When is the best treatment for pain more pain? When you’re taking part in an experiment published by a Nobel prize winner and one of the leading lights in behavioural psychology, that is.
The psychologist in question is Daniel Kahneman; the experiment described by the self-explanatory title of: When More Pain Is Preferred to Less: Adding a Better End. In the study, Kahneman and colleagues looked at the pain participants felt by asking them to put their hands in ice-cold water twice (one trial for each hand). In one trial, the water was at 14C (59F) for 60 seconds. In the other trial the water was 14C for 60 seconds, but then rose slightly and gradually to about 15C by the end of an additional 30-second period.
Both trials were equally painful for the first sixty seconds, as indicated by a dial participants had to adjust to show how they were feeling. On average, participants’ discomfort started out at the low end of the pain scale and steadily increased. When people experienced an additional thirty seconds of slightly less cold water, discomfort ratings tended to level off or drop.
Next, the experimenters asked participants which kind of trial they would choose to repeat if they had to. You’ve guessed the answer: nearly 70% of participants chose to repeat the 90-second trial, even though it involved 30 extra seconds of pain. Participants also said that the longer trial was less painful overall, less cold, and easier to cope with. Some even reported that it took less time.
In case you think this is a freakish outcome of some artificial lab scenario, Kahneman saw a similar result when he interviewed patients who had undergone a colonoscopy examination – a procedure universally described as being decidedly unpleasant. Patients in Kahneman’s study group had colonoscopies that lasted from four to 69 minutes, but the duration of the procedure did not predict how they felt about it afterwards. Instead, it was the strength of their discomfort at its most intense, and the level of discomfort they felt towards the end of the procedure.
These studies support what Kahneman called the Peak-End rule – that our perceptions about an experience are determined by how it feels at its most intense, and how it feels at the end. The actual duration is irrelevant. It appears we don’t rationally calculate each moment of pleasure or pain using some kind of mental ledger. Instead, our memories filter how we feel about the things we’ve done and experienced, and our memories are defined more by the moments that seem most characteristic – the peaks and the finish – than by how we actually felt most of the time during the experience.
Kahneman wondered whether this finding meant that surgeons should extend painful operations needlessly to leave patients with happier memories, even though it would mean inflicting more pain overall. Others have asked whether this means that the most important thing about a holiday is that it includes some great times, rather than the length of time you are away for. (It certainly makes you think it would be worth doing if you could avoid the typical end to a holiday – queues, lumping heavy luggage around and jetlag.)
But I think the most important lesson of the Peak-End experiments is something else. Rather than saying that the duration isn’t important, the rule tells me that it is just as important to control how we mentally package our time. What defines an “experience” is somewhat arbitrary. If a weekend break where you forget everything can be as refreshing as a two-week holiday then maybe a secret to a happy life is to organise your time so it is broken up into as many distinct (and enjoyable) experiences as possible, rather than being just an unbroken succession of events which bleed into one another in memory.
All I need to do now is find the time to take a holiday and test my theory.
This is my BBC Future column, originally published last week. The original is here.
Mind Hacks 
14C or 15C water is not “ice cold” and is very easy to bear indefinitely when only a hand is immersed.
I suspect you are missing a decimal point and the temperatures reported in this article are actually 1.4C and 1.5C.
Tap water from my kitchen sink is 6.1C and is easy to tolerate for one minute – no pain involved.
A pitcher of water from my refridgerator is 3.2C and is quite uncomfortable but not painful after 60 seconds. Add a few ice cubes then remove them when the water reaches 1.4C … and now I experience something akin to what is described in this article.
@Rob its a good point, and I thought it was wrong at first, so I checked it, but 14 C is how the original article reports it. There may be a missing decimal point (or it could be in Fahrenheit), but I couldn’t find a correction so I must presume that the authors meant what they wrote.
As for how painful it is, it doesn’t sound very painful, true. The water is continuously circulating so it doesn’t warm up (like your tap), and we have the reports of how painful it was so – whatever temperature – we know the pain was real
@Tom: 14F would be -10. Water doesn’t get that cold without freezing. Perhaps if they were supercooling it and circulating it, it might stay liquid.
Prolonged contact with water that cold can cause tissue damage though, so it wouldn’t have passed ethics.
The abstract does say 14C, but the whole thing only makes sense if it’s missing a decimal place. 14-15C is just slightly below room temperature, it’s tepid or maybe cool, but not cold, so that can’t be it. 14f is WAY too cold, so only 1.4-1.5C makes sense.
@Joe and @Rob, I hear what you are saying about the exact temperature. Since it can’t be verified quickly, I’ve decided to suspend thoughts on the accuracy of the temperature and focus on the point made by @Rob that participants found it uncomfortable.
I’ve just been away for two nights with my wife, two children and another same sized family. We had some good times, one place in particular getting votes as being the best – we had low expectations of it. At the end of the last day, my two boys had a bit of a falling out that engulfed us. I felt quite disappointed about the whole break during the 100+ mile journey home.
As a result the Peak-End theory has struck a chord with me. Now that it is a couple of days afterwards and the boys are getting on well it doesn’t feel as bad.
I therefore wonder if there is something in looking out our experiences in satisfying chunks. Have I just waited until a better ending? Is there a risk of self deception in the chunking path.
Very interesting indeed.
Tom, this ‘unbroken succession of events which bleed into one another in memory’ is the very same reason why time/life actually seems to run faster for older people: ‘less distinct experiences’, or in the words of W. James ‘uniformity of memory content’.
Does this study (or others like it) account for attenuation versus abrupt removal of pain stimulus? It seems to be framed more in terms of duration of pain stimulus, without regard to whether gradual or sudden removal of pain stimulus contributes to the experience. In other words, is the transition provided by gradual attenuation of pain preferable?
That’s a cool study. Kind of explains the way we look back on past events ( life experiences /relationships etc.). That is if we can assume such a relation between the memory we form of a physical sensation( pain ) to episodic memory.
Anyway, just to clear up the whole discussion about the temperature. You’d need to consider the extent of time your hand is immersed in the water. While 14 C might not seem that cold, if you leave your hand in it for a while it may get a little uncomfortable / painful. Actually, I read somewhere that exposure do degrees below 10 Celcius for less than 5 minutes can leave your hands without dexterity- eventually leading to frostbite. So it could actually be right. -Cheers. Ps. I love this blog. Thank you
My sister had a toothache – and for several days avoided seeing a dentist because of the potential for “a more intense, short burst of pain” that a dentist MIGHT cause.
Just the potential for more intense pain was enough for her to endure significant pain for some time. An example of even imagined “peak” pain being decisive?
Humans can be strange creatures.
Could it be that the 90 sec experiment was reported to less painful because the study group had more time to get accustomed to the change they experienced?
Looks like linear models do not capture the experience of pain very well. Heck, this shows that the phenomenon is not even monotone — “more” input doesn’t lead to “more” pain.
Why am I not surprised?! In fact, what ever made us think they would work — for pain or any mental phenomenon. Seems like looking for your keys where the light is better — simple models are easier to grasp and apply, but what reason is there to assume that they will be the least bit meaningful.
Some day — maybe not too far off — we (collectively) are going to have to acknowledge that consciousness and experience are _sui generis_, and will not yield their mysteries to the kinds of reductive methods that have succeeded so well with physical phenomena.