Reward Insensitivity

What it is

A deliberate action is a means to an outcome. If the outcome is devalued, the action should stop, because there is no longer anything to be gained by it. A habit works differently. It is an association between a cue and a response, laid down by repetition in a stable setting, and once it is in place the cue starts the response without consulting the outcome. Devalue the outcome and a habit carries on. Wood's phrase for this is that "the ghosts of prior rewards stick around".

That makes devaluation the one manipulation that tells the two apart. Asking people does not do it, because a person can say, truthfully, that they no longer enjoy the reward and still go on performing the action. Observing frequency does not do it either, since frequency is high in both cases. The test has to change the reward and watch the behaviour.

The paradigm comes from animal learning research, and Wood describes the standard version in Good Habits, Bad Habits (2019, p. 126). Rats are trained to press a lever for a food pellet and are then given a pellet paired with a toxin. Rats with a hundred presses of training stop pressing. Rats with five hundred keep pressing, and spit out the pellet in disgust while they do it. Wood reports the limit as well as the result: even the heavily trained rats gave up after a few minutes. Insensitivity is a matter of degree and of time, not an absolute.

In effect

A cinema and a bag of stale popcorn

The human study most often used to illustrate the idea was run in a cinema. In two field experiments reported in 2011 by David T. Neal, Wendy Wood and two colleagues, habitual popcorn eaters at a cinema were minimally influenced by their hunger or by how much they liked the food, and they ate equal amounts of stale and fresh popcorn. Two conditions mattered. They ate out of habit only when they were in the cinema, the setting tied to the past behaviour, and only when eating with their dominant hand. People with weaker popcorn habits ate because of motivations such as liking the popcorn.

Wood reports the figures in her book, and they are worth holding loosely: people without the habit ate about 70 per cent of a fresh bag and about 40 per cent of a stale one, while habitual eaters ate more than 60 per cent of either. Asked afterwards, the habitual eaters said they disliked the stale popcorn. Liking was intact and behaviour ignored it. When the habitual eaters were made to eat with their non-dominant hand, the discrimination returned, which is the second arm of the study and the reason it is more than a demonstration.

Wood corrects the popular reading of that second arm herself. Magazines reported it as a diet trick. She writes that eating with the non-dominant hand is "not a dieting technique", only a way to derail a habit.

Visits per month

Nir Eyal's Hooked defines habit twice. On page 1 it takes the cognitive psychologists' version, behaviours "triggered by situational cues" and done "with little or no conscious thought". On page 33 it offers another, that a habit is when not doing an action causes a bit of discomfort. The book's case studies measure success by return use, in the case of the workout app Fitbod about nine visits a month per user, a figure the founders supplied. That figure cannot tell a repeated choice from a habit. The book reports no devaluation test. What one would look like is a change to what the product pays, with use observed afterwards, and this site has not found one run on a consumer product.

What it does not say

It does not say that reward plays no part in habits. Rewards are how the association gets built. The property describes a mature habit, and it is the reason rewards are the wrong lever once a habit is in place, with the cue and the effort required being the levers that remain.

It does not say that people stop liking the thing. In the cinema study liking was intact, and the finding is that it did not govern the behaviour.

It does not establish that apps behave like popcorn. The cinema experiments concern a snack in one setting, and the extension to screens is an argument by analogy. The paper was read here at the level of its abstract. Wood's percentages come from her book, which gives no sample size or effect size for the cinema study, and the work is from her own lab. The principle has a long animal literature behind it and can be held as solid. The specific percentages are a single study and should be quoted as such.

Who developed it

Wendy Wood, Provost Professor of Psychology and Business at the University of Southern California, is the researcher this site associates with the idea, and Good Habits, Bad Habits (2019) is where the site meets it. The devaluation test itself comes from animal learning research, and its first use has not been traced here.

---