         victim:       opportunistic, known/nearby
   weapon:       brought to scene, removed     weapon:       found at scene, left behind
   body:         hidden, transported           body:         left in place
   ↓ inferred portrait                         ↓ inferred portrait
   intelligent, socially competent,            below-average intelligence, socially
   employed, may be married, follows           inadequate, lives/works near scene,
   the case in the media, drives a car         unskilled, sexually/socially immature

   The model claims the LEFT COLUMN (scene features) lets you read off the RIGHT COLUMN
   (the offender's personal characteristics). This inferential leap is the contested claim.

Read the diagram carefully, because the entire scientific controversy lives in that downward arrow. The left side — describing the crime scene as more or less organized — is, in principle, observable. The right side — the portrait of a person — is an inference, and the model's core claim is that the one reliably predicts the other: that scene organization is a valid index of the offender's intelligence, social competence, employment, and the rest. Stripped of the television lighting, profiling stands or falls on whether that inferential arrow is sound. (We test it in §28.4. The short version: the arrow is far weaker than the model claims, and the two "types" do not cleanly exist.)

Built atop this typology is a vocabulary you have certainly heard, and that is worth defining precisely because it is frequently used loosely. Two terms in particular do real analytic work and are owned by this chapter.

A modus operandi (MO) — Latin for "method of operation" — is the set of learned behaviors an offender uses to commit the crime and get away with it: how they gain entry, control the victim, and avoid detection. The defining feature of MO is that it is functional and therefore changeable — it serves the goal of completing the crime, and offenders refine it with experience, abandoning what fails and keeping what works. A burglar who learns that a pry bar is noisy and switches to a glass cutter has changed their MO; the function (entry) is constant, the method evolves.

A signature, by contrast, is behavior that goes beyond what is functionally necessary to commit the crime — an act the offender performs to satisfy a psychological or emotional need, not to complete the offense. Because it serves an inner need rather than a practical goal, signature is theorized to be more stable across an offender's crimes than MO, and so, in linkage analysis, it is treated as the more reliable thread for connecting offenses to a common perpetrator. The classic illustration: the method of a series of crimes may vary (different entries, different weapons), but a distinctive, non-functional ritual repeated at each scene — something done that the crime did not require — may persist, and that persistence is what an analyst looks for when arguing that one person committed them all.

🔬 Read the Evidence

text FIGURE 28.1 — "MO versus signature at three scenes" [constructed teaching example] THE ITEM Three burglary-assaults attributed, by an analyst, to one offender. Recorded at each: the method of entry, the means of controlling the occupant, and any non-functional behavior. Scene 1: pried a door, used cord found on-site, rearranged the victim's shoes in a row before leaving. Scene 2: cut a window screen, brought zip ties, again arranged the shoes. Scene 3: forced a garage, brought rope, again the shoes. THE CONTEXT The analyst is asked whether the three are the work of one person (a linkage question), not who that person is. WHAT IT SHOWS The METHOD of entry and control varies across the three (pry bar → screen cut → garage; on-site cord → zip ties → rope) — an evolving MODUS OPERANDI. One behavior is constant and is NOT required to commit the burglary: arranging the shoes — a candidate SIGNATURE. WHAT IT DOESN'T The shoe-arranging does not reveal WHO the offender is — not their age, job, or address. It is a linkage thread, not an identity. Nor is even the linkage certain: coincidence, copycat behavior, or analyst over-interpretation remain possible, and the "constant" detail may be read as more distinctive than it is. THE INFERENCE The varying MO and the repeated, non-functional shoe ritual are CONSISTENT WITH a single offender across the three scenes — a linkage hypothesis worth investigating, stated at the strength of "consistent with," never "proves the same person did all three." THE LESSON MO answers "how, functionally" and evolves; signature is the non-functional residue theorized to persist. Even at its most useful, profiling-style analysis LINKS behavior; it does not read off an identity, and it does not prove the linkage.

The MO/signature distinction is the most genuinely useful idea profiling has produced, and we will give it its due in §28.6 — linkage (deciding whether several crimes share an offender) is a more defensible task than prediction (describing who that offender is). But even linkage is an inference of consistency, not a fingerprint, and the popular version of profiling is almost entirely about the prediction the evidence cannot support. To that evidence we now turn.

28.4 Does profiling work? The evidence

We can now ask the question the television version never invites: judged as a method — by the standards this book applies to every method — does criminal profiling actually do what it claims? The honest answer, supported by the research that exists, is: there is very little good evidence that profiling reliably predicts an unknown offender's characteristics, and what evidence there is gives the practice weak and unimpressive support. Profiling is, at best, an investigative aid of modest and unproven value; it is not evidence of identity, and it should never be treated as such.

Several distinct problems converge, and it is worth taking them one at a time, because each is an instance of a validity failure you have met elsewhere in the book.

The foundational typology does not hold up. The organized/disorganized dichotomy — the bedrock of the classic FBI model — has not survived empirical scrutiny. When researchers examined large samples of crime-scene data, offenders did not sort cleanly into two types; most scenes show a mix of "organized" and "disorganized" features, and the supposed types are better described as a blurry continuum, if they describe anything stable at all. A typology that does not carve nature at its joints cannot support confident inferences drawn from it.

The inferential leap is unvalidated. Even granting that a scene can be characterized, the core claim — that scene behavior reliably predicts personal characteristics like intelligence, employment, or marital status — has rarely been tested with the kind of controlled study that could establish it, and where it has been examined, the predictive links are weak. The premise that "behavior reflects personality" so consistently that crime-scene conduct reveals a stranger's biography is, as decades of personality psychology suggest, far shakier than the model assumes: human behavior is heavily shaped by situation, not personality alone, and is notoriously inconsistent across contexts. A method built on an overstated premise inherits the overstatement.

The original method was methodologically weak from the start. The early FBI work drew its typology from interviews with a modest, non-representative sample of already-incarcerated offenders — that is, in part, the ones who got caught. There was no comparison group, no blind testing, no measurement of how often the resulting inferences were right or wrong on new cases. As a way of generating hypotheses it was reasonable; as a validated predictive method it was never established, and being interesting is not the same as being accurate.

The accuracy studies are unimpressive. The handful of studies that have tried to measure profiling accuracy directly — for instance, by comparing the predictions of self-identified profilers against those of other groups, or against the known facts of solved cases — have generally found that professional profilers do not substantially outperform comparison groups (and in some designs, non-profilers do comparably well), and that profiles tend to contain a high proportion of vague or widely applicable statements. Whatever edge expertise provides appears small and inconsistent — nothing resembling the television portrait of near-clairvoyant accuracy.

⚠️ Junk-Science Alert Beware the survivorship trick that sustains profiling's reputation. The cases the public remembers are the hits — the investigation where a profile seemed to match the eventual offender — recounted afterward, in memoirs and documentaries, with all the misses quietly dropped. This is the same selective storytelling that kept bite marks credible (Chapter 16): a vivid success, retold, stands in for a measured error rate that no one has actually computed. A method's validity is not established by anthologies of its triumphs; it is established by studies that count the failures too. For profiling, the studies that do count the failures find the method weak — and a famous success story, however dramatic, cannot outvote them. When someone defends profiling by telling you about the time it worked, your first question is the one from Chapter 1: and how often does it not, and how would we know?

There is a final, decisive point, and it is the most important practical lesson of the section: a profile that "helped catch him" did not necessarily work as a method, and may even have done harm along the way. An investigation that ends in a correct arrest can have been misdirected by its profile for weeks before unrelated evidence cracked it. The profile pointing toward the eventual offender at the end tells you nothing about whether it pointed away from him, or toward innocent people, in the middle. Worse, a confident, wrong profile is not merely useless — it is dangerous, because it can steer an investigation away from the actual offender and toward someone who happens to fit the portrait. That is not a hypothetical risk. It is exactly what happened to Richard Jewell (case-study-01), and it is exactly the risk the cold case will illustrate in this chapter's Case File: a profile is a lens, and the wrong lens does not just fail to focus — it shows you the wrong thing, confidently.

This is why courts, by and large, do not admit profiling testimony as proof that a particular defendant committed a crime. Under the admissibility standards of Chapter 5 (Daubert, Frye, FRE 702), a method offered to prove identity must have a demonstrated foundational validity — a known, acceptable error rate, testability, the things the 2016 PCAST report demanded of feature-comparison methods. Profiling, judged by that yardstick, generally fails: it has no established error rate for the identity claim, its central typology is empirically shaky, and its inferences are not reliably reproducible. Where profiling-derived testimony has been allowed at all, it has typically been confined to narrow, behavior-linkage purposes (is this series the work of one offender?) rather than offered as a portrait that points to the defendant — and even there it is contested. The profile is an investigative tool of unproven value, not a courtroom proof of who did it.

So where does profiling sit on the validity spectrum this book has built? Low. Not at the very bottom with bite marks — because parts of the broader behavioral-analysis enterprise (notably linkage via MO/signature, §28.6) have at least a coherent rationale and some empirical traction, and because profiling is rarely offered as courtroom proof of identity, so it has done less direct convicting harm than the discredited pattern methods. But low: well below DNA, toxicology, and fingerprints; below even the contested firearms and bloodstain methods, because those at least attempt a physical comparison that could be validated, whereas predictive profiling rests on a psychological premise the evidence does not support. The honest placement is "investigative aid of weak and unproven value; not a validated method for identifying an offender."

🔍 Check Your Understanding 1. Give two distinct reasons the organized/disorganized typology fails to support confident inferences about an unknown offender. 2. Why is "the profile matched the killer who was eventually caught" a poor argument that profiling works? Name the specific reasoning error (hint: think about what the story leaves out).

28.5 The Barnum effect and why profiles feel accurate

If profiling is as weak as the evidence says, a real puzzle remains: why does it feel so accurate? Investigators who have used profiles, families of victims, and television audiences all report the same uncanny sense that the profiler "nailed it." That feeling is real, and it is the single most important psychological phenomenon in this chapter — but it is a fact about the reader of a profile, not about the profile's accuracy. Its name is the Barnum effect.

The Barnum effect (also called the Forer effect, after the psychologist who demonstrated it experimentally in the 1940s) is the tendency for people to accept vague, general personality descriptions as highly accurate descriptions of themselves specifically, when in fact those descriptions are broad enough to fit almost anyone. It is the engine of horoscopes, cold reading, and fortune-telling, and it is named for the showman to whom the line "a sucker born every minute" is popularly (if dubiously) attributed — the point being that a statement crafted to feel personal while applying universally is a reliable way to manufacture a sense of accuracy where none exists.

The classic demonstration is worth picturing, because it is devastatingly simple. Give a roomful of people the identical personality sketch — "You have a great need for other people to like and admire you. You have a tendency to be critical of yourself. You have a great deal of unused capacity which you have not turned to your advantage. While you have some personality weaknesses, you are generally able to compensate for them…" — tell each person it was generated individually for them from a test they took, and ask them to rate its accuracy. People rate it, on average, as strikingly accurate — often near the top of the scale — for themselves, despite every person in the room having received the exact same words. The accuracy they feel is entirely supplied by their own minds, reading specificity into statements that contain none.

Now apply that directly to a criminal profile. A profile that predicts the offender is "a male in his twenties or thirties, with deep-seated anger, who may have had a difficult relationship with a parent, who is of at least average intelligence but underachieving, and who likely follows news coverage of the case" feels specific and insightful. It is, in fact, a Barnum statement aimed at offenders: it describes an enormous fraction of the plausible suspect pool (a great many offenders are youngish men with anger and family difficulties), it hedges with "may" and "likely," and it includes the unfalsifiable flourish — "follows the coverage" — that can be confirmed by almost anyone and disconfirmed by almost no one. When a suspect is eventually arrested, the human mind does the rest: it matches the suspect to the vague profile after the fact, counting every loose correspondence as a hit and forgetting the predictions that missed. The profile did not predict the person; the person was read into the profile by hindsight.

FIGURE 28.2 — "Barnum statement vs. specific, testable prediction"   [constructed teaching example]

  BARNUM-STYLE PROFILE STATEMENT          │  SPECIFIC, FALSIFIABLE PREDICTION
  (feels accurate; predicts almost nothing)│  (could be wrong; therefore informative)
  ───────────────────────────────────────  │  ──────────────────────────────────────
  "a male, 20s–30s, with hidden anger"     │  "left-handed; works night shifts at the
                                            │   plant on Route 9; drives a maroon sedan
                                            │   with a dented left fender"
  fits a huge share of possible offenders   │  fits very few people; checkable against a
                                            │   named suspect — and can be flatly FALSE
  "may have had a troubled childhood"       │  "has a felony conviction for arson in this
                                            │   county within the last five years"
  almost no one can be excluded by it       │  excludes nearly everyone; a database can
                                            │   confirm or refute it
  "likely follows news of the case"         │  "will telephone the victim's family on the
                                            │   anniversary from a 555-area-code number"
  unfalsifiable; confirmable after the fact │  a real prediction: it happens or it doesn't

  KEY: A statement that cannot be wrong cannot be informative. The Barnum column manufactures
       a feeling of accuracy; the specific column risks being false, which is exactly why a
       hit in that column would actually mean something. Real profiles lean heavily LEFT.

Walk the figure deliberately, because it contains the whole lesson of the section. The two columns differ in one property that this entire book has trained you to prize: falsifiability. The left-hand, Barnum-style statements are constructed — usually not deliberately, but by the nature of the task — so that they can scarcely be wrong; and a statement that cannot be wrong carries no information, no matter how accurate it feels. The right-hand statements risk being false — they name a handedness, a workplace, a vehicle, a record — which is precisely why a correct one would be informative and would genuinely narrow the field. The tragedy of profiling is that its statements live overwhelmingly in the left column. They feel like the right column, because the Barnum effect supplies the felt specificity, but they predict like the left. When you next hear a profile, sort its claims into these two columns in your head: you will find almost everything lands on the left, and you will understand at once both why it feels so accurate and why it tells you so little.

🧠 Cognitive-Bias Watch The Barnum effect rarely travels alone; it arrives with a convoy of biases that, together, make a useless profile feel like a triumph and — worse — can bend an investigation around it. Confirmation bias (Chapter 31) leads investigators to notice the suspects and facts that fit the profile and to discount those that do not. Hindsight bias rewrites the profile's accuracy after an arrest, so a portrait that pointed in several directions is remembered as having pointed at the person caught. And a profile, once issued, can seed tunnel vision: it tells the investigation what kind of person to look for, and the search obligingly finds people of that kind — sometimes the wrong ones. The safeguard is the book's recurring one (Chapter 31): treat a profile as one tentative, falsifiable hypothesis to be actively tested and disconfirmed, never as a template the evidence is bent to fit. A profile believed is far more dangerous than a profile doubted.

The Barnum effect is also, finally, the bridge between this chapter and the book's fourth theme. The CSI effect over-sells forensic science; the Barnum effect explains the psychological mechanism by which one of its products — the profile — sells itself, by recruiting the audience's own mind to manufacture the accuracy the method lacks. Understanding the Barnum effect does not just inoculate you against profiling. It inoculates you against every confident, vague, unfalsifiable claim that feels like knowledge, in a courtroom or out of it.

28.6 Where psychology genuinely helps an investigation

It would be a mistake to leave this chapter thinking psychology has nothing to offer an investigation. It has a great deal to offer — just not the thing television advertises. The honest task is to separate the genuine contributions, which are real and sometimes considerable, from the mind-reading portrait, which is not. Here, then, is the defensible list: the places psychological science earns its keep in an investigation, each stated at its true strength.

The validated assessments (the field's core). Everything in §28.2 — competency, criminal responsibility, risk — is genuine psychological science contributing to justice, governed by standards and supported by validated tools. This is not "helping an investigation" in the detective sense, but it is the real, court-admitted work of the field, and it dwarfs profiling in volume and value. Do not let the glamour of the profiler obscure that the profession's actual product is sound and indispensable.

Behavioral linkage analysis (MO/signature), used cautiously. Deciding whether a series of crimes is likely the work of one offender — using the evolving modus operandi and the more stable signature (§28.3) — is a more defensible task than predicting an unknown offender's biography, because it is a comparison of behaviors across known offenses rather than an inference about an unseen person's character. It still produces consistency judgments ("these offenses are consistent with a common offender"), not certainties, and it is vulnerable to coincidence, copycats, and over-interpretation — but as an investigative aid for prioritizing and connecting cases, it has a coherent rationale and some empirical support. Use it to link, hedge it honestly, and never let it harden into "and therefore the offender is [type]."

Geographic profiling. A genuinely more quantitative cousin of offender profiling, geographic profiling uses the locations of a linked series of crimes to estimate the most probable area of an offender's residence or anchor point, drawing on well-studied regularities in how people travel and offend close to (but not too close to) home. Because its inputs (locations) are objective and its output (a probability surface over a map) is testable and falsifiable, it is more defensible than psychological profiling — though it still produces a prioritized search area, not an address, and depends on the crimes truly being linked. It is the part of the profiling family that most resembles a real method, precisely because it traffics in measurable geography rather than inferred personality.

Investigative interviewing and the psychology of memory. Psychology's most underrated investigative contribution is how to get accurate information from people — and how to avoid corrupting it. The research on memory and suggestibility (which Chapter 32 develops for eyewitnesses, and Chapter 33 for interrogation) yields concrete, validated methods: non-leading interview techniques, the dangers of suggestive questioning, the construction of fair lineups, the conditions under which confessions become unreliable. This is psychological science with a real evidentiary payoff, and unlike profiling, it is grounded in decades of controlled experiments.

Threat assessment. Distinct from after-the-fact profiling, threat assessment asks a forward-looking, structured question — does a known, identified person (a student making threats, an employee, a stalker) pose a genuine risk of targeted violence, and what should be done? Modern threat assessment is structured, evidence-informed, and concerned with a real person whose behavior can be evaluated and managed. It is, in effect, risk assessment applied to prevention, and it shares risk assessment's relative validity.

⚖️ In the Courtroom Notice the through-line in every genuine contribution above: each concerns either a known, examinable person (competency, sanity, risk, threat assessment) or an objective, testable input (crime locations in geographic profiling; behaviors compared across known offenses in linkage). None of them is the unfalsifiable portrait of an unknown mind. That is the line a court draws, and the line you should draw: psychological evidence is welcome when it rests on examinable subjects, validated instruments, or testable data, and it is refused — rightly — when it asks a jury to convict on a profile's intuition about who an absent stranger must be. When you evaluate any "behavioral" evidence, ask which side of that line it falls on. The answer predicts both its courtroom fate and its scientific worth.

Step back and the shape of the field comes clear. Forensic psychology is a genuine and valuable forensic discipline — when it does the work it can actually do: assessing examinable people against legal standards with validated tools, linking and prioritizing cases with hedged behavioral comparison, narrowing search areas with testable geography, and protecting the integrity of the information investigations depend on. Its over-sold product, the predictive profile of an unknown offender, is the weak exception, not the rule — low on the validity spectrum, useful at best as one tentative hypothesis, dangerous when believed. The discipline's honest motto might be the book's own: it constrains and assists far more reliably than it identifies, and it does its best work on the person in the room, not the phantom at the scene.

🗂️ The Case File

The profile that pointed the wrong way. Return to the early days of the Mill Creek investigation, before the autopsy (Chapter 11) had overturned the accidental-fire assumption and well before the financial (Chapter 27), digital (Chapter 25), and trace (Chapter 24) evidence had begun to accumulate. As pressure built to make sense of a death that some in the department had started to suspect was no accident, an outside analyst was asked, informally, for a behavioral read. The analyst — working from the remote location, the fire, and the apparent absence of an obvious personal connection — produced a familiar portrait: this looked like the work of a lone drifter or stranger, someone passing through, opportunistic, unconnected to the victim's life; investigators should be looking outward, at transients and outsiders, not at the people closest to Marcus Diallo.

What that profile did. It pointed investigators away from Roy Keller — Diallo's business partner, the person with the closest financial entanglement to the property — and toward a phantom outsider who, the later evidence would show, did not exist. The profile was confident, it was plausible, and it was wrong in its direction. For a time it shaped where attention went and where it did not, and it sits in the file now as a cautionary specimen of exactly what §28.4 and §28.5 describe: a Barnum-flavored portrait ("lone, angry, opportunistic stranger") that felt like insight, fit the early facts loosely, and — most dangerously — steered the investigation, hardening into a small piece of tunnel vision before the physical and digital evidence pulled the case back on course.

The honest status. Profiling misled; excluded as proof. The profile is not evidence of anything and proves nothing about who killed Marcus Diallo; it never could. Logged in the workbook (Appendix I), it earns a single honest line and a warning label: a behavioral narrative is, at best, an investigative hypothesis to be tested and disconfirmed, and here it pointed the wrong way and very nearly cost the investigation time it could not afford. It excludes no one and includes no one. What it teaches is method, not fact: when a profile feels most compelling, that is exactly the moment to ask what it would take to prove it wrong — and to keep following the evidence that does not fit it. It is a capital mistake to theorize before one has data — and a worse one to let a theory tell the data where to look.

Conclusion

Forensic psychology is two very different things wearing one name, and telling them apart is the whole of this chapter. The working discipline — competency to stand trial, criminal responsibility, violence risk assessment — is a genuine forensic science: it examines real, present people against explicit legal standards, with structured and validated instruments that demonstrably outperform unaided intuition, and it is admitted in court every day to bounded, checkable effect. Criminal profiling — the construction of a psychological portrait of an unknown offender from the crime scene — is the field's over-sold exception: its founding organized/disorganized typology does not survive empirical scrutiny, its core inference from behavior to biography is unvalidated and rests on an overstated premise, the studies that measure its accuracy find it weak, and its persuasive power derives less from any predictive success than from the Barnum effect, the human tendency to read specificity into vague statements that fit almost anyone. Profiling sits low on the validity spectrum — above the discredited pattern methods only because its more disciplined cousins (MO/signature linkage, geographic profiling) have some traction and because it is rarely offered as courtroom proof — and it is, at its honest best, one tentative hypothesis to be tested and disconfirmed, never a finding. In the cold case it pointed the wrong way, toward a stranger who did not exist and away from the very person the financial evidence had already flagged (Chapter 27), and it stands in the file as a warning: profiling is not evidence; believed rather than tested, it is a hazard.

The chapter has advanced two of the book's themes with unusual directness. It is the purest instance yet of the CSI effect — the over-selling of forensic science — because no forensic product is more glamorized on television and less validated in fact than the profiler. And it has extended the validity spectrum into the behavioral sciences, separating the methods that examine testable subjects from the one that reads an absent mind. In the next chapter we turn to the emerging technologies — rapid DNA, microbial forensics, isotopes, and artificial intelligence — that promise the next leaps, and that we will hold to exactly the same yardstick we have just held to profiling: not "does it sound impressive?" but "what is the evidence it works, and how would we know if it didn't?"

Key Terms

• Forensic psychology — the application of psychological science and clinical practice to legal questions, chiefly the assessment of competency to stand trial, criminal responsibility (sanity), and violence risk in examinable individuals.
• Criminal profiling — the inference of an unknown offender's likely characteristics (demographics, personality, background, behavior) from crime-scene and offense evidence; an investigative aid of weak and unproven validity, not a method for identifying an offender.
• Modus operandi (MO) — the learned, functional behaviors an offender uses to commit a crime and avoid detection; because it serves the practical goal of completing the offense, it tends to evolve with experience.
• Signature — behavior beyond what is functionally necessary to commit the crime, performed to satisfy a psychological need; theorized to be more stable than MO across an offender's crimes and thus more useful for behavioral linkage.
• Barnum effect — the tendency to accept vague, broadly applicable personality descriptions as highly accurate descriptions of oneself (or of an offender); the mechanism by which an uninformative profile feels uncannily specific.

Spaced Review

1. Distinguish competency to stand trial from the insanity defense: which is present-tense and which is past-tense, what does each decide, and why is competency generally the more reliable assessment of the two? (§28.2)
2. Forensic accounting (Chapter 27) gave the cold case a financial motive tied to Roy Keller; the profile in this chapter pointed away from him toward a "lone stranger." Explain how these two pieces of the file illustrate the difference between evidence that constrains and a narrative that merely feels explanatory. (§28.4–28.6; Chapter 27)
3. Recall Locard's exchange principle and the class/individual distinction from the foundations (Chapters 1, 3). A defender of profiling says a profile "individualizes" the offender the way a fingerprint does. Name two reasons that claim is false. (§28.4; Chapters 1, 3)
4. Validity-spectrum question. Where does predictive criminal profiling sit on the NAS 2009 / PCAST 2016 spectrum relative to single-source DNA (Chapter 7) and bite-mark comparison (Chapter 16), and what specific property — present in DNA, absent in profiling — accounts for the gap? (§28.4; the spectrum from Chapter 1)
5. A detective, certain a suspect is guilty, reads a vague profile and "sees" the suspect in every line of it. Name the two cognitive phenomena at work (one from this chapter, one previewed for Chapter 31) and state the safeguard. (§28.5; preview of Chapter 31)