i haviors of children who function productively in a regular-classroom setting are ordinarily under the control of a wide variety of generalized reinforcers appropriate to that setting. Solving problems, completing assignments, and achieving success at academic endeavors generally function as powerful reinforcing events which maintain academic behavior. Such behaviors are further strengthened by praise, often accompanying appropriate academic behavior, administered by parents or teachers.1 1 It is not surprising, then, that the behavior of most school children is responsive to traditional educational procedures and methods even when no systematic efforts are directed toward gaining behavioral control.The °'acting-out&dquo; child, however, with all his accompanying academic disabilities, often misses out on these avenues of positive reinforcement which are common to the educational setting. Reinforcements for appropriate academic behavior are rarely available for him. The low probability of success or praise being associated with his academic behavior decreases the frequency of appropriate academic behavior in a spiraling process; e.g., the fewer the reinforcements, the less academic work attempted; the less work attempted the fewer the reinforcements. In addition, many of the social behaviors demonstrated by such children are aversive and thereby preclude or severely limit the probability of the child being positively reinforced by teachers or peers.When social reinforcement is available, there is some evidence which suggests that adult praise is aversive 1 R.

E FFECTIVE contingency management has become an increasingly prominent feature of human behavior management in a wide variety of natural settings. A necessary prerequisite to such programs is the identification of variables relevant to behavioral change . Such identification may entail a variety of operations, but the promotion of reliable and empirically valid measurement is an essential feature of all approaches to contingency management of behaviors. The purpose of this paper is to identify and to discuss briefly some relevant dimensions of interval recording procedures. Some socially significant behaviors result in products that may be measured after the behavior has occurred. Arithmetic computation and spelling are examples of academic behavior that can most efficiently be evaluated by counting the product. Many other classes of behavior, however, do not result in tangible products and must be measured as they occur. Behavior involving one person, such as crying, rocking, and thumb sucking, as well as various forms of social interaction, such as cooperation, talking, listening, and fighting, are examples of such behavior. Behavior possessing this general characteristic is commonly observed in a wide variety of natural settings, including the home, classroom, playground, and institutional ward. Selecting behavior for observation Selection of appropriate behavioral units is often facilitated by anecdotal recording of subject behavior under a variety of conditions. The anecdotal records consist of written descriptions of all relevant behavioral occurrences. The behavioral units selected for at NANYANG TECH UNIV LIBRARY on May 31, 2015 isc.sagepub.com Downloaded from

Hand-held digital counters are widely used to count Counter" fits comfortably in the palm of the hand and, behavior. The simultaneous recording of multiple be-after some practice, can be operated rapidly and relihavioral categories on a single subject or a single be-ably using the fingers of the same hand. This counter havioral category for multiple subjects necessitates the has been operated more than 5000 times over a threeuse of a counter with more than one channel. A com-month period without evidence of malfunction. Limitpact (5.25 by 1.5-in.) five-channel manual counter ing characteristics are the audible click that accompan-("Multi Counter" #99C9031) is available from Lafay-ies each operation and the fact that the five, three-digit ette Radio and Electronics, 111 Jerico Turnpike, Syos-channels cannot be reset independently; the reset knob set, L.I., New York 11791, for $9.50. The "Multi resets all channels simultaneously.