In this paper we examine the tradeoffs between two methods of collecting labor force data, retrospective questions versus repeated interviews.  In the case of gross change measures, such as month-to-month flows between labor force states, retrospective measures tend to understate change, while repeated interviews seem to overstate change because of response variability.  In the case of measures of experience—the amount of time spent in a labor force state within a given time frame—retrospective measures tend to be affected by recall bias, which lead to the undercounting of short periods spent in that state.  On the other hand, experience measures constructed from repeated interviews can be affected by selective attrition, since only observations with a complete set of interviews can be used.