Characteristics of behavior rating scales: implications for practice in assessment and behavioral support.

Abstract: This study examined the structure of items on commonly used behavior rating scales in order to determine their usefulness in planning and monitoring positive interventions. Fourteen forms from 9 published behavior rating scales commonly used in research and practice were selected. The items on each scale were categorized as addressing a positive action, a negative action, a lack of positive action, or a lack of negative action. Ten of the 14 scales were composed of a majority of negative action questions, which are not useful for assessing positive behaviors. Thirteen included lack-of-action questions, which indicates they do not address observable, measurable behaviors. Implications for the selection and use of behavior rating scales within the context of a proactive model of social intervention, as well as directions for future research, are discussed.

**********

Behavior rating scales are a common tool for assessing social skills and emotional functioning for both research and practice in educational settings (Heckaman, Conroy, East, & Chait, 2000). They (a) provide quantifiable information, which can be held to standards of reliability and validity; (b) provide systematically organized information; (c) are efficient to complete and score; (d) include normative data, which allow comparison of individual behaviors to that of large groups; and (e) can be used to compare ratings of different respondents or across settings (McConaughy & Ritter, 2002). These attributes contribute to the popularity of behavior rating scales. However, the popularity of a tool can be unrelated to its validity.

Assessment tools are only valid in relation to a specific purpose. Behavior rating scales are most commonly used in the diagnosis of behavioral disorders and the determination of entitlement for special programs (Hardman, Drew, & Egan, 2002; McConaughy & Ritter, 2002; McGinnis, Kiraly, & Smith, 1984). They are also used to plan behavioral interventions and support plans (Burks, 1977; Gresham & Elliott, 1990; McCarney, 1994) and to monitor behavior (i.e., the effectiveness of the interventions or support plans) over time (Achenbach & McConaughy, 1998; Conners, 1997; Kamphaus & Reynolds, 1998; Wilczenski & Ferguson, 2001; Wilson & Bullock, 1989). For example, in a survey of pediatricians, HaileMariam, Bradley-Johnson, and Johnson (2002) found that pediatricians receive rating scales for diagnosis of attention-deficit/hyperactivity disorder and evaluation of treatment four times as often as systematic direct observations of the student's behavior. These pediatricians also reported that rating scales were the most preferred method of evaluating treatment (nearly twice as many responded that they preferred rating scales to systematic direct observations).

The validity of behavior rating scales for the purposes of diagnosis and making entitlement decisions has been documented in the research literature (e.g., Martin, 1988; McConaughy & Achenbach, 1989); however, the use of these scales in planning interventions and monitoring student progress has not (Heckaman et al., 2000). The primary difference between these two types of uses is that diagnosis and entitlement decisions are normative; that is, the student is being compared to others to determine where on a continuum of social acceptability his or her behavior lies. Behavior rating scales are accurate enough to make dichotomous decisions, such as "Is this student's behavior 'normal' or not?" In some cases, additional criteria such as determination of being "at risk" for a particular problem may also be included between the two extremes.

In developing individual intervention plans, these dichotomous data may be adequate as a screening to determine whether a certain behavior should be the focus of intervention. If the student's behavior falls within the normal range, there is no need to intervene. If it falls outside the normal range, additional assessment is necessary to determine details such as the function, setting, frequency, and topography of the behavior. In monitoring interventions, however, it is not adequate to have dichotomous information. At the end of each recording period, more information is needed other than an answer to the question, "Did the intervention work?" Without more sensitive measures, there is no indication if the intervention is (a) working, but not yet fully successful; (b) not working at all; or (c) actually making the situation worse.

As such, several considerations need to be taken into account when deciding if a specific behavior rating scale is appropriate for a certain purpose. For example, legislation requires the development of proactive intervention plans (i.e., increasing positive behavior) for some students (Individuals with Disabilities Education Act [IDEA], 1997, 1999). In addition, best practice in intervention planning does not focus exclusively on reducing negative behaviors but rather on replacing them with positive behaviors (Batsche & Knoff, 1995; Gresham, 2002). If a behavior rating scale is being used to assess an increase in the replacement (i.e., positive) behavior, alignment can be lost if the items on the scale do not address positive behaviors. That is, if the questions focus on the occurrence of negative behaviors, any reduction would be rated as improvement. Therefore, if a previously hyperactive student becomes lethargic or depressed, a rating scale may actually indicate that the student's behavior has improved. In order to link the assessment to the intervention plan, the behavior rating scale must assess positive behaviors that can replace the negative behaviors and therefore be used as outcome goals.

With this in mind, it is worth noting that there are situations where the reduction of negative behaviors is an appropriate goal. Any behavior that may result in harm to the student or another person (e.g., throwing chairs) must be reduced. In such cases it is not enough to focus exclusively on increasing positive replacement behaviors. These cases require that the function of the negative behavior be determined and that potential positive replacement behaviors be derived from that analysis of the function. The resulting intervention would simultaneously address reducing the negative behavior and increasing the positive replacement behavior. For example, if the function of throwing a chair is to escape a frustrating situation or task, an appropriate goal might be to decrease chair throwing with a simultaneous goal of increasing the occurrence of the student telling the teacher when he or she is frustrated. Different types of …