A comparison of scan and focal sampling in estimating activity budgets, diet composition, and proximity patterns of a wild pair-living primate

Abstract The choice of behavioral sampling method can impact the outcome of data collection, however, few direct comparisons between methods have been made. We compared the performance of instantaneous group scan sampling (scan sampling) and focal continuous sampling with variable session durations (focal sampling) in estimating activity patterns, diet composition, and spatial proximity in seven groups of wild coppery titi monkeys ( Plecturocebus cupreus ) in Peruvian Amazonia. We used a series of paired samples Wilcoxon tests to compare daily proportions of time allocated to each type of activity/food/proximity category in each sampling method. In addition, we compared our results with those of other studies conducted on the same population of titi monkeys at other times. Focal sampling provided significantly lower estimates for moving time and significantly higher estimates for resting time compared to scan sampling, likely because scan sampling tends to give higher estimates of more conspicuous behaviors and lower estimates of less conspicuous behaviors. For diet composition, scan sampling gave similar results to other studies, while focal sampling gave significantly lower estimates for feeding on fruits and higher estimates for feeding on arthropods. The most likely reason is that focal sampling with variable session durations tends to overestimate behaviors during which a focal animal is less likely to go out of view, such as feeding on arthropods in the lower strata of the forest. Our results suggest that a pilot study comparing different methods should be conducted prior to collecting data, as not all methods are interchangeable.