Sampling Recommendations
If you are new to behavioral research, we recommend Measuring Behaviour: An Introductory Guide by Marin and Bateson (2007). This classic text provides a thorough discussion of observational behavior sampling methods beyond the scope of this article.
When designing a new project, your first task will be to create your ethogram and, importantly, determine which behaviors you would classify as states vs. events. State behaviors are those that occur for some length of time, such as resting or feeding and many other maintenance-related behaviors. On the other hand, event behaviors are those that are brief and often sudden, such as aggression or breeding behavior. Determining which category a given behavior falls into will help you determine the most appropriate sampling methodology. Typically, state behaviors are recorded using interval sampling methods and event behaviors are recorded using all-occurrences methods. These sampling methods also influence the appropriate behavioral measure to be calculated.
| Sampling Method | Ideal Use | Benefits / Challenges | Measure |
| All-occurrence | Event Behaviors | Will provide a complete record of the frequency of behavior but not the duration of time performing the behavior. | Rate (occurrences/hr) |
| Interval | State behaviors | Can provide an approximation of time spent performing a behavior but will not accurately represent brief behaviors (i.e. events). Less intensive interval sampling does allow recording of multiple behavioral categories (e.g. social proximity, shade use, etc.) | Percent of Scans |
| Continuous | State/Event behaviors | Provides a complete record of behavior but is the most intensive, thus would be best applied to record detailed behavior in a single focal animal. | Percent of Time |
