Case Study 32.2 — Marvin Anderson: The Cross-Race Effect, the Mugshot, and the Repeated Face

DataField.Dev

Case Study 32.2 — Marvin Anderson: The Cross-Race Effect, the Mugshot, and the Repeated Face

A real, publicly documented case, drawn from the public record of the conviction, the appeals, and the 2001 DNA exoneration as recounted by the Innocence Project and contemporary reporting. Where a detail is summarized rather than quoted from a primary source, it is presented as the established public account; a small number of procedural specifics are stated qualitatively where the exact record is not pinned down, and labeled as such. We treat a sexual assault soberly and confine ourselves to documented facts.

Why this case complements Cotton/Thompson

Case Study 32.1 (Cotton/Thompson) teaches the chapter through a single, sincere, repeatedly-mistaken witness and the reform that followed. This second case is chosen for a complementary angle: it isolates, with unusual clarity, the cross-race effect (§32.2) and two specific suggestive system variables — the use of a single distinctive photograph that set the suspect apart, and the repeated viewing of that same face across a photo identification and then a live lineup (§32.1, §32.3). It is also a study in how a non-DNA circumstance — which young man's photograph happened to be on file, and why — can steer an identification before the witness ever sees the array. Like Cotton, Marvin Anderson was identified with confidence, convicted, and ultimately exonerated by the DNA ground truth this book keeps returning to (Chapters 6–9). The two cases together show that the failure mode is not a quirk of one witness; it is structural.

Background: the crime and the investigation

In 1982, in Hanover County, Virginia, a young white woman was abducted and raped by a Black man who, during the assault, made remarks indicating he "had a white girlfriend." Investigators reportedly fixed on Marvin Anderson, then a 18-year-old Black man, in part because he was known to live with a white woman — a piece of case context, not identification evidence, that nonetheless shaped whose photograph the witness would be shown. This is worth pausing on against §32.3: before any lineup is run, an investigative theory determines who goes in the array, and if that theory is wrong, the entire identification is being conducted around an innocent person from the start.

Anderson had no prior criminal record. That fact produced a small but consequential procedural artifact. Because he had no record, police did not have a standard mugshot of him on file; the photograph used in the identification was reportedly a color photograph (described in public accounts as obtained from his employer or otherwise distinct in kind), while the surrounding filler photographs were standard black-and-white mugshot-type images. The result, as §32.3 ("lineup composition — the fillers") warns, was that Anderson's photo stood out from the others — a different kind of image among a set of similar ones. A photo array in which one face is presented differently from all the rest is not a clean memory test; it is a display in which one option has been visually flagged.

The forensic evidence — and the conditions

The eyewitness identification. The case against Marvin Anderson rested centrally on the victim's identification. She picked Anderson's photograph from the array, and then identified him again, shortly afterward, from a live lineup in which — by the public account — he was the only person who had also appeared in the photo array. As §32.1 (unconscious transference) and §32.6 (limits on repeated viewing) explain, this is the false-familiarity trap in its most concrete form: by the live lineup, Anderson's was the one face the witness had already seen, and a face seen before is a face that feels recognized. The "recognition" at the live lineup may have been recognition of the photograph, not of the perpetrator.
The cross-race configuration (§32.2). This was a cross-race identification — a white witness identifying a Black assailant — the configuration that the cross-race effect most reliably degrades. As the chapter stresses, this is one of the more robust findings in the field, it is not a matter of prejudice (it appears across groups and in young children), and it leaves the witness's confidence undiminished even as it lowers accuracy. A large share of the eyewitness-misidentification exonerations, including this one, involved a cross-race identification.
The standout photo and the repeated face (§32.3, §32.1). Two suggestive system variables compounded: the array singled Anderson out by the kind of photograph used, and the live lineup re-presented the one face the witness had already chosen. Each of these is a controllable feature of how the identification was collected — a system variable — and each pushed toward selecting, and then re-confirming, the same innocent man.
No usable DNA at trial. As in Cotton's case and so many others (Chapter 6), the biological evidence existed and was preserved, but the comparison that could speak to the source of that evidence — the strong, exclusionary direction (Chapter 1, §1.6) — was not available to Anderson's jury at the time of trial.

The turn: DNA, and a confession that was ignored

Marvin Anderson was convicted and sentenced to a long prison term. His case carries an additional, bitter detail that sharpens the lesson about which evidence the system trusted. Another man, John Otis Lincoln, later admitted to the crime — by the public account, even appearing in court to say so — and yet the confession did not free Anderson; the conviction stood. The system had been more persuaded by a confident cross-race eyewitness identification (the weak, inclusionary direction) than by another man's admission. It would take the strong, exclusionary direction — DNA — to break it.

Years into Anderson's imprisonment, and after his release on parole, the Innocence Project pursued post-conviction DNA testing of the preserved biological evidence. The result was the clean, near-categorical kind §1.6 describes: the DNA excluded Marvin Anderson and matched John Otis Lincoln, the man who had confessed. In 2001, on the strength of that exclusion, Anderson was exonerated; he was granted a full pardon in 2002. The confident identification that had convicted him — sincere, cross-race, made from a standout photograph and a repeated face — had been wrong from the array, and DNA was what finally said so.

Read the case against the chapter

Cross-race effect (§32.2). The core estimator variable. A white witness identifying a Black stranger faces a reliably elevated error rate that her confidence does not register — the precise pattern over-represented in the exoneration record.
Lineup composition / the standout (§32.3). A color photograph among black-and-white fillers made Anderson's image different in kind, flagging it. Good practice requires fillers that do not let the suspect stand out (§32.4, step 1); here the very format of the photo broke that rule.
Repeated viewing and unconscious transference (§32.1, §32.6). Anderson was the only person to appear in both the photo array and the live lineup. By the live lineup, his was the one familiar face — and familiarity from the procedure was available to be misread as familiarity from the crime.
The evidentiary asymmetry (Chapter 1, §1.6). A confident identification included Anderson and could not establish he was the source; another man's confession should have raised grave doubt but did not move the conviction; only a DNA exclusion — the strong direction — overturned it, and the same test included the true perpetrator who had confessed.
The system trusts the wrong number (§32.5). The case is a clean illustration of the chapter's central inversion: the system weighted a confident eyewitness identification over both a recantation-by-another and, for years, the possibility of testing. Confidence persuaded; accuracy lay elsewhere.

What the case did — and didn't — establish

It is tempting to read a case like this as proof that eyewitness evidence is worthless. That conclusion overshoots, and the chapter is careful about it. What the case establishes is narrower and more useful: that a confident identification, made under a powerful estimator variable (cross-race) and through suggestive system variables (a standout photo, a repeated face), can be flatly wrong, and that its confidence carries little information about its accuracy once the procedure has contaminated it (§32.5). It does not establish that human face memory is never useful — under good conditions and clean procedures it often is. The case is an indictment of procedure and condition, not of witnesses, and it points precisely where the chapter points: to the controllable system variables (fair fillers, no repeated viewing of the same suspect, double-blind administration, an immediate verbatim confidence statement) that could have kept the error from being manufactured.

The lesson

Marvin Anderson's case is the cross-race effect and the suggestive lineup, isolated and underlined, with the same ending as Cotton's and for the same reason. A sincere witness, a confident identification, a stacked set of conditions and procedures the witness could not see, a conviction, and — years later — a DNA exclusion that did what confidence could not. The two non-DNA accelerants here are worth carrying away: an investigative theory determined whose photo the witness would see (so a wrong theory put an innocent man in the array), and the format of that photo, plus its reappearance at the live lineup, flagged and then re-confirmed him. Both are system variables (§32.3) — controllable, specifiable, fixable. The case advances the same two themes the chapter leans on hardest: exclusion over proof (a confident identification could only include; a DNA mismatch could exclude) and the validity spectrum (a method whose core claim — "this specific person" — outran what its confidence could justify, corrected only by the one method that has earned the strong direction).

Discussion questions

Anderson's photograph was a different kind of image (reportedly color) among black-and-white fillers. Using §32.3 ("the fillers") and §32.4 (step 1), explain why this makes the array a poor memory test even before the witness's memory is considered. What rule should the fillers have satisfied?
Anderson was the only person to appear in both the photo array and the live lineup. Explain, using unconscious transference and the repeated-viewing problem (§32.1, §32.6), why the witness's "recognition" at the live lineup is ambiguous evidence — and what it might actually have been recognition of.
This was a cross-race identification. State the cross-race effect precisely (§32.2), explain why it is not attributable to prejudice, and explain why the witness's confidence is not lowered by it even though her accuracy is.
Another man confessed, yet the conviction stood until DNA testing. Using the evidentiary asymmetry of Chapter 1 (§1.6) and the confidence-trust inversion of §32.5, explain what this reveals about which evidence the system weighted most heavily — and why that ranking was backwards.
An investigative theory (that the assailant "had a white girlfriend") helped determine that Anderson's photo went into the array. Explain how a wrong theory at this stage contaminates an identification before any lineup is run, and connect it to the system-variable logic of §32.3.
Compare Anderson and Cotton/Thompson (Case Study 32.1). Both are cross-race identifications corrected by DNA. Name one feature each case isolates especially clearly (e.g., the standout photo here; the failure to recognize the true perpetrator there), and state the single shared lesson both cases drive toward (§32.6).