Autobiographical Memory and Surveys
What we remember about our past depends on how we organize our time. Kurbat, Shevell, and Rips (1998) have shown that college students recalling events from the past calendar year tend to recall more events from around the time of the beginnings and ends of academic terms. This Calendar Effect depends on the transition between school time and vacation time; just changing courses doesn't produce many extra memories. The effect also seems to be a retrieval phenomenon: Changing explicit retrieval cues can change the size of the effect. The effect also disappears when students predict events in other students' lives (rather than recalling events from their own lives). The effect does not seem to be due to event importance. Shum and Rips (1999) obtained ratings of events' importance as they occurred (by randomly beeping participants during their everyday activities). Calendar effects persist, even when importance is held constant.
Some recent research has explored autobiographical recall in survey contexts. In some longitudinal surveys, respondents have to answer questions about a number of small intervals within a larger reference period. For example, respondents may be interviewed quarterly, but must answer questions about one-month intervals within the quarter. These surveys produce a type of response error called a Seam Effect -- month-to-month changes in answers to a particular question are much larger when the data come from two different interviews than when they come from one interview. Some experiments with Fred Conrad and Scott Fricker (Rips, Conrad, & Fricker, 2004; Conrad, Rips, & Fricker, 2009) show that the size of these effects depends on the interaction between respondents' memory and estimation strategies. New work extends these findings to estimates of continuous quantities.