Animal-assisted therapy and the severely disabled child: a quantitative study - Graduate Literary Award Winner - Abstract

Journal of Rehabilitation, Oct-Dec, 2001 by Kathryn Heimlich

Adequate training is believed necessary to ensure acceptable levels of inter-rater reliability. Prior to the beginning of observations designed to determine behavior baselines, all three raters received training in the completion of the instruments utilized. Still, their opinions varied.

Several possibilities may offer insight into this phenomenon. The first, and most obvious, points to rater personality and subjectivity, or bias. Clearly, any instrument that relies on human judgment is susceptible to individual interpretation. Some raters may have a tendency to rate highly (a leniency or generosity error), a tendency to rate harshly (a severity error), or a tendency to rate everyone near the midpoint of the rating scale (an error of central tendency) (Cohen, et al., 1996). More precisely, the utilization of a Likert scale on the MOPI may have further contributed to rater subjectivity.

Contextual variables can also account for differences between raters using rating scales.

" ... different informants may have different perspectives on the subjects of evaluation based on the different contexts in which they interact with the subjects" (Cohen, et al., 1996). Although the raters attempted to observe the children in their classrooms each week, on several occasions time restraints required them to complete their observations while working directly with the children. It may be presumed that the children would behave differently in the classrooms in the presence of their peers as opposed to one-on-one contact, seated before a computer, favorite toys or gymnasium equipment. The influence of context on the behavioral assessments may account for some of the discrepancy in the raters' judgments.

A third factor with regard to the raters, addresses the amount of time and effort involved in the observations and instrument completion. Although the MOPI required a minimum of time for completion, the other two instruments were said to be time-consuming, redundant and for some students, inappropriate. Additionally, the raters were required to set aside time from fulltime work schedules to complete the observations for each of the seven students each week. The time constraints imposed on them by full caseloads, coupled with participation in this research project, may have ultimately diluted their enthusiasm and impacted their willingness to devote the time and effort required to do the job properly. This factor likely accounts for the lack of complete data on the second group of students in Trial 2.

Administrative Support

Strong administrative support of the animal-assisted therapy program and research study might have helped to alleviate some of the problems encountered with the raters. Recognition of the rater's efforts, an allowance of time for thorough training sessions and an occasional inquiry with regard to the program may have provided the impetus necessary to maintain the raters' commitment. As it happened, early in October of 1996, just as the study was about to begin, there was a change in administration. Although the outgoing director fully supported the introduction of an alternative therapy program, his successor lacked the same enthusiasm. Full administrative support may be considered an essential component of a successful program and research endeavor.