Questions about behavioral function (QABF):

Abstract

Functional assessment is a method to identify the relationships between a behavior of interest and an individual’s environment. Traditional methods for functional assessment have relied on experimental techniques in which analog sessions are designed to replicate conditions in the individual’s environment. However, these techniques can be time-consuming, require advanced training, and rely on the availability of extensive resources in the individual’s setting. Development of a brief functional assessment checklist would circumvent these difficulties and meet clinical needs for efficient assessment methods. The current study provides psychometric data for the Questions About Behavioral Function. These data include test–retest, inter-rater, and internal consistency.

Introduction

Functional assessment has proven useful for identifying environmental causes of maladaptive behaviors (Sturmey, 1996). Traditionally, experimental functional analyses were used, but they have proven to be very labor intensive. As a result, researchers have been searching for more efficient means of obtaining this type of data. Various scales have been published for some time, but unfortunately no psychometric data were collected (e.g., Functional Analysis Interview Form (O’Neil et al., 1990). Conversely, the Motivation Assessment Scale (Durand & Crimmins, 1988) has been studied but the psychometric qualities have proven to be inconsistent, and generally low inter-rater and internal consistency reliability coefficients have been reported Bihm et al 1991, Crawford et al 1992, Newton and Sturmey 1991, Sigafoos et al 1994, Spreat and Connelly 1996, Zarcone et al 1991. Nonetheless, these researchers made valuable contributions by developing first attempts at time-efficient functional analysis methods.

Given the potential value of a checklist to determine behavior function, further efforts at scale development seem warranted. As a result the Questions About Behavioral Function (QABF) (Matson & Vollmer, 1995) was designed for persons with mental retardation. Maintaining variables included in this scale are social attention, escape, tangible reinforcement, physical discomfort, and nonsocial reinforcement, and constitutes a broader range of variables than previously addressed by functional analysis checklists. In particular, under-investigated variables such as social avoidance (Taylor et al., 1994) and physical discomfort Carr 1994, Lowry and Sovner 1991 are included to make the scale more comprehensive than previous instruments.

Matson et al. (1996) presented the initial psychometric data for the QABF. Data were collected at a state institution for persons with mental retardation. Participants (N = 462) ranged in age from 13–86 and were predominantly in the profound range of mental retardation. The types of behaviors assessed varied but included typical behaviors such as SIB, aggression, and property destruction. Internal reliability and validity data were assessed. Coefficient alpha and Guttman split-half reliability coefficients were very acceptable (r = 0.86 and r = 0.91, respectively). An exploratory factor analysis with varimax rotation yielded five factors accounting for 74.5% of the variance and corresponding to the five hypothesized subscales. These initial psychometric data are promising, but further data are required to substantiate that the QABF is a reliable and valid tool for functional assessment.

The QABF has also been found to be successful in identifying the function of self-injury, aggression, or stereotypies in 84% of 398 people (Matson, Bamburg, Cherry, & Paclawskyj, 1999). Further, treatments developed from the QABF identified functions, resulted in significantly more effective treatments for these problems than treatments that were not tailored to antecedent functions identified with the QABF.

In the present research, we provide a range of psychometric data for the QABF. In experiment 1 we attempted to determine the stability of the instruments across rates and over time, inter-rater and test–retest reliability were assessed. Experiment 2 was designed to substantiate the initial factor solution identified by Matson et al. (1996), a second exploratory factor analysis was conducted.