ABSTRACT
Survey respondents are sometimes asked to report alxmt various events in the past involving different activities, persons, and time periods. For example, in the Bureau of Labor Statistics' Consumer Expenditure Interview Survey, respondents report the various types of expenditures that have been made by different members of their household over the past three months. Respondents often do not employ a systematic approach to such recall tasks on their own, and confusion or incomplete reports result (Lessler, 1989). To reduce problems of this nature, a survey designer might choose to organize the recall task hierarchically so that certain event features (time, person, activity) are nested within other event features. For example, the questions might focus on one person at a time and ask about his or her participation in various activities before switching to another person and asking about that person's participation in those activities. Alternatively, they might focus on a particular activity and ask about each person of concern before switching to another activity. An important question is whether the choice of the question sequence organization affects the recall performance of the respondents.
