Association for Behavior Analysis International

Experimental functional analysis (EFA) is considered the gold standard of assessment when evaluating the function of a challenging behavior. There are numerous peer-reviewed articles demonstrating the use of what many clinicians consider a traditional EFA and variations such as trial-based functional analysis (TBFA), latency to first response, or EFA of precursor behavior. This symposium will highlight procedural integrity as it relates to implementation of EFAs. During the first discussion, Ridgway and colleagues discuss the frequency with which authors report procedural integrity during EFAs. Additionally, data will be presented regarding the frequency with which authors report recording of inter-observer agreement on procedural integrity data collection. Rue and colleagues will follow with a review of training board certified behavior analysts (BCBAs) to implement an EFA. Specifically, the authors will present procedural fidelity data when novice BCBAs implement a traditional EFA and generalize to a TBFA. LoVullo and colleagues will present data demonstrating the impact of varying procedural integrity during a traditional EFA. The symposium will allow for a discussion of the importance of monitoring and reporting procedural integrity during EFAs and implications for practice.

The degree of procedural integrity of a functional analysis may impact the results obtained. As intervention procedures are designed based on the results of a functional analysis, the effectiveness of an intervention may be dependent on the degree to which the functional analysis was implemented with integrity. We reviewed studies on the functional analysis of problem behavior from 1961 to 2012 previously included in Beavers, Iwata, and Lerman (2013) and Hanley, Iwata, and McCord (2003) to identify the number of studies that include procedural integrity data and interobserver agreement data for procedural integrity data. Of the 110 studies reviewed thus far, 23 studies (21%) included procedural integrity data, 2 studies (2%) stated that procedural integrity data were collected, but were not reported in the article, and 2 studies (2%) reported that trainer behavior was monitored to ensure integrity, but no data were reported. Of the 23 studies that reported integrity data, eight (35%) included interobserver agreement data for integrity data. Eighteen (16%) of the 110 studies reported that the functional analysis was conducted by a trained individual, but only 8 of those studies included data on procedural integrity. Implications for practice and future research regarding procedural integrity will be discussed.

Results of a recent survey of over 600 behavior analysts indicated only 77% received formal training in the use of experimental functional analysis (EFA). Further, 63% of board certified behavior analysts (BCBAs) indicated they never or almost never used an EFA in practice. The objective of the current study was to determine if BCBAs could generalize the ability to implement a trial-based experimental functional analysis (TBFA) following training focused on traditional EFA methodology. Three BCBAs with no experience implementing EFAs acted as participants in the study. The independent variable included formal training in EFA methodology. The dependent variable was level of procedural integrity during implementation of a traditional EFA and a TBFA measured in a multiple baseline design. Participants implemented the traditional EFAs during a mock assessment. Participants implemented the TBFAs in an applied setting with clients diagnosed with autism spectrum disorder (ASD). Initial results indicate that participants can implement traditional EFA methodology with relatively high levels of integrity. Results suggest two participants demonstrated challenges maintaining levels of integrity above 80% during the demand and control conditions. Participants will likely require additional feedback to maintain procedural integrity above 80% accuracy during TBFAs. Implications for training and practice are discussed.