The Little Albert Experiment

1. Historical and Intellectual Context

The Little Albert demonstration arrived at a moment of self-conscious reinvention for American psychology. In 1913 Watson had published "Psychology as the Behaviorist Views It" in the Psychological Review, the so-called behaviorist manifesto. The piece argued that psychology should abandon introspection and the study of consciousness and reconstitute itself as a purely objective natural science of behavior — predicting and controlling what organisms do rather than describing what they report. To make that program credible, behaviorism needed an account of how complex human behavior could be built up from simple, observable processes.

Ivan Pavlov's research on conditioned salivation, becoming widely available in English through Bekhterev and others during the 1910s, offered exactly the kind of mechanism Watson needed. If a neutral stimulus paired with food could come to elicit salivation in a dog, perhaps a neutral object paired with a noxious event could come to elicit fear in a child. Pavlov was studying digestion. Watson wanted to use the same logic to claim ownership of emotion, personality, and ultimately the whole of human development.

Watson's broader theoretical commitment was a radical environmentalism. He had argued — most famously in a later 1924 statement — that, given a dozen healthy infants and the right environment, he could train any one of them to become any kind of specialist, regardless of heredity. Demonstrating that even a strong, seemingly innate emotion like fear could be wired to an arbitrary object would be a powerful prop for that worldview. The Albert study was designed to be that demonstration.

The early twentieth century was also a period when psychological work on infants and children was poorly regulated by any modern standard. Hospitals and orphanages routinely allowed researchers to test children housed in their wards. Watson worked at a teaching hospital, and the children in residence at the Harriet Lane Home were available for study. The institutional ease of access shaped the project as much as any scientific question.

2. Research Questions

The Watson and Rayner paper sets out three connected questions. First, can an emotional response — specifically a fear response — be conditioned in a human infant to a previously neutral object by pairing that object with a stimulus that already elicits fear? Second, will the conditioned response generalize to other objects that share physical features with the original conditioned stimulus? Third, how persistent is the conditioned reaction over time, and to what extent does it transfer across different environments?

Underneath these formal questions lay a programmatic claim. Watson believed that the inventory of innate human emotional reactions was very short — essentially fear, rage, and love — and that adult emotional life was built almost entirely out of conditioned attachments of these few reactions to a large variety of objects and situations. The Albert work was meant to provide an empirical exemplar of the mechanism, with the implicit suggestion that an indefinite number of phobias, dislikes, attractions, and emotional habits could be explained in the same way.

There was also an unstated practical question. If fear could be installed by conditioning, then in principle it could be removed by an opposite procedure. Watson did not perform that removal himself with Albert, but his student Mary Cover Jones would take up the implication a few years later in her work with a child named Peter, often described as the first piece of behavior therapy.

3. Method and Procedure

Baseline Testing

The experimenters first conducted what they called emotional tests. They presented Albert with a number of objects — a white rat, a rabbit, a dog, a monkey, masks with and without hair, cotton wool, burning newspapers — and observed his reactions. The published paper describes him as exceptionally placid: he reached for the animals, manipulated them, showed no fear of any of them. The baseline was the foundation of the later claim that fear of these objects was learned in the laboratory rather than imported.

The experimenters also tested what would serve as the unconditioned stimulus. They struck a four-foot steel bar with a claw hammer behind Albert's head. The sudden, loud crash startled him and on repeated presentations produced crying and avoidance. That reaction — produced reliably by a loud, unexpected noise — was the raw fear response they planned to attach to the rat.

Conditioning Trials

At about eleven months of age, Albert was returned to the laboratory for conditioning. The rat was presented and, as he reached toward it, the bar was struck behind his head. He jumped violently and buried his face in the mattress on which he was seated. A second pairing produced a similar reaction. The researchers then waited a week before resuming, to let the effects consolidate and avoid producing a chronically distressed child for follow-up testing.

On the next session, the rat was presented alone. Albert watched it cautiously and reached only tentatively. Five more paired trials of rat plus noise followed across the session. By the end he was crying at the appearance of the rat alone, turning away, and attempting to crawl off the mat. The total number of pairings was small — roughly seven — but the published account treats the conditioned reaction as established.

Generalization Probes

Five days later the experimenters tested generalization. They presented Albert with the rat, a rabbit, a dog, a sealskin coat, a package of cotton wool, the heads of the experimenters (to test whether human hair would elicit fear), and a Santa Claus mask with a white beard. The rabbit reportedly produced a strong fear reaction without any further conditioning, the dog and the fur coat produced milder reactions, and the cotton and Santa beard produced ambiguous responses. The pattern formed the basis for the claim that the conditioned fear had spread along a similarity gradient defined by furry, white-ish texture.

Environmental and Temporal Tests

Watson and Rayner also tested whether the response would transfer from the small laboratory room to a larger lecture-room setting. They reported that fear of the rat and rabbit persisted but was weaker. Roughly a month after the last conditioning trial, they conducted a final follow-up. The fear was still present but had partially faded, requiring what they called "freshening up" trials in which the rat was again paired with the noise. Shortly after that follow-up Albert left the hospital with his mother.

4. Participants and Setting

"Albert B."

The published paper describes Albert B. as the son of a wet nurse who lived and worked at the Harriet Lane Home. He was selected, the authors said, because he was healthy, unusually placid, and rarely cried. He was also conveniently available; the experimenters could test him with little disruption and no need to recruit from outside the hospital.

The Search for Albert's True Identity

For most of the twentieth century, "Albert B." remained a pseudonym whose real bearer was unknown. Beginning in 2009, Hall P. Beck and colleagues published an investigation arguing that Albert was Douglas Merritte, the son of a wet nurse named Arvilla Merritte who worked at Johns Hopkins. Beck, Levinson and Irons used hospital records, census documents, and photographs to support the identification. Douglas Merritte died at age six in 1925 from acquired hydrocephalus. Beck's group later argued, controversially, that Douglas was already neurologically impaired at the time of the experiment, casting doubt on Watson's claim that Albert was a healthy infant.

In 2014 Russell Powell, Nancy Digdon and colleagues challenged the Merritte identification on multiple grounds — age fit, description of the mother, and the neurological argument — and proposed an alternative candidate, William Albert Barger, whose family called him Albert and whose mother had also been a wet nurse at the hospital. Barger, who lived until 2007, was reported by relatives to have had a lifelong aversion to dogs and other furry animals, although his family was unaware of the experiment during his lifetime.

Both identifications have been contested in subsequent scholarship, and there is at present no fully decisive resolution. The episode is itself a lesson in how thinly archival evidence can support large historical claims, and how the desire to humanize an iconic subject can outrun the data.

The Setting

The work was carried out in the Phipps Psychiatric Clinic and the Harriet Lane Home, both attached to Johns Hopkins. Watson held the chair of experimental psychology at the university; Rayner was a graduate student working with him. Their professional relationship became personal, and a scandal over Watson's affair with Rayner led to his dismissal from Hopkins in 1920 — the same year the Albert paper appeared. Watson moved into advertising, while Rayner continued to write with him on child rearing through the 1920s.

5. Results

Acquisition

The headline result was acquisition of a conditioned fear. After roughly seven pairings of the rat with the steel-bar noise spread over two sessions, Albert reportedly cried, turned away, and tried to crawl from the rat when it was presented alone. The reaction was vivid enough to be photographed and filmed, and Watson would later show fragments of the film at lectures.

Generalization — Partial, Not Complete

The pattern of generalization was more uneven than textbook summaries suggest. The rabbit produced the strongest spontaneous fear among the test objects, with crying and active withdrawal. The fur coat and the dog produced milder reactions; the dog was at one point described as eliciting only momentary hesitation. The cotton wool and the Santa beard produced reactions described as ambiguous or weak. Watson's own notes suggest that he sometimes "refreshed" the conditioning with the noise during generalization testing, which complicates the interpretation of spontaneous transfer.

Persistence

The fear persisted across days and across the move from the small testing room to the larger lecture room, with some attenuation. At the one-month follow-up the response was weaker but still present without further pairings, suggesting that the conditioning had produced more than a transient startle.

What Did Not Happen

Importantly, the paper does not show what later summaries often imply: that any furry object reliably triggered violent fear, that the conditioning was clean and complete, or that the fear was removed. Several test objects produced little or no clear reaction, the conditioning required multiple sessions, and no deconditioning was ever performed.

6. The Researchers' Interpretation

Watson and Rayner read their data as a clean demonstration that the Pavlovian mechanism could be extended from canine salivation to human emotion. They presented the rat-noise pairings as analogous to bell-food pairings, the conditioned fear as analogous to the conditioned salivation, and the spread to other furry objects as a textbook case of stimulus generalization. The fact that the response varied in strength across test objects was treated as confirmation of generalization gradients rather than as a complication.

The broader interpretation was sweeping. Watson argued in the paper, and more aggressively in subsequent writings, that adult phobias and emotional preferences could be understood as the residue of similar conditioning episodes in early childhood. A fear of dogs in an adult might be the long shadow of a single frightening incident in infancy, with the original event long forgotten but the conditioned attachment intact. The Albert study was offered as the laboratory analog of that real-world process.

Watson and Rayner also speculated about clinical implications. They suggested, with surprising clinical caution, that fears acquired this way might be removed by procedures of reconditioning — pairing the feared object with a positive stimulus such as food, or gradually approaching the feared object in safe contexts. They were not in a position to test the suggestion with Albert because he had left the hospital. Mary Cover Jones would test it shortly afterward.

Finally, the authors used the Albert results to needle psychoanalytic interpretation. They speculated mockingly that a Freudian analyst meeting an adult version of Albert with a fear of fur coats would search for symbolic-sexual meaning, oblivious to the fact that the fear had a banal conditioning history. The Albert paper became, in this sense, a polemic in the early-twentieth-century turf war between behaviorism and psychoanalysis.

7. Modern Reanalyses and Criticisms

Ben Harris and the Textbook Distortions

The most influential modern reassessment is Ben Harris's 1979 American Psychologist paper "Whatever Happened to Little Albert?" Harris examined the original paper alongside the surviving film and the way the study had been described in introductory textbooks. He found a striking pattern of distortion. Textbooks routinely overstated the number of conditioning trials, exaggerated the cleanliness of the generalization data, omitted the "refreshening" of conditioning during generalization testing, and changed details of the unconditioned stimulus and the test objects. In some textbooks Albert had been conditioned to fear a teddy bear or a Santa beard rather than a rat; the rabbit and dog were sometimes left out entirely.

Harris argued that the Albert study had become something closer to a parable than a piece of empirical evidence. The parable served the pedagogical goal of teaching classical conditioning quickly, but it misrepresented what Watson and Rayner had actually shown.

Methodological Limitations

Modern reviewers have catalogued substantial methodological problems. There was only one participant, and his behavior was scored by the experimenters themselves with no blinded observer. The number of pairings was small. The pre-existing temperament of the child — described as unusually placid — limited generalization to fearful infants. The film record is fragmentary and was edited. The conditioning was partially "refreshed" during what was reported as generalization testing. There was no control infant matched on demographics who was exposed to the rat without the noise. There was no formal stimulus-discrimination training.

The Health-of-Albert Debate

Beck and colleagues' identification of Albert as Douglas Merritte triggered a secondary controversy: the claim that Merritte showed signs of neurological impairment in the surviving film. If true, this would mean Watson and Rayner had violated their own selection criterion (a healthy, normal infant) and might have produced a film whose emotional reactions reflected developmental pathology rather than ordinary conditioning. Powell, Digdon and others have disputed both the identification and the impairment claim. The debate is unresolved, but it has sharpened attention to how thin the demographic information in the original paper actually is.

Replication Failures and Constraints

Attempts to replicate Albert-style fear conditioning in infants have been rare for obvious ethical reasons, and the rare attempts that have been made show that human fear conditioning is more variable and constrained than the Albert paper implies. Generalization in particular is influenced by features such as the infant's temperament, the contingency of the pairings, and the specific stimulus dimensions involved. The clean narrative of "any neutral object becomes a phobia" does not survive contact with later evidence.

8. Ethical Considerations

By contemporary standards the Little Albert experiment would be unapprovable. There is no record of formal informed consent from Albert's mother. Watson and Rayner deliberately produced distress in an infant, repeatedly, for the purpose of demonstrating that distress could be produced. They planned to remove the conditioned fear but did not do so. They published photographs and filmed images that identified the child within the institutional community even if not by full name.

Defenders of the work sometimes argue that research ethics in 1920 had no formal codes for child research and that the standards of the day were not violated. This is partly accurate as a historical statement, but it understates the contemporaneous unease. Even within Watson's own career, the lack of deconditioning was noted as a regrettable gap, and the absence of any concern for Albert as an individual outside the experiment is striking in retrospect.

The Albert case became a touchstone in the post-war development of ethics codes for human subjects research. The American Psychological Association's first ethics code, published in 1953, and later federal regulations on research with minors, were partly motivated by an accumulated awareness that studies like Albert — and Stanley Milgram's later obedience work — had treated participants as means rather than as persons owed protection.

There is a quieter ethical dimension as well. Watson's affair with Rayner, his dismissal, and his later career in advertising mean the Albert study sits inside a personal story marked by professional damage. The science cannot be cleanly extracted from that context, and modern accounts increasingly acknowledge the moral weight of the silence that Albert and his mother lived inside afterward.

9. Influence on Psychology

The Behaviorist Programme

Despite — or because of — its flaws, the Albert study became a foundational reference for behaviorism. It provided a vivid, photographable, human example of conditioning, which dry rodent studies could not match. The image of a child crying at a white rat shaped the public face of the field for decades. B. F. Skinner did not need it for his operant work, but radical behaviorism inherited the Albert story as part of its origin myth.

Mary Cover Jones and Peter

The most important direct descendant of the Albert work was Mary Cover Jones's 1924 study of a three-year-old boy called Peter, who had a strong fear of rabbits. Working under Watson's mentorship at Columbia, Jones used a graded exposure procedure: she brought a rabbit progressively closer to Peter while he was eating a favored food, and reinforced calm contact through modeling by other children who were unafraid. Over weeks, Peter's fear faded and he was eventually able to touch and play with the rabbit. Jones's study, sometimes called the first behavior therapy, applied the implication that Watson had not himself tested with Albert: a conditioned fear can be reversed by counterconditioning and graded exposure.

Wolpe and Systematic Desensitization

In the mid-twentieth century, Joseph Wolpe drew on the same logic in developing systematic desensitization, in which a feared stimulus is paired with relaxation along a hierarchy of approximations. Wolpe explicitly traced his approach to the conditioning tradition that Watson and Jones had launched. From systematic desensitization grew the modern family of exposure-based therapies for phobias, panic, post-traumatic stress, and obsessive-compulsive disorder, which remain among the most strongly supported psychological treatments in existence.

Cognitive Behavioral Therapy

Cognitive behavioral therapy synthesizes behavioral techniques with cognitive restructuring, but its behavioral wing — particularly exposure and behavioral activation — is a direct line of descent from the Albert demonstration. When a clinician guides a patient with arachnophobia toward a tarantula on a desensitization hierarchy, the underlying assumption that conditioned fear can be acquired and modified is exactly the assumption the Albert paper sought to make plausible.

Concept Inheritance

Beyond therapy, the Albert study seeded several enduring teaching concepts: conditioned emotional response, stimulus generalization gradients, the distinction between innate and acquired emotional reactions, and the application of animal-learning models to human development. Even where the underlying data are now considered weak, the concepts have proved robust.

10. What the Experiment Means Today

A century on, the Little Albert experiment occupies a peculiar position in the discipline. It is widely cited, almost universally taught, frequently mischaracterized, and ethically embarrassing. It is also, in a recognizable sense, the founding image of behavior therapy — the moment at which the abstract idea that emotional learning could be the subject of natural science was given a face.

Reading the original paper today is sobering. The text is short, the methodological controls are thin, and the conclusions extend far beyond what the data could support. At the same time, the underlying idea — that a previously neutral stimulus can come to elicit an emotional reaction through its association with a strongly emotional event, and that such reactions can be modified by counterconditioning — has held up well across an enormous later literature, including modern fear-conditioning work in rodents and humans using skin-conductance and neuroimaging measures.

For students, the Albert story now teaches two things at once. It teaches the basics of classical conditioning, generalization, and the link between Pavlov's laboratory and the modern clinic. And it teaches the limits of those basics: how a small, methodologically loose study can be inflated into a textbook icon, how identification of historical participants is harder than it looks, and how the ethical conventions of research must catch up — and have, slowly, caught up — with what science is technically capable of doing to a child.

The most honest summary may be this. The Albert experiment is not the clean proof Watson wanted it to be. It is a flawed first sketch of a real phenomenon. The phenomenon — that emotional reactions can be learned and unlearned through association — has turned out to be one of the most clinically useful ideas psychology ever produced. The sketch should be remembered as a sketch, with all of its smudges, including a small child who was never deconditioned and who left the hospital with whatever residue the laboratory had created.