It seemed obvious: whenever a person makes a decision, they instinctively compare the versions of future worlds that would result from a given decision.
As a graduate student at Harvard, John L. Loeb Professor of Statistics Donald Rubin set about formalizing this deceptively simple idea. Although it was not an entirely novel concept, it was Rubin who provided the statistical mathematical notation for it. In 1986, Rubin published his “Rubin Method” and, in the process, made statistical history.
The Rubin Method is considered a hugely important discovery in the field of statistics and beyond. The method centers on the proposition that an individual person is a cause and effect unit in and of himself. Rubin uses “The Aspirin Example” to explain his theory: If you have a headache and are wondering whether or not to take an aspirin, you are implicitly comparing two worlds, not at the moment you take the aspirin, but two hours from that point. This decision has nothing to do with the change in time, but everything to do with the comparison of two potential outcomes at certain points in the future.
Rubin’s theoretical approach to statistics is unique in its rejection of the abstract. “I like working on theory and conceptual developments in statistics and I like seeing those ideas applied to actual problems,” Rubin says.
Over the last ten to fifteen years, Rubin has continued to be an innovator. The “Rubin Causal Model” gives people who were confused about how to think about effects a framework in which to consider them. Elizabeth Stuart, a graduate student under Rubin and a Teaching Fellow for his Quantative Reasoning 33 core, “Causal Inference,” attests to the flexibility of Rubin’s work. “Now [Rubin’s Model] is the standard for thinking about causal effects in economics.”
Rubin’s method has been applied to numerous fields, including economics, law, education and medicine.
