Academic journal article The Behavior Analyst Today

Comparison of the Relative Effectiveness of Different Kinds of Reinforcers: A PEM Approach

Academic journal article The Behavior Analyst Today

Comparison of the Relative Effectiveness of Different Kinds of Reinforcers: A PEM Approach

Article excerpt


In educational settings, it is desired that all students know the value of knowledge and skills and are intrinsically motivated to learn, but it is rarely the case. Therefore, teachers in school settings often have to employ extrinsic reinforcers to motivate students to learn. After the frequent use of reinforcement on good school performance, it can be expected that the act of learning will become a conditioned response, and after the teacher applies intermittent reinforcement or the student acquires natural consequence of success in education or carrier, it is hoped that the learning can become intrinsically motivated with less extrinsic reinforcement.

Reinforcement can be classified into four kinds: (a) positive reinforcement (giving positive reinforcer), (b) punishment (giving negative reinforcer), (c) punishment (withdrawing positive reinforcer), and (d) negative reinforcement (withdrawing negative reinforcer). The present study concentrated mostly on the comparison of the effectiveness of positive reinforcers including edible foods, tangible objects, activities, and tokens. Less attention is paid to the effectiveness of negative (aversive) reinforcers, but the mean effect sizes of punishment were presented for the purpose of comparison.

There have been many studies reporting success in the use of primary reinforcers to modify the behavior of participants (e.g., Forness, Kavale, Blum, and Lloyd, 1997; Kern, Ringdahl, Hilt, and Sterling-Turner, 2001; Osborne, 1969; and Williams, Koegel, and Egel, 1981).

Cameron and Pierce (1994) conducted a meta-analysis to address the issue of whether extrinsic reinforcement is harmful to the intrinsic motivation and found that rewards given for task completion or for quality of performance are not detrimental to intrinsic motivation.

According to the principle of behavior modification, in order to expect a desirable behavior to happen in the future, three conditions must be fulfilled as follows: a discriminative stimulus must be present; there must be a contingency for reinforcement of the target behavior, and the reinforcer must be able to satisfy the need of the individual. Researchers in behavior analysis have paid more attention to the third condition recently. Neef and Lutz (2001) found that the effect of more preferred reinforcers was higher than that of less preferred reinforcers. The results of Pace, Ivancic, Edwards, Iwata, and Page's (1985) study confirmed that the success of reinforcement depends on the selection of suitable reinforcement schedules and contingencies. Glynn (1970) found that the effect of self-determined and experimenter-determined token intervention on the learning of history and geography material was superior to that of chance-determined and no-token interventions.

It may be alternatively hypothesized that the effect size of an intervention would be larger when the reinforcer can better meet the needs of the participant and serve as a mechanism to increase his or her motivation.

The tool used to measure the effectiveness of different reinforcers was the percentage of data points exceeding the median of baseline phase (PEM) approach (Ma, 2006). By far, the most widely used method for measuring the effect size from single-case experimental designs is the percentage of nonoverlapping data (PND) approach proposed by Mastropieri and Scruggs (1985-1986). The strength and weakness of the PEM approach and its superiority over the PND method has been discussed by Ma (2006) and empirically confirmed by Gao and Ma (2006); Chen and Ma (2007); Ma (2009) and Preston and Carter (2009). Therefore, it was decided to apply the PEM approach to compare the relative effectiveness of different reinforcers used in the field of behavior modification.


Procedures for Locating Studies

The single -case experimental studies investigating the effect of reinforcers analyzed in this synthesis were obtained through a computer-assisted search of the relevant databases, including EBSCOhost, ERIC, and ProQuest. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.