Data were collected from 246 people on their retrospective recall of the duration of a moderate earthquake in New York state, along with estimates of arousal, alertness, and attention to media. Duration of the quake was pervasively overestimated, with significantly higher overestimates from people in the area where the most intense quake occurred. Correlation of duration estimate with reported arousal was not significant. Responses were interpreted as the result of a desire of respondents to appear as credible observers of a complex episode.At 6:07 a.m., Saturday, October 19, 1985, a minor earthquake struck the general region of Westchester County, New York. Its epicenter was pinpointed near the town of Ardsley, New York (lat. 73°50'N; long. 4O o 50'W). The quake, measuring 4.0 on the Richter scale, was felt in widely scattered portions of the Northeast in varying degrees of intensity. According to the LamontDoherty Geological Observatory in Palisades, New York, the quake could be felt for approximately 30 sec at the epicenter.Within the first few days after the quake, we began to collect data on duration estimates of the length of the quake given by those who were awake during, or those who were awakened by, the event. This was the beginning of the program to study subjective reactions to natural phenomena and to evaluate the social cognitive bases of duration estimates -duration being defined as the interval between two successive events. In this paper, we address the question of whether the lessons learned in laboratory research can be generalized to the real-life task of judging the reliability of time estimates.Estimation of duration takes place when memory is used either to associate a moment in the past with a moment in the present or to link two past events (Fraisse, 1984). An individual's estimation of the passage of time rarely coincides with the even and mechanical hands of a clock. There is great variability among stimulus modalities, between subjects, and even within a subject at different times. More intense stimuli are judged as lasting longer than less intense stimuli. These findings have been demonstrated for auditory stimulation ), visual stimulation (Goldstone, lllamon, & Sechzer, 1978, and vibrotactile stimulation (Ekman, Frankenhaeuser, Berglund, & Waszak, 1969), suggesting that there is a general relationship betweenThe authors wish to thank Mary Knoff (Mercy College) and Wes Kayson (lona College) for their invaluable help in gathering data at their colleges. We further wish to thank staff members of the Lamont-Doherty Geological Observatory who gave us a much-needed education in research on earthquakes. Reprints of this article may be obtained from the senior author of The Center for Responsive Psychology, Brooklyn College. CUNY, Brooklyn, NY 11210. stimulus intensity and duration estimates. Our own experience with eyewitness testimony over the years suggests that precise clock accuracy is the last thing on our minds when we go over the duration of past events. If the time estimate is give...