Behavioral sampling in the field: Continuous focal versus focal interval sampling
Behaviour, 137(2),153- 180, February 2000
Lisa M. Rose-Wiles, Ph.D.
I compared data collection rates for continuous and interval focal samples during a two-year, single-observer field study of white-faced capuchins (Cebus capucinus) in Costa Rica. I also compared the basic activity budgets generated by the two sampling methods, estimates of numbers in proximity, and rates at which additional ad libitum observations could be recorded. I collected 1238 hours of focal data (620 hr continuous, 618 hr interval). I found focal interval sampling to be 25% more time efficient, despite higher rate of sample loss, partly because interval samples are easier to obtain in difficult conditions. I found no evidence that interval sampling provided better opportunities for ad libitum observation than continuous sampling. Overall, the two methods yielded similar estimates of activity budgets. However, continuous sampling resulted in somewhat higher estimates of time spent eating, while interval data gave somewhat lower estimates of time spent foraging (looking for or handling food items) and moving, resulting in lower estimates of foraging success. Interval sampling also yielded slightly lower estimates of time spent vigilant. I attribute these patterns to two major effects: (1) errors of omission (missing rare behaviors of short duration) during interval samples and (2) a greater tendency toward conditional sampling bias (under-representing behaviors due to difficult sampling conditions such as rapid travel) under a continuous sampling regime.