A model of telescoping is proposed that assumes no systematic errors in dating. Rather, the overestimation of recent occurrences of events is based on the combination of three factors: (1) Retention is greater for recent events; (2) errors in dating, though unbiased, increase linearly with the time since the dated event; and (3) intrusions often occur from events outside the period being asked about, but such intrusions do not come from events that have not yet occurred. In Experiment 1, we found that recall for colloquia fell markedly over a 2-year interval, the magnitude of errors in psychologists' dating of the colloquia increased at a rate of .4 days per day of delay, and the direction of the dating error was toward the middle of the interval. In Experiment 2, the model used the retention function and dating errors from the first study to predict the distribution of the actual dates of colloquia recalled as being within a 5-month period. In Experiment 3, the findings of the first study were replicated with colloquia given by, instead of for, the subjects.Telescoping occurs when people respond with an overestimate when asked questions such as how many times they have visited the doctor during the past 6 months or how often in the past 2 weeks they have purchased a particular product (Bradburn, Rips, & Shevell, 1985;Loftus & Marburger, 1983;Sudman & Bradburn, 1973;Thompson, Skowronski, & Lee, 1988). The phenomenon is called telescoping, because it is as if time shrinks to the present, in a temporal equivalent to the shrinking of distance that occurs when an object is viewed through a telescope. Such shrinking should cause events with actual dates before the target period to be recollected as if they had taken place within the target period. Here we will show that telescoping is a result of the way the question is asked rather than a result of the compression of time.Note that no measure of time distortion is taken in survey questions that ask for frequency of occurrence in an interval. In fact, survey researchers have long noted overestimation of frequency caused by items being imported into a category when time is not involved. Mahalanobis (1946) 1 to 3 to 12 to 16 ft on a side, the estimated yield, in pounds per acre, decreased from 1,160 to 535 to 428 to 395. Similar effects were noted for rice and wheat. Mahalanobis's first hypothesis was that jute immediately outside the boundary was telescoped in for purposes of the count. In fact, the telescoping of crops near the boundary of plots did not explain the systematic biases that Mahalanobis noted; nonetheless, the observation and theory was the initial impetus for our investigation. It is plausible that in trying to answer correctly, people might have included events or plants that were near the boundary. If the question had been worded, "How many times in the last month did you , , . ," then only events older than the interval could have been imported, because events more recent than the interval would have been in the future; a systematic telescoping erro...