Best Practices in Utilizing the Changing Criterion Design

Abstract

The changing criterion design (CCD) has been a recognized format of single-case research for four decades. Published examples of the CCD have been limited and the structure of the design used in the literature has varied to a degree that might engender confusion. This review examines the structure of CCD studies published to date to identify prior implementation practices and identify best practices for future use.Keywords Single-case . Changing criterion design . Best practicesThe changing criterion design (CCD) utilizes step-wise benchmarks for manipulating a dimension (i.e., accuracy, frequency, duration, latency, or magnitude) of a single behavior already present in an individual's repertoire (Cooper et al. 2007). The design has been used with behaviors where an immediate, considerable increase or decrease may be difficult to achieve or undesired; therefore, gradual shifts toward a desired goal are applied.The CCD was first described in two books published in 1971: one a short description in a book by Hall, and the other a brief study reported by Weis and Hall which used the design to reduce daily cigarette smoking by a college student. A few subsequent studies utilized this design (i.e., Brown et al. 1974;Wallace and Davis 1974) prior to the seminal article by Hartmann and Hall (1976), which described the CCD in more detail and noted its similarity to a procedural description by Sidman (1960, pp. 254-256) in reference to the evaluation of irreversible behavior. Hartmann and Hall saw the CCD as a type of multiple baseline design that could be used with a shaping procedure. In response to the CCD as a multiple baseline design, Cooper et al. (2007) point out that the behavior being evaluated in a CCD is intentionally interdependent (unlike a multiple baseline design) which can lead to an exaggerated perception of experimental control. Conceptualizing this as an extension of the multiple baseline design adds confusion. Consequently, the CCD is best suited for the acceleration and deceleration of a behavior, but is versatile in the behaviors and populations it can be applied to. Despite this, the CCD has only rarely been mentioned as a research or intervention tool since Hartmann and Hall's article.Though not as experimentally robust as other designs (e.g., the multiple baseline design, which demonstrates experimental control through staggered, independent baselines), the CCD does achieve acceptable control through baseline logic and appropriate manipulation of design structure (see Table 1 for a simplified presentation of baseline logic of the CCD). With the CCD, a baseline serves as a measurement between the absence and presence of the independent variable, as do each criterion and its subsequent criterion. Therefore, when stable responding occurs at each step, (1) predictions about future responding can be made with each criterion, and (2) replications of intervention effect are observed repeatedly. Verification can be accomplished by varying the phase length or by reverting briefly to a previous cr...

Descriptive and normative ethical behavior appear to be functionally distinct

“…Another method to determine the influence of an independent variable manipulation is to conduct a parametric analysis (Perone & Hursh, 2013). Examples of parametric manipulations include changing‐criterion designs (e.g., Klein et al, 2017), dose–response curves in behavioral pharmacology (e.g., Branch, 1984; Calabrese, 2008), varied combinations of concurrent variable‐interval schedules of reinforcement (e.g., Baum, 1974; Herrnstein, 1970), and changing delays or probabilities in discounting research (e.g., Mazur, 1987; Rachlin et al, 1991). Parametric analyses might help the study of ethical decision‐making using a within‐subject experimental framework.…”

Randomized Single-Case Experimental Designs in Healthcare Research: What, Why, and How?

“…Barlow et al argued that the lack of treatment withdrawal throughout the course of the study is a major strength of the changing criterion design and that this feature makes the design especially attractive for clinical studies, for example in the treatment of dangerous behaviors such as self-harming [14]. Klein et al elaborated that the changing criterion design is especially valuable in situations, in which an immediate, abrupt increase or decrease in a dependent variable may be difficult to achieve or undesirable [32]. Furthermore, the stepwise changes in the frequency of the dependent variable may facilitate habitual changes in a subject's behavior [66].…”

“…Klein et al recommended incorporating "mini-reversals" into the changing criterion design [32]. Such a reversal entails reverting to a previous criterion.…”

“…If ethically and practically feasible, such reversals can greatly strengthen the demonstration of experimental control over the dependent variable. Klein et al further recommended that at least three criterion changes should be implemented for repeated observation of intervention effectiveness [32]. Regarding the minimum number of data points required per phase, clear guidelines are still lacking for the changing criterion design.…”

The Effects of Mentor Instruction on Teaching Visual Supports to Novice, Special Education Teachers

“…A classic changing criterion design administers the independent variable in steps with the goal of behaviour change in an accelerating or decelerating fashion (Klein, Houlihan, Vincent, & Panahon, 2015;McDougall, Hawkins, Brady, & Jenkins, 2006). A changing criterion design must have a treatment period long enough for the treatment to re-stabilize (Gast & Ledford, 2014, p. 347;Hartmann & Hall, 1976).…”