Skinner’s superstitious pigeons, and the replication that reinterpreted them. · published 12 September 2026
Skinner’s 1948 ’superstitious pigeon’ study is quoted as evidence for how human superstition works, but a 1971 replication by Staddon and Simmelhag recorded the whole interval instead of just the moment of delivery and found the birds’ movements converged rather than staying idiosyncratic — interim behaviour early in the gap, pecking as delivery approached — which looks less like accident amplified into ritual and more like a food-anticipation sequence pigeons already had; nobody disputes the birds did those movements, the argument is about what they meant, and it is still live.
These are the working notes, unedited. Everything the video
states as fact comes from here, and nothing else may be stated as fact.
The sections on limits and on what the finding does not show are
part of the brief, not an afterthought — see
the method.
The study: B. F. Skinner, “’Superstition’ in the Pigeon”, Journal of Experimental Psychology, 1948. Eight pigeons, each hungry, each in a box where a hopper of grain was presented for a few seconds every fifteen seconds. The delivery was on a timer. Nothing the bird did made it come, and nothing it did could delay it.
What he saw. The birds developed stereotyped, idiosyncratic movements. One turned counter-clockwise around the cage, two or three turns between deliveries. One repeatedly thrust its head into an upper corner. One made a tossing motion, as if lifting its head under an invisible bar. Two developed a pendulum swing of the head and body from right to left. In Skinner’s words, in six of the eight cases the response was defined clearly enough that two observers counted instances in perfect agreement.
His explanation: adventitious reinforcement. Whatever the bird happened to be doing when the food arrived was strengthened by the coincidence, which made it likelier to be happening at the next delivery, which strengthened it again. A behaviour bootstrapped entirely out of accident.
The leap that made it famous. Skinner drew an analogy to human superstition — the ritual before the throw, the lucky shirt — and it is the analogy, rather than the pigeons, that carried the study into every textbook. The analogy was argued in the paper’s discussion. It was not something the experiment measured.
The reinterpretation. Staddon and Simmelhag (Psychological Review, 1971) re-ran the procedure with more systematic observation, recording what the birds did across the whole interval rather than at the moment of delivery. They found the behaviour was not idiosyncratic in the way Skinner described: across birds it converged, and it divided by where it fell in the interval. Early in the gap the birds did one sort of thing, which Staddon and Simmelhag called interim behaviour; as the delivery approached they converged on pecking, the terminal behaviour. That pattern looks less like an accident being amplified and more like a food-anticipation sequence the species already has. Timberlake and Lucas (1985) later made a related argument from species-typical foraging behaviour.
Limits
Eight birds in 1948, and the disagreement is about interpretation rather than about what the animals did — nobody disputes that the pigeons performed those movements. The reinterpretation is itself contested, and adventitious reinforcement has not been abandoned as a concept; it is still studied, and the argument about how much of the superstition effect it explains is live rather than closed.
What this does not show
Not that Skinner fabricated anything, and not that the birds were trained to do these things — they were not. Nor does it mean humans are not superstitious, or that coincidence never shapes behaviour; it plainly does. The narrower point is the one worth making: the pigeon study is quoted as evidence for how human superstition works, and it was never evidence for that. It was an experiment about pigeons on a timer, and a paragraph of speculation that travelled further than the data.