Behavioral sampling in the field: Comparison of individual and group sampling methods

Sampling decisions affect the efficiency and reliability of data collection, and the appropriateness of the data for analyses of group and individual behavior. We evaluate the correspondence between interval sampling of individual behavior at high temporal density (focal interval sampling) with interval sampling of group behavior at lower density (group scan sampling) in two field studies with neotropical primates, capuchins and squirrel monkeys. The two methods provided consistent estimates of population means and variance for activity profiles, foraging activities, and height above ground. The correspondence between mean values with the two methods was greater when group sampling included individual identities than when a nominal scoring scheme was used. A group scan method (without identification of individuals) can be used alone when information is needed within a brief time, such as initial description of activity budgets of a population. Although individual identities take time to learn, data sets in which the individual is the unit of analysis provide several other kinds of analytical possibilities. We recommend use of a mixed sampling regime containing both of these elements (focal and group sampling) as a good way to minimize the time costs of data collection and as a means to evaluate reliability of data collection by a solo observer.