1 Many aspects of human behaviour impact on ecological systems. Ecologists therefore need information on changes in these behaviours and are increasingly using methods more familiar to social scientists. 2 Understanding patterns of wildlife harvesting is important for assessing the sustainability of harvests. Interviews are commonly used in which informants are asked to summarize their activities over a period of time. However, few studies have investigated the reliability of such data, the usefulness of interviews for monitoring trends, and how their information content can be maximized. 3 We carried out rapid assessment interviews with villagers in Madagascar about the quantity, timing and spatial patterns of crayfish Astacoides granulimanus and firewood collection. We compared the results with information from daily interviews with the same informants. We used mixed models to investigate how accurately people reported their activities in the rapid assessment interviews, and estimated the probability of detecting a change in harvesting from two such interviews using a Bayesian approach. 4 The interviews provided reliable information on quantities, effort, and the spatial pattern of harvesting. Simulations suggested the interviews would detect changes in catches and harvesting effort with reasonable power; for example, a 20% change in the amount of time spent crayfish harvesting could be detected with 90% power. Power is higher when the same informants are questioned in repeat interviews. 5 Synthesis and applications. Ecologists are increasingly using social techniques and it is vital that they are subject to rigorous testing to ensure robustness in trend detection. This study suggests that interviews can be used to monitor changes in harvesting patterns by resource users, but whether the power is adequate will depend on the needs of the study. To maximize the power of interviews, informants should be interviewed independently and the same informants interviewed in subsequent years.