Eugene Borgida

Eugene Borgida was a student of the social psychologist Richard Nisbett at the University of Michigan and a co-investigator in a study that became one of the book’s clearest demonstrations of how people reason about individuals. His research with Nisbett showed that learners may hear a compelling statistical fact without changing their expectations, yet rapidly revise those expectations when confronted with a surprising concrete case. The study therefore made Borgida important not as a biographical protagonist, but as part of an experimental partnership that turned an abstract claim about base rates into an observed failure of judgment.

Identity and Research Context

Borgida worked with Richard Nisbett on the consequences of the helping experiment conducted several years earlier at New York University. In that experiment, participants spoke from individual booths over an intercom about their personal lives and problems. Six people belonged to each group, but one was a stooge whose scripted account of seizure symptoms culminated in an apparent emergency and a request for help. Only four of fifteen participants rushed to help after the first request, producing a base rate of 27 percent for immediate assistance.

The experiment gave Borgida and Nisbett a way to separate general knowledge from case-specific impression. Students were asked to predict how two individuals would behave after watching videos of them. The videos were deliberately uninformative: they offered no reliable reason to think that either person would be more or less helpful than a randomly selected student. Bayesian reasoning therefore required the students to begin with the 27 percent base rate and remain close to it in the absence of useful new evidence.

The Prediction Study

Borgida and Nisbett divided students into groups with different information. One group learned only the procedure of the helping experiment and not its results. Those students predicted that both people in the videos would immediately rush to aid the stricken stranger, a response consistent with their ordinary beliefs about human nature and with the apparent kindness of the individuals they saw.

A second group knew both the procedure and the statistical outcome. The crucial finding was that this knowledge changed nothing. Their predictions were indistinguishable from those of students who had never been told that only four of fifteen participants had helped. They possessed the relevant base rate, but the filmed individuals remained psychologically exempt from it. The study thus exposed a gap between accepting a statistical result and using it to judge a particular person.

The result also showed why statistical base rates are often weaker in practical reasoning than causal information. People can use a base rate more readily when it supplies a causal story about the individual case, such as a dangerous trait in a reckless cabdriver or an unusually difficult examination affecting a student’s performance. The helping statistic, by contrast, was treated as a general fact about a population, and the students’ vivid impressions of the two individuals overpowered it.

Surprise, Learning, and Generalization

Borgida and Nisbett then reversed the teaching procedure. A new group learned how the helping experiment was conducted but was not told its overall results. After watching the same two videos, the students were simply informed that the two individuals they had seen had not helped the stranger. They were then asked to estimate the global result of the experiment.

This time the students’ guesses were extremely accurate. A surprising individual case immediately altered their understanding of the general situation. The students inferred that helping was more difficult than they had previously believed, even though they had not been given the population statistic directly. The contrast produced the study’s memorable asymmetry: subjects were reluctant to infer a particular case from a general rule, but willing to infer the general rule from a particular case.

Borgida’s contribution belongs to this reversal in the direction of learning. A surprising statistic could be repeated and discussed without changing how students imagined themselves, their friends, or strangers behaving. Two apparently nice people who failed to help created an incongruity that demanded explanation and could be incorporated into a causal story. The concrete case therefore carried more psychological force than the numerical result it represented.

Relationship to the Book’s Larger Argument

The study involving Borgida supports the broader research program on heuristics and biases developed by Daniel Kahneman and Amos Tversky. Their work described intuitive shortcuts such as representativeness, availability, and adjustment from an anchor: economical strategies that are often useful but can generate systematic and predictable errors. The helping experiment adds an educational dimension to that program by asking not only whether people err, but whether they learn from evidence that should correct the error.

Its implications are especially important for psychological teaching. The purpose of learning an experiment is not merely to remember its outcome. Genuine learning occurs when a person’s interpretation of a later situation changes. Borgida and Nisbett’s findings suggested that students may absorb a surprising result as an isolated fact while quietly exempting themselves and familiar people from its implications. Direct encounters with surprising cases are more effective because they force the learner to reconcile observation with an existing belief about human behavior.

The episode also clarifies the distinction between formal Bayesian reasoning and ordinary intuition. The proper starting point for an unidentified participant was the 27 percent helping rate; the videos supplied no diagnostic evidence that justified moving far from that prior. Yet the students’ judgments followed the vividness and apparent character of the individuals rather than the statistical distribution. Borgida’s research thus helped make visible the practical difficulty of applying probability to human cases, even when the relevant probability has already been explained.