Teacher effects on student motivation during cooperative learning:
Activity level, intervention level, and case study analyses
Ming Ming Chiu
The Chinese University of Hong Kong
National Academy of Education (U.S.) / Spencer post-doctoral fellow
February 24, 2000
In three cooperative learning studies, teacher interventions influenced student
motivation, which in turn affected group outcomes. Forty students from five 9th grade classes
were videotaped while solving an algebra problem together in groups of four. A teacher and a
teaching assistant (TA) taught these classes and intervened 54 times. Controlling for past
student achievement, higher student motivation increased cooperative problem solving success
in the first study. In the second study, increased student autonomy and greater teacher
responsiveness to students increased student motivation. Indicators of greater student
autonomy included: student-initiated interventions, % of student talk and % of teacher/teaching
assistant (T/TA) questions. Indicators of greater teacher responsiveness included % of T/TA
support and % of T/TA criticism. Also, % of T/TA closed questions and % of T/TA
compliments positively predicted student motivation. Finally, the case study discusses one
intervention in detail, showing some conditions of use, exceptions to the above effects, and
specifying how a compliment helped increase student motivation.
Teachers' effects on student motivation during cooperative learning:
Activity level, intervention level, and case study analyses
Motivation is central to student achievement. Motivated students devote more time,
effort and resources to learn and to solve problems. As a result, they perform better than their
less motivated peers (Ames & Ames, 1984, 1985; Stipek, 1988; Wentzel, 1994). Researchers
have argued that many factors can improve student motivation (e.g., a challenging complex
problem [Stodolsky, 1988], a caring classroom environment [Caballo and Terrel, 1994], etc.).
In particular, researchers have argued that both teachers (Ames & Ames, 1985; Brophy, 1983)
and cooperative learning (Caballo & Terrel, 1994; Slavin, 1990) can improve student
motivation. This article examines how teachers’ interactions with students during cooperative
problem solving increased (or decreased) students' motivation. By identifying effective and
ineffective intervention methods, educators can improve student motivation and learning. This
article presents three studies of videotape and statistical analyses of 40 students, five
teacher/teaching assistants (T/TA) and their teacher/teaching assistant interventions (TI).
This article is organized as follows. The theoretical section reviews research on
factors that improve student motivation. Three studies follow, regarding group solution
outcomes, student motivation after a teacher intervention, and a case study of a teacher
intervention. After a general discussion of the results, the paper concludes with some
limitations and recommendations for future research.
Past research showed that greater teacher engagement and student autonomy both
increase student motivation. However, these two factors can conflict if the teacher’s
assistance is unsolicited. To avoid this conflict, a teacher can let students ask for help and
Skinner and Belmont (1993) showed that teachers who devoted more time, affection
and resources to their students (rather than rejection and neglect) increased their students’
motivation throughout the school year. Likewise, researchers have shown that teacher guidance
and explanations improve student motivation (see Brophy's  review).
Giving students greater autonomy also increases their motivation. Instead of being told
what to do or how to do it, students prefer greater leeway to choose and to make decisions
(Chiu, in press). Moreover, students with greater autonomy are more motivated because their
motivation is intrinsic. In contrast, extrinsic motivation through reward structures eventually
undermines itself as students view the reward as more important than their activity (see
reviews by Brophy  and Grolnick, Ryan, & Deci ).
At the utterance level, teachers can increase student autonomy by inviting their
participation with questions rather than demanding it with commands (Chiu, in press).
Questions show a knowledge gap and give students more leeway to answer. In contrast,
commands specify actions for students to implement.
Teachers can also use closed or open questions. Closed questions have a narrow focus
and typically have one correct answer. Open questions, however, are broad, and the answers’
specificity and displays of understanding can vary widely. Researchers have argued that
teachers allow students more freedom with open questions and constrain them with closed
questions (Buzzelli, 1996; Greenberg, Woodside & Brasil, 1994). Furthermore, scaffolding
advocates (Rogoff & Gardner, 1984; Wood, Bruner & Ross, 1976) argued that teachers must
adapt their questions to the students' responses. After giving students wide leeway with an
open question, teachers can use closed questions to address students’ needs.
Excessive teacher engagement can also reduce students’ self-confidence and
motivation by encroaching on their autonomy. In particular, unsolicited teacher help can harm
students. These students may view the teacher’s help as necessary because they have low
ability (Graham, 1990). Cohen (1994) further argued that a teacher should only intervene in
student cooperative learning when the group is off-task or involved in an interpersonal conflict.
At other times, she argued, the students should retain their autonomy and rely on themselves
rather than become dependent on the teacher. By relying on themselves, the students avoid the
false failure of asking for unnecessary teacher help. Then, they can increase their confidence
by successfully solving the problem on their own.
To maintain both teacher engagement and student autonomy, a teacher should respond to
students’ requests for help. Responsiveness includes both listening to students' utterances and
evaluating them. Keller (1983) argued that teachers increase students' motivation by listening
to them and providing relevant feedback. Brophy (1983) claimed that supporting students
contingently (as opposed to undeserved praise and compliments) highlights students'
competencies and focuses their attention on their task-relevant behaviors.
Excessive praise for simple tasks and avoidance of criticism for student failures can
also have harmful effects. Graham (1990) showed that 11-12 year old students believe that
effort is inversely related to ability. She also showed that they believe praise and criticism are
indicators of effort and inverse indicators of ability. Students expect their teacher to praise
them when they are exerting effort at the limits of their abilities. A teacher who praises
students on simple tasks implies that they have reached the limits of their low abilities.
Consequently, Graham (1990) argued that teachers should not bias their evaluations toward
praise. Instead, they should criticize students for failures and avoid praise for successes on
In short, a teacher can increase student motivation by giving them some autonomy and
by responding to their requests for help.
The remainder of this article reports three sets of analyses as separate studies on group
solutions, student motivation after a teacher intervention and an intervention case study.
Study 1: Predicting Group Solutions
As discussed earlier, motivated students should be more successful than less motivated
students. Furthermore, students with past mathematics achievements are more likely to be
successful in the current mathematics task. This study tests if these hypotheses are also true for
Group motivation and past mathematics achievement predict group solution
The effects of students’ motivation level on problem solving were tested by
videotaping students’ group problem solving, transcribing the videotapes, coding the
videotapes and transcripts, and performing hierarchical regressions.
This study was part of a larger project in which we videotaped ethnically diverse
students (38% African-American, 32% Euro-American, 20% Asian-American and 10%
Latino-American) during their five algebra classes for six weeks. Two teachers and three
teaching assistants (TA) taught the classes in an urban, public high school in the United States
of America. There were one teacher and one TA per class. All teachers and TAs were Euro-
American. The teachers randomly assigned students into groups, and the students did not
receive any cooperative learning training. The teachers had taught for 10 and 11 years, and the
TAs were education doctoral candidates with 2-5 years of teaching experience. The ten
videotapes were of 40 students (ten groups of four) across five classes doing one lesson near
the end of a six-week unit on functions.
After a teacher introduced the following problem below in each class, the student
groups worked on it for 20 minutes while the teacher and TA monitored their progress with
Nintendo charges $180 for each gaming system and $40 for each video game. Sega
charges $120 for each gaming system and $50 for each video game. How many games
must a customer buy to pay less for Nintendo than for Sega? (Note: customers must buy
a gaming system before buying any video games.)
The team of teachers and researchers believed that this was a difficult problem for these
students even though they had covered enough mathematical concepts and relationships in class
to solve it. There were algebraic, graphical and tabular methods for finding the critical number
of games in which the cost is the same for either brand. Then, the correct answer is obtained by
adding an additional game so that the cost of buying Sega exceeds that of buying Nintendo.
Using equations, students could set the cost equations equal to each other and solve for
the number of games (g): 180 + 40g = 120 + 50g à 60 = 10g à 6 = g. Using a graph,
students could graph each equation and find the intersection of the lines. X is the number of
games, and y is the total cost. For Nintendo, y = 180 + 40x. For Sega, y=120 + 50x. Graphing
the two lines yields an intersection point of (x,y) = (6, 420). Thus, the cost for 6 games is $420
for both companies. With a table, students can add additional games and compute the cost until
the costs are equal (see table 1). Adding one to the number of games, 6, yields the final answer,
Insert table 1 about here
Solution score. Each group was given a score (0-3) based on their problem solving outcome:
correct solution (3), correct method (2), correct understanding of the problem situation (1), or
none of these (0). See Appendix A for coding details.
Past mathematics achievement. The students' mid-year algebra grades were used to compute
the mean grade for each group.
Student motivation. Because each student in a group must be coded for their motivation at
different times, the problem solving sessions were divided into one minute intervals. For each
minute of cooperative learning, we coded each student's motivation level as either motivated,
unmotivated, or distracting, respectively 1, 0, and -1. Motivated was operationally defined as
student behaviors that showed desire to work on and to solve the problem. They included
encouraging others to work on the problem, showing enthusiasm for the work, listening intently
to solution proposals, discussing it, etc. Students’ motivation levels were coded as distracting
if their behaviors discouraged others from working on the problem. Distracting behaviors
included denigrating the task or asking off-task questions to other group members. Finally, the
students were coded as unmotivated if they were neither motivated nor distracting, for example,
looking out the window. See Appendix A. Each group motivation score was the mean of all of
its members' motivations.
The correlation between mean mathematical grade for each group and their mean
motivation was tested for significance. Then, controlling for mathematical grade, the
significance of the correlation between mean motivation and final solution score was tested.
Videotaping, Transcription and Coding
The groups of students and the teachers were each videotaped separately. The teacher
intervention segments of the student videotapes for each class were fully transcribed for words
and gestures (McNeil, 1992). The teacher videotapes were used to triangulate transcription of
poor sound quality segments. Two people coded each variable. For student motivation, each
coder made a separate pass through each transcript for each student. For each remaining
variable, each coder made a single pass through each transcript (see Chiu, in press). Cohen’s
kappa computations tested inter-rater reliability.
All results were significant at the .05 level.
The students found this problem difficult as only half of the groups correctly solved it
(see table 2). Of the 784 student-minutes coded for motivation, students were motivated 63%
of the time, unmotivated 19% of the time and distracting 18% of the time (Cohen's kappa = .88,
p <.001). (There were 16 student-minutes that could not be coded because of poor sound
quality. None of these occurred during the teacher interventions.)
Insert table 2 about here
Mean student motivation predicted the group's final solution score. As expected, mean
mid-year mathematics grade significantly correlated with mean group motivation (r = .48, p
< .05). After controlling for grade, group mean motivation also significantly correlated with
solution score (r = .69, p < .05).
These results showed that group motivation affects the group outcome, consistent with
research showing that individual motivation affects individual outcomes (Ames & Ames, 1984,
1985; Stipek, 1988; Wentzel, 1994). Having shown that motivation was an important factor in
the group solution scores, consider the effect of teacher interventions on student motivation.
Study 2: Predicting Student Motivation after a Teacher Intervention
Properties prior to the teacher intervention can influence students’ subsequent
motivation. In particular, the students’ motivation and problem solving progress before the
T/TA intervention (TI) can affect their motivation afterwards. If students were already
motivated, they were likely to remain motivated. Also, groups who have already made some
progress can be more motivated than those that have not. However, research has shown that
past student achievement did not predict motivation at the macro-level (Crandall, 1969; Stipek
& Hoffman, 1980). This study tests this hypothesis at the micro-level.
Higher pre-TI student motivation predicts higher post-TI student motivation.
Greater pre-TI student problem solving progress does not predict higher post-TI
As discussed earlier, teacher behaviors that encourage student autonomy or are
responsive to students’ needs should increase student motivation. During cooperative learning,
teachers can wait for students to initiate discussions and avoid dominating the conversation in
both quantity (words) and quality (content).
Student-initiated interventions predict higher post-TI student motivation.
Greater quantity and quality of teacher involvement predict lower post-TI
Teachers can also choose particular speech forms to encourage student autonomy.
Teachers can ask more questions, especially open questions, rather than issue commands.
According to scaffolding advocates, open questions should precede closed questions.
Teacher commands predict lower post-TI student motivation.