Reliability and validity of using structured visual‐inspection criteria to interpret latency‐based functional analysis outcomes

Abstract

Prior research has evaluated the reliability and validity of structured visual inspection (SVI) criteria for interpreting functional analysis (FA) outcomes (Hagopian et al., 1997;Roane et al., 2013). We adapted these criteria to meet the unique needs of interpreting latency-based FA outcomes and examined the reliability and validity of applying SVI criteria to 43 previously published latency-based FA datasets. Overall, raters agreed on SVI-determined FA outcomes (98% of functions and 95% of cases) and these outcomes corresponded well to the interpretations provided by the authors of these 43 datasets (94% of functions and 88% of cases), indicating a high degree of reliability and concurrent validity. Our findings suggest that the use of SVI criteria may (a) serve as an objective aid in the identification of behavioral function(s), (b) produce high levels of agreement among expert raters, and (c) serve as a useful resource when teaching students how to interpret latency-based FA outcomes.

Exploring factors that influence the efficacy of functional communication training

“…As response latencies often share a symmetrical relation with response rates (Caruthers et al, 2015; Thomason‐Sassi et al, 2011), only considering latency to the first occurrence of behavior observed in each test session offers a valid strategy for eliminating within‐session confounds associated with procedural variability (e.g., duration of reinforcer access). Further, because response latency already has demonstrable utility to data analysis in applied settings for both functional analysis (e.g., Lambert, Finley, & Caruthers, 2017; Lambert, Lopano, et al, 2017; Lambert, Staubitz, et al, 2017; Standish et al, 2021; Sunde et al, 2022) and functional communication training (e.g., Boyle et al 2023; LeJeune et al, 2019), adoption of this strategy by practitioners may be more likely than adoption of alternatives with fewer practical functions, thereby increasing the viability of identifying predictive variables through latency measures.…”