Full text of DOI:10.1007/BF02505025

Effects of Classroom Evaluation Strategies on
Student Achievement and Attitudes
Zane Olina
Howard J. Sullivan
Classroom evaluation can have a powerful
impact on student performance and motivation
(Crooks, 1988; Natriello, 1987). In his review,
Crooks cited evidence that evaluation can pro-
vide students with knowledge of results and
corrective feedback, help students monitor their
own progress, and influence students’ continu-
ing motivation and their perceptions of their
self-efficacy as learners. From a review of 250 ar-
ticles on classroom assessment, Black and Wil-
liam (1998) reported positive effects of formative
evaluation on performance of students of all
ages and ability levels. They used the term for-
mative evaluation to refer to the evaluation of in-
struction for the purpose of improving student
performance, rather than to evaluation of an in-
structional program for the purpose of improv-
ing it during its developmental or formative
stage (Scriven, 1967).
This study investigated the effects of teacher
evaluation and the combination of teacher
evaluation and student self-evaluation on
student performance and attitudes.
Participants in the study were 189 Latvian
high school students and their six teachers.
The six teachers were assigned to one of three
treatment conditions: (a) no evaluation, (b)
teacher evaluation, and (c) self-evaluation plus
teacher evaluation. All groups completed a
12-lesson instructional program on how to
conduct experiments and produce research
reports. Students in the teacher-evaluation
group received teacher evaluation on their
initial research reports. Students in the
self-plus-teacher evaluation group
self-evaluated their reports and received
teacher evaluation on them. The no-evaluation
group received no formal evaluation
instructions.
Teacher evaluation of student work is the
most common form of classroom evaluation
cited as having positive effects on student per-
formance and attitudes. Cardelle-Elawar and
Corno (1985) found that elementary school stu-
dent performance and attitudes toward mathe-
matics improved when their teachers provided
written feedback on their homework several
times a week. Thomas et al. (1993) reported a
positive correlation between the amount of
teacher feedback on tests, quizzes and
homework assignments and student perfor-
mance in high school biology courses. Page
(1958) found in his study involving 74 secondary
school teachers that a brief written comment on
objective examinations significantly improved
student performance when compared to no
comment at all.
Students in the teacher-evaluation and the
self-plus-teacher evaluation groups received
significantly higher ratings on their final
projects than those in the no-evaluation group.
However, the no-evaluation group had more
favorable attitudes toward the program than
the other two groups, while the
self-plus-teacher evaluation group was
significantly more confident of their ability to
independently conduct future research
experiments.
Other studies have shown no effect of teacher
ETR&D, Vol. 50, No. 3, 2002, pp. 61–75 ISSN 1042–1629
61

62
ETR&D, Vol. 50, No. 3
evaluation on student performance. Stewart and
White (1976) replicated Page’s (1958) study and
reviewed 12 other replication studies, conclud-
ing that teacher comments had little or no effect
on student performance. Story and Sullivan
(1986) found that teacher comments had no sig-
nificant main effects on the continuing motiva-
tion of fifth- and sixth-grade students, but the
combination of comments and an easier task
were effective in motivating girls to return to the
same task.
receive self-evaluation training. Ross, Rolheiser
and Hogaboam-Gray (1999) reported similar
results with fourth- to sixth-grade students who
were trained in self-evaluation of narrative writ-
ing compared to their counterparts who were
not trained in self-evaluation.
Studies involving comparisons of teacher
evaluation and student self-evaluation have
shown positive effects for self-evaluation on stu-
dent continuing motivation. Salili, Maehr,
Sorensen, and Fyans (1976) had fifth-grade
Iranian students solve word anagram problems
under three different evaluation conditions. Stu-
dents in the teacher-evaluation condition were
told that their work would be evaluated by the
teacher, students in the self-evaluation condition
were told that no one would know about their
results, while students in the peer-comparison
condition were told the results would enable
them to see how well they did on the task in
comparison with their classmates. Salili et al.
found no performance differences among the
three groups. Yet students in the self-evaluation
and peer-comparison conditions showed greater
continuing motivation than students in the
teacher-evaluation condition in the form of
desire to return to the initial task. Hughes et al.
(1985) reported that students returned to a dif-
ficult task more often after self-evaluation and to
an easy task more often after teacher evaluation.
Researchers have cited several characteristics
of effective teacher evaluation. Students should
be provided with explicit evaluation criteria and
models of good work before they begin working
on a learning task (Sadler, 1989; Wiggins, 1998).
Teacher evaluation of performance is most effec-
tive when it is specific, directly related to the
task, and provides the opportunity for students
to correct their performance (Black & William,
1998; Crooks, 1988). Finally, teacher evaluation
should be used initially for formative purposes
rather than for assigning a grade (Hughes, Sul-
livan, & Mosley, 1985; Sadler, 1989).
Student self-evaluation is a second form of
classroom evaluation that may enhance student
learning. There has been considerable emphasis
in the last decade on active student evaluation of
their own work (Bransford, Brown, & Cocking,
1999; Gipps, 1994; Wolf, Bixby, Glenn,
&
Gardner, 1991). Several authors have argued
that student self-evaluation has the potential to
improve performance and motivation (Gipps,
1994; Shepard, 2000; Wiggins, 1998). Yet evalua-
tion by students of their own work is not a com-
mon practice in the classroom, and the topic is
frequently overlooked in general literature on
classroom assessment (Black & William, 1998).
Research findings on the use of teacher
evaluation and student self-evaluation to im-
prove student learning prompted the first
author of this article to investigate the potential
for these instructional strategies in her home
country, Latvia. Since Latvian independence
from Russia in 1990, the educational reform
movement in Latvia has emphasized the need to
develop lifelong learning skills, including the
ability to evaluate one’s own work. The National
Compulsory Education Standard of the Latvian
Ministry of Education and Science Center for
Curriculum Development and Examination
(LMES CCDE, 1998), as well as the Subject Con-
tent Standards issued by the LMES CCDE, em-
phasize that both teacher evaluation and student
self-evaluation are to be used as an integral part
of the teaching and learning process.
Overall, studies show positive effects of self-
evaluation on student performance and motiva-
tion across subject areas and age groups.
Maqsud and Pillai (1991) reported that South
African high school students who were asked to
self-score their tests over a course of one
semester significantly outperformed students
whose tests were scored by their teacher. In a
yearlong intervention in Portugal, Fontana and
Fernandez (1994) found that primary school stu-
dents trained in self-evaluation performed bet-
ter in mathematics than students who did not
Most Latvian teachers are not prepared for
the new expectations, and very little profes-

CLASSROOM EVALUATION STRATEGIES
63
sional development is provided for improving
teacher evaluation skills. Under the Russian
regime, all schools used the same textbooks and
a prescribed curriculum. Because of the educa-
tion reform efforts in the past decade, only Sub-
ject Content Standards have been mandatory for
schools, and teachers have been free to design
their own curricula and assessment approaches.
The authoritarian classroom management style
by teachers still dominates in most classrooms,
and there is no tradition of engaging students in
evaluation of their own work. Because of large
workloads with almost half of the teachers
teaching more than 32 direct contact hours a
week (Soros Foundation-Latvia, 2001), many
teachers also lack the time and experience to
provide students with constructive formative
feedback about their performance.
the teacher received. The students then applied
the rating scale to evaluate their written reports
and to write comments about their work. The
students were not assigned grades on their ini-
tial reports, and they had opportunities to revise
their work on their final reports. This approach
was consistent with strategies for self-evaluation
training suggested by several authors (Rol-
heiser, 1996; Sadler, 1989; Wiggins, 1998).
The following research questions were inves-
tigated:
1. Does teacher evaluation have a positive effect
on student performance?
2. Does the combination of teacher evaluation
and student self-evaluation have a different
effect on student performance than teacher
evaluation alone?
3. Does the combination of teacher evaluation
and student self-evaluation have a different
effect on student attitudes than teacher
evaluation alone?
The present study investigated the effects of
teacher evaluation and student self-evaluation
on student posttest scores, the quality of student
research reports, and student attitudes. Three
levels of evaluation were employed: (a) a no-
evaluation or control condition; (b) a teacher-
evaluation condition, and (c) a self-plus-teacher
evaluation condition. The study was conducted
in the students’ regular classes, and their regular
teachers delivered the instructional treatments.
The research questions dealing with student
performance were investigated using two dif-
ferent criterion measures, (a) scores on a posttest
covering the instructional content and (b) re-
searcher ratings of the student research reports.
The use of the research report measure is consis-
tent with the recommendation that students can
best develop self-evaluation skills when they are
given authentic performance-based learning
tasks (Stiggins, 2001; Wiggins, 1998; Wolf et al.,
1991). The instructional program and all assess-
ment instruments used in this study were field
tested earlier by the researcher with a similar
age group of students to the target population.
The teacher-evaluation component in this
study was designed based on several teacher-
evaluation strategies reported as effective by
authors of evaluation research (Black & William,
1998; Hughes et al., 1985; Sadler, 1989; Wiggins,
1998). Students were provided with specific
standards for evaluating their work before they
began working on their experiments. These
standards were first made explicit to the stu-
dents in the form of a project rating scale con-
structed for the study. The students under the
teacher-evaluation condition received teacher
comments about their work according to criteria
in the project rating scale. The students were not
assigned grades on their initial reports, and they
had opportunities to revise their work on their
final reports.
METHOD
Participants
Participants in the study were 189 Latvian high
school students from 12 classes taught by six
teachers. The average class size was 16 students.
All teachers involved in the study had com-
pleted at least four years of college and obtained
either a bachelor’s or master’s degree in educa-
tion. The teachers represented a variety of sub-
ject areas, such as language arts, mathematics,
The student self-evaluation component was
designed to be similar to the teacher-evaluation
component. Students were provided with the
same specific standards for evaluating their own
work in the form of the project rating scale that

64
ETR&D, Vol. 50, No. 3
science, and social studies. The classes were
drawn from five schools in different regions of
Latvia, representing both rural and urban areas
and varied socio-economic backgrounds. One
school was located in a city of 100,000 people,
three schools were in towns of approximately
10,000, and one was in a small town of ap-
proximately 4,000.
that they could conduct one of the two given ex-
periments or design their own. The great
majority of students chose to conduct the two
experiments provided in the program. They had
to have at least five participants in each treat-
ment group. Students had to report their experi-
ments using a two-page experiment report form
that contained spaces for introduction, method,
results, conclusions and references sections that
were broken down in several subsections. The
last part of the instructional program provided
students with detailed instruction on how to
complete the experiment report form.
Materials
A
12-lesson instructional program entitled
The Learning Explorations program was or-
ganized into six sections, each of which required
about two 40-min class periods. The program
materials consisted of a student book and a
teacher guide. The student book contained all
the information presented during instruction,
and examples of experiments, practice exercises,
and worksheets. The teacher guide included
step-by-step lesson plans on how to use the stu-
dent book, descriptions of instructional ac-
tivities, a posttest, a rating scale for student
projects, and transparencies and handout
masters. All materials used in the study were in
the Latvian language.
Learning Explorations was developed in print
form in Latvian for use in the study. The pur-
pose of the program was to introduce high
school students to the basic concepts of scientific
research and to develop their skills in designing
experiments about learning. The program was
intended for use in introductory psychology
classes or as supplemental material in other clas-
ses in which students were introduced to the
design of independent research projects. Learn-
ing Explorations was designed to teach the fol-
lowing six objectives, which are listed here in an
abbreviated form. Students were (a) to identify
the major experimental design concepts
(hypothesis, dependent variable, control group,
independent variable, treatment conditions,
constants), (b) to indicate these concepts in an
experiment scenario, (c) to detect common ex-
perimental design flaws in an experiment
scenario, and (d) to write a set of experimental
procedures and (e) summarize the results for an
experiment scenario. The sixth objective re-
quired students (f) to independently conduct an
experiment about learning and to produce a
simple written report of the results.
Criterion Measures
Four criterion measures were used in the study:
(a) ratings of the student projects, (b) posttest
scores, (c) student attitude surveys, and (d)
teacher attitude surveys.
Ratings of student research projects. During the
program, students received a two-page report
form on which to describe the results of their ex-
periments. An independent rater trained by the
researcher evaluated all student projects on the
project rating scale, a descriptive rating scale for
evaluating written research reports developed
for the study. The rater was a former classroom
teacher with considerable experience rating stu-
dent projects. The researcher rated one sixth of
all student projects to determine inter-rater
reliability. The raters were blind to the ex-
perimental condition and the school of origin for
each project. The Pearson correlation coefficient
The first part of the program introduced stu-
dents to the basic experimental design concepts
and common experimental design flaws. Stu-
dents were then provided with a problem
scenario of a fictional high school in Latvia that
had recently received funding for purchase of
new textbooks. They were asked to assist the
school’s textbook committee in developing
criteria for selecting the best textbooks. Students
were provided with two example experiments,
including complete experimental procedures
and instructions for participants, and were told

CLASSROOM EVALUATION STRATEGIES
65
for the inter-rater reliability between the ratings
of the final student projects by the independent
rater and the researcher was .90.
Figure 1
Criteria on the project rating
scale
Introduction
The project rating scale consisted of 15 items
rated on a three-point scale from 0 to 2. The 15
criteria included in the project rating scale are
listed in Figure 1. A sample item from the project
rating scale is provided below. Two points were
assigned for above average performance and 0
points for below average performance, as shown
in the example below:
1. Describes the independent variable.
2. Describes the dependent variable.
Method
3. The hypothesis is clear and specific.
4. Treatment conditions are fully described for
all groups.
5. Constants are similar for all groups.
6. There are at least 5 participants in each group.
7. Describes experimental procedures step-by-step.
8. Provides clear instructions for the participants.
9. Participants are randomly assigned to groups.
Results
Describes experimental procedures step-by-step.
❒❒ describes experimental procedures step-bystep, in-
cluding all materials and conditions (2 points)
❒❒ some steps and conditions are missing, but overall
the experiment can be carried out by following the pro-
cedures (1 point)
10. Compares mean scores for groups.
11. Evaluates the hypothesis based on the data.
Conclusions
❒❒ most steps are missing, no materials and conditions
are included (0 points)
12. Provides a feasible explanation of the results.
13. Provides suggestions for potential application
of the results.
Posttest. The posttest served as
a second
Overall Effort
criterion measure for assessing student perfor-
mance. The posttest consisted of 21 multiple-
choice and short-answer items, and had a
maximum score of 30 because some items had
multiple-point answers. The posttest items were
directly aligned with the objectives of the in-
structional program. Internal reliability of the
posttest, using Cronbach’s alpha, was .75. A
sample multiple-choice item from the posttest is
provided below:
14. The work overall is thoughtful and shows
good understanding of experimental design
concepts.
15. The work overall is accurate.
8-item attitude survey, using Cronbach’s alpha,
was .72. The 3 remaining items were open-
ended questions dealing with student likes, dis-
likes, and suggestions for improvement to the
instructional program. Four additional Likert-
type items were added to the 11-item student
survey for subjects in the self-plus-teacher
evaluation condition. These items asked about
student attitudes toward self-evaluation.
Which of the following is the factor that is changed on pur-
pose? Circle the correct answer.
a. Constant.
b. Control group.
c. Dependent variable.
d. Independent variable.
Student attitude survey. An 11-item attitude sur-
vey served as the criterion measure for assessing
student attitudes and motivation toward the in-
struction. The items were four-choice Likert-
type questions with the response choices being
strongly agree, scored as (3), agree (2), disagree (1)
and strongly disagree (0). These eight items dealt
with topics such as: Did students like the pro-
gram? Did they like conducting experiments?
Were they confident in their ability to conduct
experiments and write research reports as a
result of the program? Internal reliability for the
Teacher attitude survey. A 19-item attitude sur-
vey served as the criterion measure for assessing
teacher attitudes. Nine items asked teachers
about their delivery of the program, 6 Likert-
type items assessed teacher attitudes toward
various aspects of the instruction, and 4 open-
ended questions asked teachers for suggestions
on how to improve the instructional program.
The survey for the self-plus-teacher evaluation
group contained 3 additional items asking
teachers about their attitudes toward student
self-evaluation.

66
ETR&D, Vol. 50, No. 3
Procedures
ing and produced written reports. They formal-
ly self-evaluated their work at the same two
times (initial experimental design diagram and
initial final report) during the instruction and
used the same project rating scale as the teachers
in the teacher-evaluation condition. In addition,
the students received teacher feedback on their
written reports. Once students had completed
the self-evaluations of each product, they
handed in to the teacher their initial experimen-
tal design diagram and their initial report with
their self-evaluations of each. The teacher then
wrote her evaluation on the project rating scale
in the same manner as teachers in the teacher-
evaluation condition. Students were then able to
revise their work based on their own self-
evaluation and on the teacher evaluation.
The researcher assigned the six teachers to one
of the three treatment conditions (no-evaluation,
teacher-evaluation, and self-plus-teacher evalu-
ation). Each teacher taught two classes of stu-
dents. In order to assign teachers to treatments,
the researcher ranked all pairs of classes for each
teacher from the highest achieving to the lowest
achieving, based on the student 9th Grade
Graduation Exam scores in mathematics and the
Latvian language. The pairs of classes for each
teacher were divided into high-achieving and
low-achieving classes using a median split.
Teachers with classes from each group were
then randomly assigned to one of the three treat-
ments. All experimental groups completed the
Learning Explorations program with only the
variations in evaluation conditions described
below.
All teachers received the same version of the
instructional program. Teachers in the no-
evaluation group received no additional instruc-
tions for use of the program. Teachers in the
remaining two treatments received additional
instructions describing the evaluation proce-
dures that they were expected to complete for
their evaluation condition.
Students in the no-evaluation group con-
ducted their experiments about learning and
produced written reports. They were provided
with the project rating scale before they began
work on their experiments, but received no for-
mal feedback from the teacher. They were not
asked to evaluate their own work. Students in
this group produced the initial experimental
design diagrams before the start of the experi-
ment, wrote the initial final report after the ex-
periment, and could revise their work before
submitting their final report to the teacher.
The teachers were told that the purpose of the
research study was to investigate the effects of
student self-evaluation and teacher evaluation
on student performance. They were told that
there were three treatment groups in the study
and that each group would be doing slightly dif-
ferent things in the program as specified in the
instructions. The teachers received no additional
training on the use of the materials.
Students in the teacher-evaluation group
conducted their experiments about learning and
produced written reports. They received written
teacher evaluation in the form of feedback on the
initial version of their experimental design
diagrams before the start of the experiment and
on the initial versions of their final reports. The
teachers used the project rating scale for evaluat-
ing student work. The teachers checked the most
appropriate evaluative description for student
research projects for each of the 15 criteria on the
rating scale and wrote suggestions for improve-
ment. Students were then able to revise their ini-
tial experimental design diagrams and initial
final reports into final form by incorporating
teacher suggestions into them.
On average, teachers devoted 10 40-min class
periods to teaching the program, with a range of
9 to 12 class periods. There were no significant
differences among the treatment groups as to
the time spent on the program. The no-evalua-
tion group spent, on average, 11 class periods,
while the teacher-evaluation and the self-plus-
teacher evaluation groups spent, on average, 10
class periods. Two teachers taught the program
as part of a high school psychology course, and
the remaining four taught it during other clas-
ses. Students in all schools received a grade
based on their posttest results and research
project ratings. A day before the last class, the
students took the course posttest. On the final
day of the program, students submitted their
Students in the self-plus-teacher evaluation
group conducted their experiments about learn-

CLASSROOM EVALUATION STRATEGIES
67
reports of their experiments and completed an
attitude survey about the program.
correction was used setting alpha level at .006
(.05/8 items). The frequency of constructed
responses to the three open-ended questions on
the attitude survey was computed and is
reported later in this paper. Mean scores on the
Likert-type items and constructed response on
the open-ended items for the teacher survey
were also analyzed.
The primary researcher, a Latvian doctoral
student at an American university, traveled
from the United States to Latvia to coordinate
the study and observe its implementation.
During the study, the researcher observed two
class periods in each teacher’s classroom to en-
sure that the instructional program was proper-
ly implemented and that the evaluation
procedures for each treatment condition were
closely followed. During classroom observa-
tions, the researcher wrote down cases when the
teacher had difficulties with providing accurate
explanations and responses to student questions
or with conducting the instructional activities in-
cluded in the program. After classroom observa-
tions, the researcher conducted an informal
debriefing session with each teacher, asking
them to share their overall feelings about the
program.
RESULTS
Results are discussed below by achievement,
student attitudes, teacher attitudes, student and
teacher ratings of initial research reports, and
classroom observations.
Achievement
The mean scores and standard deviations for
both the project reports and the posttest are
shown in Table 1.
After the experiment, the researcher also col-
lected and reviewed the initial experimental
design diagrams, the initial final research
reports, and student and teacher evaluations of
these reports. The researcher counted the num-
ber of cases where initial student reports had
received lower than the maximum rating on
each criterion of the project rating scale from the
students and teachers. The researcher also ex-
amined the nature of student and teacher com-
ments.
Ratings of Student Research Reports. The table re-
veals that the mean scores for the project reports
were 14.87 (50%) for the no-evaluation group,
17.49 (58%) for the teacher-evaluation group,
and 18.10 (60%) for the self-plus-teacher evalua-
tion group.
The ANOVA conducted on the project report
scores yielded a significant overall difference, F
(2, 186) = 5.70, p < .01. Because the overall F
value was significant, follow-up Fisher’s least
significant difference (LSD) tests were per-
formed to determine whether significant dif-
ferences occurred between the mean scores for
each pair of treatments. These tests revealed that
both the teacher-evaluation group and the self-
plus-teacher evaluation group had significantly
higher scores at the p < .01 level on their project
reports than the no-evaluation group. The dif-
ference in the project report scores between the
teacher-evaluation group and the self-plus-
teacher evaluation group was not statistically
significant.
A diagram summarizing the experimental
design is shown in Figure 2.
Data Analysis
The data analysis for student performance was
carried out as two separate one-way analyses of
variance (ANOVAs), one for the student project
scores and one for the posttest scores with three
groups (no-evaluation, teacher-evaluation, and
self-plus-teacher evaluation) in each analysis.
Student attitude data were analyzed using a 3
(Treatments) × 8 (Survey Items) multivariate
analysis of variance (MANOVA) for the mean
scores on the survey items. To control for Type I
error across univariate follow-up tests for the
eight items on the student survey, Bonferroni
Posttest. The mean posttest scores were 21.87
(73%) for the no-evaluation group, 23.96 (80%)
for the teacher-evaluation group, and 22.78

68
ETR&D, Vol. 50, No. 3
Figure 2
Summary of Experimental Design
Assignment of Teachers to Treatment Groups
(Two teachers per treatment group, two classes per each teacher)
↓
Treatments
No Evaluation
Teacher-Evaluation
Self-Plus-TeacherEvaluation
• Students receive instruction
on experimental design
components.
• Students receive instruction
on experimental design
components.
• Students receive instruction
on experimental design
components.
• Students design experiments. • Students design experiments.
• Students design experiments.
• Teacher provides feedback
on experimental design
diagrams.
• Students self-evaluate and
teacher provides feedback
on experimental design
diagrams.
• Students receive instruction
on writing reports.
• Students receive instruction
on writing reports.
• Students receive instruction
on writing reports.
• Students conduct experiments • Students conduct experiments
• Students conduct experiments
and write draft reports.
and write draft reports.
and write draft reports.
• Teacher provides feedback
on draft reports.
• Students self-evaluate and
teacher provides feedback
on draft reports.
• Students produce final reports. • Students produce final reports.
• Students produce final reports.
↓
Posttest
↓
Attitude Survey (Student)
Students Submit Experiment Reports
↓
Attitude Survey (Teacher)
(76%) for the self-plus-teacher evaluation group.
The ANOVA conducted on the posttest scores
yielded a significant overall difference, F (2, 186)
= 4.11, p < .05. Follow-up LSD tests revealed that
the teacher-evaluation group scored significant-
ly higher at the p < .01 level than the no-evalua-
tion group. The differences in the posttest mean
scores were not statistically significant between
the no-evaluation group and the self-plus-
teacher evaluation group or between the
teacher-evaluation group and the self-plus-
teacher evaluation group.
Student Attitudes
The mean attitude scores by treatment for the
student responses to the eight statements on the
four-point Likert-type attitude survey ad-
ministered after completion of the instructional

CLASSROOM EVALUATION STRATEGIES
69
Table 1
Mean Project Report and Posttest
Scores by Treatment Group
instructional program. The three highest-rated
statements on the survey were “I now under-
stand how to conduct experiments” (M = 2.11),
“The program helped me learn how to conduct
experiments” (M = 2.07), and “I am satisfied
with my experiment and report” (M = 2.07). The
two lowest-rated statements were “I liked plan-
ning and conducting experiments” (M = 1.54)
and “The information was easy to understand”
(M = 1.70).
Treatment
Self +
Teacher
Evaluation Evaluation
No
Evaluation
(n = 62)
Teacher
Measure
(n = 69)
(n = 58)
Project Reports
M
SD
14.87
(5.65)
17.49
(5.83)
18.10
(5.39)
Posttest
M
21.87
(4.17)
23.96
(3.71)
22.78
(4.68)
The data in Table 2 were analyzed using a 3
(Treatment) × 8 (Survey Items) MANOVA to test
for significant differences. The overall means
were significantly different for the three treat-
ment groups, Wilks’s Λ = .723, F (16, 322) = 3.54,
p < .001. Follow-up univariate ANOVAs were
conducted on each of the eight items. Using Bon-
ferroni adjustment, each ANOVA was tested at
.006 level (.05/8 items). The analysis revealed
significant attitude differences between the
treatment groups on three of the items.
SD
Note. The maximum possible score on both the project
reports and the posttest was 30.
program are shown in Table 2. Responses were
scored on a four-point scale from 3 for the most
positive response to 0 for the most negative
response.
Posthoc LSD tests setting alpha level at .006
were performed on the three significant items.
As shown in Figure 3, students in the no-evalua-
tion group had significantly more favorable
The overall mean score across the eight Stu-
dent Attitude Survey items was 1.87,
a
moderately favorable rating indicating general
agreement with positive statements about the
Table 2
Mean Ratings on Student Attitude Survey
Treatment
Self +
No
Evaluation
Teacher
Evaluation
Teacher
Evaluation
Item
F
p
1. The program was interesting.
2.02
1.73
1.65
1.68
1.52
1.92
10.89
2.09
< .006*
ns
2. It was easy to learn from the .
program
3. The information was easy to
understand.
1.84
2.20
2.04
1.86
1.77
1.51
2.00
2.14
1.83
1.51
2.08
1.80
1.78
2.00
2.14
2.20
1.32
1.94
1.85
3.54
1.29
.55
ns
ns
4. The program helped me learn
how to conduct experiments.
5. I now understand how to
conduct experiments.
ns
6. I can now independently plan
and conduct experiments.
5.77
5.32
2.20
3.54
.006
.006
ns
7. I liked planning and conducting
experiments.
8. I am satisfied with my experiment 2.18
and report.
Overall means
1.96
< .001
* .006 was the probability level at which differences were tested for significance using the Bonferroni correction (.05/8 items).

70
ETR&D, Vol. 50, No. 3
Figure 3
Mean Differences by Treatment for Statistically Significant Student Attitude Items
scores on the item, “The program was interest-
ing,” than students in the teacher-evaluation
and self-plus-teacher evaluation groups (M =
2.02 for no-evaluation, M = 1.65 for teacher-
evaluation, and M = 1.52 for self-plus-teacher
evaluation). In addition, students in the no-
evaluation group had significantly more
favorable scores on the item, “I liked planning
and conducting experiments,” than students in
the self-plus-teacher evaluation group (M = 1.77
for the no-evaluation, and M = 1.32 for the self-
plus-teacher evaluation). On the other hand, stu-
dents in the self-plus-teacher evaluation group
had significantly more favorable scores on the
item, “I can now independently plan and con-
duct my own experiments,” than students in the
no-evaluation group and in the teacher-evalua-
tion group (M = 2.20 for the self-plus-teacher
evaluation, M = 1.86 for the no-evaluation, and
M = 1.83 for the teacher-evaluation).
on the student survey for the self-plus-teacher
evaluation group was 1.90, indicating
a
moderately favorable attitude toward self-
evaluation. The two highest-rated items were
“My self-assessment was honest” (M = 2.20) and
“My self-assessment was important to the
teacher” (M = 2.12). The two lowest-rated items
were “I liked assessing my own work” (M =
1.52), and “When I have to assess my own work,
I know what to improve” (M = 1.77).
Summary of the open-ended responses on
the student attitude survey indicated that what
students liked most about the program was par-
ticipating in and conducting their own experi-
ments, a response given by a total of 53 of the
189 students (28%) across the three groups. The
second most common response to what students
liked most was the fact that information in the
program was well presented and comprehen-
sive, a response indicated by 41 students (22%),
and third was that there were interesting and
The mean score for the four additional items

CLASSROOM EVALUATION STRATEGIES
71
Student and Teacher Ratings of Initial
Research Projects
useful experiment examples (35 students, 19%).
When asked what they liked least about the pro-
gram, 53 students (28%) indicated the use of
complex language and difficult terms, 36 (19%)
reported that they disliked too much reading
and redundant examples, and 25 (13%)
responded that they liked everything. When
asked for suggestions on how to improve the
program, 52 students (28%) suggested shorten-
ing and simplifying text, 22 (12%) recommended
including more practical experiments and exer-
cises to be conducted in class, and 12 (6%) indi-
cated that more class time should be allowed for
delivery of the program. The pattern of student
comments did not differ appreciably across the
treatment groups.
Student and teacher ratings of initial experimen-
tal design diagrams and initial research reports
were also examined. Examination revealed that
students in the self-plus-teacher evaluation
group tended to rate their initial reports higher
than did their teachers. Students in the self-plus-
teacher evaluation group assigned lower than
maximum ratings of 2 on an average of only 3 of
the 15 criteria on the initial report, while
teachers in both groups with the teacher-evalua-
tion component assigned lower than maximum
ratings of 2 on an average of 6 of the 15 criteria.
In addition, student ratings of their reports did
not contain additional comments with ideas for
improvement, while teacher evaluations con-
tained several such comments. Teacher com-
ments, for the most part, provided students with
knowledge of results, such as “You did not
describe how the treatment groups were
formed” or “You did not list all constants”.
Teacher Attitudes
The six Likert-type items on the teacher survey
were scored on a three-point scale from 2 for the
most positive response to 0 for the least positive
response. The overall mean rating for the six
teachers on the survey was 1.53, a favorable
rating indicating agreement with positive state-
ments about the instructional program. The
highest-rated item was “The material develops
student skills in designing and conducting their
own experiments well” (M = 1.83). All teachers
said that the lesson procedures in the teacher
guide were clear, that they followed the teacher
guide closely when working with the students,
and that they would use the material again in
their work with students.
Classroom Observations
Classroom observations by the experimenter
revealed that on several occasions teachers had
difficulty delivering the program optimally be-
cause of their lack of sufficient content
knowledge. In addition, it was difficult for the
teachers to properly carry out classroom
demonstration experiments described in the
teacher guide because they had not observed ex-
periments themselves. During informal debrief-
ing sessions with the experimenter after the
classroom observations, several teachers noted
that rating the student project reports was very
time consuming and that it was difficult for
them to fit this task into their busy schedules.
The two teachers in the self-plus-teacher
evaluation group thought that the program
developed student understanding of the basic
experimental design concepts in psychology to a
greater extent than did the teachers in the other
two treatment groups. When asked about their
attitudes toward student self-evaluation, the
two teachers in the self-plus-teacher evaluation
group felt strongly that self-evaluation helped
students produce higher quality experiment
reports and that they would use self-evaluation
in their teaching in the future.
DISCUSSION
This study examined the effects of teacher
evaluation and the combination of teacher
evaluation and student self-evaluation on stu-
dent performance and attitudes. Students in the
teacher-evaluation and the self-plus-teacher

72
ETR&D, Vol. 50, No. 3
evaluation conditions received significantly
higher ratings on their research projects than did
students in the no-evaluation condition. In addi-
tion, students in the teacher-evaluation condi-
tion scored significantly higher on the posttest
than the other two groups. Students in the no-
evaluation group had more positive attitudes
toward the program than did students in the
teacher-evaluation and the self-plus-teacher
evaluation group, and they reported that they
enjoyed conducting experiments more than did
students in the self-plus-teacher evaluation
group. However, students in the self-plus-
teacher evaluation group had greater confidence
about their ability to independently conduct ex-
periments in the future than the other two
groups.
tent assessed on the posttest. This information
would have been covered in the direct instruc-
tion for both groups, but may often have been
reemphasized in the teacher evaluation/feed-
back to students who did not apply it well to
their experimental designs and draft reports,
thus enabling them to learn it better prior to the
posttest. In addition, reading student initial
reports and providing feedback might have
enabled teachers to gain information on student
performance during instruction. Teachers could
then use this information to improve their in-
struction by reteaching certain concepts as
necessary, thereby helping students to improve
both student posttest scores and the quality of
their research projects.
The second research question asked whether
the combination of teacher evaluation and stu-
dent self-evaluation would have a different ef-
fect on student performance than teacher
evaluation alone. There was no strong evidence
from the study that self-evaluation when used in
combination with teacher evaluation produced
an improvement in student performance over
using teacher evaluation alone. The fact that the
addition of self-evaluation did not significantly
improve student performance could be at-
tributed to the greater evaluation expertise of
the teachers described above. Most likely,
teacher evaluations were more complete and ac-
curate than student self-evaluations, and stu-
dent self-evaluations therefore did not add
significantly to the quality of student projects.
Another reason for the lack of improvement in
student performance could be that students
were not familiar with self-evaluation as an
evaluation strategy because it is rarely used in
Latvian schools. Providing students with prac-
tice in self-evaluation, such as applying the
project rating scale to research report examples,
might help students gain a better understanding
of the use of evaluation criteria and become
more accurate at evaluating their own work.
The first research question asked if teacher
evaluation would have a positive effect on stu-
dent performance. Students who received
teacher evaluation, with or without self-evalua-
tion, produced higher quality reports of their ex-
periments than students who did not. Teachers
most likely had greater knowledge than stu-
dents about how to conduct experiments and
write reports. This knowledge would help the
teachers to provide higher quality feedback to
students than the students could generate on
their own. The better feedback, in turn, would
help the students learn what they needed to im-
prove and how to improve it. Analysis of the dif-
ferences between student and teacher ratings of
the initial research projects revealed that
teachers assigned lower than maximum ratings
on twice as many criteria on the project rating
scale as did students. In addition, teachers also
provided students with additional comments
suggesting improvements to their reports, while
student self-evaluations contained almost no
comments. Thus, the better student performance
on research reports under teacher-evaluation
conditions would appear to be due to better
evaluation and feedback provided by the
teachers.
The mean scores on the student projects were
lower than on the posttest, with overall means of
16.8 out of 30 possible for the project reports and
22.9 out of 30 possible for the posttest. Produc-
ing the research projects and reports was a more
difficult and advanced task than performing
well on the posttest. The posttest required only
The significantly higher scores on the posttest
for the students in the teacher-evaluation group
over those in the no-evaluation group may have
been due, at least in part, to information
provided in the teacher evaluation/feedback
that was directly relevant to the program con-

CLASSROOM EVALUATION STRATEGIES
73
recognition and recall of concepts and applica-
tion of knowledge to scenarios provided by the
experimenter. Planning an experiment and
producing a research report, on the other hand,
required application of knowledge to generate
new products in the form of an experimental
design and a written report. In addition, design-
ing an experiment and producing a written re-
search report required more self-direction by the
students than responding to the posttest. Stu-
dents had an especially hard time in their experi-
evaluate their own work. Their greater con-
fidence in their ability supports the findings of
other researchers that students who self-monitor
and self-evaluate their progress have higher self-
efficacy perceptions than those who do not
(Schunk, 1989; Zimmerman & Kitsantas, 1999).
The three highest-rated items on the student
attitude survey all dealt with students’ learning
how to conduct experiments and with their
satisfaction with their experiments and reports.
These items dealt quite directly with attitudes
toward desired outcomes of the program, that is,
conducting and reporting experiments, rather
than with process variables such as whether the
program was interesting, easy-to-understand, or
easy to learn from. It is encouraging that stu-
dents generally agreed that they acquired these
desirable outcomes irrespective of their treat-
ment group.
ment reports coming up with
a precise
description of experimental procedures and
writing conclusions. Several students had dif-
ficulties managing their experimental subjects
and properly implementing their intended ex-
perimental procedures.
The third research question related to the ef-
fects of the experimental treatments on student
attitudes. The attitude results revealed that stu-
dents in the no-evaluation condition were more
interested in the instructional program than
those in the other two experimental treatments
and that they liked the program significantly
better than those in the self-plus-teacher evalua-
tion condition. These results may be attributable
to the fact that the program overall was easier
for students in the no-evaluation condition.
They did not have to formally evaluate their
work and they did not receive any feedback
either from the teacher or from their own self-
evaluations regarding changes needed to im-
prove their projects. The finding that students
have more positive attitudes toward treatments
that are easier, but less effective instructionally,
is consistent with the results of studies by other
The highest-rated item by the teachers, “The
program develops student skills in designing
and conducting their own experiments well,” is
quite consistent with highly rated items by the
students indicating that the program was effec-
tive in helping them to learn to conduct and
report experiments. In addition, despite the lack
of strong evidence that self-evaluation improved
student performance, teachers in the self-plus-
teacher evaluation group showed support for
self-evaluation by indicating that it helped stu-
dents produce higher quality reports and that
they would use it in their teaching in the future.
This study revealed that providing students
with teacher evaluation and feedback in the for-
mative stages of student work and having stu-
dents incorporate the feedback into their final
products improves student performance. Yet the
teachers considered these strategies to be too
time-consuming. Development of formative
evaluation strategies that are less time-consum-
ing may be necessary to ensure their use by
teachers. One option may be to explore further
the effects of self-evaluation on student perfor-
mance in order to provide students with timely
formative feedback without overburdening the
teacher. Another option may be to have teachers
provide group feedback to the students regard-
ing the common strengths and weaknesses of
their projects during the formative stages of
their work. Emphasizing sound instruction that
researchers (Hannafin
&
Sullivan, 1996;
Schnackenberg, Sullivan, Leader & Jones, 1998).
That students in the self-plus-teacher evalua-
tion group reported significantly greater con-
fidence in their ability to independently conduct
experiments in the future than their counter-
parts in the other two groups may have been re-
lated to their involvement in the self-evaluation
process. Formal self-evaluation of their projects
may have caused these students to think that
they better knew the criteria for designing and
reporting a research project. It is possible that
they felt more in control of their learning than
students in the other two groups who did not

74
ETR&D, Vol. 50, No. 3
is aligned with objectives and providing stu-
dents with multiple guided-practice oppor-
tunities during instruction might also reduce the
need for detailed teacher feedback on student
work during the evaluation stage.
Zane Olina [olina@coe.fsu.edu] is an assistant
professor of Instructional Systems at Florida State
University, Tallahassee, and Howard J. Sullivan
[sully@asu.edu] is a professor of Educational
Technology at Arizona State University, Tempe.
This study was conducted while Dr. Olina was a
student at Arizona State.
Classroom observations revealed that, even
though teachers were provided with detailed in-
structional procedures in the teacher guide, they
did not do a particularly good job of teaching the
instructional content that, in general, was rela-
tively unfamiliar to them. The difficulties that
the teachers had delivering the instructional
program could be attributed to the fact that in-
structional methods and assessment strategies
included in the program were relatively new to
the Latvian teachers. The program included
project-based teaching requiring teachers to
facilitate student research projects. In addition,
the Learning Explorations program included per-
formance assessment as the main indicator of
student achievement, which contrasts with more
traditional assessments in Latvian schools that
focus primarily on factual, conceptual and pro-
cedural learning. Providing teachers with
pretraining on the content and use of the instruc-
tional program is one option that might enable
them to deliver instructional programs contain-
ing relatively unfamiliar content and novel in-
structional approaches more effectively.
REFERENCES
Black, P., & William, D. (1998). Assessment and class-
room learning. Assessment in Education: Principles,
Policy & Practice, 5 (1), 7–75.
Bransford, J.D., Brown, A.L., & Cocking, R.R. (Eds.).
(1999). How people learn: Brain, mind, experience, and
school. Washington D.C.: National Academy Press.
Cardelle-Elawar, M., & Corno L. (1985). A factorial ex-
periment in teachers’ written feedback on student
homework: Changing teacher behavior a little rather
than a lot. Journal of Educational Psychology, 77, 162–
173.
Crooks, T.J. (1988). The impact of classroom evaluation
practices on students. Review of Educational Research,
58, 438–481.
Fontana, D., & Fernandez, M. (1994). Improvements in
mathematics performance as a consequence of self-
assessment in Portuguese primary school pupils.
British Journal of Educational Psychology, 64, 407–417.
Gipps, C.V. (1994). Beyond testing: Towards a theory of
educational assessment. London: The Falmer Press.
Hannafin, R.D., & Sullivan, H.J. (1996). Preferences
and learner control over amount of instruction. Jour-
nal of Educational Psychology, 88, 162–173.
Limitations on the number of classes avail-
able for this study precluded use of a student
self-evaluation-only treatment that would have
permitted analysis of the unique contribution of
self-evaluation to student performance and at-
titudes. Adding a student self-evaluation group
to the design of a future study would address
this issue and permit a more complete analysis
of the effects of self-evaluation and teacher
evaluation. Discussing with students the impor-
tance of self-evaluation and providing them
with instruction on the use of an evaluation in-
strument such as the project rating scale from
this study might also help them improve their
own self-evaluations. Future research that inves-
tigates potentially productive ways to incor-
porate teacher evaluation and student
self-evaluation into classroom instruction
should help us understand the most effective
strategies for using classroom evaluation to im-
prove student learning.
Hughes, B., Sullivan, H.J., & Mosley, M.L. (1985). Ex-
ternal evaluation, task difficulty, and continuing
motivation. Journal of Educational Research, 78, 210–
215.
Latvian Ministry of Education and Science Center for
Curriculum Development and Examination. (1998).
National Standards of Compulsory Education. Liel-
varde, Latvia: Lielvards Ltd.
Maqsud, M., & Pillai, C.M. (1991). Effect of self-scoring
on subsequent performances in academic achieve-
ment tests. Educational Research, 33, 151–154.
Natriello, G. (1987). The impact of evaluation proces-
ses on students. Educational Psychologist, 22, 155–175.
Page, E.B. (1958). Teacher comments and student per-
formance: A seventy-four-classroom experiment in
school motivation. Journal of Educational Psychology,
49, 173–181.
Rolheiser, C. (Ed.). (1996). Self-evaluation . . . Helping
kids get better at it: A teacher’s resource book. Toronto,
Canada: OISE/UT.
Ross, J.A., Rolheiser, C., & Hogaboam-Gray, A. (1999).
Effects of self-evaluation training on narrative writ-
ing. Assessing Writing, 6, 107–132.
Sadler, D.R. (1989). Formative assessment and the

CLASSROOM EVALUATION STRATEGIES
75
design of instructional systems. Instructional Science,
18, 119–144.
Salili, F., Maehr, M.L., Sorensen, R.L. & Fyans, L.J., Jr.
(1976). A further consideration of the effects of
evaluation on motivation. American Educational Re-
search Journal, 13, 85–102.
Schnackenberg, H.L., Sullivan, H.J., Leader, L.F., &
Jones, E.E.K. (1998). Learner preferences and
achievement under differing amounts of learner
practice. Educational Technology Research and Develop-
ment, 46(2), 5–15.
en1.pdf
Stewart, L.G., & White, M.A. (1976). Teacher com-
ments, letter grades, and student performance:
What do we really know? Journal of Educational
Psychology, 68, 488–500.
Stiggins, R.J. (2001). Student-involved classroom assess-
ment. Upper Saddle River, NJ: Merrill Prentice Hall.
Story, N.O., & Sullivan, H.J. (1986). Factors that in-
fluence continuing motivation. Journal of Educational
Research, 80, 86–92.
Schunk, D.H. (1989). Social cognitive theory and self-
regulated learning. In B.J. Zimmerman & D.H.
Schunk (Eds.), Self-regulated learning and academic
achievement: Theory, research and practice (pp. 83–110).
New York: Springer-Verlag.
Scriven, M. (1967). The methodology of evaluation. In
R.W. Tyler, R. Gagné, & M. Scriven (Eds.), Perspec-
tives of curriculum evaluation: Vol. 1. AERA monograph
series on curriculum evaluation (pp. 39–83). Chicago:
Rand McNally.
Thomas, J.W., Bol, L., Warkentin, R.W., Wilson, M.,
Strage, A., & Rohwer, W.D., Jr. (1993). Interrelation-
ships among students’ study activities, self-concept
of academic ability, and achievement as a function
of characteristics of high-school biology courses. Ap-
plied Cognitive Psychology, 7, 499–532.
Wiggins, G. (1998). Educative assessment: Designing as-
sessments to inform and improve student performance.
San Francisco: Jossey-Bass.
Wolf, D., Bixby, J., Glenn, J., III, & Gardner, H. (1991).
To use their minds well: Investigating new forms of
student assessment. Review of Research in Education,
17, 31–74.
Shepard, L.A. (2000, October). The role of assessment
in a learning culture. ER Online [Online], 28. Avail-
able URL: http://www.aera.net/meeting/am2000
/wrap/praddr10.htm
Soros Foundation-Latvia. (2001, December). A passport
to social cohesion and economic prosperity: Report on
education in Latvia in 2000 [Online]. Available URL:
http://www.politika.lv/polit_real/files/lv/Izgp_
Zimmerman, B.J., & Kitsantas, A. (1999). Acquiring
writing revision skill: Shifting from process to out-
come self-regulatory goals. Journal of Educational
Psychology, 91, 241–250.