Why You Remember the End of a Holiday and Not the Middle
Memory is supposed to keep the best moment and the last one and bin the rest. The only study that has tested that on a real holiday found the middle was never voted down. It had gone before the vote
Key takeaways
- The only direct test of the peak-end rule on a real holiday is Simon Kemp's, with 49 students texting a happiness rating every day of a trip averaging a week. What predicted their later verdict was roughly how they had felt on average across the whole trip. The ending did more work than the best moment, the length of the trip made no difference at all, and they could not recall their own day-to-day moods afterwards (Memory and Cognition, 2008).
- Balca Alaybek and her co-authors pooled 174 separate results in 2022. The worst or best moment combined with the final moment predicts a later rating strongly, and it was matched by the flat average of every moment, which did the same job no better and no worse. A tie like that means the middle is still in the record rather than discarded.
- What genuinely drops out is the clock. Duration neglect is the robust half of the pair, and it has four known off-switches: when length is made obvious, when the experience is drawn as a graph, when the domain makes length the point, and when the experience arrives in separate chunks. A fortnight away trips all four.
- The credit is routinely misplaced. Barbara Fredrickson is first author on the 1993 paper that established duration neglect, using film clips rather than pain. Kahneman leads the cold-water study of the same year. Donald Redelmeier leads both medical papers, and the famous colonoscopy work cannot be re-run because those procedures are no longer performed without sedation.
- The commercial version is sold well past its evidence. A randomised trial of 682 patients with a deliberately gentler ending improved the remembered experience by about half a point out of ten, and barely moved whether patients returned over roughly five years. A study of more than 5,000 hotel stays across five years found first impressions mattered most, and mattered more the longer after the stay guests were asked.
Six in the morning, and the holiday starts with an alarm that feels like a punishment. Then the car park shuttle, the queue, the gate, the dead hour before boarding. You land somewhere warmer and spend the first day slightly wrong. Wrong shoes, unfamiliar coins, a supermarket you cannot read.
By the third day there is a cafe you go back to without anyone proposing it, and a walk you have now done twice. This is the part of a holiday that works, and the part nobody photographs, because nothing is happening in it, and nothing happening is what it was for.
Then the days begin folding into one another. A ruin, a beach, a market, a long lunch, another beach. Two of them rain. Somebody is short with somebody else one afternoon and it has passed by dinner. Eleven days in, you could not put them in order under oath.
The last night is booked in advance because it is the last night, and the restaurant is better than the ones before it. The last morning is a bag on a made bed, a key left on a table, a taxi that comes early. On the Monday somebody asks how it was, and you answer in one word.
Fourteen days, and the verdict is a syllable. A short answer is only manners. The odd part is that the syllable is also, as far as you can tell from the inside, what you actually think. Push on it and you get the last dinner, the day it rained, one swim. The other eleven days do not arrive, and you lived every hour of them.
The explanation in circulation says memory keeps the best moment and the final one and throws away everything between. It travels under the name the peak-end rule, it is printed in customer-experience training decks, and it is wrong in a specific way. The correction answers the question better than the rule ever did. So where do the middle eleven days go?