Functional analyses (FAs) of problem behavior typically are conducted in controlled settings to minimize potential sources of confounding. Several studies have reported that results of FAs conducted in controlled settings occasionally differ from those conducted under more naturalistic conditions, although little is known about factors that may contribute to the different outcomes. We examined correspondence between FAs conducted by staff in a clinic and those conducted either by caregivers as therapists or in the home setting. If results of the 2 analyses were dissimilar, we conducted further analyses to identify variables responsible for the different outcomes. Results showed that, in most cases, correspondence of function was observed across familiar and unfamiliar stimuli. Results are discussed in terms of implications for research and clinical practice for the evaluation of problem behavior.