Abstract: My research focused on comparing two approaches for the recall and understanding of given text material: (a) frequent testing to provide incentive motivation, and (b) outlining as a homework assignment to provide a study strategy. A total of over 300 college students were used in three experiments, conducted in an actual course involving real grades. In one of the studies, students were classified into high, middle and low groups on prior academic achievement (GPA), and in another, students were classified as high, middle, and low procrastinators based on Procrastination Scale scores. In one condition, tests were given on a weekly basis. These were designed to minimize transfer to three regular examinations, which were used as criterion measures of achievement. In the second condition, weekly homework assignments required students to identify, define, and elaborate on chapter main points. Frequently tested students, overall, were found to outperform homework students on examinations in all three studies. The difference was found to be based on dramatically large differences (10% or a full letter grade) among low GPA students (in one study), and among high procrastinators (in another) favoring students in the frequent test condition.
I
had been doing research on enhancing students' academic engagement and
achievement in college (e.g., Tuckman, 1990; Tuckman & Sexton, 1990, 1991,
1992) and had begun to suspect from my results that many poorly performing
students were knowledgeable about study skills and self-regulation strategies,
but simply failed to use them. Typical low achievers did not seem to spend the
necessary time and effort, or apply themselves in the necessary way, to get the
information they were learning in their classes into their long-term memories.
What they seemed to lack was not intellect or skill, but drive.
I
had come to the conclusion from my own work and that of others that achievement
required aptitude and motivation, and motivation required a combination of
attitude, strategy, and drive (Tuckman, 1999).
I had been able to improve both attitude (Tuckman & Sexton, 1991) and
strategy (Tuckman, 1992), but had not found a way to increase drive.
I had begun to think about what it might be that finally drove a student
to get information into long-term memory, when it hit me: tests!
Tests motivate students because they create the opportunity or necessity
to achieve success or avoid failure. In
that way, tests provide an incentive to learn.
They are a source of incentive motivation.
Pintrich
and Schrauben (1992) review a large body of research that suggests that (1) the
value of an outcome to the student affects that student's motivation, and (2)
motivation leads to cognitive engagement, such engagement manifesting itself in
the use or application of various learning strategies.
If not directly taught to these students, the use of these strategies
would suggest that they already exist in the student's repertoire.
Thus, if enhancing incentive value or incentive motivation by itself
actually improves outcomes, the explanation is likely to be that effective
cognitive engagement has occurred.
Incentive
theories of motivation (e.g., Rotter, 1954, Rotter, Phares and Chance, 1972;
Overmier and Lawry, 1979, Petri, 1998) suggest that people will perform an act
when its performance is likely to result in some outcome they desire, or that is
important to them. For example, in anticipation of a situation in which a person
is required to perform, that person may expend considerable effort in
preparation because of the mediation provided by the desire to achieve success
or avoid failure. That desire would
be said to provide incentive motivation for the person to expend the effort.
Hence, I theorized that a test, as a stimulus situation, would provoke students
to study as a response, because they link a behavior to a goal, namely the
achievement of a successful grade or avoidance of a failing grade on that test.
Putting in the effort to study for the test, which means attempting to
get potential test content into long-term memory, therefore, would be the result
of incentive motivation.
It
was now necessary for me to devise a way to test the hypothesis that tests
prompted better achievement, and that they did so because they motivated
students to get what they were being taught into their "heads" (or
long-term memories). If I could
show that, then it would also imply that they already knew appropriate learning
skills and strategies.
I
was fortunate to have available to me a number of sections of the same course,
covering the same material, and using the same textbook, and examinations.
This was ideal, since I felt that academic motivation and performance
could be better studied in real classes involving real consequences (i.e.,
grades) than under simulated conditions. I
needed something to compare frequent testing to that (a) caused students to
spend equal time working on course materials - controlling for time-on-task, (b)
could be graded, (c) did not motivate students to get information into their
long-term memories - the essential feature of tests, and (d) enhanced study
skills - as an alternative explanation. What
I came up with as a comparison to frequent testing was required homework,
namely: to outline each chapter by identifying the 20 most important concepts,
and provide a textbook definition and elaboration (i.e., a description in one's
own words) of each.
Methods
I
did essentially the same study three times; (1) on a five week segment of the 15
week course, in which I also included a third group that got neither tests nor
homework; (2) on the whole course, where I included only tests and homework
conditions, I required all students to keep a log of time spent on coursework,
and I compared the results for high, middle, and low GPA students; (3) exactly
the same as the second time except I measured students tendency to procrastinate
(using the Procrastination Scale; Tuckman, 1991), and compared results for high,
middle, and low procrastinators.
Subjects.
Numbers of subjects in each study ranged from 110 to 120. They were
juniors and seniors in college, all preparing to be teachers as either a major
or a minor, having an average age of 21, two-thirds women, and about 92 percent
white. They were enrolled in three
sections (in the first study) or two sections (in the second and third study) of
an educational psychology course required for teacher certification.
All sections met once a week (on consecutive days) at the same time of
day, covered the same content (learning theories), used the same textbook, and
were taught by the same instructor. They
were randomly assigned to treatment conditions. A comparison of the sections
used in each study on age, gender, and scores on the verbal portion of the
College Level Academic Skills Test (CLAST) showed them to be equivalent, thus
satisfying the requirements for a quasi-experimental design.
Correlations between CLAST verbal scores and achievement in this course
have been found to be about 0.5 (Tuckman, 1993).
Treatments.
(1) Frequent test condition. One class in each study was given a 15
minute, seven-item, completion-type test at the beginning of each class period
(i.e., prior to instruction), covering the textbook chapter assigned for that
week. Following the test, students exchanged papers, and the answers were gone
over by the instructor so that students could grade one another's tests.
Students were informed that the average of their grades on these tests
would count toward their final grade as much as one of the three major course
examinations.
(2)
Homework condition. One class in each study was given the homework
assignment of identifying the 21 most important terms in the assigned chapter,
and preparing a definition of each term along with a one-sentence elaboration of
each. A list of about 28 terms was
predetermined for each chapter, and students' choices had to fit this list.
If a term chosen by a large percentage of students was not on the list,
it was added. The text included no
glossary, but many signals so that term identification was not difficult.
For example, in the chapter on reinforcement theory, a key term would be
"reinforcer," its definition would be "something that increases
the likelihood of occurrence of the response it follows," and an
elaboration might be "getting something good to eat after doing my
homework." Students were given some training before they started and after
having done two assignments. They
were also given feedback on all aspects of each assignment so their proficiency
would improve. Each assignment was
graded (A, B, or C) based on number of correct terms included, correctness of
definitions, and appropriateness of elaborations. These grades were averaged and counted as much toward the
final grade as the average of frequent test grades in the testing condition.
In
the first study only, there was also a Control condition, that is, a class given
only lectures on the chapters (no frequent tests, no written homework).
Moderator
Variables. In the second study, grade point average (GPA) was used as a
moderator variable, as was procrastination tendency in the third study.
For GPA the ranges were high (3.6-4.0), medium (2.9-3.5), and low
(2.0-2.8), while for procrastination they were divided into tertiles.
Dependent
Variable. One (in the first
study) and three (in the second and third study) 50-65 multiple-choice item
tests, matched to instructional content, were given to measure achievement
outcomes. The tests had a K-R
reliabilities of .82-.87. While
virtually all of the questions on this test related to key terms, a key feature
of the TDE homework assignments, they measured comprehension rather than factual
recall. In other words, they were
higher-order or conceptual questions. Students
were typically asked to identify the concept that fit a given example or the
example that fit a given concept. For
example, "According to the PREMACK PRINCIPLE, which of the following
reinforcers would be most appropriate for the given group? a. Money for adults;
b. Tokens for inner city children; c. Playing for third graders; d. Praise for
teenagers." (The answer is c.
since the Premack principle applies only to activity reinforcers and playing is
the only activity among the four choices.)
This conceptual feature made them equally unlikely to favor the frequent
test condition either, since the questions on the frequent tests were all
completion-type measures of factual recall.
Moreover, frequent test questions were intended to measure care taken in
reading the chapter, and so focused on details and specific points.
As a result there was minimal overlap between questions on the
examinations and on the frequent tests.
Results
and Discussion
This
is what I found. In all three
studies, students in the frequent testing condition earned significantly higher
examination grades than students in the homework condition (and in the control
condition in the first experiment). The
difference ranged from a whole grade (a B compared to a C) to a third of a grade
(a B- compared to a C+).
In
the first study, the final achievement test results for the three approaches
were as follows: frequent test mean=82.8 (sd=9.3), homework mean=71.6 (sd=9.4),
control mean=66.9 (sd=12.6). The
ANOVA for condition yielded F=21.69 (df=2/106), p<.001. Cell comparisons by Newman-Keuls test showed the frequent
test approach to yield significantly better results (p<.001) than either of
the other two conditions, while homework exceeded the control at the p<.10
level. The effect size is near or
above 1.00 for each comparison with the frequent test results.
What
was considerably more dramatic and revealing were the results for students at
different GPA levels (in the second study) and different degrees of
procrastination (in the third). Regarding
GPA, there were no differences in examination grades between frequent tests and
homework conditions for students with high (3.6-4.0) or middle (2.9-3.5) GPAs.
However, for students with low GPAs (2.0-2.8), those in the class given
frequent tests averaged 76% (B- on the grade scale) across the three tests
compared to 66% (C-) for those in the homework class, a highly significant
difference of 10 percent on a 100-point scale.
In fact, the low GPA students given frequent tests did so well they
outscored the middle GPA students in both the homework and frequent test
conditions. These results are shown
in Figure 1.
For
procrastination level, the findings were even more dramatic. High
procrastinators in the frequent test condition not only significantly outscored
high procrastinators in the homework condition (by over 10%, or a full grade),
they outscored both middle and low procrastinators in both homework and test
conditions. These results are shown in Figure
2.
The
overall results clearly showed that frequent tests worked better than homework
in improving achievement, but this was primarily because of its dramatic effect
on students with low GPAs and high procrastination tendency (characteristics
that may, in fact, reside in the same students). This suggests that a low GPA does not necessarily reflect
either low academic ability or the absence of study skills, but the lack of
motivation to do timely study in such a way to internalize course content.
Similarly, high procrastinators have their inability to get started to
blame for their poor academic performance.
In regard to time, students in the frequent test condition actually
reported spending somewhat less than homework-condition students.
An
issue relevant to the interpretation of the results is the extent to which the
frequent tests worked, not as an incentive to study, but as a direct training
aid or targeted study guide for the achievement exams.
However, the nature of the items was quite different, the former being
factual/short answer and the latter conceptual/multiple-choice.
Moreover, the achievement exam questions were as similar to the homework
assignments as to the frequent tests. Further,
from the empirical perspective, if the frequent tests were merely study guides,
they should have helped students at all three GPA levels, particularly those
middle and low. While they did help
the low GPA students substantially, they had no effect at all on the middle GPA
students. This suggests that the
frequent tests enhanced motivation for students who have typically performed
poorly to get content into long-term memory rather than merely targeting for
them what to study.
Another
issue is one of exactly what it was that the students in the frequent test
condition were motivated to do that differentiated them from students in the
homework condition, since both tests and homework assignments were graded, and
students in the homework condition actually devoted more time to the task than
their counterparts spent studying for tests. My speculation is that homework
involves using the text to prepare a document on paper, thereby providing no
motivation to store text information in long-term memory.
Tests, on the other hand, require such storage in order to have access to
the information while taking the test. Therefore, the incentive motivation based
on the grade for test takers would be linked to the motivation to store
information in long-term memory while, for homework doers, it would not.
My work should send a message to students.
Even if professors don't give frequent tests (although my findings
suggest that they should), students should test themselves on a regular basis to
insure that they are getting what they learn into long-term memory.
It they wait till the midterm, there will be too much information to
process in too short a time. Procrastination
may be what causes students to end up with a low GPA.
References
Overmier,
J.B. & Lawry, J.A. (1979). Conditioning and the mediation of behavior. In
G.H. Bower (Ed.), The psychology of learning and motivation (Vol. 13).
New York: Academic Press, pp. 1-55.
Petri,
H.L. (1998). Motivation: Theory, research and applications (4th ed).
Belmont, CA: Wadsworth.
Pintrich,
P.R. & Schrauben, B. (1992). Students' motivational beliefs and their
cognitive engagement in classroom academic tasks. In D. Schunk & J. Meece
(Eds.), Student perceptions in the classroom: Causes and consequences.
Hillsdale, NJ: Erlbaum, pp.
Rotter,
J.B. (1954). Social learning and clinical psychology. Englewood Cliffs,
NJ: Prentice-Hall.
Rotter,
J.B. Chance, J.E. & Phares, E.J. (1972). Applications of a social
learning theory of personality. New York: Holt, Rinehart & Winston.
Tuckman,
B.W. (1990). Group versus goal-setting effects on the self-regulated performance
of students differing in self-efficacy. Journal of Experimental Education,
58, 291-298.
Tuckman,
B.W. (1991). The development and concurrent validity of the Procrastination
Scale. Educational and Psychological Measurement, 51, 473-480.
Tuckman,
B.W. (1992). The effect of student planning and self-competence on
self-motivated performance. Journal
of Experimental Education, 60, 119-127.
Tuckman,
B.W. (1993). Motivational components of college students' performance and
productivity. Paper presented at the annual meeting of the American
Educational Research Association, Atlanta, GA.
Tuckman, B.W. (1996). The relative effectiveness of incentive motivation and prescribed learning strategy in improving college students' course performance. Journal of Experimental Education, 64, 197-210.
Tuckman, B.W. (1997). Using tests as an incentive to motivate procrastinators to study. Journal of Experimental Education, 66, 141-147.
Tuckman,
B.W. (1999). A tripartite model of motivation for achievement: Attitude-
Drive-Strategy. Paper presented in the Symposium: Motivational Factors
Affecting Student Achievement-Current Perspectives, at the Annual Meeting of the
American Psychological Association, Boston, 1999.
Tuckman,
B.W. & Sexton, T.L. (1990). The relation between self-beliefs and
self-regulated performance. Journal of Social Behavior and Personality,
5, 465-472.
Tuckman,
B.W. & Sexton, T.L. (1991). The effect of teacher encouragement on student
self-efficacy and motivation for self-regulated performance. Journal of
Social Behavior and Personality, 6, 137-146.
Tuckman,
B.W. & Sexton, T.L. (1992). The effects of informational feedback and
self-beliefs on the motivation to perform a self-regulated task. Journal of
Research in Personality, 26, 121-127.
Copyright 2000-2008 The Ohio State University. All rights reserved.