The hot-cold empathy gap

The clearest demonstration is a study of smokers published in 2008 by Michael Sayette, George Loewenstein and colleagues. Ninety-eight smokers were asked a deliberately concrete question rather than a rating: what is the minimum amount of money you would accept to delay smoking by five minutes in a session a week from now, when you will have gone twelve hours without a cigarette. One group answered it while already craving, having abstained for twelve hours and while holding a lit cigarette. Another answered having smoked normally that day, holding a roll of tape as a neutral object. Everyone then returned a week later, craving, and named the figure again.

The group who had answered while not craving named a higher figure the second time, rising on average from 8.18 dollars to 9.33. They had underpredicted what access to a cigarette would be worth to them. The group who had answered while already craving showed no such rise, and if anything drifted slightly the other way, from 9.53 to 8.77. The direction is the finding. The size is about a dollar, which is real, statistically significant, and modest, and worth holding in mind before anyone stretches it.

Loewenstein had named the mechanism years earlier, in work on what he called visceral factors: hunger, thirst, pain, fear, anger, exhaustion, desire. These states have two properties. They feel intensely pleasant or unpleasant, and while you are in them they change what looks worth doing. Thirst does not only feel bad, it makes water more valuable than food. Anger does not only feel bad, it makes retaliation look reasonable.

The gap has been mapped in both directions. Cold to hot is the one most people meet: calm, you cannot picture the pull, so you walk into situations unprepared and are then surprised by yourself. Hot to cold is the mirror: while aroused you cannot imagine wanting anything else, which is why decisions made in the heat feel permanent and look strange a day later. Loewenstein also showed it runs between people, not only within one, so someone who is calm systematically underestimates what a state is doing to somebody else's behaviour.

The honest limits. Much of the framework is Loewenstein's own programme, though independent groups in Amsterdam and at Caltech have found the pattern in pain, restraint and neural studies, and it has held across several unrelated domains over nearly thirty years, which is why it grades as holding up rather than debated. What varies is size, and the smokers study is a fair guide to how modest it can be. How large the gap is depends on the state, its intensity, and how far ahead the prediction reaches, so treat the direction as reliable and any specific number as local to its study. The smokers study also carries a design limit its authors state plainly: the craving group was both nicotine deprived and looking at a lit cigarette, and the study was not built to separate those two causes. And the same study's self-reported urge measure, unlike its money measure, did not produce a significant result. The evidence is also concentrated in visceral states, and whether an ordinary work pressure like a deadline or a persuasive room produces the same effect is an extension of the finding rather than a demonstrated part of it.

Every rule your company runs on was written cold. The hiring bar was set in a planning session, calmly, by people imagining a hypothetical candidate. The spending limit was agreed in a quarterly review with no live deal on the table. The policy on discounting was written by someone who was not, at that moment, three days from the end of the quarter with a client saying they need one more concession.

Then all of them get tested hot. The hiring bar meets a candidate the panel liked in the room. The spending limit meets an opportunity that will apparently disappear by Friday. The discount policy meets a deal that closes the number. And the people who break the rule are usually the same people who wrote it, which looks like hypocrisy and is not. The version of them that wrote it had no working model of the version that would face the moment.

Which changes what a rule is for. If breaking rules were a discipline problem, better rules and firmer reminders would fix it, and they mostly do not. What works instead is arranging the decision so the cold version keeps some authority over the hot one. A written floor before the negotiation begins. A scorecard filled in before the candidate walks in. A named person who has to approve the exception, ideally someone not in the room. None of that is about willpower. It is about making sure the calm version of you left something behind that the excited version has to walk past.

There is a personal reading too, and it is the uncomfortable one. The advice you give a friend is cold-state advice, which is precisely why it tends to be better than what you do. You are calm about their situation and hot inside your own. So the gap between what you tell people and what you actually do is not weakness or dishonesty. It is two different states producing two different answers, and only one of them ever gets tested. The move that turns cold judgment into something that survives the moment is implementation intentions. The related failure of mispredicting how a future outcome will feel is affective forecasting.

Source: The smokers study: Sayette, Loewenstein, Griffin and Black, Exploring the Cold-to-Hot Empathy Gap in Smokers, Psychological Science, volume 19, 2008, pages 926 to 932. The framework and the term are Loewenstein's, set out in his 1996 work on visceral influences on behaviour and reviewed in Hot-cold empathy gaps and medical decision making, Health Psychology, volume 24, 2005.

Predictably Irrational

Dan Ariely, 2008

Ariely ran one of the sharpest tests of this gap with Loewenstein, asking people the same questions in a calm state and an aroused one and getting very different answers. The book covers it alongside the other quiet ways our judgment shifts without our noticing.