LIBRA R Y
Michigan S are
University

    

Mail

OVERDUE FINES:

. ‘ 25¢ per do per item
.1 4‘3“}; 5 momma LIBRARY MATERIALS:

,. ~ ‘3. , ,y ’ Place in book return to remove
. I ‘3',” . charge from circulation records

 

leafiot’s WOO

 

 

 

THE EFFECTS OF FIXED AND ASCENDING CRITERIA
ON ACHIEVEMENT, ATTITUDE AND
STUDY EFFICIENCY IN MASTERY LEARNING

By
James Anthony D'Albro

A DISSERTATION

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Administration
and Higher Education

1980

ABSTRACT

THE EFFECTS OF FIXED AND ASCENDING CRITERIA
ON ACHIEVEMENT, ATTITUDE AND
STUDY EFFICIENCY IN MASTERY LEARNING

By

James Anthony D'Albro

Selection of the most appropriate criterion scores
for criterion-referenced testing under mastery learning
is uncertain because there are no procedures for sorting
through the many approaches for setting such scores.
Implementation strategies for nearly all the approaches
are also lacking. This research was directed toward solving
the problem of setting the criterion in the mastery
learning strategy.

Thus. the overall purpose of this research was to
identify a criterion which yielded the best achievement
throughout a quarter course in greenhouse management while
requiring the least amount of study time and maintaining the
best student attitudes.

A different criterion level was set in each of

three 50-minute mastery classes. The criteria were used in

James Anthony D'Albro

conjunction with a mastery learning strategy. The criteria
were 80% fixed, 90% fixed, and ascending (80% for the first
unit and increasing 5% each unit test until 90% is reached,
additional units were graded at 90% of total points). A
fixed criterion was one which had the same standard

applied to each of the five unit tests in the quarter.

The textual material of the mastery strategy in
greenhouse management was divided into sections containing:
instruction for completion, objectives for each unit, a
set of review questions. and the lectures given by the
instructor.

Mastery in this research was defined by three
elements: instruction, grades, and testing. In order to
reach mastery of a unit of study, the students had to
attain the minimum criterion set for each instructional
unit in a given treatment group. Whenever the criterion
was not met on the first attempt of any test, the student
was provided with additional instructional assistance
and permitted to attempt mastery a second time. There was
a total of five unit tests, each test being given at the
end of a two week unit of instruction.

The control group received the same statement of
objectives as the mastery groups. They were lectured on
each unit of study, and were given the same test questions

as the mastery groups.

James A. D'Albro

The effect of the treatments was measured by a
multiple choice achievement test, an attitude scale, and
the total time spent on study as reported by students.

The groups taught under the mastery strategy
attained a significantly higher level of achievement than
the control. The study also showed that setting a higher
criterion or gradually raising the criterion did not yield
higher achievement than a lower criterion. The (90%)
fixed criterion produced less efficient study scheduling
than other criteria without any gain in achievement over
a lower criterion. The ascending criterion did not produce
the increase in achievement over the other criterion groups
that might have been expected. Student's attitudes toward
the course were also less positive when they were pushed
to meet higher levels of criterion.

The findings suggested that the 80% fixed criterion
could produce the best learning while maintaining the
most productive student attitudes. Moreover, this
criterion yielded the best student learning in the least

amount of total study time over the quarter.

DEDICATION

To My Wife D'Anne
For her patience, understanding,

and constant support

ACKNOWLEDGEMENTS

The author wishes to acknowledge the assistance
and advice given to this research by the following
committee members:

Dr. Van C. Johnson

Dr. Max R. Raines

Dr. J. Lee Taylor

Dr. Stephen L. Yelon

The author wishes to express his sincere thanks to
Dr. Stephen Yelon who spent a great deal of time editing
and suggesting areas of improvement.

A special recognition is extended to Dr. Robert
Smidt for his assistance on the use of statistics and
computers for this research. Also, Mr. Charles Strong
is recognized for his assistance on the technical writing

of this dissertation.

iii

TABLE OF CONTENTS
Page
LIST OF TABLES . . . . . . . . . . . . . . . . . . . vii
LIST OF FIGURES. . . . . . . . . . . . . . . . . . . ix

Chapter
I I INTRODUCTION. I I I I I I I I I I I I I I I

Statement of the Problem. . . . .
The Instructional Philosophy of Mastery
Learning . . . . . . . . .
Purpose of the Research . . . .

Importance of the Research. . . . . . . 9
Research Questions. . . . . . . . . . . 11
Research Hypotheses . . . . . . . . . . 11
Hypotheses Regarding Achievement. . . . 12
Hypotheses Regarding Attitude . . . . . 15
Hypotheses Regarding Time Spent on
Instruction. I I I I I I I I I I I I 18
Overview of Literature Survey - . . . . - 22

II. LITERATURE SURVEY . . . . . . . . . . . . . 23

Introduction. . . . . . . . . . . . . . . 23
Research Regarding Mastery Learning . . 25
Strategies Used to Set Criteria . . . . 27
Factors to Consider for Setting

Criteria . . . . . . . . . . . . 33

Setting the Level of the Criterion. . . . 34

Summary of Research Regarding Mastery
Learning . . . . . . . . . . . . . . 37
Study Time Needed to Attain Criterion . . 41
Summary of Study Time Needed to Attain
Criterion. . . . . . . . . . . . . 44
Student Attitudes and Learning for
Mastery. . . . . . . . . . . . . . . . 44
Summary of Literature on Student
Attitudes I I I I I I I I I I I I I I “8
Summary of Literature Survey and Relation
to Research Questions. . . . . . . . . 48

iv

Chapter

IIII PILOT STUDYI I I I I I I I I I I I I I I I
IntrOduCtion I I I I I I I I I I I I I I
Population and Sample. . . . . . . . . .
Course Material and Instruction. . . . .
Course Evaluation. . . . . . .
Other Types of Study Aids Used in the

Course. . . . . . .

Validity of the Achievement Test . .

Item Analysis of Unit Tests. . . . .

Reliability of the Achievement Test. .

Summary and Conclusion of the Validity
and Reliability of the Achievement

TeStI I I I I I I I I I I I I I I I
Development and Assessment of the Attitude
MeaSLlre I I I I I I I I I I I I I I I

IV. RESEARCH DESIGN AND PROCEDURES . . . . . .

Introduction . . . . . . .

Experimental Design. . . . . . . . . . .
Independent Variables. . . . . . . . .
Control of Internal Validity of

Treatments. . . . . . . . . . . . .
Dependent Variables. . . . . . . . . .

Procedures . . . . . . . . . . . . . . .
Population and Sample. . . . . . . . .

Treatments . . . . . . . . . . . . . .

Instrumentation and Data Collection. . .

Data Analysis. . . . . . . . . . . . . .

V. ANALYSIS OF THE RESULTS. . . . . . . . . .

Introduction . . . . . . . . . . . . . .
Analysis of Covariates . . . . . . . . .
Grade Point Average. . . . . . . . . .
Statistical Analysis of the Chi Square
Test of the Elective-Required

Covariate . . . . . . . .
Regression Analysis for the Age
Covariate . . . . . . . . . . . . .

Summary of Analysis of Covariates. . . .

Interaction of Criterion by Repeated
Measures on Mean Achievement . . . . .

Interaction of Criterion by Repeated

Measures on Mean Attitude. . . . . . .
Interaction of Criterion by Repeated

Measures on Mean Study Time. . . . . .
Summary. . . . . . .

Limitations of the Results on Attitude
and Study Time . . . . . . . . . . . .

93

95
96

96
102

106
110

114

Page
Chapter
VI. DISCUSSION, CONCLUSIONS AND RECOMMENDATIONS. 115

Introduction- - - - - - - - - . . . - - - 115
Experimental Design . . . . . . . . . . 118
The Sample of the Research. . . . . . . 119
Method of Data Collection . . . . . 122
The Importance of the Covariables . 123

Discussion of the Analyses of the Results 124
Discussion on the Results of Achievement 124

Discussion on the Results of Attitude . . 127

Discussion on the Results of Total Study

Time I I I I I I I I I I I I I I I I I 13]-

Conclusions . . . . . . . . . . . . . 137

Summary of Conclusions. . . . . . . . . 139
Recommendations and Further Questions . . I40

APPENDICES I I I I I I I I I I I I I I I I I I I I I 143

Appendix A. Course Objectives . . . . . . . . . 143
Appendix B. Edward's Criteria for Selecting
Attitude Statements . . . . . . . . 152
Appendix C. Attitude Survey A . . . . . . . . . 154
Attitude Survey B . . . . . . . . . 156

BIBLI OGRAPHY I I I I I I I I I I I I I I I I I I I I 158

vi

LIST OF TABLES
Page

Table 1. Table of specifications for direct assessment
of objectives covered on five unit achievement tests
in the pilot study. See Appendix A for number
corresponding to objective. . . . . . . . . . . . . 58

Table 2. Item analysis in % correct for each
question for all unit tests in Greenhouse
Management. *Items of less than 50% correct
response were revised . . . . . . . . . . . . . . . 62

Table 3. Reliability coefficients calculated by the
Livingston formula for criterion-referenced tests.
Values are shown for each unit test taken by
students during the pilot study . . . . . . . . . . 66

Table 4. The variable matrix is shown. The multiple
dependent measures are shown for each time for each
experimental variable . . . . . . . . . . . . . . . 76

Table 5. Prerequisite profile of students subjected
to the criterion treatments . . . . . . . . . . . . 82

Table 6. Additional course work taken by the students
in the stated criterion treatments. . . . . . . . . 82

Table 7. Analysis of variance for grade point
average 0 o o o o o o o o o o e o o o e e o o e o e 92

Table 8. Means of the four experimental groups on
grade point average . . . . . . . . . . . . . . . . 93

Table 9. Summary of data from the four experimental
groups based on a 7:3 ratio (e1ective:required).
Ratio was obtained when the pilot study was made. . 94

Table 10. Summary of data from the four experimental
groups based on 41:15 ratio of the observed totals. 94

Table 11. Statistics for regression analysis for the
age covariate . . . . . . . . . . . . . . . . . . . 95

Table 12. Univariate results for the criterion by
repeated measures interaction on achievement. . . . 98

Table 13. Mean achievement scores for each unit test
for the groups under study. . . . . . . . . . . . . 99

vii

Page

Table 14. Univariate ANOVA for the comparison
of the mean achievement of the criterion groups to
the contrOlI I I I I I I I I I I I I I I I I I I I I loo

Table 15. Univariate ANOVA for the comparison of the
mean achievement of the ascending criteria to the
fixed criteria . . . . . . . . . . . . . . . . . . . 101

Table 16. Mean attitude scores for each unit test
for the groups under study . . . . . . . . . . . . . 105

Table 17. Mean study time in hours for each unit for
the groups under study . . . . . . . . . . . . . . . 110

viii

LIST OF FIGURES

Page
Figure 1. Mean attitude score plotted over time
for each group under study. . . . . . . . . . . . 103
Figure 2. Mean study time plotted over time for
. . 109

each group under study. . . . . . . . . . . .

ix

CHAPTER I

INTRODUCTION

Statement of the Problem

Educators are faced with the same methods of
teaching and evaluating students that they used in the past.
Typically a lecture hall is used to assemble a class. The
instructor faces the class and begins an hour of talking
about a subject. The instructor then designs a test to
determine what has been remembered and/or understood. The
test is graded with the intention that most students get a
'C' grade. The assignment of grades is based on the
following assumptions: if students are normally distributed
with respect to aptitude for some subject and all students
are given exactly the same instruction, then achievement
measured at the completion of the subject will be normally
distributed. Aleamoni (March 1979) states that a
distribution of student grades follows a normal curve so
that there are 3 percent A's, 13 percent B's, 68 percent
0'3, 13 percent D's and 3 percent F's. But Block (1971)
suggests that American education must turn away from this
traditional method of teaching and evaluating. Schools must

provide successful and rewarding learning experiences for

2

most students, not just a few. He suggests that criterion-
referenced testing under the mastery strategy offers the
greatest potential for students.

Mastery learning (Block, 1971) offers a powerful
new approach to teaching which can provide almost all
students with the successful and rewarding learning
experiences now allowed to only a few. Block (1971)
suggests that 75 to 90 percent of the students can reach
the same high level of achievement as the top 25 percent
do under traditional group-based instructional methods.
Group-based instruction is teaching of a group of students
at a set hour in a set room.

Further, the mastery approach includes procedures
that are primarily designed for use in the group—based
instructional situation, where the time allowed for learning
is relatively fixed (Bloom,l968). Bloom's mastery strategy
minimizes the time a student needs to learn. Therefore,
most students can master the material within the calendar
instructional time available.

The next section explains the essential philosophy
of the mastery strategy. It also discusses some of the
important features of this innovative method.

The Instructional Philosophy of Mastery Learning

Mastery learning is an instructional philosophy and

an associated set of ideas about instruction. This

philosophy asserts that under appropriate instructional

conditions most students can learn what they are taught
(Block, 1974).

There are several procedures in the mastery
learning strategy which have made the above mentioned
philosophy a reasonable instructional approach (Bloom, 1974).

First, the idea of mastery does not require a
normal distribution of grades from A to F. Instead, it
suggests that each student should be given enough time to
master the subject matter being taught. Thus, most students
can get an 'A' grade.

Second, and more central to the mastery learning
strategy is the use of feedback and corrective procedures.
Bloom (1974) has stated that there are a variety of
procedures to provide practice and feedback. They are tests,
homework and workbooks. Brief diagnostic progress tests
has proved to be the most useful. The test shows what
the student has learned from a chapter, a unit or some
other learning sequence. It also is valuable feedback to the
student and the instructor on what aspects of the learning
unit are weak, needing correction and further study.

Third, the diagnostic test must have some value
which defines competence or mastery. The instructor can
select a certain number of points correct out of a total or
a percentage grade to define the needed level of competence.
The benefits of declaring points or a percent is that it
clearly relates achievement to the degree of mastery of

what is set out to be learned. It provides a standard

4

measure of achievement. Consequently, students are not
competing against one another. The manner by which levels
of mastery have been determined is arbitrary. Herein lies
a problem.

The philosophy has been converted into procedures
for grading and collecting of data upon which to set a
grade. As yet there is no sound basis for deciding whether
or not the student can be considered a 'master'. A master
is a student who has met or exceeded the criterion score
set by the instructor and therefore has learned to a
sufficient degree.

The evidence for a decision of mastery must be made
on the basis of grades. Therefore, the problem is the
setting of a certain criterion to determine grades. The
criterion represents the absolute performance standard
against which the sufficiency of each student's learning can
be evaluated and graded (Block, 1971). This standard should
indicate the specific amount of skills a student must show
before he or she can be judged to have mastered the skills
taught. The standard also indicates how well the student
has learned. In that respect, the instructor knows exactly
how much each student has learned. This is unlike a
relative standard of grading which judges students in
relation to others and not in relation to the course content.

An example of an absolute standard is given by
Bormuth (1970) and Glaser and Nitko (1970). They state that

criterion referenced tests are absolute and that these

5

standards define what proportion of a well defined body of
content and behavior the student is expected to learn.
Thus, a test with a standard of 85% suggests that a student
must show competency of the content to that level.

According to Block, (1970) evaluators have ignored
the problem of defining the criterion in an objective way.
They have developed increasingly sophisticated data
gathering instruments and procedures. But not one valid
technique has been stated for defining the criterion in an
objective manner. Hence, the degree of mastery of many
students is being misjudged.

For example, suppose a student's learning is
misjudged due to a poor criterion. The student may have to
review material already learned and in so doing waste
valuable study time. This time could have been spent
studying material of a more advanced stage. A continuation
of this defect in evaluation may eventually lead the student
to a poor attitude toward a subject, a major and even to
school itself. As Block (1970) states, this is most
unfortunate because accurate indications of the sufficiency
of a student's learning are crucial to his/her cognitive and
attitudinal outcomes.

Hambleton et a1. (1978) in their review of criterion
levels state that the matter of the determination of
criterion scores seems unclear. Further, they state that
there are no procedures for sorting through the numerous

approaches for determining the criterion scores in order to

6

select the most appropriate one for a given situation.
Implementation strategies for nearly all of the approaches
are also lacking. Hence, if we are to solve the problem of
setting the criterion in the mastery learning strategy, a
full scale directed effort must be undertaken to research
the problem further. This research attempts to partially
solve that problem.

In conclusion, the best available information on
setting a criterion suggests that criteria are set in an
arbitrary manner. It is up to an instructor to determine
the level of the criteria. Therefore, this research
addresses itself to the problem of the determination of a
criterion score used in the mastery strategy. The research
effort seeks to find an empirical basis for the criterion.
The score can then be implemented when the mastery strategy
is used. With this intention, the following purposes are
stated.

Purpose of the Researgh

Instructors have selected their criteria for
mastery without an explicit theory or any evidence suggesting
that those chosen over others are superior in fostering
student development (Block, 1970). Block (1970) goes on to
say, therefore, it is entirely possible that criteria
selected may not represent the best learning and attitudes
as other criteria.

The research proposed has been developed to correct

the major problem of setting a criterion within the mastery

7

strategy. Based on this research some criteria may be
decided as a result of evidence.

Four steps were taken within this research to
remedy the problem of setting a criterion. First, students'
achievement results from criterion referenced tests were
logged. These tests were diagnostic tests taken during
the course.

Second, the setting of the criterion was based on
a test performance which produced the highest achievement.
Therefore, the instructor can evaluate what has been
learned and how well it has been learned.

Third, the time to attain the criterion was within
a period allocated for the course. In addition, the time
to learn the content was efficient for the student.

Fourth, the attainment of high levels of achievement
on criterion tests did not sacrifice the attitude of the
student toward the course or major.

In an effort to correct the problem of setting a
criterion, the following purposes of this research are
stated:

1. It is expected that having a criterion as a goal
will be of a benefit to students to achieve each unit of
study in the most efficient manner. To accomplish this
one may use fixed or ascending criteria to evaluate the
performance of students on various units of study. Fixed
criteria are absolute standards set prior to testing. They

do not change from one test to the next. For example, an

8

80% criterion is the standard which all students are judged
on each unit test. In the case of an ascending criterion,
a first test may be assigned an absolute standard of 80%.
Each successive test will have a new and higher percent
standard by which achievement is judged.

It was hoped that by setting criteria on each test
we would be able to identify levels which, when maintained
throughout the learning, encourage students to learn
adequately and score well on a criterion measure (Bormuth,
1969 and Block, 1970). Thus, the attainment of the criterion
will indicate to the instructor that most course objectives
were learned by the students. In addition, the instructor
will be assured that the student has acquired a sufficient
amount of course information. Therefore, the student will
be judged competent in the subject.

2. The positive attitude of students may increase or
decrease in response to the difficulty of a criterion. For
example, a very high criterion of 90% may force students to
reach that level. But, it may also cause the attitude of
the student to decrease significantly. Since attitude may
change, it will be the purpose of the research to identify
a criterion which when used throughout the course, will
produce the best student attitude toward the course.

3. Since the efficiency of learning can be
interpreted as the total time it takes to learn a skill or
series of skills in a unit of study, it was also the purpose

of this research to manipulate the criterion or standard

9

used to judge students to see if students can be made to
reduce their time to learn over the duration of the course.

In summary, the intention of the research was to
establish a criterion which can produce the greatest
efficiency and achievement without a sacrifice of the
student's attitude toward the course. In addition, the aim
of high achievement will also foster greater efficiency of
study as manifested by time invested in the learning of
content. If these objectives are attained, then a partial
answer will be given as to a basis for setting the criterion
under the mastery strategy.

Finally, Block (1970) suggests that if instructors
can choose adequate performance levels, then educational
'programs might become more effective. Sound criteria make
this possible. So that the research is not understated,
the following statements are made to emphasize its
importance.

Importance of the Study

This study was important for the following three
reasons. First, the study provided a basis of setting
criterion levels for a mastery program. Thus, one can base
a criterion on experimental evidence. Through the
implementation of an improved method of criterion selection,
the student may be able to attain high levels of achievement.

Second, the proper selection of criterion is keyed
closely to the attitudes of the student. There must be

knowledge of a criterion which can produce a high level of

10

achievement without sacrificing the attitude of the student.
Block (1972) states that "the attitudinal changes which do
occur raises the important question of whether in pushing
some students to attain very high levels of performance
throughout their learning we are not, in fact, promoting
their intellectual development at the expense of their
feelings toward the material learned."

Lastly, the criterion selected must be attainable
within the time allowed for the course. In order to meet
the objective of an attainable criterion, the student must
be molded into more efficient behavior. It is suggested
that there may be particular criterion levels whose
attainment early in the sequence will progressively increase
the amount of material achieved per time later_in the
sequence. Students who learn under the mastery strategy,
therefore, may eventually be able to achieve their required
criterion level in the same amount of instructional time
that should ordinarily be expected of students who learn
under non-mastery conditions. Non-mastery refers to group-
based instruction whereby a student is instructed and
evaluated on his/her performance relative to others in the
class, and where curves are established to define a spread

or range of scores from'A'to'F'.

11

Research Questions

The need for this research has been stated in the
above sections. In general, the research is addressed
to the kinds of variations of presenting the criterion to
students in order to yield the greatest achievement. In
addition, in what ways can criteria be presented to students
so that they maintain a high achievement and a positive
attitude toward a subject? Lastly, the cognitive learning
of the course content should be done with a minimum amount
of time.

In an effort to provide a comprehensive answer to
a selection of a criterion, the research seeks to investigate
the following questions:

1. Does one criterion produce more achievement
than another?

2. Does one criterion produce better student
attitude than another?

3. Does one criterion produce more efficient
study scheduling than another?

Research Hypotheses

From the research questions for this study, the
following hypotheses were drawn. In each case, the
selection of criterion was tested for its effect on
achievement, attitude and total instructional time needed

to learn each unit of study.

12

Hypotheses RegardingiAchievement

The overall hypothesis of the interaction of the
treatments and time for the achievement dependent variable
was: There will be an interaction between treatments and
time for mean achievement. The direction and the magnitude
of the interaction for each treatment group is stated below,

1. The ascending criterion group will have a
progressively higher score on achievement for each unit
test over the period of the quarter. It is expected that
early success on unit tests should motivate students to
succeed later in the quarter when the criterion is at its
highest level of 90%.

2. The 90% fixed criterion group will have a
progressively lower score on achievement for each unit test
over the period of the quarter. This group is required to
reach a high level of achievement from the beginning of the
quarter. We should expect early frustration in an attempt
to attain this high level. There may be a loss of motivation
to succeed to a high level later in the quarter if early
failures are encountered.

3. The 80% fixed criterion group will have the
next lowest but a moderately stable score on mean achievement
for each unit test over the period of the quarter. The
relative ease of reaching a low level of criterion should
produce little change in achievement score from unit to unit.

4. The control group will have the lowest mean

achievement score of any group for each unit test over the

13
period of the quarter. The students of this group are
graded on a straight percent, 90%, 80%, 70%, 60% and 50% of
the total score. With only one testing at each unit test,
we should expect a distribution of scores of'A'to'FY.
Furthermore, the distribution should produce an average
grade of 'C'. This outcome is unlike criterion-referenced
testing. Testing is done to bring most students to the
highest level of achievement. Alternative learning aids
are used to assist learning when mastery is not reached.
Retesting is used to re-evaluate student learning. Students
of the control group are not given a chance at remediation
and retesting.

5. The ascending criterion referenced group, the
80% and 90% fixed criterion-referenced group will receive
a higher score on a measure of achievement than students in
the control class. This difference will exist because a
greater number of students will attain mastery of each of
the tests taken. This is so because students under the
mastery teaching method are required to attain mastery or
reach the prescribed criterion for each unit of study before
advancing to the next unit. Therefore, more students
should reach an 'A' under the criterion based grading.

6. The ascending criterion group will receive a
higher mean achievement score than a fixed criterion group
of 80% or 90%. The ascending criterion starts at 80% and
increased by 5% in each successive unit until 90% is reached.

It then remains stable at 90%. The fixed criterion (80% or

14

90%) remains the same for each test. The ascending criterion
is predicted to be better than the 90% level because

students will find it easier to achieve mastery early in

the quarter. It is assumed that the early success creates

a positive attitude toward the subject under this condition.
To know that one can pass the tests early should motivate
the students to work hard to pass tests with higher
achievement levels later. Also, the maintenance of a
particular high performance level later in the subject is
less threatening and approached with an expectation of
success. The 80% fixed criterion group should be equal to
the ascending group early in the sequence. As the ascending
group finds it more difficult later in the sequence, the
mean difference on the achievement will become more apparent.
The ascending group will score significantly higher than

the 80% group since the level of achievement for the 80%
group is lower.

7. The 90% fixed criterion group will get a
significantly lower score on achievement than the ascending
group. From the beginning, the level of achievement is set
very high. Students will probably feel that this is an
unreasonable expectation to meet. They will probably have
much frustration in an attempt to score to a high level set
for the course. Under these conditions, we can expect
students to become discouraged early in the quarter. The
early disappointment over failure to score properly will

probably discourage students to try to score higher

 

 

 

15

in the quarter.

8. The 80% fixed criterion group will get a
significantly lower achievement score than the ascending
group. Since the achievement level of the 80% group is set
so low for the whole term the students need not score as
high as the ascending group to obtain mastery or an 'A'
grade.

Finally, the 80% fixed criterion group will receive
a lower mean achievement score than the 90% fixed criterion
group. A 80% criterion from start to finish is set so low
that students of this criterion do not have to score as high
as the 90% criterion group. Therefore, the average
achievement score should be much different.

Hypotheses Regarding Attitude

The overall hypothesis of the interaction of the
treatments and time for the attitude dependent variable was:
There will be interaction between treatments and time for
mean attitude. The direction and the magnitude of the
interaction for each treatment group is stated below.

1. The ascending criterion group will have the most
positive mean attitude over time. This group is expected
to have a progressively more positive attitude because they
will attain the stated criterion for the earlier unit tests
without much difficulty. The early success should motivate
students to succeed later when the unit tests have become

more difficult to attain. The continued success should

16

produce the higher positive attitude toward the course
later in the quarter.

2. The 80% fixed criterion group will have the
next most positive attitude over time. The relative ease
of attaining criterion for each unit should produce a high
positive attitude. The attitude change is expected to be
moderate over time since the level of criterion can be
attained without much effort.

3. The 90% fixed criterion group will have a
progressively more negative attitude over the period of the
quarter. While the attitude of the 90% group may start as
positive as the other groups, the attitude is expected to
become more negative as the criterion level continues to be
difficult to attain. The author is not suggesting that it
is impossible to attain very high levels of achievement but
that in pushing students to do so may produce a negative
response in attitude toward the course over a period of time.

4. The control group will have the most negative
attitude toward the course over the period of the quarter.
This group should have the most difficult time of any group
in trying to succeed in a course which has a straight
percentage method of grading. The students have no
opportunity for remediation and retesting. We should expect
the greatest frustration over difficulties in attaining
a desirable score on each unit test.

5. The mastery students under criterion referenced

testing will have a higher mean score on a measure of

1?

attitude than students in the control class. One would
expect a more positive attitude under mastery since those
students would be given more opportunity to attain a high
level of achievement. One should also expect students to

be threatened less by a course which does not seek to promote
a standard distribution of grades from 'A' to 'F'.

6. The ascending criterion group will have a
significantly higher mean score on a measure of attitude
than the 80% or 90% fixed criterion groups. Early success
on the unit criterion-referenced tests of the ascending
group will build confidence. In addition, it is predicted
that students achieving early in the quarter are likely to
report that they are learning well. Thus, it is reasonable
to suggest that a positive attitude toward learning will
stimulate students to master the next higher criterion
level. Students will not turn away from the subject matter
as they might if success on tests is hard or impossible to
attain from the beginning.

7. Students in the 90% fixed criterion group will
have a significantly lower mean score on a measure of
attitude than the ascending group. The score which this
group must attain from the beginning is set very high. There
will probably be a great deal of frustration in trying to
reach such a high level of achievement on unit tests. The
frustration will probably be increased by the restudy and
retesting that must be done in an attempt to succeed on the

achievement tests. Because of this, we can expect students

18

to become much more negative in attitude earlier in the
quarter. Since the high standard is fixed until the end of
term, the attitude will remain negative.

8. Students in the 80% fixed criterion group will
have a significantly lower score on the measure of attitude
than the ascending group. The ease at which the 80% group
can achieve each unit test will probably cause no
significant change in attitude toward the course. The
ascending group will also have success on the same unit
test as the 80% group when the criterion is low. But, the
ascending group is faced with more difficult levels of
criterion late in the quarter. Therefore, the early success
should motivate students to try harder to succeed later.

As the ascending group continues to achieve on its tests, a
more positive attitude should be noticed.

Lastly, the 80% fixed criterion group will have a
significantly higher mean score on the measure of attitude
than the 90% fixed criterion group. While one might be able
to push students to attain a very high level of performance
throughout their learning, students of the 90% group may
develop a negative attitude toward the material learned.
This is why we should expect significantly higher positive
attitude at the lower criterion.

Hypotheses Regarding Time Spent On Instruction

The overall hypothesis of the interaction of the

treatments and the time for the time spent on studies

dependent variable was: There will be an interaction between

19

treatments and time for the mean time spent on studies.
The direction and magnitude of the interaction for each
treatment group is stated below.

1. The ascending criterion group will spend
progressively less time per unit on studies over the period
of the quarter. In the beginning of the course, students of
the ascending group will have an opportunity to adjust to
the course when the criterion is low. Once they have
established themselves under the lower criteria, the study
time of students will be less and less for each successive
unit. It is suggested that students who are successful
early in the quarter will not find it necessary to over-
study to ensure that learning is complete.

2. The 80% fixed criterion group will spend an equal
amount of time per unit on studies over the period of the
quarter. A criterion of 80% is easy to reach. Once
students realize that it takes little time and effort to
obtain the desired score for each unit, they will spend an
equal amount of time on their study of each unit to assure
success.

3. The 90% fixed criterion group will spend
progressively more time per unit on studies over the period
of the quarter. A Criterion of 90% for each unit test is a
very difficult standard to reach. If students are to assure
themselves of reaching the desired criterion, a greater
amount of time must be spent on studies. Any failure to

reach the stated criterion level for a unit test will

20

indicate to the student that one must study more for the
next unit test to achieve a criterion of 90%. The students
will tend to add much more study time in an attempt to make
sure that they have learned. Therefore, this pattern of
study habit is very inefficient.

4. The control group will spend the most time per
unit on studies over the period of the quarter. The
opportunity to do very well in a course will depend upon
performance on each test. The understanding of the content
must be complete on the first attempt of each test. There
is no chance for remediation and retesting. In an effort
to be as complete as possible on the understanding of the
course content, the students of the control group will spend
a great deal of time in learning. This situation will produce
inefficient study scheduling by students.

5. The criterion-referenced groups will spend
significantly less mean time on study than the control
group. The time spent on study can be considered a measure
of study efficiency. When there is less study time spent
to master a particular unit, the time spent on study is
considered more efficient. There may be a particular
criterion level whose attainment early in the sequence of
study will progressively increase the amount of material
achieved per unit time later in the course. Thus, students
learning under criterion-referenced testing may be spending
less time in study for tests. The students of the control

group should be expected to spend a constant amount of time

21

through their learning. Consequently, the time spent on
study will be greater for the control group than the
criterion-referenced groups.

6. The ascending criterion group will spend
significantly less mean time on study than the 80% or 90%
fixed criterion groups. Students of the ascending group
will have an opportunity to adjust to the course early in
the quarter. Once they have established themselves under
the lower criterion, their study time will be less and less
for each successive unit. In further support of the
hypothesis, it is suggested that students who find success
early in the course will not find it necessary to over-study
to make sure that they have learned. The overall effect will
be to shape the student into an efficient pattern of
studying. Therefore, it is predicted that the ascending
criterion offers the greatest opportunity to provide
students with greater efficiency of study. This is
accomplished by a gradual incline to a more difficult criterion.

7. The 90% fixed criterion group will spend
significantly more mean time on study than the ascending
group. Students in the group with 90% fixed criterion will
have a very high level of achievement. Therefore, it can be
expected that there will be more time spent on initial
learning. Additional time will also be needed for re—study
and re-take tests. Overall, the time to learn to an

adequate level will be greater than the ascending group.

22

8. The 80% fixed criterion group will spend
significantly more mean time on instruction than the
ascending group. Since the level of achievement for the
80% group is fixed at 80% of the total points, we can expect
students to spend the same amount of study time for each
unit during the quarter. The ascending group is expected
to decrease in study time over the same quarter of
instruction as the 80% fixed group. Therefore, the mean
score on study time will be greater for the 80% fixed group.

Overview of Literatgre Survey

The next chapter will review the literature of
mastery as it specifically relates to research on mastery
learning and to the criterion setting procedures. Since this
chapter has involved attitudes and the study time of the
student, it is relevant to explore the literature of mastery
learning with regard to these subjects. This will be done
to develop background of data for specific criterion setting
research procedures which produce best learning in a
relatively short amount of time without sacrificing the

attitude of the student.

CHAPTER II

LITERATURE SURVEY

Introduction

There are many innovative possibilities to foster
the learning process. Mastery learning is a specific
method of the general mastery strategy which can be used
to implement this process. The general mastery strategy is
defined by two essential features. One, the course content
is segmented into a number of relatively short, self contained
units. Students are tested on each unit.

Second, students are expected and required to meet
a predetermined criterion or level of mastery before
progressing to the next unit and its test.

The basic assumption of the mastery approach is
that almost all students can and will learn. To meet the
assumption, a set of procedures have been established. The
first is that mastery entails the formulation of a set of
instructional objectives that all students are expected to
achieve to a particular mastery performance standard. The
second procedure is the breakdown of a course into a
sequence of smaller learning units where each unit typically

covers several course objectives.

23

24

The third procedure is the construction of brief
progress tests called formative evaluation instruments for
all learning units. These tests are typically ungraded,
but in some cases they may be used as the basis for the
final grade. The resultant grade indicates whether the
student has or has not achieved the course objectives to
the appropriate level.

The final procedure is the preparation of
alternative learning materials for students who have not
attained mastery of the objectives of the unit. These
alternatives teach the objectives in a way different than
the teacher's lecture presentation.

The procedures used in this thesis were those
reviewed by Block (1971). Briefly, Block reports that a
subject is chosen and broken down in a specified number of
units. Preferably, the subject is one requiring convergent
thinking; that is, it has a definite body of knowledge upon
which a group of experts can agree. Objectives are
specified in a behavioral sense so students know what is
expected. Ideally, the units of study build on one another.
In some cases, courses may not be in a hierarchical order
but are broken into units by subtopic. The students are
asked to master each unit of study at a specific criterion
level. The grading is, therefore, absolute in that it
depends upon a level of attainment of criterion and not the
class average or a curve generated from relative groups of

students. When students do not reach mastery, a wide range

25

of procedures is initiated to help students study the same
unit material and correct deficiencies. Students are then
allowed to retake a unit test for mastery. The various
unit tests represent formative tests; that is, tests which
are not used for grade but tests which are used to inform
the student of deficient areas. The summative test is used
at the end of the course to put together all that has been
mastered. This is the test for a grade. Bloom (1971) states
that this method of learning for mastery has allowed up to
90% of the students in a particular class to achieve an 'A'
grade.

Since this thesis centered around mastery learning
and the use of criteria, the concept of mastery learning
was reviewed in some detail.

Research Regarding Mastery Leagnigg
Block (1971) reviewed the results from approximately

 

40 major studies on mastery learning. All these studies
have been done under actual school conditions. They have
involved all levels of education and in subjects ranging
from arithmetic to philosophy to physics. Block states that
these major studies have shown that 90 percent of the
mastery learning students have achieved as well as 20
percent of the non-mastery learning students. Several other
studies not reviewed by Block (1971) are reported below in
this review.

In 1968, Amthor compared two classes of a course

in descriptive geometry at the college level. Both classes

26

were presented with identical instruction but differed in
the type of evaluation or learning strategy used. One class
was taught the content of the course in lecture. They were
tested on the content one time and awarded letter grades

'A' through 'F'. The students of the other class were taught
under a mastery learning strategy. This strategy was
explained earlier in this chapter. The results of Amthor's
study were reported in terms of the number of students who
received a grade of 'A' in each of the classes. The results
show that 23 of the 29 students (about 80%) received a grade
of 'A' for the mastery learning treatment while only 11 of
the 63 (17.46%) received an 'A' in the 'A' through 'F' non-
mastery graded system. Foth (1973) reports that an improved
version of his mastery learning program in soil science at
Michigan State University produced a grade of 'B' or better
for 90% of the students; 70% achieved a grade of 'A'. In
general, research by Foth found that between 70% and 80%

of students received an 'A' instead of the 95% proposed by
Bloom and Block. In further support of mastery learning,
Wentling (1973) finds that high school students enrolled

in General Automobile Mechanics obtain significantly higher
mean achievement scores for both immediate achievement

(test a day later) and retention (same test given three
weeks later). A study conducted by Johnson, Gnagey and
Chesbro (1970) contradicts the research of Foth, Amthor and
Wentling. They used the mastery method whereby students

were tested over the materials covered in lectures, texts

27

and outside readings. One group had to make a score of
80% on weekly quizzes for mastery. They were required to
retest if unsuccessful until they passed. A second group
was given the same four 60-item unit exams and a
comprehensive final examination and assigned letter grades
on the first try. A third group received no weekly test
but spent the time discussing the material. None of the
groups showed any increase in learning as reflected by
examinations covering the material. The students were alike
in their learning. The research of Johnson et al. was the
only research which contradicted the positive results of
mastery learning. Nevertheless, the evidence is over-
whelmingly weighted in a positive direction for improved
achievement under the mastery learning procedure when all
of its aspects are used to teach a course.

Strategies Used To Set Criteria

The literature of the past 20 years has not reported
much research on the basis for setting of criterion levels.
Instead, it has produced a controversy on the validity of
setting criteria.

The controversy has centered around Ebel's (1971)
objection on the general meaningfulness of criteria of
achievement. He states that criteria must not represent the
interests, values and standards of just one teacher, but
they usually do. This is true because teachers have not
taken the time to come to a consensus about criteria.

Therefore, according to Ebel, they lack validity and useful

28

meaning. Block's (1971) rebuttal to Ebel is not strong and
direct. Instead, he contends that the setting of an
absolute level insures that each student completes his/her
learning before advancing to new information. How high a
level of achievement or what knowledge is to be acquired is
not answered.

With the exception of experimental papers by Block
(1970) and Carlson and Minke (1975), much of the rationale
used to set criterion levels for mastery learning has been
subjective. In this regard Bloom (1971) states that a
necessary condition for mastery is the setting of absolute
performance standards. Block (1971, 1974) remarks that
there are no hard and fast objective rules for setting
criteria. But criteria must be set to use as the basis for
grades in order to reflect attainment of those standards.

One broad suggestion regarding a strategy used to
set criteria (Bloom, 1971, Block, 1971 and Millman, 1973)
is to set realistic performance standards for each school
or group in cooperation with teachers and administrators.
The teachers and administrators would inspect test items
to determine the minimum number of items that students
must answer correctly in order to be considered in a
"mastery state." A variation of this suggestion is proposed
by Millman (1973). Test items are sorted into meaningful
clusters. The clusters may correspond to the objectives of
the course. Experts in the field determine the criterion

score for each cluster of items. "Mastery status" could

29

be assumed for students whose test performance on test
items in each cluster met or exceeded the corresponding
criterion score.

Another educational approach for setting the
criterion is Millman's (1973) approach. Millman (1973)
suggests two procedures. One deals with setting the
criterion so that a predetermined percentage of a group of
students pass. This procedure is inconsistent with the
philosophy of mastery. The philosophy asserts that students
should be encouraged to achieve optimum learning of the
stated course objectives. A second procedure is to
administer a test to students who have already mastered the
material. The criterion is chosen as the raw score
corresponding to a chosen percentile score. Hambleton et
al. (1978) state that this procedure has its limitation but
they do not state why it is limited.

A third approach for setting criteria is that grades
for the following year might be based on grading standards
arrived at the previous year if parallel examinations are
used. Specifically, Block (1971) states that scores which
earned students learning under non-mastery condition 'A's'
and 'B's' might be useful mastery grading standards. Based
on Block's suggestion, Hapkiewicz and Foth (1973) have
reported that scores which earn students 'A's' or 'B's'
in a previous term when grades were assigned on a curve were
specified as the standard for students in mastery learning

courses. A scale was developed from previous course grades

3c

in Soil Science 210 at Michigan State University (Hapkiewicz
and Foth, 1973). The scale was more rigorous than most
previous scales used since no one received a grade point
average of 4.0 with less than 88%: whereas some students
received a 4.0 with only 84% when grades were based on a
curve under the non-mastery system.

A fourth approach is suggested by Hambleton et a1.
(1978). They claim that, in general, criterion scores
probably should be based on psychological and educational
considerations, but in some instances statistical
considerations can be brought to bear on the problem of
setting a criterion score. Several statistical procedures,
which they reviewed, are stated in the following paragraphs.

Huynh and Perney (in press-see Hambleton et a1.
1978) suggest a method of estimating criterion scores.
Test performance data for a group of students on a series
of unit tests plus test scores from a "referral task" are
needed to start their algorithm for criterion score and
domain score estimation. 0n the basis of an initial
classification of students into mastery states, determined
by data obtained from the referral task, a score and domain
scores for the last unit in the sequence can be obtained.
The criterion score and mastery determination from the
last unit will then serve as the "referral task" data for
the second to last unit. The process is continued until
scores and domain score estimates are available for all

students on each of the unit tests.

31

According to Hambleton et a1. (1978) the practical
value of Huynh and Perneys' method of estimating criterion
scores is unknown. The method of Huynh and Perney appears
to have several problems. It assumes all items in a unit
test to have equal difficulty. It requires the existence
of an independent measure of performance, to which they
referred in their work as a "referral task." There must
be the proper sequencing of units. Also, there is a
subjective assignment of students into mastery states based
on the referral task.

Berk (1976) proposes a relatively simple procedure
for selecting a criterion score. The method requires the
selection of instructed and uninstructed groups of students.
Instructed students are those who have received "effective"
instruction on an objective to be assessed. Effective
instruction involves a qualitative judgement about the
mastery of an objective by students. Uninstructed students
are those who have not received instruction on an objective.
They are also tested to see if they have mastered the
objective.

Generally, the distribution of instructed and
uninstructed student scores, ranging from zero to i, where
i is the number of items on the test, can be divided by a
series of criterion scores into two general categories:
masters and non-masters. According to Berk, a criterion
which produces the greatest frequency count of students

at or above the criterion identifies the groups generally

32

considered masters. Those below the criterion are considered
non-masters. Since it is assumed that the students of the
instructed group are 'true masters“, these students are,
specifically put into two classes: True masters (TM), and
false non-masters (FN). Similarly, the students in the
uninstructed group are classified as false masters (FM) and
true non-masters (TN). The classification just mentioned

is expressed in a box form below.

Criterion Classification
Instructed (I) Uninstructed (U)

 

 

 

 

 

 

Predicted
:1; Masters Type II Error
.23 (PM=TM + FM) True Masters False Masters
H+>0 (TM) (FM)
ocsua
13.3 no
-H#4§
3E3: Type I Error
3:;“5 Predicted False Nonmasters True Nonmasters
<J~’ Nonmasters (FN) (TN)
(PN=FN + TN)
Masters Nonmasters
(M=TM + FN) (N=FM + TN)

For clarity, the cells of the above box are identified by a
classification term for the instructed and uninstructed
students. In practice, the probability of scoring on the
test is placed within each cell of the box. The
probabilities of the four classifications can be obtained
by simply expressing the frequency count of the number of
students placed within each classification as proportions

of the total sample. For example, the proportion of true

33

masters equals the true masters divided by the total number
of instructed and uninstructed students in a sample. The
optimum criterion score is the one that maximizes the
proportion of correct classification. The proportion of
correct classification for a particular criterion is equal
to the proportion of students in the instructed group
assigned to a mastery state (TM) plus the proportion of
students in the uninstructed group assigned to a non-mastery
state (TN). We may then assign the optimum criterion to
a particular course.
Factors to Consideg_§gg_8etting Cgiteria

There are several factors to consider for creating
criteria. These factors are proposed by Millman (1973).
Recently, they have been reviewed by Hambleton et a1. (1978).
The factors are:

1. educational consequences

2. psychological and financial costs
3. errors caused by guessing

4. item sampling

The educational consequences involve setting higher
criterion scores for fundamental or prerequisite skills.
Millman (1973) states that skills that are not prerequisite
to others may not require criteria at all. He suggests the
higher criteria are needed with prerequisite courses to make
sure that students are well prepared for advanced courses.
Setting the criteria too high may prove wasteful of teacher

and student time and resources.

34

A consideration of psychological and financial
costs led Millman (1973) to suggest that a low criterion
score should be set when remediation costs are high. In
situations with lower remediation costs or with higher
costs associated with false-positive errors, (a marginal
pass that is not a pass) high levels of a criterion should
be considered.

Errors caused by student guessing may lead to not
classifying certain students as masters. For this reason
Millman (1973) states that there may have to be a correction
for guessing to adjust the criterion score. How to make
the correction was not treated.

The errors introduced by item sampling is a bias
resulting from systematically disregarding some of the types
of questions and some content in the domain of test items
measuring an objective. Knowledge of this bias has not led
Millman to any conclusion as to a method of correction.
Perhaps, the bias can be minimized by careful consideration
to test construction or a clear and concise writing of the
objectives for a course.

Setting the Level of the Criterion

As a result of Block's research (1970, 1972), there
may be an objective rational for establishing a criterion.
His research shows that when one wants a great deal of
learning, selection of high criterion (95%) is appropriate,
but when there is a greater concern for the attitude of

students, a lower criterion should be set. In his case, 85%

35

was used. Furthermore, Block states that a criterion
between 85% and 95% can be chosen which gives a desirable
blend of achievement and attitudinal outcomes. It is
pointed out by Block that the above results must be
interpreted cautiously until they can be reproduced with a
much larger sample, on a longer learning sequence and in a
variety of subjects.

Carlson and Minke (1975) worked with 147 students.
Three consecutive lO-week night classes of Survey of
Psychology, at the University of Hawaii were used in the
study. Carlson and Minke tested three experimental criteria;
that is, ascending from 60% to 90% and two fixed criteria
of 80% and 90% for all quizzes. The students were informed
that a lO-item multiple choice quiz on each unit was to be
mastered at a stated criterion level. The stated criteria
for final grades were based on the total number of units
passed.

The description of the above research differs from
the research of the author in the following ways.

1. The ascending criterion of the author started
at 80% and increased by 5% until 90% was reached. The
criteria were applied to each of 5 unit tests. Carlson and
Minke started at 60% and increased by 10%. The criteria
were not applied consecutively to each of 28 units. For
example, units 2, 8 and 9 were graded at 70%.

2. The research of the author was done in the same

quarter. Carlson and Minke used three consecutive quarters.

36

3. Each unit in the research of the author had to
be passed before the student could take the next unit test.
All units had to be mastered. The final grades of Carlson
and Minke were based on completion of units. For example,
15 units had to be passed for a grade of At No requirement
was specified for any order of units to pass.

Carlson and Minke found that the highest criterion,
90%, produced the lowest number of high course grades and
passing grades. This is contrary to Block's (1970) notion
that a 95% criterion level produces the greatest learning.
The best performance was shown by the 80% fixed group for
final grades, passing grades per unit test and cumulative
units passed per student. One should expect a high passing
rate for the lower fixed criterion. Students of the 80%
fixed group did not have a criterion level as high as others
to master each unit. The sixty to nenety percent ascending
group required fewer attempts overall to master unit
quizzes than did either 80% or 90% groups. Carlson and Minke
believe that the early success felt by students helped to
reduce frustration which may have been felt by the other
groups. The effect may be positive reinforcement to continue
to perform. This was evidenced by a tendency of the 60% to
90% group to make fewer errors than the 80% group on units
when the criterion was 80%. Carlson and Minke (1975)
believe that the enhanced performance on tests of the 60% -
90% group was a result of 'shaping'. That is. the high

level of performance later is brough about by a gradual

37

increase in the mastery criteria. However, the 60% ~ 90%
groups passed fewer units on the first take when the
criterion was 90% when compared to the 90% fixed group.
This does not fully support the concept of shaping. Instead,
it appears that students might have come to "motivational
ceiling." They could have decided that a lower grade level
was good enough. Carlson and Minke suggest that the
ascending group may have reinforced less than optimal study
habits which persisted and retarded performance on later
units. This explanation supports a "motivational ceiling"
effect suggested by the author.

Summary of Research Regarding Mastery Leagning

In summary, achievement gains under the mastery
strategy appear to be well documented in the literature.
Students who learn under the mastery strategy with criterion
referenced testing achieve and learn more than students in
non-mastery classes.

In addition, the mastery strategy produces a greater
number of grades in the 'A' and 'B' category than under
other methods of instruction. The reason is that students
are asked to achieve to a certain level of performance. The
performance level is designated as the standard which
demonstrates the best learning of course content. Attainment
of that level is awarded an 'A' grade. The level of
criterion is keyed to the best understanding of the
objectives of the course. When students fail to reach

criterion, they are asked to use alternative learning systems

38

(test, tutor, audiotape, etc.) to study those lecture
objectives which are not clear.

There does not appear to be a concrete strategy
for setting criterion scores. What is available in the
literature? There are several educational and statistical
approaches. Educationally, groups of experts in a field
in cooperation with administrators may decide upon the
relevance of a criterion to the content of a course.
Additionally, the prerequisite status of a course may
influence the experts and administrators on how difficult
the standard should be.

Criterion scores may be set by using scores which
earned students 'A's and 'B's under a non—mastery approach
to teaching. This is even better if parallel examinations
are used for the mastery groups.

A criterion score may be set so that a pre-determined
number of students pass. This method is inconsistent with
the philosophy of mastery and criterion-referenced testing.
The philosophy is that students must be evaluated on their
absolute performance on the stated course objectives. The
instructor sets the level of achievement so that the
instructor can evaluate whether or not the stated objectives
have been met by the student. All students must be given
the opportunity to master the course content.

Lastly, there are a few statistical procedures which
may help to set criterion scores. All the procedures are

based on classifying students into two categories: masters

39

and non-masters. One statistical approach is to subjectively
assign students to the categories of mastery and non-
mastery. This is based on independent measures of
performance and a series of unit tests. From this data,
the criterion score is determined for the next group of
students. On the other hand, Berk suggests that a criterion
score should be determined by using the test scores from
samples of instructed and uninstructed students. A series
of criterion scores are used to determine which criterion
produces the greatest number of scores for students at or
above the criterion. These students are masters of the
test. The instructed group of masters are classified as
'true masters' while the uninstructed group is classified
as 'false masters.‘ The instructed and uninstructed groups
who do not pass the test are classified as 'false non-
masters' and 'true non-masters' respectively. The optimal
criterion is selected according to the estimated
probabilities of correctly classifying students. The
probabilities are obtaineiby dividing the number of scores
for each classification by the total number of students.
These probabilities are proportions of the total sample.
The proportion of students in the instructed group assigned
to mastery plus the proportion of students in the
uninstructed group assigned to non-mastery equals the

optimal criterion for a particular course.

40

Several factors should be considered before creating
criterion scores. They are:

1. educational consequences

2. psychological and financial costs
3. errors caused by guessing

4. item sampling

In terms of educational consequences, Millman
suggests that prerequisite coursesshould have criteria.

The criteria should be higher than other courses. The
higher standard assures that students are well prepared
for advanced courses. Courses which are not prerequisite
probably do not need criteria.

When psychological and financial costs are high,
low criterion scores should be set. When costs are low,
creating higher criteria may be more reasonable.

Millman (1973) suggests that errors caused by
student guessing must be corrected by some method. Without
the correction, students may not be classified as masters.
No method of correction is proposed by Millman (1973).

The last factor to consider for creating criteria
is item sampling. A bias may result from disregarding some
of the types of questions and some content in the domain of
test items measuring an objective. While the problem is
recognized by Millman (1973), no suggestion on correction
of the problem is proposed. The author suggests that it
may be corrected by careful consideration of test

construction. Also, precise writing of objectives may

minimize the problem.

41

Research has also been directed toward setting the
level of the criterion. The research reported in this paper
does show that there is a criterion which produces a
desirable blend of achievement of course objectives and
attitude outcome. Block believes that a criterion between
85% and 95% produces this blend. On the other hand, Carlson
and Minke propose that ascending the criterion in a course
(start low and increase to a high) is better because it
produces a student with a better attitude later in the
course. A high level of performance later is brought about
by a gradual increase in the mastery criteria. This is not
fully supported by the research of Carlson and Minke.
Students taught by a fixed criterion of 90% achieved more
than the ascending group when their criterion was 90%.

The researchers suggest that the ascending group may have
reinforced less than optimal study habits. This persisted
and retarded performance on later units.

Study Time Needed to Attain Criterion

Block (1972, 1974) reports that there is little
doubt that the mastery group characteristically requires
additional time and help to bring them to the particular
criterion established by an instructor. Furthermore,

Block (1974) states that in order to bring 80% of the
students to the level of achievement attained by 20% of
the students under non-mastery conditions, 10 to 20%
additional out-of-class study time is needed for

certain students. Wentling (1973 found that immediate

42

and delayed achievement was significantly higher for a

fixed criterion group of 80% but that the amount of time
spent on instruction was 50% greater for this group as
compared to non-mastery students. Wentling's measure

of time was to have all students keep a record of all

time spent upon instruction and testing for each unit.
Perhaps, the efficiency of the students' study habits can

be increased, eventually decreasing the time spent on
instruction and testing. An indication of this is noted

by Block (1974). He observed that students under mastery
varied in the beginning quite a lot with regard to extra
time needed for mastery. As a term progressed, the students
became more alike in their learning efficiency as measured
by time devoted directly to the learning effort. Block '
(1971) and Glaser (1968) suggest that perhaps this initial
difference in study time is due to aptitude levels and that
these levels are less obvious when time is varied for
individuals. This led Carroll (1970) to conclude that each
student has a time to attain the criterion and that the time
to learn is the aptitude of each student. Mastery strategies
offer the needed variation for each student to come to
criterion and to gain in learning efficiency.

Some literature has showed that under certain
conditions mastery can be achieved successfully in shorter
periods of time. In other words, as the mastery criterion
level is increased, the time to master the unit decreases.

Block (1972) attempted to test this relationship by assigning

43

different criterion levels (65%, 75%, 85%, 95%) to groups
of students. These levels are the percent correct of total
score needed to be considered as a pass and therefore
mastery. Block measured total amount of learning time as
an indication of efficiency of studies. The total time to
learn included textbook learning and time spent on
correction and review for each unit. It is concluded by
Block that all of the mastery groups spent more learning
time than the non-mastery treatment group. Also, the 75,
85 and 95 percent treatment groups spent the same total
amount of learning time. In the same amount of time, the
95% group achieved more course content than the 85% group
and the 85% group achieved more course content than the

75% group. This situation indicates that the 95% group
learned more efficiently than the 85% and the 85% group
learned more efficiently than the 75% group. These results
can only be taken as tentative because Block's sampha

was very small, the learning sequences were short and the
age of the subjects and the course matter used to make the
study was very limited in scope. Several students had
dropped out of several of the treatment groups which would
tend to bias the results. Also, the control group and the
treatment groups were in the same classroom, which may have
created a competitive atmosphere among the groups, thus

causing another treatment bias.

44

Summary_of Study Time Needed to Attain Criterion

In summary, specification of ways to reduce the time
spent on study and testing appears to be inconclusive in
the literature. Variables reported are so difficult to
control that the knowledge of their effects on time use
may not be known for a long time. The motivation of the
student, the quality of instruction, the prerequisite
background of the students and the previous study habits of
the students are a few of the factors which create confusion
when investigating time to learn and efficiency of study
under the mastery strategy. Additionally, the method of
collection of the data adds difficulty to the problem. Some
individuals allow students to report data while others
observe the time used to study and/or test. Also, the kind
of data collected varies. Researchers have used total time.
Total time is the amount of time used to read textbooks,
study for tests, do study projects, take tests, and do
correction and retakes on unit tests. Others were using
time spent by the student on instruction to complete a unit.
In order to equate the allocation of the time on unit tasks,
the latter is preferred.

Student Attitudes and Learning_for Mastery

It would be best to define the subject of attitude
as influenced by mastery before it is surveyed. Bloom (1971)
described attitude as a general disposition to regard
something in a positive or negative way. It is a feeling

which attracts one toward a subject or repels one away.

45

Many researchers have come to a common conclusion;
that is, when a student does well in a subject and more
generally in school over time, he/she.continues to develop
a positive attitude toward the object. In a general sense,
Bloom (1971) has also made these conclusions regarding
attitude in school. First, if a student develops a negative
(or positive) attitude toward school, it may include the
subjects, the teachers and staff. It may also include the
whole idea of school and school learning. Second, different
amounts of failure (or success) may be needed for different
students to develop this negative or positive attitude
toward school. It is a matter of degree. All individuals
who accumulate sufficient experiences of failure (or success)
will at some point develop negative or positive attitudes
toward school. Third, the degree of certainty of attitude
formation is likely to be much greater for negative
attitudes and repeated evidence of inadequacy. Last, other
variables determine whether the school and school learning
is viewed as positive and favorable, e.g., values of parents,
peer group attitudes, meaningfulness of schooling for the
individual's career aspirations. Bloom (1971) concludes
that in order for a student to view himself in a positive
way, he/she must be given opportunities for rewards. Mastery
learning provides the necessary reassurance and reinforcement
to assure a positive attitude.

Wooford and Willoughby (1968) measured attitude in

order to predict scholastic behavior. Seventy-two students

46

of general psychology answered a 40-item sentence completion
attitude scale which measured attitudes toward two specific
factors: instructor and the course. Two general factors
measured were college and life. Scholastic behavior

measures were absences, tardiness and course grades. They
concluded that the best predictor of this scholastic behavior
is the composite attitude scores (instructor, course,

college life). Of equal or greater interest was the finding
that course grades were significantly related to the attitude
toward the course but not significantly related to attitude
toward college.

Neidt and Hedlund (1967) found that student attitudes
toward a particular learning experience become progressively
more closely related to achievement in the learning
experience as the period of instruction progressed. In this
case, attitudes remained very course specific, but as
mentioned earlier, a continuation of negative or positive
success could lead to a general negative or positive
attitude about school.

Reports by Harris et a1. (1969), Neidt and Hedlund
(1967),and Sheppard and MacDermot (1970), indicated that
there is a correlation between high success and high
positive attitude. Harris et a1. (1969) and Sheppard and
MacDermot (1970) suggest that a positive attitude is a very
significant asset of the mastery strategy. According to the
latter authors, students are systematically led to success

on units of study and a course in general.

47

Block (1970, 1972) suggests that attitudes may
become negative when high achievement is established under
mastery. Block tested attitudinal changes by using 91
eighth graders who were taught a three-unit sequence on
matrix arithmetic. Sixteen students in each of four classes
were assigned to mastery treatments (four students per
treatment). Each treatment helped the student to reach a
particular performance level, for example, to attain either
65, 75, 85, or 95 percent of the material in each unit.

The percent of material attained was the student's score on
diagnostic-prescriptive unit tests. Other students in each
class were assigned to a non-mastery treatment. They were
not required to attain any particular performance level.

Block's results indicated two important points.
First, students of the mastery treatments had a significantly
higher attitude score toward arithmetic than the non-mastery
group.

Second, the achievement scores and attitude scores
toward matrix arithmetic increased up to the 85 percent
performance level. The achievement scores of 95 percent
performance level also increased while scores on the 24 item
attitude questionnaire showed a decline in attitude toward
matrix arithmetic. Therefore, a mastery strategy which
forces attainment of very high achievement scores may

eventually cause a decrease in attitude toward the subject.

48

Summary of Literature on Student Attitudes

In summary, student attitudes are correlated to
academic achievement. When students succeed in a course,
they generally develop a positive attitude toward a subject
and school. A different amount of success or failure is
needed by each student to develop a positive or negative
attitude toward school or his/her course work.

Attitude formation is likely to occur with greater
certainty when failures in course work are continually
encountered in school.

Student attitudes toward a particular course is
closely related to course grades. Student attitudes toward
a course become progressively more closely related to
achievement in the course as the term advances.

When instructors set criterion levels in mastery
learning, they should consider the attitudinal outcomes
of students. Unreasonably high levels of performance can
lead to negative attitudes toward the course. If this
becomes a consistent pattern, the negative attitude of the
student can extend to his/her major or even the school.

Summary of Literature Survey and Relation to Research
Questions

While it is clear that gains in achievement are
possible under the mastery strategy, it is not clear what
criterion score determines the greatest learning.

Thus, the major concern of this research is to search

for a criterion score which will produce the best

49

learning on each test. The results of the hypothesis written
on achievement will be used to determine the selection of
criterion which produces relatively greater achievement. It
would be a step toward setting a minimal level of performance
that students should be required to maintain throughout
their learning. A minimal level in this case is one which,
under certain conditions,is the best learning of course
content that we are able to produce. If this research
hypothesis is supported, it will then partially answer the
question of which criterion level produces the best learning.
A consideration of setting a criterion score leads to two
other dimensions mentioned in the survey of literature.
They are student attitudes and the amount of time spent on
studies.

It is certainly important that instructors find
ways to produce the highest learning of course content.
There are indications though, that influences on learning
can also influence student attitude. Thus, the second
major research question was: Is there a criterion score
which will produce relatively positive feelings toward the
subject? This question points to a relationship between
achievement and attitude. The indication in the literature
is that a student's achievement in a course in turn
influences his/her attitude toward a subject. It seems
reasonable that the student's perception of his/her learning
adequacy should influence his/her academic attitudes. The

student's ability to maintain particularly high criterion

50

levels would likely produce a positive perception of his/her
learning adequacy (Block, 1971). Attitudes should contribute
to learning. In turn, learning should positively reinforce

a predisposition toward a certain attitude. Thus, the
results of the hypotheses on attitudes may answer the
question of whether or not a particular criterion can cause

a relatively positive attitude toward the course.

Time spent on studies also appears to be related to
achievement. The few studies on the subject indicate that
maintenance of particular criterion scores has an effect on
study time on a task. General statements about the effects
are inconclusive because the research is conflicting. Some
reports such as Block (1970) and Wentling (1973) show more
time spent on instruction under criterion-referenced testing
while others such as Block (1974) report decreased time
spent on the task. The literature is further complicated
by results which show different criterion scores producing
the reduction in time spent on instruction.

Furthermore, it may be that student attitudes
influence their use of time. There appears to be a
"motivational ceiling" developed by students with regard to
time and difficulty of criterion. It may be that students
feel that a particular grade is good enough and no further
effort is necessary. This is supported by the research of
Carlson and Minke. Their ascending criterion (60, 70, 80,
90%) did not fully shape students to succeed at the highest

criterion as compared with a group maintained at 90%

51

throughout the term. It is suggested by Carlson and Minke
that the lower criterion in the beginning may have gotten
the students off to less than efficient study habits. Later,
this inefficiency was maintained instead of improving.

The inconclusive nature of the research on study
time under mastery makes it a likely target for further
study. The relationship between the maintenance of a
particular criterion throughout learning and the time
students need to learn must be studied in greater depth.

The lack of evidence on study efficiency under mastery has
made the following question an important part of this
research. Are there criteria which help to make a student
more efficient in his/her studies as the quarter progresses?
In this study, efficiency is expressed as time needed to
learn a unit of study. Time is defined as homework, textbook
study, note study, extra reading assignments, tutoring and
any other time directly spent to learn the content of a unit
of study. The results of the hypothesis on study time will

partially answer the question just mentioned.

CHAPTER III

PILOT STUDY

Introduction

There are three questions to be answered by this
research. First, what criterion will yield the best
learning?

Second, what criterion will yield the best attitude
toward a course?

Lastly, what criterion will yield the best study
efficiency?

In order to find an answer to the above questions,
three criteria and a control were used as treatment
variables. The three criteria were 80% fixed, 90% fixed
and ascending. A fixed criterion remained the same for each
test of the quarter. An ascending criterion increased in
percent over the quarter. In this research, the ascending
criterion was 80%, 85%, 90%, 90% and 90% for each test
respectively. Mastery under the criterion-referenced testing
is defined by a score equal to the multiplication of the
criterion percent by the total possible points for each test.

The control group was treated on a straight

percentage of total points. Thus, each test would have a

52

53
percent of 90%, 80%, 70%, 60% and 50%.

The dependent variables of achievement, attitude
toward the course, and student study time were the measured
variables used in the study.

First, the achievement score for each test was
used to analyze the effect of the treatment variables. The
mean achievement scores would be compared to determine group
differences.

Second, the attitude scale was used to analyze
the attitude of the student toward the subject as related
to the treatment variables. The mean attitude score was
compared to determine group differences.

Lastly, the study time of the students in each
treatment group was reported by students for each unit.

The time for course studies in this research was defined
as total hours directly related to learning of content.

The pilot study was undertaken to assess the validity
and reliability of the achievement test. Also, the attitude
scale was constructed during this part of the research.

The reliability of the attitude was also calculated.

Lastly, the pilot was used to evaluate the objectives and the
textual material to be used in the study. The pilot did

not evaluate the dependent measure of study time to be used
in the research because the time would be reported by

students during the treatment part of the research.

54

Population and Sample

The population of the pilot study consisted of
seventeen junior and senior state university students at the
California Polytechnic State University, San Luis Obispo,
California. The students were enrolled in a Greenhouse
Management course during the Summer Quarter of 1976. Eighty
per cent of the students were seniors. The average age of
the students was 22 years. Twelve students in the pilot
were transfer students. Most students can be classed as
elective students. Elective students chose the course
freely.

The pilot group was not informed that this was a
preliminary research study. To do so may have caused
students to act in an unnatural way towards the course.
This is typically referred to as the Hawthqrne Effect. The
group might have done better on the measurement instruments
because they knew that they were being studied.

.Qggrse Material and Instruction

The objectives (see Appendix A) were passed out to
the group of 17 students in a Greenhouse Management course,
Ornamental Horticulture 323-01. The students were informed
that the objectives were related to the lectures, handouts,
and assignments, and that unit tests were derived from the
lectures, handouts and assignments. The lecture material
or course content was primarily disseminated orally by the

lecture instructor during the scheduled hours for the course.

55

The handouts supplemented the lectures. When desirable,
students were able to use several greenhouse management
texts as references.
Course Evaluation

Students were informed that grades were derived
from a single administration of each of the five unit tests.
Each unit test was related to a defined amount of content as
represented in the objectives for each unit. The unit tests
were assigned a score of 90% in order to receive an 'A'
grade. Grades of 'B', 'C', 'D', 'F' are rated at 80%, 70%,
60%, 50%, respectively. Final grades were computed by an
average of all unit tests taken. Averaging the scores
seemed to be more of an incentive to do better on the
individual unit tests than any other method. Students may
have felt that they had a better chance fora higher final
grade when they were tested on a smaller amount of course
content.

Other Types of Study Aids Used in the Course

Students were encouraged to use the recommended texts
listed in the written introduction of the course. These
texts related to all of the objectives at one time or the
other. There were also numerous agricultural extension
bulletins made available to students as the need arose.

Lastly, the students were instructed that there were
numerous human research personnel located at the state
university. These people could help students in the understand-

ing of any lecture content. Students were also reminded thatiheir

56

lecture instructor was available for out of class tutoring
of any lecture content.
Validity of the Achievement Tegt

Ebel writes that the standards for Educational and
Psychological Tests and Manuals delimits three kinds of
validity for tests: content validity, criterion-related
validity and construct validity. Criterion related validity
determines the extent to which scores of a test provide
useful estimates of a student's knowledge in a subject. In
order to validate the scores, one would need a test generally
accepted as or known to be valid. Such a test was not
available. Therefore, this kind of validity was not used.

Construct validity is the degree to which test
scores measure particular psychological traits. Some of
these traits are creativity, anxiety and practicality. The
terms are the constructs being validated. This study did
not measure any psychological trait which would rate
students on a particular construct.

For the purposes of the pilot study, content
validity was used. The study was interested in the extent
to which the content included in the unit achievement test
was a balanced and complete sampling of the knowledge,
skills and understanding the instructor was attempting to
develop in the course (Ebel 1973, Erickson and Wentling
1976). Therefore, the content validity was determined by
comparing test content with the instructional objectives

for the course (Erickson and Wentling, 1976).

57

The comparison was done by the author. Other
individuals were asked to assess the validity of the content,
but all declined. The basic argument was that the author
was the one who best knows the content. This is reasonable
because the course used in this study had been taught by
the author six times prior to this study.

The author also had made numerous test questions
in the past on course content in greenhouse management.
Lastly, the industry experience as a greenhouse and personnel
manager have added expertise to relating evaluative
instruments to course content.

According to Ebel (1973), there is no commonly used
numerical expression for content validity. It was determined
by a thorough inspection of the items of the test by the
author. In order to do this inspection, a table of
specifications for direct assessment of student performance
objectives was used. This was patterned after Erickson and
Wentling (1976). The table on the next page shows the
objectives by unit with a check-off system for identification
of objectives which were included in the achievement test.

As is shown, there was a high agreement between the
objectives and the achievement test for the first four units
of study. Objective 2 under the heading 'Define Management'
and objective 3 under the heading 'Describe types of
business ownership' have been deleted. These objectives
were dropped because they were never part of the course

content. Therefore, they were never written into the unit

58

Table l . Table of Specifications for direct assessment of
objectives covered on five unit achievement tests in the
pilot study. See Appendix A for number corresponding to
objective.

 

 

Objective by Unit Objectives
Included in
Achievement
Test
I. Define Management +
1
2
3 +
4 +
Describe types of Business Ownership
1 +
2 +
3 . . . '
Describe different Market1ng Set-Ups
l +
2 +
3 +
)4,
Apply Market set-ups to Business
1 +
2 +
Diagram Organization Flow Charts
1 +
2 +
3 +
4 +
II. Identify Different Recruitment Procedures
1 +
Describe Orientation Procedures
1 +
2 +
3 +
Plan Training Procedures
1 +
2 +
3 +

 

59

Table l . CONTINUED

 

Objective by Unit Objectives
Included in
Achievement
Test

 

III. Estimate Production Peaks
1
2

3
L,

Calculate Year Around Crop Rotations
1

2

Calculate Number of Plants/Pots

1

2

+-++-+

+.+

-++-+

3
IV. Schedule Year-Around Crops

-++-+

Describe Profitability

[—1
+

 

60

tests. Objective 4 under the heading 'Describe different
marketing set-ups' was covered in future administrations of
the achievement tests.

Unit 5 of the pilot study had the greatest deficiency
of coverage of objectives by the achievement test. For this
reason, objectives were rewritten to match the course
content. The complete set of rewritten objectives is
included in Appendix A.

In conclusion, there is no way at this time to
quantify the area of content validity. It can only be
stated that this analysis as presented in table 1 does
represent high validity of content for four of the units of
study. The fifth unit had low content validity.
Consequently, this unit was rewritten to conform with the
course content.

Item Analysis of Unit Tests

In order to make each item in each unit test as
clear as possible, an analysis of items was undertaken.

The available procedures for item analysis of criterion-
referenced tests require two administrations of a test.
Since this pilot study tested students only once, it was not
possible to use those procedures.

A computer program at the California Polytechnic
State University scores true-false and multiple-choice. It
also prints a percent which indicates the number of students
who answered the item correctly. Since there was no other

way to identify poor items, the author decided to use the

61

percent of those who answered the item correctly.

This procedure provided another opportunity to
evaluate the test items closely.

Any item answered incorrectly by more than 50% of
the students was critically evaluated on the following
points: Table 2 shows the results of the item analysis.

1. ambiguity

2. poor grammatical structure

3. more than one answer to an item
4. irrelevancy to the objective

The above points were selected by the author as necessary
for clarity of each test item. Each test item was rewritten
when it appeared to fail one or more of the points.

The use of the above four-point criteria for item
analysis may not fit the philosophy of criterion-referenced
testing. Criterion—referenced measures relate an individual's
performance to an achievement level which indicates the best
performance. The goal of an instructor is to bring each
student to a point of optimal performance (mastery of
content) without regard for relative group comparisons.
Therefore, the type of item analysis just described can be
argued as inappropriate for criterion-referenced test
procedures. The four-point criteria just described have
value to criterion-referenced tests. A low response of
correctness for test items in the pilot only alerted the
author that there may be something wrong with the items.

It was another check on the measurement tool which was used

62

Table 2 . Item analysis in % correct for each question for
all unit tests in Greenhouse Management. *Items of less
than 50% correct response were revised.

 

 

Question No. I II III IV V
1 65% 41%* 35%* 59* 100%
2 88 24* 24* 94 12*
3 94 71 24* 47* 71
4 94 88 71 82 47*
5 88 82 71 100 35*
6 100 71 82 71 65
7 64 100 82 76 53
8 94 47* 82 59 47*
9 82 94 94 82 76

10 94 71 41* 94 12*
11 29* 88 71 82 76
12 41* 76 94 59
13 82 59 88 82
14 18* 41* 65 76
15 88 35* 59 100
16 94 71 82 88
17 100 47* 88 88
18 100 41* 88 76
19 47* 88 94
20 94 47* 88
21 71 71
22 41 65
23 88 94
24

82 35*
* 64

N
Kn

63

in the research. Items were reviewed and changes were made
on the above four point criteria. No data were logged as to
what changes were made. The revised questions appeared in
the instrument which measured achievement.

The analysis was based on the best judgement of the
author. The reason for the analysis can not be defended any
more than what was stated. Erickson and Wentling state that
other data and personal judgement should play key roles in
ultimate decisions about item retention and revision. They
do not offer any suggestions on criteria to use.

Reliability of the Achievement Test

Since it was important to know how consistent the
various unit tests are, it was important to measure its
reliability. Erickson and Wentling (1976) define reliability
as the degree to which an instrument (test) provides a
trustworthy or consistent measure of whatever it does
measure. If an instrument has high reliability it is highly
consistent in its measurement. Reliability of a test can be
determined by comparing student scores on two administrations
of a test. The tests can be the same tests or similar
tests. Also, comparison of two halves of a single test can
be done. Usually the halves are created by separating the
odd and even numbered test items into two groups. The
scores of each half are compared. The pilot study consisted
of only one administration of one form of each unit test.
Therefore, reliability estimates are obtained by comparing

two halves (odd and even items) of the test.

64

It was inappropriate to use the usual formulas for
determining reliability coefficients because these formulas
are used for norm-referenced tests when variability in
scores is desirable. In the case of criterion-referenced
testing, the variability of scores is minimized since most
students are expected to reach a criterion score. In this
pilot, 90% was the criterion. For this reason, it is
suggested by Ebel (1973), Hambleton et a1. (1978) and
Erickson and Wentling (1976) that formulas developed recently
for criterion-referenced testing be used. Livingston
developed a formula for estimating the reliability of
criterion referenced measures. This formula was used to
arrive at reliability coefficients for this research. It

is written as follows:

 

r = rXXsz + (x - o)2
co
Sx2 + (X - C)2
Where rcc = criterion referenced reliability
rXX = any one of the classical estimates of reliability
8x2 = observed score variance
X = observed class mean

C = criterion score set for the class
This formula is an adaptation of the classical
formula for estimating reliability (rxx) of a test. The

reliability (rcc) of the unit criterion-referenced test is

65

expressed as a coefficient. A high coefficient indicates
that the variance of the correlated test is due to the
measure and what it is intending to assess. For example,

a reliability coefficient of 0.75 indicates that 56.25%
(the square of 0.75) of the common variance is due to the
test. Thus, the coefficient assists in answering the
question, "Would students obtain similar scores on the same
test if the students were to be retested?"

The criterion score (C) of the formula is set by
an instructor for a course. In this pilot, 90% was the
criterion. The average or mean (X) was derived from actual
scores of an unit test. The score variance (8x2) and mean
was obtained from a computer program at the California
Polytechnic State University. The variance (8x2) denoted
a measure of dispersion or spread from the mean.

If the criterion score (C) equals the observed class
mean (X), Livingston's formula is the same as classical
reliability (rxx). The further the criterion score (C)
deviates from the mean (I), the higher the criterion-
referenced reliability. As shown in Table 3, the estimated
reliability coefficients vary from 0.6900 to 0.9849. What
can be inferred from these estimates? Generally, the
coefficient provides a quantitive estimate of the accuracy
of the test itself. A coefficient of 0.6900 for unit test
IV means that the test questions have 48.53% of their

variance in common. The highest coefficient, 0.9849 for

66

unit test II, has 97% of its variance in common. Borg and
Gall (1971) state that coefficients ranging from 0.65 to
0.85 are accurate enough for most test purposes.
Coefficients over 0.85 indicate a close relationship between
the two variables correlated. We can be confident that a
very good relationship exists between variables correlated.
If we were to measure a student's level of achievement on
future administrations of the same test, we would expect the

test to give similar results.

Table 3 . Reliability coefficients calculated by the
Livingston formula for criterion-referenced tests.
Values are shown for each unit test taken by students
during the pilot study.

 

UNIT TEST RELIABILITY COEFFICIENT PER CENT OF COMMOM
VARIANCE
I 0.7500 56.25%
II 0.9849 97.00%
III 0.9674 93.58%
v 0.9050 81.90%

 

Summary and Conclusion of the Validity and Reliability of
the Achievement Test

In conclusion, the individual achievement tests were
a useful measurement of the domain of knowledge. According
to a subjective review, the content validity of the tests

correlated highly with the objectives of the course. This

67

should be so since criterion-referenced tests should be
keyed very Closely with stated course objectives.
Additionally, a high amount of confidence was placed in the
unit tests to provide a consistent assessment of the
knowledge, and skills, being measured in future
administration of the tests. This was reflected by the
medium to high reliability coefficients calculated by the
Livingston formula for criterion-referenced test reliability.
Development and Assessment of the Attitude Measure

The researcher constructed the attitude measure
during the pilot study. After the measure was fully developed
and assessed, it was used as part of methods and materials
to complete this research. The attitude measure has been
developed by the summated ratings method (Edwards, 1957)
described in the next paragraph.

In order to develop the attitude measure, a self-
made Likert scale described by Edwards (1957) was used.
Two hundred individuals were asked to express their feelings
about ornamental horticulture by writing three favorable
and three unfavorable statements about ornamental
horticulture. One neutral statement was also requested.
From this response, 50 favorable and 50 unfavorable state-
ments were selected with the assistance of Edwards' 14-point
criteria. (See Appendix B.) The 100 statements were
randomly placed on pages with a response set reading:

strongly agree, agree, neutral, disagree, and strongly

68

disagree. The response set received a weighting of 5, 4, 3,
2, 1 so that a statistical analysis could be made. The
highest weight was given to the 'strongly agree' term when
the attitude question was favorable. Conversely, the
'strongly disagree' term was given a 5 weight when the
statement was unfavorable. The 100 statements were scored,
and a value of 't' was calculated for all statements. The
attitude scores and value of 't' was computed by a computer
program located at the California Polytechnic State
University. Following Edwards'approximate rule of thumb,

a value of 't' equal to or greater than 1.75 indicated a
significant statement. Therefore, the author selected as
many statements as possible which had the greatest 't' value.
The value of 't' is a measure of the extent to which a
given statement differentiates between the high-scoring and
the low-scoring groups.

As a result of the summated ratings methods of
attitude scale construction, it was possible to develop two
attitude surveys which most likely gave high response values
to favorable and unfavorable statements. Each survey had 22
statements, 11 favorable and 11 unfavorable. All statements
were randomly placed on the final survey forms.

The expected value for a strongly positive response
to the survey was 110 (22 statements times a weight of 5
for 11 strongly agree and 11 strongly disagree statements).

On the other extreme, the lowest score of 22 was obtained

69

when all statements with a value of l were picked . A
neutral response is 66.

Since a second administration of the newly developed
attitude surveys was not possible, an odd-even split half
coefficient of internal consistency was calculated for the
two forms of the survey. The attitude surveys are shown
in Appendix C. The first form had an estimate of
reliability of 0.6298 while the second form is 0.559. In
1966, Barker reported a 0.709 coefficient of correlation for
a self-made attitude scale toward school guidance. He
considers this value as a preliminary estimate of the
alternate form reliability of the scale. In general, Mehrens
and Lehmann (1973) state that attitude scales have
reliabilities around 0.75. Borg and Gall (1971) indicate
the low, medium,and high reliabilities for 18 reported
attitude scales to be 0.47, 0.79 and 0.98 respectively. Borg
and Gall (1971) state that coefficients around 0.50 (25%)
common variance may be a crude estimate of what is being
predicted. Based on the available references, the calculated
reliability coefficients of the attitude surveys were fair.
The reliability accounted for only 31.24% to 39.66% of the
common variance. This may result in a considerable restraint
on what we find. It could be that if we get no differences
in attitude later that it is due to the low reliability.
Also, any significant differences among scores must be

cautiously evaluated. With a low reliability, a very high

70

level of significance must be used to show that the
magnitude of differences among scores is a true difference.
In the absence of any established attitude scale in
ornamental horticulture, these attitude scales served as a

crude estimate of attitudes.

CHAPTER IV

RESEARCH DESIGN AND PROCEDURES

Introduction

The research was designed so that the following
three questions could be answered:

1. Will one criterion produce better student
achievement than another?

2. Will one criterion produce more favorable
student attitudes toward the course than another?

3. Will one criterion produce more efficient
study than another?

In order to partially answer the above questions,
several criteria and a control were used as treatment
variables. The criteria were 80% and 90% fixed criteria
and one criterion called ascending. A fixed criterion
remained the same for each unit test. The ascending
criterion started at 80% for unit one and increased 5% for
each successive unit until 90% was reached. The percent
of any criterion was used to calculate the minimum score
out of a total score which defined mastery of a unit of
study.

The control class was not assigned a required

criterion level. Instead, the class was graded on a straight

71

72

percentage of total points for each unit. The percentages
were 90%, 80%, 70%, 60%, and 50%.

The effect of the treatments was measured by self-
made multiple-choice achievement testson each of five units.
The scores of the achievement testswere used to calculate
averages for each unit for each treatment group. The
averages were used to determine group differences.

The effect of the treatments was also measured by a
self-made attitude scale. The score of the attitude scale
represented the relative degree of positive attitude toward
the course in greenhouse management. The average attitude
for each unit for each treatment group was used as a
comparison among groups.

Lastly, total study time was used as a measure of
the effect of the treatments. The study time was reported
by students at the beginning of each unit test. The study
time represented all time directly spent on studies in
greenhouse management. The average total time on studies
for each unit was used to make comparisons among treatment
groups.

From the research questions for this study, the
following hypotheses were drawn. In each case, the selection
of criterion was tested for its effect on achievement,
attitude.and total instructional time to learn each unit of

study.

73

1. There will be an interaction between treatments
and time for mean achievement.

2. The control group will have the lowest mean
achievement score of any group for each unit test over the
period of the quarter.

3. The ascending criterion-referenced group, and
the 80% and 90% fixed criterion-referenced groups will all
receive a higher score on a measure of achievement than
students in the control class.

4. The ascending criterion group will receive a
higher mean achievement score than a fixed criterion group
of 80% or 90%.

5. The 90% fixed criterion group will get a
significantly lower score on achievement than the ascending
group.

6. The 80% fixed criterion group will get a
significantly lower achievement score than the ascending
group.

7. There will be interaction between treatments
and time for mean attitude.

8. The mastery students under criterion-referenced
testing will have a higher mean score on a measure of
attitude than students in the control class.

9. The ascending criterion group will have a
significantly higher mean score on a measure of attitude

than the 80% or 90% fixed criterion groups.

74

10. Students in the 90% fixed criterion group will
have a significantly lower mean score on a measure of
attitude than the ascending group.

11. Students in the 80% fixed criterion group will
have a significantly lower score on the measure of attitude
than the ascending group.

12. The 80% fixed criterion group will have a
significantly higher mean score on the measure of attitude
than the 90% fixed criterion group.

13. There will be an interaction between treatments
and time for the mean time spent on studies.

14. The criterion referenced groups will spend
significantly less mean time on instruction than the control
group.

15. The ascending criterion group will spend
significantly less mean time on instruction than the 80%
or 90% fixed criterion groups.

16. The 90% fixed criterion group will spend
significantly more mean time on instruction than the
ascending group.

17. The 80% fixed criterion group will spend
significantly more mean time on instruction than the
ascending group.

Experimental Design

This study took the form of an experimental design

with multiple treatments. As shown in Table 4,the variable

matrix for this study was a two-way design having five

75

repeated measures. The time or unit of study (see Table 4)
was crossed with each treatment group. Since groups received
different treatments, the students were nested within a
treatment. There was an unequal number of students in

class and all students in a class were used in the research.
The number of students in the 80% and the 90% fixed criterion
groups, 90% criterion fixed group, the ascending criterion
group, and the control was 16, 10, 13 and 17 respectively.

No students were dropped from the course and all students
were used in the research.

Independent Variables

The treatments were the independent variables of the
research. They are identified in the variable matrix of
Table 4 as the experimental treatment variables.

The treatments were criteria and the control.
Criteria were defined by three levels; that is, 80% of total
points for each of five units, 90% of total points for each
of five units and an ascending criterion. The ascending
criterion started at 80% of total points and increased 5%
for each unit test until 90% of total points was reached.

A control group was used in the experiment to
determine the significance of the treatment levels over the
traditional method of teaching as used in this study. This
control group was graded on a straight percentage of total
points. The percentages were 90% for an 'A', 80% of a 'B',
70% for a 'C', 60% for a 'D' and 50% and below for an 'F'.

76

Table 4. The variable matrix is shown. The multiple
dependent measures are shown for each time for each
treatment variable.

TIME 1 TIME 2 TIME 3 TIME 4 TIME 5
(Unit 1) (Unit 2) (Unit 3) (Unit 4) (Unit 5)

 

 

 

 

 

m a: a: H: a:
m c: a: m m
0 <3 C) o o
um um cm: om om
m Um04 mm» mm» mm» mmm
E4 on on on on on
2 sun: EDD 90: so: 90:
Treatment m 2025-): 202$ 202$ 202$ 202$
Var1ab1e g gm Em gm Em gm
9: mmz mmz mmz mmz mmz
0) >90 >30 :wxa >=o >90
me me hfla me me
nap-«g HHE HHEIJ HHE HID-lg
326-! 3:9 LEE-IE :29 LEE-I
06H QBH mea 09H UBH
<vm4 <<e <<e <<e <49
Group 1 $1
80% Fixed :
Low S.
Criterion 1
Group 2 Si+l
90% Fixed :
High S2.
Criterion 1
Group 3 S .
+
80 to 90% .21 1
Low to High é ,
Criterion 31
Group 4 S -
1+1
Control .3
S41

 

 

 

 

 

 

77

Control of Internal Valigity of Treatments

The four treatment groups, shown in Table 4 did
not have random assignment of subjects. Rather, groups
were randomly assigned to the above mentioned treatments.
Since subjects were not randomly assigned, selection is
considered a threat to internal validity. In order to
control for this source of invalidity, analysis of covariance
was used in the statistics of the research. Hence, age,
grade point average,and manner of student selection of the
course were considered as covariables. It should be stated
here that a class profile was also made to determine the
statistical significance of group differences for the
covariables mentioned. These results are presented in
Chapter V.

Another concern of internal validity arises from
the multiple testings of the subjects. This is the effect
of taking a test upon the scores of a later test. Since the
achievement tests for each unit of study were different,
there was not any problem with this threat to internal
validity. 0n the other hand, the attitude scale may be
remembered by students. For this reason, two forms of the
attitude scale were used to obtain data. With two forms of
an attitude scale, each form will not be reused until four
weeks have passed. This should minimize an effect of one
testing on the other.

There are two factors to consider for external

validity. One is the possible artificiality of the

78

experimental treatment and the students knowledge that they
are involved in an experiment. The other is the multiple
treatment interference. The former factor had been
eliminated by not revealing any knowledge of the experiment
to any student. The fact that the students were being
tested and graded differently from what they were familiar
was explained as the approach used by the particular
instructor in the Greenhouse Management course. Repeated
measures of the attitude scale and the collection of data
on time spent on studies was explained as a tool being used
by the instructor for self-evaluation of the course. The
latter factor may have some effect on generalization of the
experiment. No students in the same class received different
treatments. But, the effect of students talking to each
other outside of class was considered as a possible threat
to generalization. Therefore, there was a chance that
student attitudes might have varied because students were
discussing the method grading. The researcher did not
control for this possibility. There are no other major
concerns for the validity of the experiment.
Dependant Vapiables

The dependent variable of achievement on each unit
test was the number of correct responses out of a total
possible points. Mastery was achieved when students reached
the criterion level assigned to a class. The achievement
unit test was administered during a lecture hour following

the end of a unit of study. Each unit of study was two

79

weeks long. A second administration of the test was done
when a student failed to receive a score which defined
mastery. This was done by arrangement out of class with
the instructor.

The attitude score was administered at the end of
each unit achievement test. Thus, there were five times
when attitudes were measured.

The third dependent variable, total time for study,
was reported in writing by students. The data was written
onto a standard reporting form. The form was collected
at the beginning at each unit test.

An objective test was used to measure achievement
on each unit. The number of correct responses for each
mastery level treatment group was logged for each unit.
Each student score out of the total possible score was used
in the analysis. The individual score permits means to be
calculated and compared among other treatment groups.

Attitude toward the course was the third dependent
variable measured. This variable was assessed by a self-
made attitude scale as described in Chapter 111. Two forms
of the attitude scales were used. This was done to minimize
the chance that students might remember how they responded
on a previous test. Form A of the attitude scale was
administered immediately after the end of unit test one,
three and five. Form B was administered immediately after

unit two and four. The attitude scale was administered

80

only after the first try of each unit test.

Time spent on instruction and testing was logged
by each student for each unit. The time log was collected
as a ticket to take each unit test. In this way, the
author was assured of getting the time log. At the time of
collection, the log was checked for proper recording of
minutes and hours.

The covariables of age, grade point average,and
manner of student selection of the course (elective vs.
required) will be used to control for initial differences
among groups.

Procedures
Popp1ation and Sample

The following description is a representation of
the type of students used in the study. They are described
in great detail so that other researchers could reconstitute
a similar group of students and so that generalizations
may also be made to a larger population. The population of
the research consisted of third and fourth year university
students in Ornamental Horticulture at the California
Polytechnic State University, San Luis Obispo, California.
About 80% of the students were seniors.

The age of the students ranged from 21 to 24 and
about 31% were women.

The students were first-time (native) high school
graduates and transfer students from community colleges

throughout California. Seventy point nine percent of the

81

students were transfers. This factor was not considered to
bias the study since all students had been at the California
Polytechnic State University for several years.

Most students selected this course as an elective
as compared to a program requirement. Of 56 students in the
entire study, 72.72% had elected to take the course.

The prerequisite background did not vary among
classes which are involved in the research. All students
had taken courses in Fundamentals of Ornamental Horticulture.
Table 5 shows the percent distribution of students who have
taken the prerequisite courses. The percent of other major
courses is also shown. The other major courses are shown
so that the experimental groups can be typed precisely.

The control group and the ascending experimental group

both had about 31% of the class with the pot plant
prerequisite while the 80% Fixed and 90% Fixed experimental
groups had 18.75% and 20% respectively. Only the 80% Fixed
and 90% Fixed showed any background in cut flower production.
While there were differences among the experimental groups
with regard to the additional course work, the difference
did not influence the research. The course information was
self-contained and was taught with the Fundamentals of
Ornamental Horticulture as the only prerequisite course.

In an effort to further type the population with
respect to additional course background, students were
asked to check courses taken in accounting, business law

survey and other business or management. Table 6 on the

82

Table 5 . Prerequisite profile of students subjected to the
criterion treatments.

 

 

 

Treatments
Prerequisite Control Ascending 80% Fixed 90% Fixed
Courses N=l7 N=13 N=l6 N=10
Fundamentals 100% 100% 100% 100%
Pot Plant
Production 31.25% 31% 18.75% 20%
Cut Flower
Production 0 0 12.5% 40%

 

 

Table 6 . Additional course work taken by the students in
the stated criterion treatments.

 

 

 

Treatments
Background Courses Centrol Ascending 80% Fixed 90% Fixed
Taken N=l7 N=13 N=l6 N=10
Business Law 100% 76.92% 93.75% 70%
Accounting I 68.75% 38.46% 62.5% 60%
Accounting II 37.5% 23.07% 37.5% 20%

Other Business/
Management 25% 15.38% 31.25% 60%

 

 

83

next page portrays the courses taken by each treatment
group. Most students in any of the treatment groups have
taken business law survey before entering Greenhouse
Management. All other background courses varied quite a
bit among the various experimental groups. In all but the
ascending treatment group, the next greater background
course completed was Accounting 1.
Treatments

A different level of criterion was used in each
50-minute mastery class. The criteria were used in
conjunction with a mastery learning strategy. The criteria
were 80% fixed, 90% fixed, ascending (80% for the first unit
and increasing 5% each unit test until 90% is reached.
Additional units were graded at 90% of total points). A
fixed criterion was one which had the same standard applied
to each of the five unit tests in the quarter.

The textual material of this mastery strategy in
Greenhouse Management (Ornamental Horticulture 323) was
divided into sections. The sections contained:

1. instruction for completion

2. objectives for each unit

3. a set of review questions and

4. the lectures given by the instructor

Mastery in this research was defined by three
elements: instruction, grades and testing. In order to reach
mastery or achieve an 'A' on the achievement test for a unit
of study, the students had to attain the minimum criterion

set for each instructional unit in a given treatment group.

84

Whenever criterion was not met on the first attempt of any
test, the student was provided with alternative instructional
assistanCe for a particular unit of study. The alternatives
were, but not limited to, tutors, a different text on the
subject, library readings, a restudy of notes or a review
of lectures on audiotapes. The second attempt for mastery
of a unit of content was considered the final try. There
was a total of five unit tests with each test being given
at the end of a two week unit of instruction.

The last treatment group or non-mastery group was
the control for the research. The non-mastery treatment
group received the same objectives as the mastery groups.
They were lectured on each unit of study and given the same
test questions as the mastery groups.

There were several fundamental differences between
the mastery groups and the control group. First, the
control group had only one try on each examination. They
were not required to reach any particular level of
achievement. The earned points on the first try of any
unit was the grade for that unit. The score of the
achievement test was obtained on a straight percent of
total points: that is 90%, 80%, 70%, 60% and 50% of total
points for each unit test.

Second, the control group did not receive any
benefits of remediation. The results of the unit test were
shown to the students at the next class meeting. At that

time, students had a brief opportunity to review the test

85
and check any incorrect answers. But students were not
exposed to alternative learning aids to assist in a better
understanding of the course content of a unit.

Lastly, student scores on the first and only attempt
on each unit test represented the ability of students to
more or less learn and retain knowledge in a fixed period of
time. This was unlike the criterion groups which were
given additional opportunity to understand the course content.
Retesting evaluated the improved learning status of students.

Since treatments were applied to an entire class
of any treatment,multiple treatment interference was
eliminated. Although, it was recognized that future
administrations of the achievement test within each class
will have an effect on students. This was considered a
legitimate carry-over effect since one of the primary
hypotheses states that a certain criterion score when
applied over a period of time will produce a change in
attitude of the student toward the subject. In addition,
the time the student needed for instruction and testing
was altered as a result of maintaining a certain criterion
throughout the course of study. Therefore, the relation of
treatment to attitude and time needed for study and learning
could not be eliminated.

Instrumentation and Data Cp1lect19p

The measure of achievement was taken using a

instructor-made unit test which covered five-two week

sequences in Greenhouse Management. The instructor-made

86

tests contained a number of multiple choice items that were
keyed to the objectives for each unit of study. The test
was administered to the entire experimental treatment group
during a class period. The students responded to the items
by marking its answer on an opscan computer sheet. The
sheet had the corresponding letters to choices given on the
printed multiple choice test. There was only one answer
for each question. The total number of correct responses
was the individual's score. There is a minimum number of
correct responses needed to reach the particular stated
criterion.

There were five teacher-made tests for each of the
five units of instruction. The five tests were the same for
all the treatment groups. Five additional tests for each
unit were available for students who did not reach the
specified criterion on the first attempt. The second test
was administrated by arrangement outside of class. In the
event of a second try, the highest score was accepted as
the students score. No further achievement testing was
done after the second attempt to attain criterion. The
control group receive only one attempt on each test.

The individual scores of students for each unit was
entered as data into the computer program. The analyses
phase of the computer program compared means for each unit
test for each group to test for differences.

While the researcher did collect data on the number

of retests taken by each group, this information was not

used in this study.

87

The attitude measure was taken by a self-made
attitude scale developed by the summated rating method as
described by Edwards. The attitude scale was administered
to each treatment group immediately after the achievement
test was taken. The students responded to the attitude
scale by placing a mark on an opscan computer sheet. The
students used a response set of strongly agree, agree,
neutral, disagree, strongly disagree to answer each of 22
items on the attitude survey. When the survey was analyzed,
the response set received a weighting of 5, 4, 3, 2, 1.
When the item on the survey was a favorable item, the
highest weight of 5 goes to the strongly agree term.
Conversely, when the item was unfavorable, the highest weight
of 5 goes to the strongly disagree term. In this way, a
person with a strongly positive attitude obtained the
highest score of 110 (22 statements multiplied by 5. The
individual scores of each student from each treatment group
was entered on computer cards and used in the analysis. The
average attitude score per unit was used for comparison
among treatment groups.

The measure of total time spent on instruction was
taken as a student-supplied record of all time spent to
study for each course unit. Students received a log sheet
for each unit. Upon this sheet was written the category of
study (for example, read text, studied notes, tutored with
instructor), and the time spent on each category by day.

The log sheet was then returned to the instructor at the

88

time of taking the achievement test. To be assured of
receiving this data from each student, the time log sheet
acted as a ticket for the test. No time log means no test
was taken until the sheet was returned. There were no
instances where students failed to bring the time log. The
total hours of study time per student from each treatment
group was the third datum entered on computer cards and
used in the analysis. The average total time per unit was
used for comparison among treatment groups.
Data Analysis

The attitude survey was hand calculated by applying
the weights of 5, 4, 3, 2, 1 to the appropriate response
set for each item on the attitude survey. The weights were
totaled and the total for each student tabulated for key-
punching onto cards. The time reported by students also
was hand totaled for each student for each unit. The total
was tabulated for keypunching. The total points correct for
each achievement score for each student was keypunched onto
cards with the corresponding attitude score and time score.
Each keypunched card contained the three dependent variables
scores of each student for the five units of study in
Greenhouse Management. The keypunched cards were analyzed
via the California Polytechnic State University IBM 360.
A multivariate analysis of covariance with repeated measures
with multiple dependent measures was the general program
used. Since there was an unequal number of subjects in each

treatment group, a specific program called the Finn Program,

89

Version IV for multivariance was used in the actual analysis.
The alpha level for significant differences among means of

dependent variables was set at 0.05.

CHAPTER V

ANALYSIS OF THE RESULTS

_ntroduction

Data from the achievement scores, the attitude
scores and study time were collected for the statistical
analysis of the experimental test.

The achievement score for a unit test was the
number of responses correct out of the total possible score.
The score of each student for each unit test was used in
the analysis so that mean achievement scores could be
calculated. The mean achievement scores were compared to
determine any group differences.

The attitude score for each unit for each student
was used so that mean attitude could be computed. The scores
were used for a comparison of attitudes for each group.

The total study time for each student for each unit
was entered into the analysis so that a mean study time
could be computed. The mean study times were compared for
any significant group differences.

The grade point average, the required-elective
factor,and the age of the student were collected as covariable

data to be used in the analysis. The required-elective

90

91

factor is defined as the way the student chose the course.
One or more of these covariables were thought to influence
the outcome of the experimental tests.

Fifty-six students were involved in the study. All
students who were enrolled in the classes took part in the
experimental test. All the classes were within the same
academic quarter. Four distinct and separate classes were
used to test the experimental variables and the control.

The experimental variables were the levels of
criterion for each class and the control. Two classes had
fixed criteria of 80 and 90 percent. In each of these
classes, the student had to attain a score equal to the
fixed percent of the total points possible. The third
class had a level of criterion which ascended. The level
started at 80% of total possible points for unit one. The
criterion was increased 5% for each successive unit until
90% was reached. The criterion remained at 90% until the
end of the quarter.

The control class was graded on a straight percent
of total points: that is, 90%, 80%, 70%, 60% and 50% of
total points. This was done for each unit test. Since
the control class was not required to reach any particular
level of achievement, the earned points on the first try of
any unit was the grade for that unit.

The experimental conditions were repeated five times
at intervals of two weeks apart. The study lasted for the

full length of an academic quarter.

92

The data was run on an International Business Machines
360 at the California Polytechnic State University Computer
Center employing the Finn Program, Version IV.

Analysis of Covapiates

Grade Point Averggg

As is shown in the analysis of variance, Table '7 .
there were no differences among the grade point averages of
any of the experimental groups. The necessary F ratio is so
low that statistical comparison of means was not reported.
The means are illustrated in Table 8 . So, it was not

necessary to use grade point average as a covariate.

Table 7 . Analysis of variance for grade point average.

 

 

Source of Required F
Variation Mean Square D.F. Observed F 5%
Total 0.2389 55

Groups 0.0328 3 .131 2.78

Error 0.2508 52

 

 

93

Table 8. Means of the four experimental groups on grade
point average.

 

 

 

Experimental Grade Point Average
Control/Non-mastery 2.8488
Ascending Criterion 2.7685
80% Fixed Criterion 2.8763
90% Fixed Criterion 2.8710

 

 

Statistgcal Ana1y§is of the Chi quare Test of the Elective-
quuired Covariate

Students within this research were classified as
elective students or as those who were required to take the
course. It was necessary to make a statistical check to
determine if the ratio of the elective to required
classification was the same for each class. If they were
not the same, then it would be necessary to use this data
as a covariate.

The Chi Square test of the frequency counts of those
who chose the course as an elective and those who must take
it indicated that the ratio of the various groups were alike.

Tables 9 and 10 illustrate the results.

94

Table 9 . Summary of data from the four experimental groups
based on a 7:3 ratio (elective:required). Ratio was
obtained when the pilot study was made.

 

 

 

Source D.F. Chi Square
Total 4 0.6475
Pooled 1 0.2755
Heterogeniety 3 0.3720

 

 

Table 10 . Summary of data from the four experimental groups
based on 41:15 ratio of the observed totals.

 

 

 

Source ' D.F. Chi Square
Total 4 0.3926
Pooled 1 0.0000
Heterogeniety 3 0.3926

 

 

The Chi Square for a ratio of 7 to 3 (elective to
required) indicated a probability of about 95% that the Chi
Square value of this size or larger could come from a
homogeneous set of samples from a single population. When
tested against an observed ratio of the totals (Table 10),

the probability is 99% that the population has a ratio of

95

41:15, elective to required respectively. Also, the
observed ratio shows that there is a 95% probability that
the groups were drawn from the same population of students.
Therefore, this covariable was not used.
Regyession Analysis_fppthe Age Covariate

A regression analysis was used to determine the
relationship of age and the dependent variables of student
achievement, student attitude and total time it took to
study each unit. The regression analysis within the Finn
Program showed age not be significantly correlated with any
dependent variable. The results of the regression analysis

are shown in Table 11 .

Table 11 . Statistics for Regression Analysis for the age
covariate.

 

 

 

Dependent Square Multiple P
Variable Multiple R R F Less Than
Achievement 0.0097 0.0987 0.5016 0.4821
Attitude 0.0030 0.0546 0.1527 0.6976
Study Time 0.0153 0.1235 0.7906 0.3782

 

Since age was not significant, it was not used as

a covariate.

96

Summapy of Ana1ysis of Covariateg

The analysis of variance for the mean grade point
average of the four treatment groups indicated that the
group were alike on this measure. Therefore, this covariate
was omitted from the analysis.

The chi square test was used to investigate the
differences in the proportion of students in each treatment
group who chose the course as an elective and those who must
take it. All groups had the same proportion of students.
Therefore, this covariate was omitted from the analysis.

The age covariate was introduced into the multi-
variance program. The results of the regression analysis
of the program showed that age had a very little correlation
to any dependent variable. Therefore, age was meaningless
as a covariate.

The following sections describe the multivariate
analysis of variance for the dependent variables of
achievement, attitude,and study time. First, the analysis
was done to determine any interaction of criterion and time.
Second, a failure to get interaction permitted investigation
of the differences between groups.

The statistical results are presented for
achievement, attitude and total study in that order.

Interaction of Criterion bijepeated Measures on Mean
Achievement

The overall hypothesis regarding the interaction of
criterion by repeated measures on mean achievement was:

There will be an interaction between treatments and time for

97

mean achievement. The results of the multivariance test
of interactions indicated an F-ratio of 0.8097 with a
probability of 0.6400. The initial decision was to reject
the overall hypotheses for interaction.

In addition to this hypothesis. several specific
hypotheses were written for the achievement dependent
variable. They are:

l. The ascending criterion group will have a
progressively higher score on achievement for each unit test
over the period of the quarter.

2. The 90% fixed criterion group will have a
progressively lower score on achievement for each unit test
over the period of the quarter.

3. The 80% fixed criterion will have the next
lowest but a moderately stable score on mean achievement for
each unit test over the period of the quarter.

4. The control group will have the lowest mean
achievement score of any group for each unit test over the
period of the quarter.

In order to investigate the above hypotheses, the
univariate results were used. The results are presented in
Table 12 . According to Cooley and Lohnes (1971) and Finn
and Mattsson (1978), one may examine the univariate results
for significant interactions when the initial test is not
significant. As shown in Table 12 , the results of the

univariate analysis of variance did not show significance

98

at a probability of 0.05. Therefore, there was no
interaction of the criterion by repeated measures on mean
achievement. As a result of this data, the overall and
specific hypotheses noted above for the interaction were

rejected.

Table 12 . Univariate results for the criterion by repeated
measures interaction on achievement.

 

 

 

Error Hypothesis Univariate P
Variable Mean Square Mean Square F Less Than
Linear 3.2978 8.5954 2.606 0.0615
Quadratic 2.8181 1.8935 0.6710 0.5732
Cubic 3.1516 0.2024 0.0642 0.9786
Quartic 2.5938 2.322 0.8955 0.4498

 

 

The results of the main effects were examined next
in order to determine any significant difference among
treatment groups.

The first hypothesis regarding the achievement
score on each unit test was: The criterion treatment groups
will receive a higher score on the measure of achievement
than the control class.

An analysis of variance was used through the Finn

program, Version IV,in order to analyze this hypothesis.

99

The results of group means and the analysis are shown in
Table 13 and Table 14 respectively.

The F statistic for the analysis of variance was
significant for all comparisons to the control. Therefore,
the hypothesis was accepted. Students taught under the
mastery strategy with criterion-referenced testing in this
research did score significantly higher in mean achievement
for each unit test than students of the traditional classroom

approach.

Table 13 . Mean achievement scores for each unit test for
the groups under study.

 

 

 

 

Control Treatment Groups
Group 80% 90% Ascending
Unit
Test S.D.* S.D.* S.D.* S.D.*
1 21.07(1.62) 22.26(1.31) 22.13(1.37) 21.88(l.40)
2 17.93(2.79) l9.45(1.58) l9.50(2.84) 18.65(3.37)
3 18.38(3.l9) 19.79(2.56) 21.25(2.36) 19.88(1.98)
4 19.63(3.93) 21.40(2.50) 23.18(0.96) 22.54(2.69)
5 18.6l(2.52) 19.37(2.60) 19.90(1.37) 21.31(1.32)

*S.D. is the abbreviation for standard deviation. All
standard deviations are in parantheses.

 

 

100

Table 14 . Univariate ANOVA for the comparison of the mean
achievement of the criterion groups to the control.

 

 

Experimental Error Hypothesis Univariate P
Group Mean Square Mean Square F Less Than
Ascending 7.01036 127.9246 19.6744 0.0002
80% Fixed 7.03457 133.6831 19.0037 0.0025
90% Fixed 6.70051 72.9300 10.8842 0.0002

 

The univariate analysis of variance was used to
determine differences in the mean score of achievement for
each unit test among the criterion groups. The analysis
was carried out by comparing each of the fixed criterion
groups to the ascending group. The hypotheses for these
tests were the following:

1. The ascending criterion will receive a higher
mean achievement score than the fixed criterion group of
80% or 90%.

2. The 90% fixed criterion group will get a
significantly lower score on achievement than the ascending
group.

3. The 80% fixed criterion group will get a

significantly lower achievement score than the ascending

group.

101

4. The 80% fixed criterion group will receive a
higher mean achievement score than the 90% fixed criterion
group.

The results recorded in Table 15 indicate that the
ascending criterion is no different than the 80% fixed or
90% fixed criterion on mean achievement of the unit tests.
The decision was to reject hypothesis 1.

The individual univariate tests made for hypothesis
1 permitted decisions to be made on the other hypotheses in
question. Since the ascending group was no different on
mean achievement from the 80% or 90% fixed group,then no
further analyses were required for the other hypotheses.
Since the previous statistic showed that the groups were
alike on mean achievement for each unit test than differences
did not exist between the other comparisons. Therefore,

hypotheses two, three, and four were rejected.

Table 15 . Univariate ANOVA for the comparison of the mean
achievement of the ascending criteria to the fixed

 

 

 

criteria.
Error Hypothesis Univariate P
Treatment Mean Square Mean Square F Less Than
Ascending To
80% 7.4849 13.1187 1.7527 0.1967

Ascending To
90% 8.1068 .4457 0.0550 0.8169

 

 

102

Interaction of Criterion by Repeated Measupes on Mean
Attitude

The overall hypothesis regarding the interaction of
criterion by repeated measures on mean attitude score for
each unit of study was: There will be an interaction between
treatments and time for mean attitude. The mean attitude
scores for each group under investigation are summarized in
Table 16 .

The F-ratio for the multivariate analysis was
2.9124 for 12 and 129.93 degrees of freedom with a
probability of .0014. The test of significant interactions
for the multivariate test was significant. The decision was
to accept the overall hypothesis.

The significant F value for the criterion by time
interaction indicated a different attitudinal response to
the course depending on the treatment (level of criterion
and the control). There was a change in direction of the
attitude as well as a change in magnitude of the attitude
score depending upon treatment group. Since the initial
interaction was significant, the means of all groups were
plotted. Figure 1 illustrates the trend of this plotting.
The graphed results were visually examined to analyze
trends which occurred over time.

The first specific hypothesis for interaction was:
The ascending group will have the most positive mean
attitude over the period of the quarter. The ascending
criterion group showed an increase in attitude at the close

of unit two,but it was not the most positive Change. While

MEAN ATTITUDE S C ORE

95

90

85

75

103

 

 

80% fixed // \\
. , / \ /‘\
I Ascend1ng——\\_,. \\ // \\
1’”! \\\ \ // \
M, \\ ' \\.
. / . \ //:\\\
Contro '\ \\ ,/ \\.
e\ \\ l//
90% fixed ' ‘ / '
\ \\ // ./ \ o

.. .\ \V/ / \.\

 

 

i 2 3 4 5
UNIT TEST

Figure 1. Mean attitude score plotted over time for
each group under study.

104

the attitude of students fluctuated up and down for the
remainder of the quarter, the trend in attitude showed an
overall decline. Therefore, the hypothesis was rejected.

The second hypothesis for interaction was: The 80%
fixed criterion group will have the next most positive
attitude over the period of the quarter. The pattern of
change of positive attitude increased much faster than the
ascending group by unit two. The attitude at this time
was the most positive of any group (see Figure 1). After
unit two, attitudes fluctuated up and down as the ascending
group but not as sharply. Even with these fluctuations,
the 80% criterion group continued to show the most positive
attitude. Therefore, the hypothesis was rejected.

The next hypothesis for interaction was: The 90%
fixed criterion group will have a progressively negative
attitude over the period of the quarter. A progressively
negative attitude was defined as a general decline in
attitude over the quarter. Overall, the trend was for a
decreasing student attitude toward the course throughout the
quarter. The attitude of students was less negative at unit
four, but this correction ended when attitude became more
negative again at the end of the quarter. Since the
attitude of the 90% group, in general, was progressively more
negative over time, the hypothesis was accepted.

The last hypothesis was related to the control
group. It was: The control will have the most negative

attitude toward the course over the period of the quarter.

105

The control did not react in the expected direction.

Instead, the attitude of the students toward the course was

relatively positive and unchanged throughout the quarter.

This is easily seen in Figure 1.

This was unlike the

criterion groups which showed increasing or decreasing

positive attitudes depending upon group.

for the control group was rejected.

Table 16 .

groups under study.

The hypothesis

Mean attitude scores for each unit test for the

 

 

 

 

Control Treatment Grou 8
Group 80%—_——____—90%_‘p Ascending
Unit
Test S.D.* S.D.* S.D.* S.D.*
1 88.47(8.40) 85.31(9.88) 85.80(5.98) 87.3l(7.33)
2 85.94(6.32) 92.63(9.21) 84.80(4.7l) 89.46(6.94)
3 83.94(5.63) 87-63(7-53) 75.90(ll.98) 79.46(7.61)
4 85.82(5.35) 90.63(9.02) 82.20(l3.52) 86.62(8.06)
5 83.82(7.09) 87.13(8.36) 78.80(10.39) 83.46(7.66)

*S.D. is the abbreviation for standard deviation.

standard deviations are in parantheses.

All

 

 

Since there were significant interactions, the

results of the main effects were meaningless.

hypotheses on mean attitude were ignored.

All further

106

Interaction of Cpiterion by Repeated Measures on Mean Study
Time

The overall hypothesis regarding the interaction
of criterion by repeated measures on mean study time was:
There will be an interaction between treatments and time
for the mean time spent on studies.

The results of the multivariate test of interactions
indicated a F-ratio of 2.0762 for 12 and 129.93 degrees of
freedom. The initial test for the interaction was
significant at a probability of less than 0.0227.

The significant F value for the groups by time
interaction on the measure of time spent on studies indicated
a different response to the amount of study time reported
depending on treatment level (criterion level and control).
Since significance was noted for the overall hypothesis,

a specific hypothesis is presented below for each treatment
group. The means concerning the treatment group are
summarized in Table 17. The means were plotted (Figure 2)
so that a visual examination of the trends in study time
could be analyzed. Based on this analysis, one could accept
or reject the subhypotheses.

The first hypothesis was for the ascending group.

It was: The ascending criterion will spend progressively less
time on studies over the period of the quarter. Progressively
less time is defined as a gradual decline in the amount of
study time over the quarter.

As shown in Table 17, the ascending group spent

107

progressively less mean time in hours on studies throughout
most of the quarter. As graphically illustrated in Figure
2, the mean study time did level off somewhat by the end of
the 10 week term (unit five). This result supported the
hypothesis. Therefore, the hypothesis was accepted.

The second hypothesis was: The 80% fixed criterion
will spend a steady amount of time on studies over the
period of the quarter. The 80% group did not respond as
expected. As shown in Figure 2, the reported total time
spent on studies for the 80% group was about the same for
unit one and unit two. After this point, total study time
declined but the initial rate of decline from unit two to
unit three was not as sharp as the ascending group or the
90% group. After unit three, students maintained a low
amount of study time. Furthermore, the results indicated
that the study time was about the same for unit four and
five. Since the results did not support the hypothesis,
it was rejected.

The 90% criterion-referenced group was expected
to spend more time on studies throughout the quarter.
Therefore, the third hypothesis was: The 90% fixed criterion
group will spend progressively more time on studies over
the period of the quarter. Progressively more time is
defined as a gradual incline in the amount of study time
over the quarter.

As reported for unit one, the students of the 90%

group began the quarter with the greatest amount of time

108

spent on studies. After that unit, the 90% fixed group
declined rapidly in mean study time on subsequent units
until unit five. The reported results (Table 17) show

that the mean study time increased at unit five. The rate
of decline for the quarter was the sharpest of any treatment
group. While the 90% fixed criterion reacted in the manner
just described, they did not respond according to the
expectation of the hypothesis for the 90% fixed criterion.
Therefore, the hypothesis was rejected.

Lastly, the control group was examined on the basis
of the following hypothesis: The control group will spend
the most time on studies over the period of the quarter.
The control group had somewhat of an erratic behavior on
study time. A decreasing trend in study time was noted in
the beginning of the quarter,but the time the students
spent on studies rapidly increased at unit three,followed
by a large decrease for unit four. After unit four, the
time spent on studies increased slightly. In addition,
the results were similar to those of the 80% group. These
results are shown in Table 17 and graphically presented
in Figure 2. Since the general trend of the control group
was to spend less time on studies over the quarter, the

hypothesis was rejected.

MEAN STUDY TIME

10.»

 

109

 

  

\

Control

\. . ’-A
. / \ . \."—-”.—
80% f1xed

i 2 3 4 5
UNIT TEST

Figure 2. Mean study time plotted over time for each

group under study.

110

Table 17. Mean study time in hours for each unit for the
groups under study.

 

 

Treatment Groups

 

 

Control Group 80%’ 90%“ Ascending
Unit
Test S.D.* S.D.* S.D.* s.0.*
1 4.69(2.80) 4.49(2.56) 8.96(4.ll) 7.05(3.94)
2 4.35(3.26) 4.53(2.33) 7.66(7.43) 6.70(4.05)
3 6.23(4.96) 3.35(1.90) 5.13(4.84) 4.85(3.41)
4 1.96(1.53) 2.3l(l.07) 4.15(3.43) 3.12(1.72)
5 2.72(1.97) 2.74(1.35) 5-31(3-23) 3-36(l-64)

*S.D. is the abbreviation for standard deviation. A11
standard deviations are in parantheses.

 

 

Summary

The statistical analysis of the study has been
presented in this chapter. Measures of the students
achievement, attitude toward the course,and study time were
taken for each of five units. The measures were analyzed by
a multivariate analysis of variance, Version IV, Finn

Program. The computer facilities of the California
Polytechnic State University at San Luis Obispo were used

during the analyses phase.

111

There were no significant interactions between
criterion and time for the measure of achievement. The
achievement scores of the criterion groups were significantly
higher than the control throughout the term. There were no
differences in mean achievement among any of the criterion
groups.

Five hypotheses were written for the interaction
of criterion and time on mean attitude toward the course.
The overall hypothesis indicated that there was a
significant interaction. The following hypotheses were
written in order to present the expected direction and
magnitude of a response by a particular treatment group
within the interaction. The decision for each hypothesis
is also presented.

1. The ascending group will have the most positive
mean attitude over the period of the quarter. The decision
was to reject the hypothesis.

2. The 80% fixed criterion group will have the
next most positive attitude over the period of the quarter.
The decision was to reject this hypothesis.

3. The 90% fixed criterion group will have a
progressively negative attitude over the period of the
quarter. The decision was to accept this hypothesis.

4. The control group will have the most negative
attitude toward the course over the period of the quarter.

The decision was to reject this hypothesis.

112

Since many of the attitude hypotheses were not
supported, the following summary of the actual responses is
presented.

The 80% fixed criterion group showed the greatest
initial increase in attitude toward the course from unit one
to unit two. During the same time, the ascending group
showed an increase in positive attitude,but to a lesser
extent. The higher positive attitude was not maintained
throughout the course by either group. Instead, the attitude
was more or less positive until the end of the term. The
fluctuating pattern was much greater for the ascending
group than the 80% fixed group. While student attitude did
fluctuate for the 80% fixed group, the students of this
group maintained the most positive attitude throughout the
term.

The 90% fixed criterion group had a progressively
more negative attitude over the quarter. The pattern was
interrupted at unit four. At that time, the student
attitude became more positive,but this increase was not
continued. Instead, students returned to being negative
in their attitude toward the course at the end of the quarter.

The students of the control group were neutral in
attitude toward the course. This was maintained throughout
the quarter.

Five hypotheses were stated for the interaction of
criterion and time on mean time spent on study. The overall

hypothesis showed that the interaction of criterion and time

113

was significant. Four specific hypothesis were written

to state the expected direction and magnitude of a response
by the particular treatment groups within the interaction.
They are shown below with the decision for each hypothesis.

1. The ascending criterion will spend progressively
less time on studies over the period of the quarter. The
decision was to accept this hypothesis.

2. The 80% fixed criterion will spend an equal
amount of time on studies over the period of the quarter.
The decision was to reject this hypothesis.

3. The 90% fixed criterion will spend progressively
more time on studies over the period of the quarter. The
decision was to reject this hypothesis.

4. The control group will spend the most time on
studies over the period of the quarter. The decision was
to reject this hypothesis.

Many of the above hypothesis were not supported.
Therefore, a summary of what did happen is presented below.

Overall, the criterion groups spent a decreasing
amount of mean time on studies throughout the quarter. The
ascending and the 90% fixed criterion groups showed the
greatest decrease in time spent on studies. This was very
obvious for the measurements taken at unit two, three and
four. For the same period of time, the 80% fixed criterion
decreased in mean study time but the decrease was not as
rapid. The results of the unit five indicated that the

ascending criterion and the 80% fixed criterion had not

114

changed much in mean study time from unit four. On the
other hand, the students of the 90% group showed a marked
increase in time invested in studies.

Overall, the students of the control group were
spending a lesser amount of time on studies by the end of
the quarter. But students did report an unusually large
amount of study time for unit three.

Limitations of the Results on Attitude and Study Time

Since the reliability of the attitude scale was low,
there are serious questions as to what the results on
attitude represent. For example, the fluctuating pattern
of attitudes throughout the quarter may have occurred from
other factors which confound the results. Therefore, the
attitude findings should be viewed with a great deal of
caution.

The total time as reported by students may not
reflect the relative efficiency of students. If students
do poorly on a unit test, they may take more time to study
for the next unit test in an effort to succeed. Therefore,
there is a possibility that the study time data may reflect
an over—reaction to the poor test results of a previous
test. This over-reaction may have led to a greater amount

of necessary study time.

CHAPTER VI

DISCUSSION, CONCLUSIONS AND RECOMMENDATIONS

Introduction

Schools can provide a successful learning experience
for most students. The use of criterion-referenced testing
under the mastery strategy offers the greatest potential
for these students. While this potential is present,
it has been hampered by the lack of a sound basis for
deciding whether or not a student can be considered a master.
A master is a student who has met or exceeded the criterion
score set by the instructor. Thus, the student has learned
to a sufficient degree. As yet there is no objective
manner for setting the level of the criterion which yields
the best learning.

If the problem of setting the criterion level
remains unsolved, the degree of mastery of many students
will continue to be misjudged. The specific amount of
skills a student must know cannot be adequately evaluated.
Also, the instructor cannot adequately judge how well
students have learned.

In an effort to solve this problem, this research

was addressed to the kinds of variations of presenting the

115

116

criterion to students in order to yield the greatest
achievement. Also, the variations of the criterion were
presented to students in order to produce the best positive
attitude toward a subject. Lastly, the purpose of the
research was to identify a criterion which produced
effective study scheduling throughout the term. Thus,
cognitive learning could be done with a minimum amount of
study time.

In order to provide a comprehensive answer to the
problem of setting the best criterion, the research sought
to investigate the following questions:

1. Does one criterion produce more achievement
than another?

2. Does one criterion produce better student
attitude than another?

3. Does one criterion produce more efficient
study scheduling than another?

The research questions were analyzed by a series
of hypotheses. A summary of these hypotheses are presented
below. They are:

1. There will be an interaction between treatments
and time for mean achievement.

2. The ascending criterion-referenced group, the
80% and 90% fixed criterion-referenced group will receive
a higher score on a measure of achievement than students

in the control class.

117

3. The ascending criterion group will receive a
higher mean achievement score than a fixed criterion group
of 80% or 90%.

4. The 90% fixed criterion group will get a
significantly lower score on achievement than the ascending
group.

5. The 80% fixed criterion group will get a
significantly lower achievement score than the ascending
group.

6. There will be an interaction between treatments
and time for mean attitude.

7. The mastery students under criterion-referenced
testing will have a higher mean score on a measure of
attitude than students in the control class.

8. The ascending criterion group will have a
significantly higher mean score on a measure of attitude
than the 80% or 90% fixed criterion groups.

9. Students in the 90% fixed criterion group will
have a significantly lower mean score on a measure of
attitude than the ascending group.

10. Students in the 80% fixed criterion group will
have a significantly lower score on the measure of attitude
than the ascending group.

11. There will be an interaction between treatments
and time for the mean time spent on studies.

12. The criterion-referenced groups will spend

118

significantly less mean time on instruction than the control
group.

13. The ascending criterion group will spend
significantly less mean time on instruction than the 80%
or 90% fixed criterion groups.

14. The 90% fixed criterion group will spend
significantly more mean time on instruction than the
ascending group.

15. The 80% fixed criterion group will spend
significantly more mean time on instruction than the
ascending group.

Experimental Design

The experimental design had multiple treatments
which were crossed with the five repeated measures. Since
the separate classes received different treatments, the
students of each class were nested within a treatment. The
number of students in the 80% fixed criterion group, the 90%
fixed criterion group, the ascending criterion group and
the control was 16, 10, 13, 17 respectively.

The treatments were criteria and the control.
Criterion was defined by three levels. The levels were 80%
and 90% of total points for each of five unit tests and an
ascending criterion which started at 80% of total points for
unit one and increased by 5% for each successive unit test
until 90% was reached. The last treatment was the control.

This group was graded on a straight percent of 90%, 80%, 70%,

119

60% and 50% of total points on the first and only try of
each unit test.

There were three dependent variables in the research
design. They were the achievement test, the attitude scale
and the total study time as reported by the student.

The dependent variable of achievement on each unit
objective test was the number of correct responses out of
total possible points. The achievement unit test was
administered during a lecture hour following the end of
each two week unit of study. A second and final
administration of a unit test was administered by arrangement
outside of class.

The dependent variable of attitude was measured at
the end of the first try of each unit achievement test.

The dependent variable of total study time was
collected from the students at the beginning of each unit
achievement test. Additional study time was collected from
students who restudied in preparation for the second and
final try of any achievement test.

The Sample of the Regearch

The sample of the research consisted of 56 third
and fourth year university students in Ornamental
Horticulture at the California Polytechnic State University,
San Luis Obispo, California. About 80% of the students
were seniors. The age of the students ranged from 21 to 24

years and about 31% were women.

120

The students of the sample were first-time high
school graduates and transfers from community colleges
throughout California. Transfers were 70.9% of the
sample.

Most students selected this course as an elective
as compared to a program requirement.

Finally, all students had the necessary pre-
requisite course of Fundamentals of Ornamental Horticulture.

Four separate classes are chosen for the research.
Since individual students could not be randomly assigned
to separate classes, the class itself had to be randomly
chosen for each experimental group and the one control
group. All the students in each class are used in the
research.

In order to conduct this research, the mastery
strategy was employed. Students are given a complete set
of objectives which delineated the content of the material
to which the course was addressed. In addition, the course
was broken down into segments or units of study. The units
covered two weeks of course material before any unit test
was given. The unit test was administered in the classroom
at the end of each unit. In the event students do not
master the evaluative instrument, correctives were offered
to the student. These correctives included additional
study of notes and/or textbook, further readings in

textbooks which are recommended references, listening to

121

audio—tapes related to the course and tutoring assistance

on material which was not understood. The student or
students who were classified as non-masters were retested

to determine if they have succeeded in the new understanding
of the course content of a unit.

Lastly, teaching for mastery demands that a
criterion be set which defines whether or not the student
can be declared as a master for part of all of the course.
One class was assigned a criterion of 80% for each unit
test: one class was assigned 90% for each unit test. The
last experimental group was assigned a criterion which had
an increasing standard. This group started at 80% and rose
5% each unit until 90% is reached. The response by the
students to the different criteria offers the opportunity
to explain the objective basis for setting the criterion
under the mastery strategy as used in this research. The
response was measured by an achievement score, an attitude
of the student toward the program and a total study time
invested in studies.

The control class was the last treatment group.
This group received the same objectives as the criterion
groups. They were given the same lectures on each unit of
study and the same test questions as the criterion groups.
They differed in several ways. First, they were not
required to achieve a level of criterion which defined

mastery.

122

Second, students were not given remediation for
their learning inadequacies. Third, retesting was not done
to re-evaluate student performance.

Method of_Data Collection

The measure of achievement was taken using an
instructor-made multiple-choice unit test which covered
five two-week sequences in Greenhouse Management. The test
questions were keyed to the objectives for each unit of
study. The total number of correct responses was the
individual's score. The analysis phase of the computer
program compared means for each unit test for each group
in order to determine group differences on achievement.

The attitude measure was taken by a self-made
attitude scale developed by the summated rating method
as described by Edwards (1957). The attitude scale was
administered to each treatment group immediately after the
achievement test was taken. A student with a strongly
positive attitude could obtain a score of 110. A strongly
negative attitude was measured at 22. A neutral attitude
was measured at 66. The average attitude score per unit
was used for comparison among treatment groups.

The measure of total time spent on studies was
taken as a student-supplied record of all time spent to
study for each unit. When additional study was required
as a result of not reaching mastery on the first test of

any unit, the additional time for that study was also

123

reported. This time was added to the rest of the time each
student spent on studies for each unit. The average total
time per unit was used for comparison among treatment groups.
The Importance of the Covariables

The covariables: age of the student, method of
course selection and grade point average were selected as
possible factors which could bias the data. It was thought
that these variables may influence the results of the
research. Therefore, the treatments may not be the sole
variable influencing the results. In order to avoid this
problem, the covariables can be removed statistically so
that the effects of the treatments can be analyzed.

The results indicated that students were no
different on age, grade point average and method of course
selection. The results may be generalized to students
of an age group of 21 to 24 years. Furthermore, the
results may be generalized to a group of students whose
range of average grade points was 2.76 to 2.87 out of a
#.00. Thus, the entering achievement level of the students
in each group was the same.

Lastly, all the groups had 70 per cent of the
students selecting the course as an elective. Therefore,

a sample with a ratio of 7 to 3 will be likened to this
sample. If the ratio would have favored a required selection
of the greenhouse management course, then one should expect

those students to respond differently to the attitude survey.

124

It seems reasonable to suggest that students who must take
a course have a different attitude toward the course than
students who want to take it. Furthermore, there is a
chance that students who are required to take a course will
be less motivated to achieve to a level that they are
capable. These students are often classified as under-
achievers. It also follows that underachievers probably
spend less time on their studies. With this in mind, the
study time data of this research may be confounded.

In summary, the samples of students in each of the
experimental groups involved in this research are
homogeneous with regard to age, grade point average and
method of course selection. As a result, one can expect
to draw students from such a population and get similar
results.

Qigcusgion of the Analyses_9§_the.Re§plts
pi§cussion on the Rggults of Achievement

The use of any criterion as used in the mastery
strategy of this research did produce higher achievement
scores than the control group. These findings should be
viewed in relation to the similarities and differences of
the criterion groups and the control group.

The criterion groups and the control were similar
with regard to the following procedures. The groups were
given instructions for completion of the course, objectives
for each unit and a set of review questions for each unit of

study. All instruction was done in class to an entire group.

125

The differences between criterion groups under the
mastery strategy and the control may have produced the
higher achievement for the criterion groups. First, the
mastery strategy of this research required students to reach
a particular level of achievement. For two of the three
mastery groups, the levels of achievement were 80% and 90%
of total points for each of five unit test. The third
group had an ascending criterion which started at 80% and
increased 5% for each successive unit test. Therefore,
attainment of an achievement score, which defined the
criterion, would indicate that the best learning of course
content has taken place.

Second, the significantly higher achievement scores
for the criterion groups may also be attributed to the
benefits of remediation which was employed in the mastery
strategy of the research. If students failed to reach a
certain criterion level, then certain alternative learning
resources were prescribed to students so that a better
understanding of the material might occur. These alternative
learning resources were additional reading materials on a
particular course objective or objectives, a review of
audio-tapes for a particular segment of a unit, tutoring
assistance or a review of the class notes of a student.

Lastly, the strategy of mastery permitted further
assessment of the adequacy of students' learning on a unit

of study in which there was a deficiency. Therefore,

126

retesting was done to evaluate the improved learning status
of a student.

A comparison among the criterion groups without the
control sought to answer the following question: Does one
criterion produce more achievement of the course content
than another? The results of this research have shown that
none of the treatments using variations of criteria were any
different from each other on achievement. Accordingly,
all students were alike in their learning of course content.

One may interpret these results as meaning that
the mastery approach of this research was able to bring
students to a minimum level of performance of the 80% fixed
criterion. A visual examination of the means in Table 13
shows that the averages were similar to the group subjected
to an 80% criterion. Setting a criterion any higher than
80% did not produce higher achievement. Therefore, it may
not be necessary to set a higher criterion under the
circumstances and subject matter described.

In summary, criterion-referenced testing under the
mastery strategy produced higher achievement than the
control group. The reason for the higher achievement may
be credited to learning under the mastery strategy. The
strategy is designed to bring most students to a better
understanding of the course content. In order to meet this
goal, students were presented with the results of their test

so that deficient areas could be noted. After this,

127

students were given the necessary alternative learning
resources which permitted further study of objectives not
understood. After additional studying was completed,
students were retested to evaluate their overall under-
standing of the objectives of a unit of study.

A comparison of the criterion-referenced groups
showed that the groups were alike on achievement of unit
tests. An examination of the means (Table 13 ) for all
criterion groups has revealed that the group means were
like the 80% group. Setting higher levels of criterion
would appear to be unrealistic under the mastery strategy
conditions of this research.

Discussion on the ResulthQf Attitugg

Attitude of the student was introduced into this
research as a dependent variable because it was thought
that certain criteria may influence the attitude of the
students toward the course. Therefore, it was necessary
to answer the question: Does one criterion produce better
student attitude than another? The results have shown that
there was a significant interaction between criterion and
the repeated measures on student attitude toward the course
in greenhouse management.

When the relatively unchanged attitude of the
students of the control was compared with the ascending
criterion and the 90% fixed criterion, it appears that the

overall attitude of the students of the ascending group and

128

the 90% fixed group were becoming more negative over time.
Furthermore, the attitude fluctuated up and down throughout
the quarter. One should have expected that students of the
ascending criterion and the 90% fixed criterion who were
given the opportunity to do better in their course work
should have had a better attitude about the course in which
they were enrolled. This was not the case in this research.
The setting of a high criterion either from the beginning

of the course or gradually working up to it as in the case

of the ascending criterion may have produced much frustration
in an attempt to achieve at such a high level. This idea
was further supported by the lack of significant differences
among the achievement of the scores of any of the criterion
groups. Even though the students were alike on achievement,
the students who were pushed to attain high achievement
scores reacted by becoming more negative toward the course.
It may be that students though that the higher criterion was
an unreasonable expectation of academic success in greenhouse
management. Therefore, they became more negative as the

term progressed.

The 80% fixed group showed fluctuations in its
attitude toward the course. While attitude fluctuated, it
was the most positive attitude of any group. The relative
ease of attaining criterion for each unit may have produced
a high positive attitude toward the course. Additionally,

the students were probably not threatened with the prospect

129

of achieving to an unreasonably high level. Their high
academic standing was very secure and therefore there may
not have been any reason for students to get a negative
attitude.

The fluctuating pattern of attitude for all the
variations of criterion may have resulted from a difference
in the difficulty of one test or another. The tests were
constructed with the intention of one not being more
difficult than another, but there was no measurement of
difficulty. One might argue that if the difficulty of a
test influenced the results, that attitude score could have
been influenced by the relative ease by which criterion
groups could reach a level of performance. This could
account for the degree of fluctuation of attitude throughout
the quarter by the criterion groups. When a criterion was
easier to reach, the attitude tended not to change as much.
This argument is weak because the control group did not
react to any apparent difference in test difficulty.
Therefore, the author suggests that the results were more
likely due to the level of criterion and the requirement
that students reach it.

Alternatively, various personal circumstances of
students on the day of the test might have changed their
attitude. But this argument is even weaker than the above
because all students in the different criterion groups

could not have had simultaneous good or bad days.

130

Furthermore, the control group would have reacted in a
similar fashion. But the control did not show fluctuations
in attitude as did the criterion groups.

The results on attitude were further complicated
by the low reliability of the attitude measure. There was
a chance that the variations in attitude or the lack of
variation was due to an inconsistency of what the scale
intended to measure. For example, the unchanged attitude
of the control group could have happened from a lack of
accuracy of the scale to detect attitudinal change over time.
On the other hand, the fluctuations of attitude in the
criterion groups may have resulted from the scale measuring
something other than attitude toward the course. For
example, students might have felt that the method of
instruction under which they were taught was unreasonable.

In summary, the overall trend in student attitude
toward the course varied according to the treatment group.
By setting a low criterion of 80% the attitude of students
was the most positive of any treatment group. The 80%
fixed criterion produced the best achievement in the course
without sacrificing the positive attitude of the students
toward the course. It may be that the relative ease at
which students achieved on unit tests produced a positive
attitude toward the course.

The most negative attitude toward the course over
time was noted for the 90% fixed criterion group. Also, the

general trend for the ascending attitude was to be

131

progressively negative in attitude over time but not as much
as the 90% fixed criterion. It may be that when students
were pushed to attain high levels of achievement from the
beginning of the quarter or gradually working up to high
levels of achievement, the effect was to produce a negative
student attitude toward the course in greenhouse management.
Also, if students felt threatened by a criterion which was
difficult or impossible to attain, the response may have
been a negative perception of the course.

Discussion on the Resglts of Total Study Timg

Since students of the mastery strategy were
subjected to additional study and retesting when their
performance on a unit test did not reach a level of criterion
set for a treatment, a dependent measure of total time on
studies was made. This total study time is a measure of
relative efficiency among the treatment groups. Thus,
the research sought to answer the following question: Does
one criterion produce more efficient study scheduling than
another?

A significant criterion by time interaction on the
dependent measure of study time indicated that there were
differences in the amount of total study time among the
treatment groups through the quarter. The differing amounts
of time accounted for the change in direction of the plotted
total study time as shown in Figure 2 .

The trend was for an increasing efficiency of

132

studies which was noted by a general decline in study time
for each successive unit. While efficiency increased over
time, the amount of time spent on studies varied with
treatment groups.

As the results generally have shown, the 90% fixed
criterion and the ascending criterion declined in study
time over the quarter. But the time spent on studies was
still more than the 80% criterion or the control. The
students of 90% group spent the greatest amount of time over
the quarter. This was followed by the students of the
ascending group. The decline in total study time did not
persist for the 90% fixed group. Instead, there was an
increase in study time at the end of term.

The trend by the 90% fixed criterion group is
explained in this manner. The students of the 90% group
were faced with a very high level of criterion. Because of
this high level, the students apparently put more time into
studies than any other treatment group in an attempt to
reach the criterion set for each unit test. But a higher
level of achievement was not reached as a result of adding
additional hours of study as compared with the other
criterion groups. There were no significant differences on
achievement among criterion groups as has already been
stated earlier.

The increase in total study time at the end of the

term might have resulted from students being insecure about

133

doing well on the last examination of the quarter. Since
this group may not have been sure about their ability to
reach the criterion score at the end of term, more hours
were used to study. Thus, while there was a trend to use
less total time to study by the students of the 90% fixed
group, the study efficiency was not maintained at the end of
the term.

The ascending criterion and the 80% fixed criterion
were alike on the level of criterion set in the beginning
of the term. However, the ascending criterion spent more
total time on their studies at that time. But the total
study time was less than the 90% fixed group (see Figure 2 ),
While the level of criterion was easier to attain at that
time, the students of the ascending criterion may have
felt that it was necessary to obtain the highest achievement
score possible early in the term. Faced with the
uncertainty of doing well later in the quarter when the
criterion would be higher, the students may have studied
alot more in order to do their best on the earlier unit
tests. Therefore, additional hours of study were spent in
a hope of learning as thoroughly as possible.

When the level of criterion for the ascending group
was increased to 90% at unit three, the ascending criterion
group became similar to the 90% fixed group on total study
time. Subsequent to unit three, the ascending group showed

further decreases in total study time and finally a leveling

134

off at the end of term. On the other hand, the 90% fixed
group was tending to increase in total study time during
the latter period of the quarter. Since the ascending group
was able to spend less time than the 90% group when the
level of criterion was kept at 90% then the ascending group
probably became more efficient later in the quarter. These
results can be interpreted in the following way. When the
level of criterion for the students of the ascending group
was lower, there was an opportunity to become adjusted to
studying for the course. Once their study routine was
established, the students did not have to put in any more
time than necessary in an attempt to reach the level of
criterion set for the course. This is graphically
illustrated for these groups in Figure 2. Thus, the study
time levelled out at a low point and stayed that way until
the end of term. It appears that gradually ascending the
criterion has some benefit in producing more efficient
study.

Except for the unusually great amount of time spent
on studies at unit three by the control, the 80% criterion
group and the control were similar on the decreasing time
spent on their studies throughout the quarter. The 80%
group had a significantly higher achievement than the
control with about the same amount of total study time over
the quarter. Thus, the students of the 80% criterion were

more efficient in total time spent on studies.

135

Also, the lack of significance among the scores of
the criterion groups suggested that the 80% criterion is
the best criterion to set in order to produce the best
achievement in a time-saving manner.

In summary, there was a general trend for a
decreasing amount of total time spent on studies over the
quarter. While the time decreased as the quarter progressed,
the amount of time spent on studies varied with the
treatment group.

The 80% fixed criterion group had the least amount
of time spent over the quarter. Also, the 80% fixed
criterion group was generally like the control on total
study time over the quarter. Since the 80% group achieved
more than the control class over the quarter with the same
amount of study time, the students were more efficient in
their study scheduling.

Also, the reported results of no difference among
the achievement scores of the criterion groups suggests
that the 80% group was the most efficient of any criterion
group in their total study time. This is graphically shown
in Figure 2 .

On the other hand, the 90% group spent the greatest
amount of time on studies even though it tended to decline
over time. The decline in total study time did not last
until the end of term. After unit three, there was a

general trend to increase the time spent on studies in order

136

to achieve as much as the other criterion groups. The
students studied a great deal for each unit test in an
attempt to reach the high level of criterion set for the
course. Even though a great deal of time went into studying
for each unit, the students did not achieve any more than
the other criterion groups.

The ascending criterion group began the quarter
with the second greatest amount of time spent on studies.
It was at that time that the level of criterion (80% for
unit one) was the same as the 80% fixed criterion. Even
though the criterion was the same, the ascending group
apparently spent more time in an effort to achieve a high
achievement score. Since the criterion of the ascending
would become higher later, the students might have felt
compelled to get the highest possible score early in the
quarter.

As time passed in the quarter the total time
students spent on studies became less and less. Later in
the quarter, total study time became more like the 80%
fixed criterion group. The latter trend suggests that the
students had a chance to establish effective study scheduling
early in the term. Later in the quarter, it was not
necessary to use any more time than necessary to attempt
to score well. If this was not the case than the ascending
criterion should have reacted more like the 90% fixed

criterion later in the quarter. This group had a trend

137

upward in total study time toward the end of the quarter.
CONCLUSIONS

The findings of this research have led to the
following conclusions. The conclusions are presented for
the dependent measures of achievement, attitude toward the
course and total time spent on studies in respective order.

1. The conditions of criterion-referenced testing
under the mastery strategy of this research produced
significantly higher achievement when compared with the
control. Those conditions included the following:

a. The course was divided into units of study.

b. Each unit of study had specified objectives.
These objectives stated what the student is expected to
learn.

0. Testing was done to evaluate the knowledge
learned for each unit.

d. The student was evaluated relative to the
performance on a unit test. Therefore, a level of criterion
defined mastery or the adequacy of learning on a unit.

e. Students were given an opportunity to review
their test and note a deficiency in learning.

f. Those who fail to master the content were given
alternative learning resources. In this way further study
could be done on the content of a unit.

g. Retesting was done to determine if mastery of

content has been reached.

138

2. Under the mastery conditions of this research,
achievement was the same regardless of the level of criterion.
These findings have shown that higher achievement was not
produced by setting a higher criterion throughout the term.
In addition, a gradual increase from 80% to 90% (ascending
criterion) did not yield any more learning than the other
groups. Based on these findings, it appears that the
level of criterion to set for instruction like that in this
study should be 80%. Above that point, students were unable
to master any more material.

3. Because the control group had a relatively
unchanged attitude over time as compared with the decreasing
positive attitude of the ascending criterion and the 90%
fixed criterion groups, it can be concluded that students who
were pushed to attain high levels of criterion may become
negative in their attitude toward the course as time
progresses. This trend was not shown by the 80% fixed
criterion group. The higher positive attitude of the
students of the 80% group throughout term suggest that it may
be better to set the criterion lower to produce the best
achievement and the most positive attitude toward the course.
The lack of significant difference among achievement scores
of the criterion groups suggests that the push to attain
higher levels may have produced a decline in positive

student attitude toward the course.

139

4. The setting of a high fixed criterion (such as
90%) forced students to spend a great deal of time over the
quarter in an attempt to reach the high level set for the
course. This was of no avail since students were not able
to reach the high level of achievement set for the course.
The criterion may be an unrealistic standard to set under
the conditions of mastery as used in this research.

5. There is no reason to believe that gradually
raising the criterion under mastery learning has any
advantage in producing more achievement in a time-saving
manner than a lower fixed criterion. The ascending group
eventually decreased in average total study time to
approximately the averages of the 80% fixed group in order
to achieve as well as the 80% group. Thus, it may be
better to set a relatively low criterion (such as 80%).
This produced the best efficiency of time spent on studies
throughout a quarter.

6. In general, the similar trends and total study
averages between the 80% and the control suggests that
setting an 80% fixed criterion strategy may be more
efficient in producing higher achievement under mastery
learning than the control or non-mastery group.

Summary of Conclusions

In summary, the conditions of criterion-referenced

testing under the mastery strategy produced significantly

higher achievement when compared with the control.

lQO

The research also indicated that setting a higher
criterion or gradually raising the criterion did not yield
higher achievement than a lower criterion. The student
attitude toward the course was also less positive when they
were pushed to attain higher levels of criterion. Also,
higher fixed criterion (90%) produced less efficient study
scheduling than other criteria without any gain in
achievement over a lower criterion. While the trend in the
ascending criterion (start at 80% and increase 5% to 90% and
then hold at 90%) was for increasing efficiency, there was
not any commensurate gain in achievement over the other
criterion groups.

The findings of this research suggested that the
80% fixed criterion could produce the best learning without
sacrificing the attitude of students toward the course.

In addition, a criterion which was set at 80% yielded the
best student learning in the least amount of total study
time over a quarter.

RECOMMENDATIONS AND FURTHER QUESTIONS

The following recommendations and questions for
further research are based upon experiences of this research
project. They are:

l. The setting of a relatively low criterion of
80% produced the best achievement. However, there may have
been an adverse effect of the low criterion. The students

may not have learned enough material in the quarter. To

141

overcome the problem of the learning of an inadequate amount
of material, a different criterion may be necessary to assure
that an adequate amount of material is learned. Would a
criterion of 85% yield the best learning?

2. The degree of difficulty of a particular
criterion may be expressed as the number of students who
master units of study. This type of data may be useful for
setting an optimum criterion. Therefore, the following
question may be answered: Is there a criterion which produces
the greatest number of masters?

3. Data on the number of retests taken by students
at the various levels of criterion should be taken. This
may suggest the relative difficulty of a criterion. In this
way, the following question may be answered: Is there a
criteria which produces the lowest number of retests among
students?

4. The attitude scale in future research should be
revised to include a subset on attitudes toward the method
of instruction under which the students are being taught.

In this way it may be possible to measure the attitude of
students toward the use of criteria under the mastery
strategy. Would the use of criteria produce attitude
changes in students who are taught by the mastery strategy?

5. An attitude scale with a low reliability may

produce results which are not a consistent measure of

142

student attitudes toward the course. An attitude scale with
a minimal reliability of 0.75 should be used to assure
accuracy in measurement of student attitude. With higher
reliability, one would have a great confidence in the
predictability of the results. Consequently, the following
question may be answered with greater accuracy: Is there a
level of criterion which produces the best attitude toward

a course taught under the mastery strategy?

APPENDI CES

APPENDIX A

Course Objectives

Greenhouse Management (one quarter)

Define management

1.

Given the term management, the learner will define
management as developed from the class discussion.

Given the term management, the learner will define the
term as it is correctly identified to line management,
maintenance, production management and sales management
as stated in the class.

Given the term manager, the learner will define the term
as it was developed in class.

Given statements or components of statements for
management or manager, the student will identify the

statement as it applies to management or manager.

Describe the types of business ownership

1.

Given the terms sole proprietorship, partnership and
corporation and without reference material, the learner
will describe them in writing as to include the
considerations, liabilities and limitations of the

terms with 100% accuracy.

143.

144

Given the terms general partnership and limited
partnership and without reference material, the learner
will describe as to liabilities and limitations of the
terms with 100% accuracy.

Given the article terms for corporations and partner-
ships, the learner will define in writing that term with

100% accuracy and as presented in class.

Describe different ornamental horticulture marketing set-ups

1.

Given the terms wholesale house, pool system, cooperative
and direct sales on a work sheet, the learner will
describe in writing each term as it is applied to the
marketing of ornamental horticulture crops according to
lecture.

Given the terms wholesale house, pool system, cooperative
and direct sale, the learner will list secondary
advantages to the ornamental horticulturist in columnar
form for each term according to lecture.

Given a statement or phrase which best describes the
marketing set-up, the learner will choose the best term
from a group which identifies the statement. This will
be with 100% accuracy.

Given a key phrase as marketing channels, the learner
will diagram the channels for the sale of ornamental
horticulture crops according to lecture and references

used to develop lectures.

145

Apply marketing set-ups to ornamental horticulture firms

1. Given the marketing set-up terms and the ornamental
horticulture firms, the learner will identify each firm
with the appropriate marketing set-up term with lOO%
accuracy.

2. Given the marketing term 'auction selling' and without
reference material, the learner will apply this
marketing procedure to the sale of cut flowers and
potted plants in the U.S. as presented in class.

Diagram organization flow charts for various ornamental

horticulture business types

1. Without reference material, the learner will diagram
the three types of business organization in a
hierarchical manner labeling each level within the
diagram with 100% accuracy.

2. With the aid of the above diagram, the learner will
diagram with arrows the manner in which communication
and responsibility flows with 100% accuracy.

3. Given the concept of communication within business, the
learner will explain in writing how it is best achieved
in business. This will be done according to class
discussion and the text on Greenhouse Management.

4. Given the concept of responsibility within business, the
learner will explain in writing:

a. its role in the organizational charts

b. its relationship to communication

146

0. its relationship to business activities within the
organization
This will be done according to class discussion and the

text on Greenhouse Management.

END OF UNIT I

Identify different recruitment procedures

1.

Without reference material, the learner will identify
procedures to follow in order to recruit the following
employee types:

a. unskilled worker

b. foreman

c. assistant manager

d. truck driver

e. production manager

f. agricultural economist

Describe the orientation procedures

1.

Without references, the learner will write reasons for
orientating workers according to lecture notes.
Without references, the learner will write items which
represent areas of orientation according to lecture
notes.

Without references, the learner will state in writing
the relationship between orientation and business

efficiency according to lecture notes.

Plan training procedures for ornamental horticulture workers

1.

With references and job forms, the learner will prepare

147

a list of steps for getting ready to train employees

including shortcut steps for the employee for an actual

job according to the procedures of the Agriculture

Education Department of the University of California at

Davis.

Given references, job forms, greenhouse facilities,

tools, soil, plants, etc., the learner will plan an

actual training session with a member of the class as

the trainee. The training session will be judged on a

scale of l to 5 for each of the following:

a. effectiveness of training

b. training under actual conditions

0. clarity of training

d. preparing the worker

e. preparing the job

f. the steps listed on the one-page leaflets given in
class

Given an on—the-job problem situation, the learner will

identify the problem as it relates to orientation or

training. This will be done according to handouts and

lecture notes.

END OF UNIT II

Estimate of production peaks for ornamental horticulture

crops

1.

Without references, the learner will calculate the

expected date of peak production by random counting of

l48

selected shoots or flower buds or stages of growth with
90% accuracy. (Accuracy is based on previous crop
records.)

With the necessary references, the learner will graph or
write the dates of estimated peak production for a one—
year period with 100% accuracy.

Given references, the learner will estimate the date of
flowering of Easter Lilies according to the standard
leaf and bud counting procedures in print.

Without references, the learner will state in writing
the stages of plant growth which decide when crops will
peak in accordance with the stages shown in laboratory

on living plants.

Calculate year around crop rotations for selected crops

1.

With references and crop rotation forms, the learner will
calculate year around pot plant rotations for:

a. the appropriate number of crops

b. utilizing 365 days of the year

c. for specific holiday periods

d. utilizing 85% of the space each month

All forms to be returned in an appropriate, specified
period of time.

Given references, a list of cut flower crops or nursery
crops, holiday names and dates, the learner will calculate
the correct planting dates, pinching dates and bloom

dates based on the first expected holiday bloom within

a two-day accuracy.

149

Calculate number of plants/pots

1.

Without references, the learner will calculate the

number of pot needs in area to the exact value as
computed by the prescribed formula.

Without references, the learner will calculate the

number of pots in the width and the number of rows in

the length. This is in accordance with Weiss formula.
Without references, the learner will calculate the

number of plants in a bed according to formulas presented

in class with lOO% accuracy.

END OF UNIT III

Schedule year around ornamental horticulture crops

1.

Given reference sources, the student will write a
schedule for different nursery crops. The schedule is
correct when the crops are available at the same time.
Given references or cues, the student will schedule in
writing a floral crop so that the crop is available:
a. on the specific holiday or peak date

b. on a weekly, biweekly or monthly schedule

c. during a season

Without references, tight spacing will be calculated
when it is necessary to get maximum utilization of space.
With the use of assigned references, the student will
write a chart showing:

a. start and stop dates for each crop

b. total time of production in days, weeks or months

150

c. the holiday or season for each crop
The chart is correct when it conforms with all

references and commercial practices.

END OF UNIT IV

Compile cost estimates for greenhouse construction

1.

Given references, forms and greenhouse blueprints, the
learner will compile a cost estimate for all necessary
costs of construction material by using a local lumber
company as an estimator with 90% accuracy.

Given references, forms and the names of greenhouse
owners, the learner will compile a cost estimate for
heating systems, cooling systems, all plumbing, all

electricity and growing tables with 90% accuracy.

Analyze selected total cost of production

1.

Given references, the learner will write a comparison of
selected cost of production for different regions of the
country. This will be correct when a chart shows the
items with the cost for each region.

Given references, the learner will state in writing the
ratios or % which indicate the financial condition of
the company. This is correct when it is in accordance
with the Operating Cost Studies of the Horticultural

Research Institute.

Define financial term

1.

Without references, the learner will write definitions

for terms in accordance with the operating cost studies

151

of the Horticultural Research Institute.

Compute cost of production for selected crops

1.

Given a list of crops, cost of production materials,

reporting forms and previous assignments, the learner

will compute the cost of production by using the

reporting form for each crop until all information

for each crop is recorded as requested on the reporting

forms. The forms are to be returned in an appropriate,

specified period of time. All cost figures must be

within three decimal place accuracy.

Given the computed cost of production for selected

crops, the learner will compute:

a. the selling price for each crop in order to break
even

b. the total cost per sq. ft. per crop

c. the total cost per sq. ft. of production area per
year

d. the gross receipts of all crops per sq. ft. per year

e. the net receipts of all crops per sq. ft. per year

to the nearest .01 dollars

Describe profitability of crops

1.

Without references, the learner will write ways to
manage profitability for ornamental horticulture crops

according to Perkins and Levins.

END OF UNIT V

1.

7.

8.

9.
lo.

11.

APPENDIX B

Edward's Criteria for Selecting Attitude Statements

Avoid statements that refer to the past rather than to
the present.

Avoid statements that are factual or capable of being
interpreted as factual.

Avoid statements that may be interpreted in more than
one way.

Avoid statements that are irrelevant to the psychological
object under consideration.

Avoid statements that are likely to be endorse by almost
everyone or by almost no one.

Select statements that are believed to cover the entire
range of the affective scale of interest.

Keep the language of the statements simple, clear and
direct.

Statements should be short, rarely exceeding 20 words.
Each statement should contain only one complete thought.
Statements containing universals such as all, always,
none and never often introduce ambiguity and should be
avoided.

Words such as only, just, merely and others of a similar

152

12.

13.

14.

153

nature should be used with care and moderation in
writing statements.

Whenever possible, statements should be in the form
of simple sentences rather than in the form of
compound or complex sentences.

Avoid the use of words that may not be understood by
those who are to be given the completed scale.

Avoid the use of double negatives.

APPENDIX C

Attitude Survey A

INSTRUCTIONS: Mark your honest feelings concerning
Ornamental Horticulture as a field of study or profession.
Do this by responding to the following statements. Use
the answer sheet provided. Do not answer on the statement
page or on your lecture answer sheets.

Some of the statements reflect agreeable attitudes or
feelings: some reflect disagreeable attitudes or feelings.

You are to rate each statement according to HOW agreeable
or HOW disagreeable an attitude or feeling it has on you.
When you cannot decide one way or the other, you may mark
undecided.

a b c d e
Strongly Strongly
agree Agree Undecided Disagree disagree

l. It is great to work outdoors.

2. As a profession, Ornamental Horticulture adds much
beauty.

3. The work in Ornamental Horticulture has a lasting
benefit to people.

4. Trained monkeys can perform Ornamental Horticulture
jobs.

5. Ornamental Horticulture makes people become

depersonalized.

15L:

9.
10.
11.
12.

l3.
l4.
l5.
l6.
l7.
l8.
19.

20.
21.
22.

155

Working with a person's hands in Ornamental Horticulture
gives enjoyment.

Fruit is a better gift for people.

(Ornamental Horticulture is of little productive value

to society.

Gardening can be done by homeowners without training.
Physical labor is too much in Ornamental Horticulture.
Ornamental Horticulture upsets nature.

Enjoyment in the beauty of life is provided by
Ornamental Horticulture.

Ornamental Horticulture saves the environment.
Ornamental Horticulture provides for a beautiful world.
It is enjoyable work.

Ornamental Horticulture causes people to kill flowers.
Floral goods make people feel good.

It is unnecessary to society.

Ornamental Horticulture as a profession takes too much
time.

Ornamental Horticulture is a luxury.

Ornamental Horticulture communicates gifts of nature.
Accomplishments in Ornamental Horticulture are very

rewarding.

APPENDIX C

Attitude Survey B

INSTRUCTIONS: Mark your honest feelings concerning
Ornamental Horticulture as a field of study or profession.
Do this by responding to the following statements. Use

the answer sheet provided. Do not answer on the statement
pages or on your lecture answer sheets. Write your student
number on the answer sheet.

Some of the statements reflect agreeable attitudes or
feelings; some reflect disagreeable attitudes or feelings.

You are to rate each statement according to HOW agreeable
or HOW disagreeable an attitude or feeling it has on you.
When you cannot decide one way or the other, you may mark
undecided. Note the responses under each letter.

a b c d e
Strongly Strongly
agree Agree Undecided Disagree disagree

1. Ornamental Horticulture manipulates nature.

2. Ornamental Horticulture is unaware of new concept.

3. The conditions of work are bad.

4. Ornamental Horticulture makes people work long hours.

5. Ornamental Horticulture is for dummies.

6. Ornamental Horticulture is a trivial field of study.

7. There is satisfaction in Ornamental Horticulture from
visual accomplishments.

8. Ornamental Horticulture is a very rewarding profession.

156

9.
10.
11.
12.
13.
14.
15.
l6.
l7.
l8.
19.
20.
21.

22.

157

Ornamental Horticulture
People are able to work
Ornamental Horticulture
Ornamental Horticulture
Ornamental Horticulture
Ornamental Horticulture

Ornamental Horticulture

is very pleasing work.

in open air.

creates a psychological lift.
is very stimulating to people.
is great for the health.
destroys essential land use.

creates beauty.

Working with a living plant form is satisfying.

Regardless of the field,

plants are satisfying.

Living plants make living more bearable.

Ornamental Horticulture

is for stupid people.

A degree is not worth anything for the job.

Ornamental Horticulture

people.

gives little prestige to

Unskilled workers can do the jobs in Ornamental

Horticulture.

BI BLIOGRAPHY

l.

2.

10.

ll.

BIBLIOGRAPHY

Aleamoni, Lawrence M. Why Is Grading Difficult?
National Association of Colleges and Teachers of
Agriculture Journal, March 1979. Vol. XXIII, No. l.

Amthor, W.D. Is There Merit in a Pass-Fail Grading
System? Paper presented at the meeting of the Americam
Industrial Arts Association, Minneapolis, April 1968.

Barker, Donald G. Development of a Scale of Attitudes
Toward School Guidance, Personnel and Guidance Journal,
June 1966.

Berk, R.A. Determination of Optional Cutting Scores
in Criterion-Referenced Measurement. Journal of
Experimental Education, 1976. Vol. 45, 4-9.

Block, J.H. Mastery Learning, Theory and Practice.
Holt, Rinehart and Winston, Inc., 1971.

. Schools, Society and Mastery Learning.

 

Holt, Rinehart and Winston, Inc., 1974.

. Student Learning and the Setting of

Mastery Performance Standards, Educational Horizons,
1972, pp. 183-191.

. The Effects of Various Levels of
Performance on Selected Cognitive, Affective and Time
Variables. Unpublished Ph.D. dissertation, University
of Chicago, 1970.

 

 

. Criterion-Referenced Measurements:
Potential. School Review, February 1971, pp. 289-298.

 

Campbell, D.T. and Stanley, J.C. Experimental and
Quasi Experimental Designs for Research, 1973. Rand
McNally College Publishing Company.

Carlson, J.G. and Minke, K.A. Fixed and Ascending

Criteria for Unit Mastery Learning, Journal of
Educational Psychology, Vol. 67, No. l, 1975, pp. 96-101.

158

12.

13.

14.

150

16.

17.

18.

19.

20.

21.

22.

230

159

Carroll, John B. Problems of Measurement Related to
the Concept of Learning for Mastery, Educational
Horizons, 48, No. 3, 1970, pp. 71-80.

Cooley, W.W. and Lohnes, P.R. Multivariate Data
Analyses, 1971. John Wiley and Sons, Inc.

Ebel, R.L. Criteria Referenced Measurements:
Limitation. School Review, February 1971, pp. 282-288.

Educational Development at Michigan State University,
Analysis of the Mastery Instructional Model, No. 5,
Spring 1973, Office of the Educational Development
Program, Michigan State University.

Edwards, Allen L. Techniques of Attitude Scale
Construction. Prentice Hall, Inc., Englewood Cliffs,

NJ, 1957.

Finn, J.D. and Mattsson, I. Multivariate Analysis in
Educational Research. Applications of the
Multivariance Program, 1978. International Educational
Services, Chicago.

Foth, Henry D. A Mastery Learning Program in Soil
Science, Journal of Agronomic Education, Vol. 2,
November 1973, pp. 65-68.

. Improving Learning with Mastery

Learning, Department of Crop and Soil Science,
Michigan State University, East Lansing, MI 48824

 

Glaser, Robert. Adapting the Elementary School
Curriculum to Individual Performance, Proceedings of
the 1967 Invitational Conference on Testing Problems,
Princeton: Educational Testing Service, 19 8, pp. 3—36.

Gray, William M. A Comparison of Piagetian Theory and
Criterion-referenced Measurement, Review of Educational
Research, Spring 1978, Vol. 48, No. 2, pp. 223-249.

Hambleton, R.K., Hutten, L.R., and Swaminathan H.

A Comparison of Several Methods for Assessing Student
Mastery in Objective-Based Instructional Programs.
Journal of Experimental Education, 1976, 45, 57-64.

Hambleton, R.K.; Swaminathan, H.: Algina, J.: Carlson,
D.B. Criterion-Referenced Testing and Measurement: A
Review of Technical Issues and Developments, Review of
Educational Research, Winter 1978, Vol. 48, No. 1,

pp. 1' 7o

160

Hapkiewicz, Walter G. and Foth, W.G. Can All Students
Learn What They Are Taught? Educational Development
Program Report, Michigan State University, East
Lansing, MI, No. 35, January 1973.

Harris, T.C., Kiefert, J.T. and Darby, M.D. Attitudes
Expressed by Students Toward a Beginning Course in
Educational PsychologY. Journal of Educational
Research, Vol. 62, No. 8, April 1969.

Johnson, J.J., Gnagey, W.J. and Chesbro, P.M. The
Effectiveness of Applying the Concept of Mastery to the
Teaching of Educational Psychology. Paper presented

at the meeting of the American Educational Research
Association, Minneapolis, March 1970.

Malott, R.W., Editor. Research and Development in
Higher Education: A Technical Report of Some Behavior
Research at Western Michigan University, Summer 1971.

Mehrens, William A. and Lehmann, Irvin J. Measurement
and Evaluation in Education and Psychology, Holt,
Rinehart, Rinehart, Winston, Inc., 1973.

Merrill, M.D., Barton, Keith and Wood, L.E. Specific
Review in Learning a Hierarchical Imaginary Science,
Journal of Educational Psychology. 61, 1970, pp. 102-109.

Millman, J. Passing Scores and Test Lengths for
Domain-Referenced Measures. Review of Educational
Research, 1973, Vol. 43, No. 2.

Neidt, Charles 0. and Hedlund, Dalva E. The
Relationship Between Changes in Attitudes Toward a
Course and Final Achievement, Journal of Educational
Research, Vol. 61, No. 2, October 1967.

Sherman, J.G. Application of Reinforcement Principles
to a College Course. Paper presented at the annual
meeting of the American Educational Research
Association, New York, 1967.

Wentling, T.L. Mastery Versus Non-mastery Instruction
with Varying Test Item Feedback Treatments, Journal of
Educational Psychology, Vol. 65, No. l, 1973, pp. 50-58.

Wentling, T.L. and Erickson, Richard C. Measuring
Student Growth, Allyn and Bacon, Inc., 470 Atlantic
Avenue, Boston, MA, 1976.

35.

161

Wofford, J.C. and Willoughby, T.L. Attitudes and
Scholastic Behavior, Journal of Educational Research,
Vol. 61, No. 8, April 1968.

General References

Borg, W.R. and Gall, M.D. Educational Research, 2nd.
Edt., 1971, David McKay Company, Inc.

Cox, D.R. Planning of Experiments, 1958. John Wiley
and Sons, Inc.

Finn, J.D. Multivariance, A Fortran IV Program,
Version 4, June, 1968. Department of Educational
Psychology, State University of New York at Buffalo.

Little, T.M. and Hills, F.J. Statistical Methods in
Agricultural Research, 1972. University of California,
Agricultural Extension, AXT-377.

Steel, R.G.D. and Torrie, J.H. Principles and
Procedures of Statistics, 1960. McGraw—Hill Book
Company, Inc.

Wilson, P.R. and Sookpokakit, S. Multivariate Casebook:
A Guide for Use of Finn's Multivariance Program in
Processing Univariate and Multivariate Analysis of
Variance, Covariance, and Repeated Measures Designs,
Occasional Paper No. 31, August, 1978. Office of
Research Consultation, College of Education, Michigan
State University, East Lansing, Michigan.

"111111111111111711“