Framing: The Asian Disease Problem

The problem is presented as a public health emergency. An unusual disease is expected to kill six hundred people, and two programs have been proposed. Program A will save two hundred people. Program B has a one-third probability of saving all six hundred and a two-thirds probability of saving no one.

Seventy-two percent of respondents chose the certain option, Program A.

A second group received the identical problem with the programs described differently. Under Program C, four hundred people will die. Under Program D, there is a one-third probability that nobody will die and a two-thirds probability that six hundred will die.

Seventy-eight percent chose the gamble, Program D.

A and C are the same program. B and D are the same program. Nothing changed except whether the outcome was counted in lives saved or lives lost, and the majority preference reversed.

Tversky and Kahneman published this in 1981, and it is the cleanest demonstration in the literature of a claim that sounds abstract until you see it done: preferences are not retrieved from storage, they are constructed at the moment of choice, and they are constructed against whatever reference point the description supplies. The saved frame puts the respondent in the domain of gains, where prospect theory predicts risk aversion, so the certainty of two hundred saved is attractive. The die frame puts them in the domain of losses, where the value function is steeper and people become risk-seeking, so the gamble that might avoid all four hundred deaths is attractive. Same facts, opposite reference point, opposite behavior.

The immediate objection is that this is a laboratory vignette with undergraduates. It is not confined to the laboratory. A Cochrane review of framing in health communication finds attribute and goal framing effects in real clinical messaging. Surgeons and patients respond differently to a treatment described as having a ninety percent survival rate than to the same treatment described as having a ten percent mortality rate. Tax policy, criminal justice statistics, employment figures and military casualty reporting are all routinely available in both frames, and the choice between them is rarely accidental.

What makes framing the most important mechanism in this part of the guide is that it does not require a single false statement. Everything said in both versions of the disease problem is true. There is nothing to fact-check. This makes framing invisible to the ordinary defenses people apply to persuasion, which are almost all oriented toward detecting falsehood — and it makes control of the frame far more valuable than winning the argument inside it. The argument is conducted on terrain the frame has already shaped.

This is why the political consulting industry, examined later in this guide through Frank Luntz's work, is largely a vocabulary industry. Estate tax and death tax refer to the same levy. Undocumented immigrant and illegal alien denote the same person. Regulation and red tape, pro-choice and pro-abortion, enhanced interrogation and torture. In each pair the fact is constant and the reference point is not.

The countermeasure is mechanical and takes about ten seconds. When a choice is put to you in one frame, write the other one. State the same numbers as losses if they were given as gains, as percentages if they were given as counts, as absolute figures if they were given as relative ones. If your preference flips, your preference was not about the world. It was about the sentence.