Classic Studies: Watson & Rayner and Bandura

Every topic in Edexcel A-Level Psychology has two classic studies prescribed by name, and for Learning Theories the two are Watson and Rayner (1920), the "Little Albert" study, and Bandura, Ross and Ross (1961), the "Bobo doll" study. These are not merely illustrative examples to be mentioned in passing; they are examinable in their own right, which means you must know each one in the classic-study framework — its aim, method and procedure, its results, and the conclusion the authors drew — and you must be able to evaluate each on methodological and ethical grounds. This lesson treats both studies at that depth and then does the thing the specification most rewards: it compares them. The two were chosen deliberately, because between them they demonstrate the two great mechanisms of the learning approach — Watson and Rayner conditioned an emotional response by classical conditioning, while Bandura, Ross and Ross produced a behaviour by observational (social) learning. Studying them side by side crystallises the difference between learning by association and learning by watching, and shows how psychology moved from a purely behaviourist model to a cognitive-behavioural one across the four decades that separate them.

---

Spec Mapping

This lesson addresses the Edexcel 9PS0 — Paper 1, Topic 4: Learning Theories requirement to study the topic's two classic studies in full: Watson and Rayner (1920) and Bandura, Ross and Ross (1961). You are required to know the aim, procedure/method, results and conclusions of each study, and to evaluate each in terms of its methodology (control, validity, reliability, sampling, generalisability) and its ethics (protection from harm, consent, the treatment of participants). Because Edexcel names these two studies together, you should also be able to compare them — what each demonstrates about learning, and how the mechanism, method and findings of one differ from the other. In assessment-objective terms, you should be able to describe each study accurately (AO1), apply the classic-study framework and the topic's conditioning and social-learning concepts to interpret them (AO2), and evaluate their methodological strengths and weaknesses and their ethical status (AO3).

Connects to…

• Classical conditioning and explaining phobias (earlier in this topic): Watson and Rayner is the direct evidence base for the two-process model of phobias — Albert's acquired fear of the white rat is the classical conditioning of an emotional reflex, and his fear of other furry objects is stimulus generalisation, so this study is where the theory of phobia acquisition is grounded in data.
• Social learning theory (earlier in this topic): the Bobo doll study is the empirical foundation of Bandura's theory of modelling, imitation and identification; where the SLT lesson taught the theory, this lesson examines the study in the classic-study framework and evaluates it as evidence.
• The ethics of psychological research and animal studies (later in this topic): both classic studies raise sharp ethical questions — the deliberate creation of fear in an infant, and the deliberate exposure of nursery-age children to an aggressive model — which link forward to the topic's treatment of research ethics and to the BPS Code of Human Research Ethics.

---

The Classic-Study Framework

Before the two studies, it helps to fix the framework Edexcel expects you to use for any classic study. Every study can be reported and evaluated under the same headings, and marking rewards answers that are organised this way rather than as an undifferentiated narrative.

Element	What it means	What to look for
Aim	The purpose — what the researchers set out to find	A precise statement of the question or hypothesis
Method	The overall research strategy	Experiment, observation, case study; design (independent/repeated); controls
Procedure	Exactly what was done, step by step	Sampling, apparatus, sequence of conditions, measures
Results	The findings, quantitative and qualitative	Key data, comparisons between conditions
Conclusion	The interpretation the authors drew	The theoretical claim the results are said to support
Evaluation	The critical appraisal	Methodology (validity, reliability, sampling, generalisability) and ethics

A recurring exam skill is to keep results (what was observed) separate from conclusions (what it is taken to mean): a great many marks are lost by students who report an interpretation as if it were a finding. With that framework in place, we turn to the first study.

---

Study 1 — Watson and Rayner (1920): "Little Albert"

Aim

John B. Watson, the founder of behaviourism, and his research assistant Rosalie Rayner, set out to demonstrate that a fear response could be conditioned in a human infant by classical conditioning — that is, that an emotional reaction to a previously neutral stimulus could be learned by association, exactly as Pavlov had conditioned salivation in dogs. Watson wanted to show that emotions were not mysterious inner states but conditioned reflexes, and to challenge the psychoanalytic assumption that fears reflected unconscious conflict. A secondary aim was to see whether such a conditioned fear would generalise to similar stimuli and whether it would persist over time.

Method and Procedure

The study is best described as a controlled observation conducted in the form of a single-participant (case study) experiment. The participant was an infant known as "Albert B.", around 11 months old at the start of conditioning, the son of a wet-nurse at the hospital where Watson worked. Watson and Rayner first established that Albert was a calm, "stolid and unemotional" baby who showed no fear of a range of stimuli.

1. Baseline (pre-conditioning) tests. Albert was presented with a white rat, a rabbit, a dog, a monkey, cotton wool, burning newspapers and masks. He showed no fear of any of them, reaching out towards the rat with curiosity. Separately, the researchers established the one thing that did reliably frighten him: a loud noise made by striking a steel bar with a hammer behind his head produced crying and distress. In conditioning terms, the loud noise was the unconditioned stimulus (UCS) and the fear/crying it produced was the unconditioned response (UCR); the white rat was the neutral stimulus (NS).
2. Conditioning trials. When Albert was about 11 months old, the researchers presented the white rat and, just as he reached to touch it, struck the steel bar behind his head. This pairing of rat and noise was repeated across two sessions a week apart — approximately seven pairings in total.
3. Testing the conditioned response. The rat was then presented alone, without the noise. Albert now reacted with fear — he cried, turned away and tried to crawl off the table. The white rat had become a conditioned stimulus (CS), and the fear it now produced was a conditioned response (CR).
4. Testing for generalisation. Albert was presented with other stimuli that shared features with the rat. He showed fear of a rabbit, a dog, a fur coat, cotton wool, a Santa Claus mask with a white beard, and even Watson's own hair — demonstrating stimulus generalisation to white, furry objects.
5. Testing for persistence. After a gap of a little over a month, and after Albert had not been exposed to the stimuli in the interim, he was tested again. The conditioned fear had weakened somewhat but was still present, showing that the learned response endured over time.

The sequence is summarised below.

graph TD
    subgraph BEFORE[Before conditioning]
        A1[White rat = NS] --> A2[Curiosity, no fear]
        A3[Loud noise = UCS] --> A4[Fear/crying = UCR]
    end
    subgraph DURING[During conditioning ~7 pairings]
        B1[Rat + Loud noise<br/>repeatedly paired] --> B2[Fear/crying = UCR]
    end
    subgraph AFTER[After conditioning]
        C1[White rat alone = CS] --> C2[Fear/crying = CR]
        C3[Rabbit, dog, fur, mask] --> C4[Fear = generalisation]
    end
    BEFORE --> DURING --> AFTER
    style A2 fill:#27ae60,color:#fff
    style C2 fill:#e74c3c,color:#fff
    style C4 fill:#e67e22,color:#fff

Results

The findings were clear-cut. Before conditioning, Albert showed no fear of the white rat and interacted with it happily. After the pairings of rat and noise, the rat alone reliably produced fear. That fear generalised to a range of other white, furry stimuli, and it persisted over a period of more than a month, albeit in a somewhat diminished form. In the language of the topic: a neutral stimulus had been turned into a conditioned stimulus for fear, the response had generalised, and it had not fully extinguished over the delay.

Conclusion

Watson and Rayner concluded that an emotional response — specifically fear — can be conditioned in a human being through classical conditioning, in exactly the way Pavlov had conditioned a physiological reflex in dogs. They argued that many, perhaps most, human fears and phobias are learned through association rather than being innate or the product of unconscious conflict, and they took the study to support a thoroughgoing behaviourist account of emotion. The study became the empirical cornerstone of the learning explanation of phobias and, more broadly, of Watson's claim that experience — not inner life — shapes behaviour.

---

Evaluating Watson and Rayner (1920)

Methodological Evaluation

Consideration	Point
Control	Conducted under controlled conditions with a clear baseline, standardised UCS (the struck bar) and systematic testing, giving reasonable internal validity — the fear plausibly was produced by the conditioning.
Objectivity of measurement	The dependent variable (fear) was recorded by observation of crying and avoidance, some of it filmed; but coding of "fear" was subjective and there was no independent inter-rater check reported, weakening measurement reliability.
Sample	A single infant. This makes generalisation hazardous: Albert may have been unusually fearful or unusually placid (Watson described him as unemotional), and no individual differences can be assessed from an n of 1.
Ecological validity / mundane realism	The conditioning was contrived and repeatedly reinforced under laboratory conditions; everyday phobia acquisition may differ, and some later attempts to replicate infant fear conditioning had mixed success.
Lack of extinction phase	Albert was removed from the hospital before Watson could attempt to decondition the fear, so the study leaves the induced fear in place — a methodological loose end as well as an ethical one.

A key strength is the study's demonstration that a full classical-conditioning sequence — acquisition, generalisation and persistence — could be produced in a human under controlled conditions, which is why it remains a foundational demonstration. The dominant weakness is the single-participant design: causal confidence about this infant is reasonable, but the leap to "human fears in general are conditioned" rests on one case.

Ethical Evaluation

The study is one of the most frequently cited examples of ethically unacceptable research, and evaluating it well means being specific rather than merely disapproving.

• Protection from harm. The researchers deliberately induced distress and fear in an infant — the central ethical objection. Modern codes require that participants leave a study in no worse a state than they entered it; Albert plainly did not.
• Failure to decondition. Because Albert left the hospital, the conditioned fear was never removed. Watson and Rayner had discussed methods for removing it but did not carry them out, so the study left a child with an induced fear (later scholarship has debated Albert's identity and fate, but the point of principle stands).
• Consent and the vulnerability of the participant. An 11-month-old cannot assent, so the ethics turn on parental consent and on whether it was adequately informed; historical accounts question how fully the mother understood what would be done.
• Historical context. It is fair to note that the study predates the BPS and APA ethical codes, so Watson and Rayner did not breach guidelines that did not yet exist. That contextualises but does not excuse the harm: the study is retained in the curriculum partly because it illustrates why such codes became necessary.

---

Study 2 — Bandura, Ross and Ross (1961): "Bobo Doll"

Aim

Albert Bandura, with Dorothea Ross and Sheila Ross, aimed to test whether aggression could be learned through observation and imitation of an adult model — the central claim of what would become social learning theory. Specifically, they set out to test four hypotheses: (1) children who observed an aggressive model would imitate that aggression, even in the model's absence; (2) children who observed a non-aggressive model, or no model, would show little aggression; (3) children would be more likely to imitate a same-sex model; and (4) boys would be more aggressive than girls, especially in physical aggression. The study thus tested both whether observational learning of aggression occurs and what moderates it.

Method and Procedure

The study was a laboratory experiment with an independent-measures design and multiple independent variables (model behaviour: aggressive/non-aggressive; model sex; child sex). The sample was 72 children (36 boys, 36 girls) aged 3 to 6 years (mean age about 4 years 4 months) from the Stanford University nursery. Crucially, the children were pre-tested for aggression by nursery staff and the experimenter and matched on baseline aggressiveness before being allocated to conditions, so that the groups were equivalent to begin with.

1. Allocation. Children were divided into an aggressive-model condition, a non-aggressive-model condition and a control condition (no model). Within the model conditions they were further split by whether the model was male or female and matched for baseline aggression.
2. Modelling phase. Each child was brought individually into a room and seated at a table with attractive toys (potato prints, stickers). In a corner of the room were a mallet, a Bobo doll (an inflatable, weighted clown that rights itself when struck) and other toys. In the aggressive condition, the adult model, after a minute of assembling toys, turned to the Bobo doll and attacked it in distinctive, scripted ways: laying it on its side and punching it, striking it on the head with the mallet, throwing it in the air, kicking it, and uttering specific verbal aggression — phrases such as "Sock him in the nose", "Hit him down", "Pow", "Kick him". In the non-aggressive condition, the model played quietly with the other toys and ignored the Bobo doll.
3. Aggression-arousal phase. Each child was then taken to a second room with attractive toys but, after beginning to play, was told the toys were being kept for other children — a mild frustration deliberately induced so that any subsequent aggression could not be attributed simply to prior contentment, and to ensure a baseline level of arousal across conditions.
4. Test for delayed imitation. The child was taken to a third room containing both aggressive toys (a Bobo doll, a mallet, a dart gun) and non-aggressive toys (a tea set, crayons, dolls, a ball). The child was observed for 20 minutes through a one-way mirror, with behaviour recorded at 5-second intervals by observers using predefined behavioural categories: imitative physical aggression, imitative verbal aggression, imitation of non-aggressive verbal responses, partially imitative behaviours (e.g. sitting on the doll, using the mallet on other toys), and non-imitative aggression (novel aggressive acts not modelled). Inter-observer reliability was checked, with high agreement reported.

Results

The results supported the observational-learning hypothesis strongly.