Memento MedicaHistory of medicine & the mind
All stories

Psychology, sampling bias & the path away from diagnosis · Los Angeles, 1953–1973

Evelyn Hooker and the Blind Pairs

Psychology had built its picture of gay men largely from patients and prisoners. Evelyn Hooker recruited men living outside institutions, hid their identities, and asked experts to distinguish the records.

6 minute read · 1,259 words · Published November 11, 2026

An anonymous mid-century woman researcher compares two stacks of unlabeled psychological test records.
AI conceptual illustration of blinded psychological comparison in the 1950s. It is not Evelyn Hooker, her office, or an actual participant record.

Content note: Historical psychiatric pathologizing, discrimination, coercive treatments, and criminalization; no graphic imagery.

Two stacks of test records sat without names. One came from gay men, one from straight men. Could experts tell which was which?

In the 1950s, the dominant clinical literature treated homosexuality as a disorder. Its evidence often came from people already in psychiatric care, prison, or other institutions. Psychologist Evelyn Hooker asked what that sampling choice had built into the conclusion.

With help from Los Angeles gay community networks, she recruited thirty gay men who were not patients and matched them with thirty heterosexual men. She administered three projective tests, removed identifying information, and gave the protocols to expert judges. The judges could not separate the groups reliably or rate one as less adjusted.

The study did not, by itself, remove homosexuality from psychiatry’s diagnostic manual. Its methods and limits were contested, and projective testing has since lost much of its authority. But Hooker changed the population under observation—and made a circular diagnosis harder to defend.

A diagnosis built from patients

When researchers study only people who have entered treatment, the sample already contains distress, conflict, or coercion. In mid-century research on homosexuality, that distortion was often ignored. Findings from psychiatric patients or incarcerated people were generalized to gay people as a class.

The logic could close around itself. Homosexuality was classified as illness; clinicians then examined people referred because of homosexuality; the difficulties visible in those clinical records were treated as evidence that homosexuality itself caused disorder. Social hostility, secrecy, and legal danger rarely received equal weight.

Hooker’s intervention began with who counted as evidence. Her friend Sam From, a gay former UCLA student, challenged her to study gay men who were living ordinary lives outside clinics. Through community organizations and personal networks, she gained access to volunteers whom conventional researchers seldom reached.

The work carried risks for participants. Same-sex intimacy was criminalized, employment could be lost, and police raids were common. Confidentiality was therefore a condition of meaningful research, not a decorative ethical promise.

Sources [1] · [2] · [5]

Thirty matched pairs

Hooker received a National Institute of Mental Health grant in 1953. She recruited thirty gay men and thirty heterosexual men, matched by age, education, and intelligence. None was in therapy at the time, a design choice aimed directly at the patient-sampling problem.

Each participant completed the Rorschach inkblot test, the Thematic Apperception Test, and the Make-A-Picture-Story test. These were projective methods: psychologists interpreted stories and perceptions as indirect evidence of personality and adjustment.

Hooker removed information about sexual orientation and presented the records to clinicians with expertise in the tests. The blinding mattered. If a judge knew which record belonged to a gay man, prevailing theory could shape the reading. Anonymous protocols forced the diagnosis to confront its own instruments.

The design was small and depended on men willing and able to volunteer. It did not represent every gay person, women, transgender people, or the full range of race and class. Its strength was narrower: it created a fairer comparison than the clinical samples used to label an entire population.

Sources [1] · [2] · [3]

The experts cannot sort them

Two independent Rorschach experts rated overall adjustment. According to the American Psychological Association’s summary, each placed two-thirds of both groups in the three highest categories. Asked to identify which protocols came from gay men, the judges performed no better than chance.

A third expert reviewed the story-based tests. Again, adjustment ratings did not differ significantly between the groups. Hooker published the results in 1957 under a title that used the terminology of its period: “The Adjustment of the Male Overt Homosexual.”

The conclusion was deliberately bounded. The study did not prove that every participant was free of difficulty. It challenged the claim that homosexuality, by itself, could be detected as psychological disturbance in otherwise comparable men.

That distinction loosened a central diagnostic assumption. If expert clinicians using respected instruments could not distinguish the blinded records, then the presumed pathology was not appearing where the theory said it should.

Sources [1] · [3] · [4]

A historical tool, not a modern gold standard

Projective tests were influential in mid-century clinical psychology, but their reliability and validity have been heavily disputed. Modern readers should not convert Hooker’s use of them into an endorsement of every inference once drawn from inkblots or picture stories.

In this study, however, their prestige made the negative result strategically important. Hooker used tools the profession respected to test a belief the profession also respected. When the judges could not sort the records, the failure belonged to the claim as measured by its own accepted methods.

The study would not meet every modern expectation. The sample was modest, recruitment was not random, demographic coverage was narrow, and statistical practice has changed. Those limitations narrow the claim without erasing the comparison.

The deeper methodological lesson survives: a diagnosis can reflect the route by which people enter a dataset. Change the sampling frame, blind the evaluator, and an apparent difference may disappear.

Sources [1] · [3] · [5]

From one paper to a professional change

Hooker continued studying gay communities and became an important adviser on sexuality research. Her work circulated as gay and lesbian activists increasingly challenged psychiatric authority, demanded evidence, and described the harms created by diagnosis.

In December 1973, the American Psychiatric Association removed the general diagnosis of homosexuality from DSM-II and replaced it with the narrower category Sexual Orientation Disturbance. A later “ego-dystonic” category persisted until 1987. It would be inaccurate to credit one researcher or one experiment with these changes: activism, other research, internal debate, and changing professional experience all contributed.

Hooker’s study supplied a durable piece of evidence within that larger movement. It made it harder to defend pathology as an observable fact rather than a conclusion derived from clinical selection and social prejudice.

The history also warns against treating professional reform as instant liberation. Diagnostic revisions did not end discrimination, coercive practices, criminal penalties, or stigma. They changed one institutional statement whose effects had reached courts, families, employers, and treatment rooms.

Sources [2] · [5] · [6]

A network that made research possible

Recruiting outside clinics required trust that a university title could not supply by itself. From and other community contacts helped potential participants judge the request, while Hooker had to protect identities in a period when disclosure could bring arrest or dismissal.

That dependence complicates the heroic-researcher story. Hooker designed and published the study, but the comparison existed because gay men accepted risk and a community network supplied access. Their names remain absent from the paper for reasons bound to the danger the study was meant to expose.

Sources [2] · [5]

What blindness did to authority

Blinding did more than reduce a technical bias. It temporarily removed the social label that had organized the profession’s expectations. The expert still held authority, but that authority had to operate on the record rather than on knowledge of the participant’s identity.

This is why the paired protocols remain such a powerful object. They show how a fairer procedure can reveal that certainty belonged partly to the examiner’s frame. The result did not make psychology value-free; it made one source of expectation visible and testable.

Sources [1] · [3]

The record with the label removed

The physical object at the center of Hooker’s work was ordinary: a protocol, copied and stripped of identity. Its anonymity created an unusual test of authority. Experts could still interpret every mark and story, but they could no longer let the category announce itself first.

That design did not ask gay men to prove exceptional health. It asked whether comparable people looked inherently more disturbed when the evaluators did not know whom they were judging.

The answer was limited but disruptive. In those matched records, using those respected tools, the profession’s experts could not find the difference its diagnosis presumed.

Hooker’s contribution was therefore both human and technical. She reached people whom clinical research had excluded, protected their identities, and arranged the evidence so that a label could not do the examiner’s work.

Sources [1] · [2] · [3]

More stories from Memento Medica

For more interesting content, follow Memento Medica on YouTube.

Follow Memento Medica

Written and illustrated with AI assistance for Memento Medica. The lead image is a conceptual reconstruction, not an archival photograph, a likeness of Evelyn Hooker, or a documented UCLA office. Factual claims were checked against the references listed below.

Sources & what we can know

This original account uses the historical paper title only as a source title. It does not invent participant stories or Hooker’s private thoughts, treat projective tests as modern diagnostic standards, or claim that one study alone changed the DSM. Follow @MementoMedicaOfficial for more independently sourced stories from the history of medicine.

  1. American Psychological Association: Being Gay Is Just as Healthy as Being Straight

    APA research summary, rechecked November 11, 2026. Used for sample, tests, blinding, ratings, and careful interpretation.

  2. APA Monitor: The Myth Buster

    APA historical biography, rechecked November 11, 2026. Used for Hooker’s relationship with Sam From, grant, social context, and influence.

  3. PubMed: The Adjustment of the Male Overt Homosexual

    Original 1957 paper record, rechecked November 11, 2026. Used for title, publication, and primary-study scope.

  4. Taylor & Francis: The Adjustment of the Male Overt Homosexual

    Original article record and DOI, rechecked November 11, 2026. Used to verify publication details.

  5. NCBI Bookshelf: Lesbian, Gay, Bisexual, and Transgender Health

    National Academies context, rechecked November 11, 2026. Used for research history, sampling limits, and diagnostic change.

  6. APA Foundation: History of DSM and Homosexuality

    APA Foundation historical chronology, read November 11, 2026. Used to distinguish the 1973 DSM-II revision from later removal of narrower categories.