Supplemental Instruction

Gen M. Ramirez
California State University, Long Beach
 

Introduction and rationale

 After 10-15 years in which the trend in academic support had been to focus on high-risk students, higher education in the 1980's began more critically examining the established approaches toward working with underpreparation and underachievement.  Except for developmental courses, whose prolongation was defined by the academic calendar, most interventions with struggling students remained sporadic and brief.  Students used the services of learning centers (study skills instruction, tutors) when they, their instructors, or their advisors perceived an immediate need, and the services rendered generally were focused accordingly.  With the exception of study skills ó if the instruction was applied ó these resources often got students through the conscious bottleneck and produced success in the course or assignment involved, but they became necessary once again as each subsequent difficulty arose.  The goal of learning assistance and of its multiple components, including tutoring, is usually seen by the practitioners as equipping students to be independent, successful learners; its perception and use by students, however, is typically so narrow and fragmentary that that goal cannot always be achieved.

 As summarized by Martin et al. (1982), the traditional approaches to servicing high risk students became questionable in the light of four arguments:

1. the insufficiency of standardized testing to accurately identify high-risk students (students may do better or more poorly than what test scores would predict)

2. the impossibility of remediating skill deficits quickly  enough to avert a failure pattern (attrition begins some six weeks into a semester)

3. the reluctance of high-risk students to acknowledge difficulties and to seek assistance initially or consistently

4. the stigmatization of learning centers as remedial programs because of their aggressive outreach to high-risk students

 In the 1980's, Missouri educators at both the university and community college levels began to disseminate information about their innovative approaches in response to the above and similar factors.  Their focus was no longer on high risk students but instead on high risk courses, the latter defined as "those traditionally difficult, entry-level courses wherein student D and F rates and withdrawals exceed 30 percent of course registrants."  (Blanc et al., 1983)  The intervention they designed to impact attrition and failure rates was designated "Supplemental Instruction."  It included the presentation of review sessions given repeatedly

[page 78]

throughout the week whose emphasis was on content review combined with the "model[ing] of thinking and languaging behavior" appropriate for the field while student competency in reading, reasoning and study skills was also increased.  Attendance was voluntary.  Leaders, presented as being themselves "students of the subject," prepared by attending course lectures and completing assigned readings; they offered the review sessions three or four times weekly for the convenience of participants.

 According to its developers, SI distinguished itself from other intervention modes in two ways:  the shift in emphasis from high-risk students to high-risk courses, and the organization of services on an outreach rather than a drop-in basis (i.e., delivery of instruction outside the Student Learning Center) in classrooms and in association with established curriculum.

 Perhaps the most signficant distinctive of the program is its foundation on a cognitive development theory.  SI is one of the few intervention models which, practiced properly, recognizes the critical need for developing the learning and thinking skills basic to content mastery.  In that respect, SI endorses the "new assumption about the purpose of education" articulated by such scholars as Malcolm S. Knowles (1981), that being to produce autonomous lifelong learners.  The primary objective of the program is skill development and, only secondarily, the review of a common subject matter content to which these skills are applied.

 Blanc et al. (1983) note, very appropriately, that a significant proportion of entering freshmen (their estimate is 50%) lack reasoning skills at the formal (abstract) operational level defined by Piaget and Inhelder.  Those still working at the concrete level struggle to process unfamiliar information presented through the abstract media of lecture and textbooks, focusing on detail and surface content rather than on synthesis or application of concepts.  Because basic courses sometimes evaluate memorization rather than the assimilation of concepts, students operating at the preformal level might do well in these courses if they identify known "pegs" to which they can anchor information or find other effective memory techniques.  However, they will fail wherever they are required to demonstrate an integration or application of their knowledge at introductory or advanced levels.

A.  Program Description

  The SI program model is designed to give students the opportunity to address what they presume to be their content-centered needs, but in fact provides a strong emphasis on developing higher level cognitive skills in those who lack them.  Although course and discipline materials serve as a common vehicle for skill instruction and development in problem or deficient areas,  SI leaders give priority attention to reasoning and questioning skills.  They're trained to assess the quality of student questions and responses to identify those whose thinking levels would limit their mastery of new concepts.  At the same time, they pursue the development of traditional study skills such as study reading, notetaking, memory, test-taking, and time management.

 As described in the model, leaders range from professional staff in the Learning Center to student assistants.  Their basic operational guidelines are six:

[page 79]

1. attendance at every class lecture (or emergency replacement by a substitute)

2. presentation of every scheduled SI session (modeling consistency and providing students the security of regularity)

3. a supportive posture toward the professor (helping students cope and adapt, rather than validating criticism)

4. encouragement of all questions, serious attention to them, and modeling of appropriate ways to think about course content

5. provision of feedback about student difficulties to the professor whenever opportunities arise

6. availability to students to discuss course material beyond the SI classroom or meeting time.

B.  Evaluation

 The University of Missouri-Kansas City, which initiated its SI offerings in 1978, disseminated early years' results in 1983.  That they attempted an impact analysis of the program is itself unusual among such interventions, because the complexity of factors that can affect student outcomes is usually considered an impediment to accurate measurement.  However, the attachment of SI sessions to common performance obligations and to measured achievement allows outcomes to be examined in ways not usually possible with the diverse populations served by Learning Centers.

 Their findings (Blanc, et al., 1983), for students who attended from 1-25 sessions (with a program average of 6.5 sessions per semester), showed significant improvement attributable to SI.  In order to account for motivation as a factor in SI participation, they compared SI students to two control groups (one "motivational" population reporting high interest but unable to participate due exclusively to scheduling conflicts and the other the remaining nonparticipants).  As indicated by the following chart, SI students did significantly better with respect to GPA in the target course, semester GPA, and their attrition ("D", "F", "W") rates in the class.

  [page 80]

They also found reenrollment at the university to be significantly higher for previous SI participants (73.2%) than for non-participants (60.0%).  Though actual numbers varied considerably, the proportional improvements in both grades received and reenrollment rates for high risk and high achieving subgroups were comparable.

[page 81]

The tracking of an individual course (economics) from 1978-1980 showed a gradual decrease in the attrition rate paralleling a gradual increase in SI participation:

 A replication study was undertaken at Anne Arundel Community College (Maryland) in fall 1986 with a history course.  With that group, the sample size was much smaller, but the results were consistent with the pattern established at the university.  They were more dramatic in that for both the target course and the entire semester's grades, SI meant the difference between success (2.5 in history, 2.6 overall) and academic probation level performance for non-participants (1.6 both in history and overall).  The attrition rate in the course for SI participants was only 16% compared with 55% without it (Wolfe, 1987).

 Any number of factors might account for the beneficial impact of SI programming.  The UMKC group have identified the following four as influencing higher levels of student academic performance:

[page 82]

1. a proactive approach where the service precedes difficulties

2. direct attachment to specific courses so that skills instruction has immediate application

3. student perception of the program as enrichment (for all students) rather than remediation (for the underprepared)

4. promotion of high degrees of student interaction and support (that produces peer study groups and facilitates mainstreaming non-traditional students).

 Program developers point out, as an additional advantage, that SI is more economical to administer than tutoring, the traditional service it most closely approximates.  At UMKC, it was estimated that the cost of servicing 106 students by one leader during two semesters was $12 per student, substantially more service at lower cost than what one individual could have provided these students in a conventional program.  Martin and Blanc (1981) describe the program as "cost-effective, in terms of both staff allocation and budget allocation."

C.  Adaptation of the Model

 Both for its philosophy and for the results it's been found to produce, SI is an ideal program model.  Based on very sound principles of cognitive development, if instruction carries this emphasis, the changes not only in the target course, but in general student achievement certainly qualify it as the most economical and most effective known approach to academic support.  Beyond that, it's a democratic program, i.e., one that offers something for virtually any college student, as distinct from the traditional developmental or remedial interventions provided for students who are defined by an institution as deficient or disadvantaged; SI fails to apply that stigma.

 SI is unquestionably a model that deserves wider recognition than it has received.  Many campuses, without any knowledge of the SI model, have developed adjunct and paired courses that serve a similar purpose through a more formal arrangement.  But for many, tutoring and other conventional approaches to academically supporting students, which continue to have their place, are the only instructional support services available.

1. Rationale:

 Although staff at hundreds of institutions have been trained to implement it, the number of campuses actively providing SI programs is still a very small proportion of the nation's colleges.  This is due in part to difficulties encountered when local needs or constraints do not seem to be effectively served by the model.  If a large, urban state university in California is typical, the problem lies not in the educational model, but in the noneducational campus realities that preclude its effective adoption.  The discussion that follows is offered as an example of the adaptation process required to tailor the prototype to local objectives or obligations.  For community colleges, residential campuses, or other types of institutions, there is a different constellation of special considerations on which an adaptation must rest.

[page 83]

 In 1985, a campus Trio director attended SI training at UMKC and returned to campus to initiate a local effort parallel to the traditional tutorial offerings.  During each semester of 1985-86, five high-risk classes were chosen for SI adjuncts supported by federal (Trio) funds:  economics, calculus, physics, government, and biology.  These were either general education courses or prerequisites to the campus's most popular majors at that time (business and engineering).  For those who participated, results were very good, equal to those reported in the UMKC data.  But after three semesters of introducing the model, those to whom administrative responsibility had been given recognized inherent problems that demanded attention.

 Funding was a significant factor.  As is typical of many institutions, tutorial funds were not general but categorical.  This meant that not only was the primary target population for SI the supporting program's own students, but their students had to be involved in SI sessions in order for the expenditure of funds to be justified.  The way SI advertised its services was through the target high-risk course, so every student in the class was offered access.  In practice, attendance at SI sessions was sporadic and inconsistent.  But the greatest response was from traditional students, not from the categorical program students, and the trickle of participants swelled dramatically just before each exam.  For the faithful few attending regularly, the skill development was aborted by the temporary throng seeking only help with content and refusing anything else.  And because the majority were not categorical students at all, many scheduled sessions were paid for questionably; furthermore, attendance patterns raised the issue of whether leaders should be paid for time committed to SI, even if they had no students present.

 From a legal perspective, the funding issues had to be resolved.  Pragmatically, because of both funding and the campus concern for assisting underprepared and under-achieving students, attendance questions required new strategies.  And politically, administrators faced the all-too-common reluctance of advisors to accept and endorse a departure from the well-established tutorial model:  despite the move toward SI and drop-in tutoring for general education courses, they continued to advocate a return to scheduled weekly appointments for all of their students requesting help in any course they might be taking.

2. Adapted model:

 The result of these deliberations was an adaptation of the SI model to meet local needs.  The task, as defined, was to effectively service high risk students enrolled in high risk courses but, like the UMKC model, achieve a mix of that target group with traditional students, including high achievers.

 So in Fall 1987, 13 sections of SI 060 were introduced as a one-unit non-baccalaureate adjunct, each section attached to a designated high risk target course.  That single element successfully addressed the attendance problem.  One reason categorical students hadn't participated consistently or at all was their failure to make provision for these sessions in their schedules.  The SI class was scheduled to meet either the hour immediately following or immediately preceding the target course; students enrolled for it and, thereby, made a time commitment to attend.  Although administrators could control the availability of seats in SI offerings, it was critical that students get into the proper section of the target course; those in participating academic assistance programs (EOP, SSS/Trio, minority engineering and busi-

[page 84]

ness, student athletes) whose advisors were directing them into an SI section received priority enrollment.  This provision was a practical necessity, but it also actively involved program advisors in placing students into SI sections during advisement.  Listing SI courses in the Schedule of Classes allows any student able to register in the correct target course to enroll simultanously in SI if he/she desires.

 Enrollment in SI courses now averages 15-35;  the number of target courses supported each semester has grown from the original five to about 35.  With semesterly fluctuations, academic assistance program students average 50-60% of enrollment.

 Besides student commitment, there are some other practical benefits to this arrangement.  Although non-baccalaureate, this course gives students studyload credit toward financial aid or other full-time enrollment obligations.  As a course, it also generates FTE and therefore funding; SI sections are now taught by senior tutors (funded by categorical state and federal allocations) and by teaching assistants (paid through the university's instructional funds).  It is significant that the campus has taken some fiscal responsibility for offering a service originally limited to soft monies for a prescribed population.

 To select target courses readily, an attempt was made to isolate some of the characteristics of a high risk course on this particular campus.  Though over time many of the same sections became regular target classes (due to known high risk, the supportive posture of instructors and the preference of advisors in special assistance programs), the following elements served initially to aid target course selection:

1. large lectures, minimizing faculty-student interaction

2. courses found difficult based on their technical nature (such as the sciences) or based on the unfamiliarity of our target population with these disciplines (economics, advanced math).

Final grades in target classes served as evidence that those chosen were truly high risk.  In Fall 1990, for example, 13 of 34 courses analyzed (38%) had average grades for non-SI students below 2.0, and in Spring 1991, 15 of 30 courses (50%) had probationary averages for non-SI enrollees.  Using UMKC's criteria for high risk courses, these target sections definitely exceed the 30% "D", "F", & "W" rate.

 Though the SI courses carry one unit of studyload credit, they actually meet the same number of hours as the target class and on the same days.  Regularizing the time devoted to this activity to match actual course time was desirable conceptually:  it got students into a pattern of reviewing and processing course content directly before or after each lecture.  It was also a necessity for working effectively with numbers far greater than the traditional SI study group of four to eight.

3. Staffing:

 The need to use only student leaders and no professional LAC staff has carried its own complications.  It was a fiscal necessity, because fully-occupied Learning Skills

[page 84]

Professionals were not available, and augmenting their numbers was not an option.  The Learning Skills and Writing Center staff have always been available to SI instructors for classroom presentations or consultation just as they are to any university faculty member, but university students alone have filled the ranks as SI leaders.

 Student SI Leaders have been recruited through a variety of networks including the recommendations of target course faculty and the corps of outstanding tutors within the LAC.  They are required to hold GPA's of 3.0 or higher in their discipline and overall, present three letters of recommendation from faculty they've worked with, and support the policies and procedures of the SI program.  Those selected receive extensive presemester training on study skills and classroom techniques, as well as periodic inservices during the semester and clinical supervision while on the job.  Their assignment involves attending all target lectures, preparing for and conducting all scheduled SI class meetings, holding an office hour each week, meeting as scheduled with supervisor(s), reporting student attendance and performance (on target course exams), and administering mid-term (program) and final (university) faculty evaluations to their classes.  They also assign grades of Credit/No Credit to SI enrollees.

 As a group, these individuals are very competent in their disciplines, but their consciousness of their own or of model study skills varies.  Some have a strong sense of how they study and what students need to know in order to really master the content;  their objectives are as broadly based as the SI model's (cognitive development).  Others, despite the study skills training imparted to them, would define their objectives as helping students get good grades in the target course.  They find it difficult to abandon their own discipline-priority enough to superimpose a study skills approach on the curriculum.  Feeling pressured to review and work with every detail of the target professor's lectures, they focus so heavily on content that students may do well in the course, but remain at the preformal stage of cognitive development despite their participation in SI.  So while the involvement of student leaders serves many purposes, it places an additional supervisory demand on the program administration to assure maximum effectiveness.

4. Program evaluation:

 The enrollment of the approximately 35 SI sections each semester is about 600 students, with an average of 50-60% of them coming from academic assistance programs.  To some extent, this skews the distribution toward the lower end of the UMKC population group studied, where the benefit of SI often carried otherwise probationary students into a satisfactory grade range.  Similarly, in some of this university's courses, the difference between SI participation and non-participation is a clean passing grade of "C" rather than a probationary level grade:

Cases like these are more likely the result of the inherent level of difficulty of the course (or of the target professor's teaching), but in some cases they can be a consequence of student underpreparation.

 In another group of courses, there are some dramatic improvements in grade outcome for those availing themselves of the SI course:

This pattern was found to be more typical of target courses in which the average grade is a solid "C" or above, where reasonably accessible instruction is combined with good SI intervention.

 And in still another profile, the greatest difference attributable to SI is the retention of students within a target course that has tremendous attrition rates.  Introductory Biology (BIOL 200), for example, began with an enrollment of 180; of those, only 97 (54%) completed the course and received a grade.  Only 55 of the 137 students not enrolled in SI (40%) persisted to the end.  But 42 of the original 43 SI students (98%) completed the course.  The average course grade was 1.82, with SI students doing only minimally better, but their willingness to continue their effort in the face of discouraging grades and the obvious attrition rates around them was dramatic.  Thus even in a very high risk class, the SI contribution is significant, though not as dramatic in improving GPA.

 It must also be noted that not every SI class population did better overall than the performance of the target class as a whole.  In some instances, they did less well than the target group.  Those anomalies are logically the result of levels of underpreparation or lack of effort on the part of the SI students or an ineffective approach by the SI Leader.

 Additionally, differences became apparent in how diverse subpopulations perform with the benefit of SI adjuncts.

 If 1990-91 is a typical year, freshmen do better than more advanced/experienced students.  For example, freshmen collectively in target classes averaged (a probationary) 1.94

[page 87]

overall, but 2.33 in SI (+0.39).  Juniors, by comparison, earned a GPA of 2.19 overall, with modest improvement over that--2.33--with the benefit of SI (+0.14).  It was assumed in selecting a great majority of 100-level general education courses to receive SI adjuncts that student need at that level was the greatest and that introducing them to good study skills early in their enrollment had the greatest potential payoff.  These data seem to support the first hypothesis.  The longitudinal impact of SI on retention and performance is now under review, using a random sample of SI and target-only students from the same academic year.

 Similarly, the data from 1990-91 (with program outcomes very close to those measured the previous academic year) indicate that students from academic assistance programs benefit considerably more than traditional students.

 The 1990-91 overall annual SI grade impact of +0.29 is a composite of two factors:

 +0.22 received by traditional students in SI compared with their peers not in SI and

 +0.44 earned by SI students from academic assistance programs in SI compared with their non-participating peers

 (level of significance: 0.01 using z-test for all three comparisons).

As these figures indicate, program students benefit twice as much (the improvement in their GPA is twice as much) as traditional students in SI.

 
Figure 1

Graphically, the performance of traditional students in SI over their target peers appears as a fairly normal distribution of differences in grades (with the 0.22 mean improvement over the norm established by peers not in SI).  However, the performance of high-risk students from academic assistance programs is both shifted to a higher mean and skewed in distribution showing that a greater proportion of students surpassed the overall SI mean of 0.29.  (See Figure 1.)

 More strikingly, in 45 of 62 target classes (73%) analyzed, program students enrolled in SI did better than their non-SI peers;  and in 34 of these 62 courses (55%), as a group they surpassed the GPA earned by all target class students not in SI (84% of these being traditional students).  (See Figure 2 and Figure 3.)  Level of difficulty in these target classes is best documented by the fact that in

[page 88]

half of the 34 courses in which program SI students did better than any other group, their average grade in the target course was still below 2.0!  It must also be noted that there were 17 target courses (of 62, or 27%) in which program SI students did not do as well as their peers who weren't enrolled in SI.  There is sufficient variation among these courses that the results cannot be attributed primarily or exclusively to student preparation, effort or SI leadership.

 
Figure 2-------------------- Figure 3

In general, then, high-risk/underprepared students tend to benefit more from SI than traditional students in the immediate application of their SI instruction.  A number of reasons were  suggested to explain that pattern.  The greater skill or academic background needs of program students may make them more reliant on SI classroom instruction for their performance.  Or the style of teaching typically found in many of these classes may be more beneficial to that group than to differently prepared, traditional students.

Because the answer to this question would be a strategic planning tool for future semesters, this matter received prompt attention.  The only factor found to be statistically significant in explaining the divergent results was the extent to which critical thinking and study skills were primary elements.  Interestingly, in a sampling of course sections where good SI Leaders whose primary emphasis was on content were compared with counterparts in the same disciplines who subordinated content to critical thinking and study skills, both program and traditional students improved measurably in the skills classes over non-SI peers.  Here, still, the program students did even better than mainstream students, but both groups benefited.  In the content-bound sections, program student performance was mixed compared with their non-SI peers, while traditional students in SI generally matched or did less well than their peers not in SI.  It may be that some content-emphasis SI classes are little different from a tutorial, which is less needed by traditional students but known to benefit under-prepared students.  A course contributing to cognitive development, on the other hand, benefits all students and has long-term measurable and immeasurable impact.

[page 89]

5. Fiscal considerations

 Certainly the cost of this program, compared to other interventions, is minimal.  Based on the average number of semesterly sessions attended by every enrolled student, the per student cost for a whole semester's instruction is about $45, which on a student assistant salary scale would buy seven to eight hours of tutoring (as compared to the 45 semester hours of SI).  That figure includes the training and supervision of student leaders, their 10 hr./wk. salary as either (undergraduate) student assistants or (graduate) teaching assistants, acquisition of textbooks for them, and a portion of the administrative cost for overseeing the SI component.  Servicing large numbers of high risk students for that amount seems a very judicious and efficient use of resources.

 Funding for this particular adapatation of the model currently comes from two sources:  about half are categorical monies allocated from state and federal educational equity programs to the LAC for tutoring and SI;  the other half is the university's contribution in the form of the Coordinator's position and 25-30 teaching assistanceships from instructional funds.

D.  Conclusions:

 Both published outcome data and direct experience suggest  that Supplemental Instruction is the direction for the future, not at the expense of other learning center programming, but certainly as a significant element within a campus's planning.  There are a number of reasons why this appears to be the case.

 Fiscally, as state economic resources continue to shrink and the cost of educating individuals remains on the rise, institutions are forced to seek effective ways of stemming attrition.  Every academic failure is costly, whether that comes in the form of repeated courses, replacement of a lost individual with a new recruit/applicant/enrollee, or simply prolongation of one student's campus career because of academic difficulties.  The cost of failure is greater than the cost of SI.

 Educationally, if Blanc et al.'s assessment of cognitive development levels is accurate (more than likely, it's conservative), SI is the only academic support model to effectively address the need for critical thinking and cognitive development with sufficient continuity to have substantial impact.  If secondary schools are not preparing students adequately in this area, whether they see it as their proper role or not, it falls to higher education to provide resources for this purpose.  Preformal cognitive levels may prove adequate for introductory college courses in areas where factual information predominates or where an instructor's testing mode is very basic.  But it will not permit success in those disciplines, courses, or classrooms where higher level thinking skills are essential.  The SI leader's modeling of those skills as well as his/her direct efforts to inculcate such skills and approaches in SI students are valuable ways of addressing that need in order to produce a population of capable, successful college students.

 Finally, there are responsibilities implicit in campus's admission policies and in their affirmative action efforts.  Published data continue to show very poor retention and graduation rates for historically underrepresented student groups, except in those institutions where

[page 90]

they are the majority.  The fact that many of these students enter institutions only under special action has been cited as a justification for poor outcomes:  i.e., given that they would not have been accepted at the university, the fact that 20% or 40% succeed is a positive outcome.  However, it appears to be the moral obligation of institutions which seek out and enroll these students to aggressively enhance their success by making them competitive, not by lowering standards.  Dropping expectations or letting them fail is reinforcing the false assumption (affirmed throughout their education) that it is they, not the instruction given them, who are deficient.  Approaches that single out underprepared students for special help and make them dependent on that help perpetuate such stereotypes.  If attrition rates remain high despite traditional interventions, the latter are either inadequate or ineffective for the majority of the students they're designed to serve.  In contrast, the SI model is working for these students;  furthermore, it avoids segregating and stigmatizing them.  As a program, SI offers benefits for any student, for the high achiever as well as for the struggler, giving either one cognitive tools and their refinement for long-term benefit.  It replaces the message of traditional academic supports ("you're incapable" or "you'll fail on your own") with a note of confident encouragement:  "This program offers college students the secrets of academic success!"
 

References

Blanc, Robert A. et al. (1983).  Breaking the attrition cycle: The effects of Supplemental Instruction on undergraduate performance and attrition, Journal of Higher Education, 54, 80-90.

Caswell, R.  (1991).  Adjunct class sessions:  Assisting underprepared students to achieve academic success, Teacher Education and Practice, 6(2), 73-74.

Knowles, M. S.  (1981).  Preface.  In D. Boud (Ed.), Developing student autonomy in learning.  London: Kogan Page.

Lundeberg, M. A.  (1990).  Supplemental instruction in chemistry, Journal of Research in Science Teaching, 27(2), 45-55.

Martin, D. C. & Arendale, D. R.  (1990).  Supplemental Instruction:  Improving student performance, increasing student persistence, Eric Document 327103, 14 p.

Martin, D. C. & Blanc, R.  (1981). The Learning Center's role in retention: Integrating student support services with developmental instruction, Journal of Developmental and Remedial Education, 4(3), 2-4 & 21-23.

Martin, D. et al.  (1982).  Supplemental Instruction - A model for increasing student performance and persistence. nd

Wolfe, R. F.  (1987).  The Supplemental Instruction program:  Developing learning and thinking skills.  Journal of Reading, 31(3), 228-32.