UC Santa Barbara

Applied behavior analysis (ABA) interventions are delivered using a tiered-service model, in which supervisors with higher levels of experience, education, and certification oversee clinicians who provide direct intervention for individuals with Autism Spectrum Disorder (ASD). Despite the base of empirical support for ABA constructed over several decades, little attention has been paid to the ways in which supervision is implemented among service providers. The majority of current standards have been issued by the Behavior Analyst Certification Board (BACB) with additional recommendations from researchers in the field, yet most of these guidelines have not been empirically validated. The small body of existing research on supervision practices indicates that supervision implemented within the context of university-based research centers leads to improvements in clinician fidelity as well as social communication behaviors for children with ASD, and clinicians report high overall satisfaction with supervision. Findings from previous correlational studies suggest that the amount of supervision provided and supervisor characteristics (i.e., certification and level of experience) predict client outcomes over time, and that Perceived Supervisor Support (PSS) serves as a protective factor against clinician burnout and turnover intention. Yet little is known about specific practices or supervisor behaviors correlated with ratings of PSS, which is an important step for understanding how supervision can be improved in both efficiency and efficacy. Furthermore, with the exception of findings indicating that newer clinicians need more frequent supervision in order to reach or maintain fidelity, there is a dearth of research indicating how supervision should be individualized according to characteristics of clinicians and clients and other contextual factors.The primary goal of the current study was to build upon the current body of literature regarding supervision of ABA services for individuals with ASD by examining the relationship between practices used for supervision sessions and clinician satisfaction with supervision. To address this research question, a cross-sectional survey was designed based on BACB- and researcher-developed guidelines for supervision and supervisory practices. 125 clinicians working in clinic- and community-based ABA service agencies in California were recruited to complete an online two-part survey, inquiring about (1) supervision practices and satisfaction with “typical” sessions, and (2) supervision practices and satisfaction with sessions delivered through videoconferencing (VC) during the Coronavirus Disease 2019 (COVID-19) pandemic. Two separate exploratory factor analyses were conducted using variables from part 1, resulting in a three-factor model for supervision practice (i.e., supervision activities, supervisor proficiency, and clinician evaluation processes) and supervision satisfaction (i.e., satisfaction with supervision content, satisfaction with perceived level of support, and dissatisfaction with supervisory relationship). Similar composite variables were created using data from part 2 to assess changes with practices and satisfaction during the transition to agency-wide VC-delivered supervision sessions. The predictive nature of select demographic and work-related variables were assessed using Multivariate Analysis of Variance (MANOVA) and multiple regression analysis. Several variables were significant predictors of supervision practice. Supervision hours per month and frequency of individual and group supervision meetings predicted higher supervision practice scores for part 1, while frequency of individual meetings predicted higher supervision practice scores for part 2. Supervisor performance evaluation in a high number of specific categories predicted higher supervision activities and supervisor proficiency scores for parts 1-2. Weekly/semi-weekly supervision frequency predicted higher supervisor proficiency scores for part 2. In addition, a few variables were significant predictors of clinician satisfaction for parts 1-2. Supervision practice scores predicted higher satisfaction scores, and having at least 3 years of experience predicted higher satisfaction with perceived level of support. A number of common themes from open-ended comments were identified with suggestions for improving supervision sessions, supervisor behavior, and training/supervision topics in each delivery format. In absence of experimental validation for many of the established supervision guidelines and recommendations, these research findings provide a preliminary foundation of correlational evidence supporting the importance of specific supervision practices and how they predict clinician satisfaction with supervision, which can help prevent burnout and turnover intention. In addition to supervision intensity, supervisor preparation, communication, performance feedback, evaluation, and use of activities that provide opportunities for self-assessment, collaboration, and continued training were associated with higher-quality supervision practice. Ratings of supervision activities and supervisor proficiency increased when supervisors solicited detailed evaluation of their own performance from the perspective of their supervisees. Greater use of these supervision practices was associated with higher clinician satisfaction with supervision sessions. Also, while clinicians with higher levels of experience were more satisfied with the perceived level of support from their supervisors, greater use of supervision practices identified in the current model predicted higher satisfaction scores for clinicians regardless of experience level. Thus, modification of specific supervision practices may help prevent levels of increased burnout and turnover among newer clinicians. These results also show the effectiveness of widespread VC-delivered supervision and offer insight into which practices are particularly important for remote supervision (e.g., conducting sessions on a weekly basis), limitations of VC sessions (e.g., lower ratings of helpfulness and engagement with supervision activities), and aspects of remote supervision that require further study.