ARPobservation provides a set of tools for simulating the sort of data generated during direct observation of a behavior over time. It works by first simulating a behavior stream based on an alternating renewal process, given specified distributions of event durations and interim times. The following distributions are currently supported:

• Exponential
• Gamma with fixed shape parameter
• Mixture of two gammas
• Weibull with fixed shape parameter
• Uniform(0,x)
• Constant (degenerate)

Different methods of recording data can then be applied to the simulated behavior stream. Functions are provided for the following recording methods:

• continuous duration recording
• event counting
• momentary time sampling
• partial interval recording
• whole interval recording