Educational Technology Research Section

Teaching the Low-Performance
College Student: An Inexpensive
Behavioral Technology That Works

Roger D. Ray
Michelle Salomon
 

Author Notes
Roger D. Ray, Department of Psychology, Rollins College; Michelle Salomon, Orlando FL. A substantially different version of this paper was first prepared by Michelle Salomon and presented for consideration as an undergraduate honors thesis supervised by the senior author. Salomon is presently a graduate student in the Counseling and Human Development program at Rollins College. The author may be contacted via internet at: rdray@rollins.edu.
 
 

A ten year investigation of behavioral contingency contracting to improve student skills in a sophomore-level undergraduate course on Learning was conducted across 17 offerings of the course. Points of contrast between this contract system and previously reported contracted learning arrangements include a systematic effort to shape a broad array of desired student behaviors across a variety of relevant settings. A token-economy point system was used to implement contracts which had been collaboratively developed by the instructor and the class as an integral part of the course's content development process. Course engagement under contractual arrangements was introduced just after mid-term and was completely voluntary. The average midterm (pre-contract) grade of contracting students represented the 15th percentile of the class, suggesting the contract's appeal to those students in greatest need of its intents. Under contractual conditions, these students increased their test scores by an average of 17.3 percentage points for all tests in the course, including those which had been taken prior to contracting and had to be made up. In comparison, students opting not to contract improved only 4.8 percentage points on tests, despite starting at performance levels which clearly allowed for greater improvement. Contracting students' grades at course end averaged a full letter grade higher than their mid-term grades (from C- to B-), while non- contracting students showed no change between mid-term and final course grade levels (both a B). In contrast to much popular hype and moderate evidence regarding the positive effects of modem and expensive computerized technological innovations in education, behavioral contracting technologies involve very little cost and produce highly effective results.
 
 

Hardly a day goes by that colleagues aren't heard lamenting their poor fortune in having to teach the world's worst students. Whether at lunch, waiting for stragglers at committee meetings, or especially at faculty meetings where curricular issues are discussed, the conversation hasn't changed much over many years. Complaints about student shortcomings run the gamut from poor attendance, little or no evidence of interest in class (including a lack of sophisticated note-taking and an absence of verbal participation), a lack of effort in assigned reading (or, when certified that reading is attempted, too many deficiencies in reading comprehension skills), and either lethargy or moans of protest in response to all assignments meant to carry the student into libraries or social conversations on academic subjects. One even hears of instructors' efforts to address the lower 15-20% of their students by passive, if not active, encouragement for these students to drop the class.

Our institutional strategies for addressing the low-performance student seem to be either: (a) student intimidation until eventual elimination, rather than education; or (b) massive efforts at educational remediation through development of large professional support units which specialize in every deficiency imaginable, including centers for writing, quantitative reasoning, learning disabilities, study skill development, and career development. The financial costs implied in either of these approaches are enormous, whether in lost revenues or direct expenditures. There are also human costs borne by the low-performance student, who is often viewed as a failure or one who needs a lot of special attention to compete with peers.

It was in acknowledgment of such problems in teaching low performance students that the senior author began to search more earnestly for educational technologies that directly and effectively address student learning problems. While I believe, with many others, that some of the best of modern computer-based educational products are likely to help in encouraging better preparations and in facilitating more effective learning by students, I pause to reflect on what investment costs to institutions will be involved. While prices for today's computer technologies are continuing to drop, they are still relatively expensive. Computer technologies also continue to change rapidly, with hardware often becoming obsolete within increasingly shorter periods of time. Are all of our institutions of higher learning going to be capable of keeping pace with such requirements of capital investment, depreciation, and rapid replacement costs? Wouldn't it be nice to find an effective technological solution to student learning problems that was cheaper and more accessible to all faculty in all disciplines?

As it turns out, one such technology has been on our doorsteps for nearly three decades. It is not only much cheaper than electronic approaches, it is much easier for novice instructors to master and implement. Unfortunately, its simplicity is probably its downfall. It isn't glitzy. In fact, modern bandwagon movements towards cognitive sciences within psychology make it almost as passé; in appearance as yesterday's 286 PC or Mac Plus. But just as those machines still perform the task they were designed to perform, behavioral contingency management programs can still transform students' behavioral repertoires.

The following report is based on ten years of personal investigation into such a contingency management technology. The investigation was conducted across many offerings of a single course and in a relatively systematic fashion, when one considers the problems of doing empirical research within ongoing classroom environments. As noted above, the technology in question relies upon simple behavioral principles which stress the need to arrange straightforward reinforcement contingencies for desired student behaviors. What is unique about this program in relation to the dozens of other demonstrations and investigations of contingency management using behavioral contracts is the range of behaviors targeted, the broad specification of settings involved, and the mode of contract introduction and implementation. To more fully appreciate these subtle and important innovations, we will consider briefly some relevant literature. Following that, we will present empirical data generated by the investigation and reflect on their significance for addressing problems with low performance students.
 

Contingency Contracting in Education
One of the earliest and most systematic, albeit somewhat brief, statements on how to best apply behavioral contingency principles within a classroom setting was published by Homme (1970). While Homme's approach influenced numerous elementary and special education programs, he never seems to have captured the imagination of many college teachers, despite earlier publications by Malott and Svinicki (1969) and McMichael and Corey (1969) demonstrating highly effective applications of behavioral approaches at the college level. This may be due to the fact that Homme's book is very clearly aimed at elementary teachers. Most of its programmed instructional questions and many of its suggested reinforcement dynamics reflect applications for a young population. For college students, programs which offer "Do ten arithmetic problems correctly and then we will watch the first four minutes of this Popeye cartoon" (p. 20) are obviously inappropriate, even when chemistry is substituted for arithmetic and soap operas are substituted for cartoons. First of all, the simplicity of such an approach seems somewhat "beneath the dignity" we feel we should accord the assumed maturity we hope college students bring to the classroom (although that assumption is not always confirmed by student behaviors). But even beyond this impediment to considering contingency management strategies, college teachers tend to think in terms of managing subject matter, not classrooms. Perhaps this is why few college teachers are familiar with other inspiring demonstrations of the effective use of behavioral technologies (e.g., O'Leary & O'Leary, 1972), even if they do stress the elementary classroom.

Nevertheless, upon careful reading, Homme reviews virtually all of the fundamentals of good behavioral contracting, no matter what age student is involved. His ten basic rules for contingency contracting say it well:

Rule 1. Reward contracted behavior immediately.

Rule 2. Call for and reward small approximations in initial contracts.

Rule 3. Reward frequently with small amounts.

Rule 4. Call for and reward accomplishments rather than obedience.

Rule 5. Reward the performance after it occurs.

Rule 6. Be fair in contract requirements.

Rule 7. Be clear about the terms of the contract.

Rule 8. Be honest in the contract.

Rule 9. Be positive about the contract.

Rule 10. Use contracting as a method very systematically.

There is certainly nothing in these rules that suggests elementary school children are the only ones suited for behavioral contingency contracting

The major difference between Homme's approach and the few college classroom demonstrations cited earlier lies in their assumptions of initial behavioral repertory. For example, Malott and Svinicki's implementation of contingencies for the introductory psychology student makes little or no effort to shape behaviors where there are deficits. Like most college programs stressing personalized systems of instruction (PSI) (e.g., Ferster, 1968; Keller, 1968; Johnston & Pennypacker, 1971; Binder, 1988; West, Young, & Spooner, 1990), only the end behavior of oral or written test-taking was targeted for reinforcement. While such behaviors as study time have been investigated (Born & Davis, 1974), they have not typically been targeted directly for contingency management [see Bristol & Sloane (1974) as a notable exception].

Unfortunately, the really low performance student may not yet have developed the necessary study skills, as opposed to merely not wanting to invest the necessary time, to easily meet typical mastery and/or fluency PSI testing standards. Even though most available data clearly indicate many introductory students can respond to personalized instruction by accomplishing the requisite mastery and/or fluency of the material, they may need either several test retakes and make-ups or intensive peer or supervisor management to do so. It is under these circumstances that we should recognize the educational relevance of some very important principles discovered by modern behavioral analysis.

Based on his classic experiments investigating how cats learn to escape from puzzle boxes, Thorndike became one of the first researchers to articulate the importance of arranging consequence for behavior (his well known Law of Effect). However, it was B. F. Skinner who recognized and improved upon an important procedural element of Thorndike's research. Thorndike always placed a cat into a puzzle box and waited for trial and error processes to result in the animal learning the appropriate escape response. Skinner, on the other hand, saw the possibility of managing this acquisition process by rewarding any and all behaviors which may only approximate the final required responses. This not only made the learning acquisition process much more deliberate, it accelerated it considerably. This process is called response shaping (e.g., Skinner, 1951; Catania, 1992), and it relies upon a general task analysis of all component behaviors viewed as a continuum which begins with what a subject can do and, by specifying a series of successive response approximations, ends with the behavior that the learning manager (teacher) desires.

Thus, shaping procedures arrange for reinforcement contingencies to be sufficiently dynamic as to begin by reinforcing behavior that is most likely for the subject to manifest. As this response class increases in likelihood, other related response classes also increase due to the phenomenon of response induction (Catania, 1992). This implies that one may gradually shift reinforcement contingencies to these related responses along the continuum of successively approximating behaviors until the desired, but initially absent, behavior is created from the dynamic shifts in reinforcement contingencies.

Translated into the above student repertoire problem with regard to an absence of effective study skills, shaping principles imply that, if we target the requisite component study behaviors for developmental shaping, test performance might well take care of itself and a greater generalization of study skills applied to other learning environments might be expected.

This is, of course, what Homme had in mind when he encouraged teachers to aspire to shape students to a complete independence of manager- controlled contracting altogether (which really involves the process of stimulus fading -- another successive approximation process which targets the generalization of learned behaviors into a wider variety of stimulus control settings than the setting in which shaping initially took place). He states:

Interestingly, the most common college use of contingency contracting beyond PSI (which almost always is applied in introductory courses) seems to start where Homme has aspired to bring his students at his transitional contract's conclusion That is, they fail to recognize any developmental continuum between beginners and experts. Thus, beyond introductory students, the most frequently used form of multiple-behavior contracting at the upper-division college level seems to target the Independent study" course format (Barlow, 1974; Berte, 1975; Chickering, 1975; Duley, 1975; Feeny & Riley, 1975; Wald, 1978). While this very broad contracting technique is highly effective for well-motivated and self-directed students, it doesn't seem a likely candidate for developing skills in students populating the lower 15-20% of most upper-division college courses.

It was within this context that the senior author began, during the middle 1970s and through the present, to explore contracted learning of a different sort. By the early 1980s, exploratory variations became somewhat solidified and an effort was made to collect data more systematically on the educational effects noted informally in previous years. The following is a report of what was discovered from these more systematic efforts.