Peak-end rule
Claim assessed: We remember an experience mostly by its most intense moment and its end, not by its length.
Published · how we work
In short
The peak-end rule says that when we remember a pleasant or unpleasant experience, we judge it mainly by two moments: the most intense one and the final one. How long it lasted matters little, a phenomenon called duration neglect. This leads to something paradoxical: an unpleasant experience can be remembered as less unpleasant if a somewhat less unpleasant stretch is added at the end, even though there was more suffering in total.
What it says
The cold water. Kahneman, Fredrickson, Schreiber and Redelmeier (1993) asked students to keep one hand in water at 14 °C for 60 seconds. With the other hand they did a longer trial: the same 60 seconds at 14 °C, then another 30 seconds while the water warmed slightly to 15 °C, still unpleasant. The long trial thus contained all the discomfort of the short one, plus some more. When asked which one they would repeat, 22 of 32 students (69%) chose the long trial. The authors cite earlier studies in which a combination of the worst moment and the final one predicted the overall evaluation, while duration mattered very little.
The colonoscopy. Redelmeier, Katz and Kahneman (2003) tested the idea in a Toronto hospital with 682 patients. For half of them, chosen at random, the instrument was left still for a few minutes at the end of the colonoscopy before being removed. The procedure lasted about one minute longer, but the final minutes were less painful. These patients remembered the procedure as less unpleasant (4.4 versus 4.9 on a 10-point scale) and ranked it lower among other unpleasant experiences. In the following years, they came back slightly more often for a repeat colonoscopy.
For complex experiences, the rule weakens. The classic studies used simple experiences, such as pain. Strijbosch et al. (2019) tested the rule with a film watched in virtual reality that went through positive and negative emotions. With 40 participants, the memory right after the film was best predicted by the most intense moment, but the ending didn’t matter. A week later, the best predictor was the average emotion over the whole duration. The authors conclude that the rule can’t be generalised to experiences closer to real life, and suggest that, in a rich experience, many moments contribute to the memory, not just two.
Example
A two-week holiday was mostly good, but the last day was a long journey with lost luggage. Years later, the holiday is told mainly through the best evening and the trip home. The example is an illustration, not a study cited here.
How to apply it
The steps below are a practical way to apply it that we propose, based on the studies cited.
- When you organise something, look after the ending. A meeting, a course or a visit that ends well will be remembered better.
- When you judge a past experience, also ask how long it lasted. Your memory shortens or stretches the experience without your noticing. For repeated decisions, write down how it was at the time.
- Don’t rely on the rule for complicated experiences. For an experience with mixed emotions, the ending mattered little, and a week later the average mattered most.
- In medicine, a gentler ending may help, but it raises questions. Armitage (2025) argues that deliberately prolonging a painful procedure without the patient knowing is hard to justify ethically.
Limits and nuances
- Most of the evidence comes from simple experiences. Cold water and colonoscopy have a single dimension, pain. For more complex experiences, the rule, especially the part about the ending, works less well.
- The effects are real but not huge. In the colonoscopy study, the difference in memory was about 10%, and the difference in return rates small.
- Not everyone reacts the same way. In the cold-water study, 11 of 32 participants felt a small or no drop in discomfort in the last 30 seconds, and some chose the short trial.
- We couldn’t read a recent meta-analysis. The meta-analysis by Alaybek et al. (2022) on the peak-end rule is paywalled; we didn’t use it.
- What we read. Kahneman et al. (1993), from the copy on the site of a University of Zurich institute, Redelmeier et al. (2003), from the York University repository, and Strijbosch et al. (2019), from Frontiers in Psychology, we read in full. For Armitage (2025) we read the abstract. We haven’t read Redelmeier and Kahneman’s 1996 study on memories of colonoscopy.
Sources
- Daniel Kahneman, Barbara L. Fredrickson, Charles A. Schreiber, Donald A. Redelmeier (1993). When more pain is preferred to less: Adding a better end
- Donald A. Redelmeier, Joel Katz, Daniel Kahneman (2003). Memories of colonoscopy: A randomized trial
- Wim Strijbosch, Ondrej Mitas, Marnix van Gisbergen, Miruna Doicaru, John Gelissen, Marcel Bastiaansen (2019). From experience to memory: On the robustness of the peak-and-end-rule for complex, heterogeneous experiences
- Richard C. Armitage (2025). Prolonging medical procedures to exploit the peak-end rule: An ethical analysis
See also: Negativity bias, Serial position effect