Background
According to the peak-end rule, the retrospective evaluation of an experience was predicted best by taking the most extreme emotional valence rating (peak effect) and the emotional valence experienced during the final moments of the experience (end effect). However, inconsistent findings that support the predictive value of only peaks, ends, or the average of peaks and ends.
Aims
The inconsistent findings may arise because previous research tested either the short-term or long-term effect of the peak-end rule and never considered both in one study. Hence, the present research tested the peak-end rule in real-life circumstances in short-term and long-term time frames (1-day and 10-day). In addition, since most of existing studies used emotional valence as the single predictor of emotional experience, the present research also included the emotional arousal, as well as the axis that combing the information of both valence and arousal as predictors of emotional experience.
Method
138 undergraduates attended a meeting before Web-based daily data collection. They were asked to complete their daily diary based on the Day Reconstruction Method for 10 days.
Results
The data were analyzed using HLM 8.1. In the day level analysis, the results shows that the retrospective evaluation of life satisfaction for the previous day was predicted by both the most extreme emotional rating (valence, HAP) and the emotional rating experienced during the last (two) episodes (valence, HAP). However, the peak effect was more robust than the end effect. In the 10-day level analysis, the results shows that retrospective evaluation of life satisfaction for the previous 10 days was predicted only by the emotional rating experienced during the last day (valence, arousal, HAP), and not by the most extreme emotional rating within the 10 days.
Conclusion
According to above findings, both the peak- and end-effects were found in the time frame of 1-day. However, the end-effect was more important than the peak-effect in the time frame of 10-day.