common mom; menus m
KEG-R KAEN, THAILAND 'ro DETERMINE
r5513 TESTING mus

A Dissertation
for “10 Degree of pk. D.
MICHIGAN STATE UNIVERSITY
Sor-Wasna Pravalpruk
1974

   
 

 
 

LIBRAR Y g
Michigan State
University

,'\\\ ‘
F"!- ‘ - ~ ‘f- . . 1 ‘1
_\ _ i , I . \. "- . .- \ ~ . . '
J 1a-

This is to certify that'the -—

  

n '1 ‘ thesis entitled

COMPARISON AMONG TEACHERS IN KI‘ION KAEN, THAILAND '
' TO DETERMINE THEIR TESTING NEEDS ‘

presented by

Sor-Wasna Pravalpruk

has been accepted towards fulfillment
of the requirements for

Ph. D. degree in Education

Major professor

Date W
/

0-7839

 

 

OVERDUE EINES ARE 25¢ PER DAY
PER ITEM

Return to book drop to remove
this checkout from your record.

 

V7

 

 

ABSTRACT

COMPARISON AMONG TEACHERS IN KHON KAEN, THAILAND

TO DETERMINE THEIR TESTING NEEDS

By

Sor-Wasna Pravalpruk

In-service programs in educational measurement, with emphasis
on test construction.procedures, i.e., test planning and curriculum
analysis, test item editing, item analysis, and reporting test scores,
have been given to teachers in various parts of Thailand by the Ministry
of Education and the Test Bureau at the College of Education, Prasarnmitr.
It was the purpose of this study to find out which group of teachers in
Khon Kaen province actually needs in-service training and in which areas
of measurement the need is greatest. This study also yields some
followbup information on the effects of the earlier in-service programs.
Two types of needs, actual and perceived needs, were investigated.

A Likert type questionnaire was constructed to measure the
perceived needs. The items were a random sample of the tasks usually
involved in the process of test construction. The respondent was asked
to rate each.task.from.1 to 5, where 1 meant most needed
and 5 not needed at all. The first edition of the questionnaire had
eight items for each of the areas of in-service training for a total of
thirty-two items. The actual needs were measured by a random sampling

of test items measuring knowledge on educational measurement similar

Sor-Wasna Pravalpruk

to those items used in introductory measurement courses at Michigan
State University. There were thirty-two items, eight each under the four
areas of the in—service program. Both questionnaire and test were tried
out and the reliabilities were found to be .87 and .41 respectively.

The revised editions of the questionnaire and test had twenty-
eight items and twenty items respectively. They were sent together with
the demographic data sheet and introduction letter to 400 teachers
who were randomly selected from eight strata in Khon Kaen province.

The stratification was based on the three variables: location of

school university area or out of university area; levels of teacher

education holders of teacher certificate and lower or holders of higher teacher
certificate holders; and teacher experience less experience or more

experience using eight years as the dividing line. Responses were

received from teachers forming an equal sample size for each of

the eight strata.

Responses were coded and data cards were produced. All
canned programs and FORTRAN programs were run on the CDC 6500 at
the Computer Center at Michigan State University.

The results of the descriptive information from the demo-
graphic data sheet indicated that the teacher-to-students ratio from
the sample was lower than the ratio from the population (1:31 and
1:33 respectively). The correlation between total actual needs and
total perceived needs was very low, i.e., .004. The correlation
between the number of teachers in the respondent's school with
perceived needs was low and negative at -.069 and with actual needs
at .160.

The test of homogeneity of the variance-covariance matrix

Sor-Wasna Pravalpruk

was not significant for the four subscales of the actual needs but

was significant for the perceived needs. The univariate repeated
measures analysis was applied to the actual needs and the multivariate
test to the perceived needs. The three way factorial design (2x2x2)
with four repeated measures were analyzed by the FINN program. The
design was crossed and balanced.

The analysis of the actual needs showed that the interaction
among location of schools, level of teacher education, teaching
experience, and the subject matter in measurement was significant.
Descriptive follow-up showed that the less experienced teachers
holding only the teacher certificate who worked within the university
area and the more experienced teachers holding the higher teacher
certificate who also worked within the university area had the most
contrasting profiles. The four way interaction was explained by the
significance of the location by teacher education by teacher experience
interaction on the contrast variable of Test Planning and Reporting
Test Scores. The three-way interaction was not significant when Test
Planning was analyzed individually, but the interaction was significant
for Reporting Test Scores alone. In the presence of the four-way
interaction, the subject matter by the level of teacher education
by experience interaction, the subject matter by experience interaction,
the subject matter by location interaction, the subject matter main
effect, the location by experience interaction and the teacher
education main effects were significant at a = .05. All interactions
were presented in graph.form.

0n perceived needs, only three null hypotheses were rejected.

The subject matter by location interaction was studied along with

Sor-Wasna Pravalpruk

the subject matter main effect. For teachers in the university area,
the more experienced teachers had perceived needs ranging from highest
to lowest in the following order: Item Analysis; Reporting Test

Scores; Test Planning; and Item Editing. But for those outside the
university area, the perceived needs on Item Editing was second,
followed by Reporting Test Scores and Test Planning. In general, the
Item Analysis had the highest perceived needs. The other null hypothesis
rejected was the location of schools by teaching experience interaction.
The interaction was disordinal which could be explained by the
observation that less experienced teachers within the university area
had lower perceived needs than the more experienced teachers within

the university area but less experienced teachers outside the university
area had higher perceived needs than more experienced teachers outside
the university area. The difference between perceived needs of all
(both inside and outside the university area) of the less experienced

teachers was smaller than all of the more eXperienced teachers.

COMPARISON AMONG TEACHERS IN KHON KAEN, THAILAND

TO DETERMINE THEIR TESTING NEEDS

By

Sor-Wasna Pravalpruk

A DISSERTATION

Submitted to

Michigan State University

in partial fulfillment of the requirements

for the degree of

DOCTOR OF PHILOSOPHY

Department of Counseling, Personnel Services, and Educational Psychology

1974

ACKNOWLEDGEMENTS

I wish to express my sincere appreciation to my Advisor and
Chairman of my study, Professor William.A. Mehrens for his assistance
and suggestions throughout my program of study and during the preparation
of this dissertation. To all the members of my doctoral committee, I
am grateful for their contribution of encouragement and insights. To
Dr. Archibald B. Shaw, who was kind enough to spend many hours in
correcting my English. To Drs. Maryellen McSweeney and Willard warrington
for their help in the development and improvement of the study.
Appreciation is also conveyed to Nitra Somswasdi, my sister,
for her help in distributing and collecting the questionnaire. To
Dr. Kowit Pravalpruk, my husband, for his help in analyzing the data.
His constant support and encouraging me to persist with my study were
very much appreciated.
I acknowledge with appreciation the support of the Thai
Government which allowed me to pursue my doctorate study. Special
thanks are also extended to Terry Nafisi-Movaghar who typed with
patience this dissertation.
Finally, for their encouraging, I wish to extend my gratitude

to my parents, my brothers and sisters, and to Puttaporn, my daughter.

ii

TABLE or CONTENTS

Chapter
I The Prom-m O O O O O O O O O O O O O O O I O O 0

Introduction .r. . . . . . . . . . . . . . .
Need for the Study . . . . . . . . . . . . .
Thai Educational System . . . . . . . . . . .
Purpose of the Study . . . . . . . . . . . .
Hypotheses . . . . . . . . . . . . . . . . .
Definition of Terms . . . . . . . . . . . . .
Overview . . . . . . . . . . . . . . . . . .
II Review of the Literature . . . . . . . . . . . .
The Importance of Classroom Testing . . . . .
The Qualifications of Teachers as Evaluators
In-Service Programs in Testing and
Evaluation . . . . . . . . . . . . . . . .
Summary . . . . . . . . . . . . . . . . . . .
III Procedure . . . . . . . . . . . . . . . . . . . .
Population . . . . . . . . . . . . . . . . .
Sampling Procedure . . . . . . . . . . . . .
Instrument . . . . . . . . . . . . . . . . .
Data Collection . . . . . . . . . . . . . . .
Design . . . . . . . . . . . . . . . . . . .
Analysis . . . . . . . . . . . . . . . . . .

Summary . . . . . . . . . . . . . . . . . . .

iii

10
11
12

15

18
24
27
27
28
30
34
35
38

39

Chapter
IV Analyses and Results . . . . . . . . . . . . . .
Descriptive Data . . . . . . . . . . . . . .
Homogeneity of Variance-Covariance Matrices
Repeated Measure Analysis on Actual Needs .
Repeated Measure Analysis on Perceived Needs
Additional Analysis . . . . . . . . . . . .
Summary . . . . . . . . . . . . . . . . . .
V Summary and Conclusions . . . . . . . . . . . .
Summary and Discussion . . . . . . . . . . .
Conclusions and Implications . . . . . . . .
Recommendations for Further Study . . . . .
Appendix A The Instrument (First Edition) . . . . . . .
Appendix B The Instrument (Final Edition) . . . . . . .
Appendix C Pairwise Profile Analysis on Actual Needs .
Appendix D Multivariate Analysis of Variance

on Four Subscales on Actual Needs

Bibliography . . . . . . . . . . . . . . .

iv

. 46

. 79

.103

.124

.126

.128

Table

3.1

3.2

4.1

4.2

4.3

4.4

4.5

4.6

4.7

LIST OF TABLES

Total Number of Teachers in Khon Kaen Classified

by Location of School, Class of Teacher

Certificate and Years of Teaching Experience . .

Design of the Three Independent Variables:

Location of Schools, Levels of Teacher Education,

and Teaching Experience . . . . . . . . . .
Averages and Standard Deviations of the Number
Teachers and Number of Students in Sampled

Schools Classified According to

Eight Groups of Teachers . . . . . . . . . .
Inter-Correlations among Number of Teachers,

Number of Students, Perceived Needs,

and Actual Needs . . . . . . . . . . . . . .
Variance-Covariance and Correlation Matrices

of Four Test Scores . . . . . . . . . . . .
Variance—Covariance and Correlation Matrices

of Four Perceived Needs Scores . . . . . .
Means and Standard Deviations of Four Test

Scores (Actual Needs) . . . . . . . . . . .
Univariate Repeated Measured Analysis on

Test Scores . . . . . . . . . . . . . . . .
Means and Standard Deviations of

Four Perceived Needs Scores . . . . . . . .

of

29

36

42

43

47

48

49

51

61

Table

4.8

4.9

4.10

4.11

4.12

Univariate and Multivariate Repeated

‘Measured Analysis on Perceived Needs . . . . . .
Comparison on Perceived and Actual Needs

between Teachers Who Attended and Did Not Attend

the In-Service Program . . . . . . . . . . . . .
Comparison on Perceived and Actual Needs

between Teachers Who were Favorable and

Not Favorable to the National Testing Policy . .
Comparison on Perceived and Actual Needs

between Male and Female Teachers . . . . . . . .
Comparison on Perceived and Actual Needs

among Teachers Who Taught Different Grade Levels

vi

63

68

69

7O

71

Figure

II

III

IV

VI

LIST OF FIGURES

Graph Presentation of Cell Means for Subject Matter,
Teacher Education, and Experience (Actual Needs) .

Graph Presentation of Cell Means for Subject Matter
and Teaching Experience (Actual Needs) . . . . . .

Graph Presentation of Cell Means for Subject Matter
and Location of Schools (Actual Needs) . . . . . .

Graph Presentation of Cell Means for Location of
Schools and Teaching Experience (Actual Needs) . .

Graph Presentation of Cell Means for Location of
Schools and Subject Matter (Perceived Needs) . . .

Graph Presentation of Cell Means for Location of
Schools and Experience of Teachers

(Perceived Needs) . . . . . . . . . . . . . . . .

vii

55

57

58

59

64

65

CHAPTER I

THE PROBLEM

Introduction

 

Man has always been concerned with measurement and evaluation.
From the earliest beginnings of society, men have measured the abilities
of other men and have recognized the existence of differences in the
abilities possessed by different individuals. As human society has
grown in size and complexity the recognition of individual differences
has increasingly been reflected in the structure of societies.

In school, the teacher employs a central role in the testing
and evaluation process. The teacher must make judgments, measure,
appraise and report. To do this he must use tests, examinations,
inventories and other types of measuring instruments and devices. He
must know hOW’tO construct such instruments for his own purposes, and
to use and interpret the results of others. He must know something
about the analysis and interpretation of scores on tests.

Classroom testing not only provides an expedient method for
grading students, it also gives the student a far more valid indication
of what the teacher considers important than do statements made in the
textbook or by the teacher. Because it provides the basis for grades,
the test indicates unmistakably to the student what should be learned or
what is expected. Tests can also provide the teacher with needed
information on precisely what students are learning and what they are

not learning.

Need for the Study

Tests have become an integral part of our academic life.
Teachers evaluate students from the time they are in kindergarten
until they leave high school or college. Although decisions are often
made on the basis of a student's scores, little, if any, thought is
given to the qualifications or skill of the teacher as an evaluator.

It is essential that we recognize the value that test results may
play in the life of the student.

Teachers occupy a central role in the testing and evaluation
process. Teacher-made tests are frequently the major basis for evaluating
the progress of students. They may be better than good commercially
prepared achievement tests, in the sense that they may be more
relevant to a teacher's particular objectives and pupils. Unfortunately,
many teacher—made tests do not serve their function optimally because
very often teachers have not been given adequate preparation in this
area of competence. Two surveys done about 20 years ago (Noll, 1955;
Allen, 1956) were in substantial agreement in showing that an intro-
ductory course in measurement is not generally required by State
Departments of Education for a teaching certificate. Most institutions
preparing teachers offer an introductory course in educational
measurement but comparatively few require it for a recommendation for
a teaching certificate. More recent studies done by Goslin (1967) and
Roeder (1972) lead to the same conclusion, that the problem of preparing
teachers in measurement and evaluation is real, clear, and substantial.
Many teachers do not seem to be adequately prepared in this respect,
and apparently they recognize their inadequacy.

If we accept testing and the use of test results as a part

of the teachers' role, efforts need to be made to make sure that this
part of their role is not carried out in a context of naivety about
tests and what they measure. The role of teachers in testing is too
important to be left to chance.

In many systems, in-service courses are offered to teachers
from time to time. A varying number of sessions is held according
to the other demands being made on teachers. This type of program
might be useful to help school personnel become more familiar with

the theory and principles of measurement and evaluation.

Thai Educational System

The educational system in Thailand is quite different from
the United States. Historically, the Thai educational system has had
a strong connection with Buddhism, the main religion in Thailand.
The wet (Buddhist temple) was the only place available for youngsters
to be educated formally and the Buddhist monks were the only teachers.
The subject matter was Buddhist philosOphy, history, and theology.
Vocational education was either taught by a skillful person in that
particular job or it was passed along within the family. Only a
comparatively few youngsters and only boys, were given an opportunity
to be educated in schools. Education for most was just on-the-job
training or part of the living in the family occupation. Most of
the boys learned to do as their fathers did, and the girls learned
only the skills of house-work and home crafts. Therefore, testing
for most was in the form of tests of performance and evaluation was
just "it is the time for a birdie to fly alone."

In 1871 the first school was established in the old palace

in Bangkok, the capital city of Thailand. King Rama V (Chulalongkorn)
supported the school, and formal education became the royal family

and royal cousins' affair. Going to school was aimed towards becoming
a royal officer in one of the civil services. Learning was to read,
to write, and to have the social graces needed to get along in the
King's services. Schooling proved to be such a success and so useful
to the development of the kingdom that the King proclaimed that schools
be establishedthroughout the country within a few years. By 1887
there was at least one school in each province. The curriculum

was basically the three R's. A British-type school was implemented.
.From the beginning and down to today, youngsters go to school three
quarters of the year. There is no school during the summer. There
are tests at the end of each quarter but the final examination at

the end of the school year is the basis for promotion to the next
grade. Early examinations were either in the form of class recital
and individual recital or in essay form when the recital examination
was too time consuming. When the Ministry of Education was founded

in 1892, a nation-wide system of examinations was put into practice.
For a long time, all tests were in essay form.

In 1954 the objective test item was introduced. The
initial effort was credited to M.L. Tuy Chumsai. With very little
knowledge and understanding of the principles involved, teachers
began to use objective tests. The lack of knowledge caused the
production of bad objective test items.‘ The first textbook on
educational measurement in the Thai language was written by Chaval
Paeratakul and published in 1963. Two years later, a nation-wide

inservice program in educational measurement was pr0posed and carried

out by Chaval Paeratakul and the staff of the Test Bureau in the
College of Education at Prasarnmit in an attempt to improve the
quality of the teacherdmade tests. During this period, the Test
Bureau constructed four standardized tests for grades 4 and 7, two
in the subject Thai Language and two in Arithmetic. The tests and
their norms were published in 1971.

In 1965 in-service workshops were scheduled for 5 days,
and covered these tapics: purposes of measurement; curriculum analysis
as a blueprint for test construction; types of test items; item analysis;
scores and norms; and reporting the test result to parents and
authorities. The morning sessions included lectures by specialists
from the College of Education and the Department of Teacher Training.
The afternoon sessions were practicums. Teachers were assigned to
one of the small groups working on bluprints, writing test items,
performing item analysis, and reporting standard scores. In the
past five years, the training has emphasized test construction and
writing test items according to Bloom's Taxonomy.

The quality of the test is critical because the evaluation
that follows can slow down a student's progress through school by
one year. If a student fails the final examination at the end of
the school year he has to be in the same grade for another year with
some unfavorable attitude and the label "repeater." Since education
in Thailand is in effect a selective system, less than 7 percent of
the first graders will finish early secondary school (grade 10).
Entrance examinations are used to select students who want to go on
to upper secondary school (grades 11 and 12). Another entrance

examination at the end of grade 12 is used for admission to

universities. Good tests are needed in the selection process so that
educational wastage is minimum,

In school, a student is mostly recognized and rewarded by
teachers, parents, friends, and community on his cognitive abilities.
Those cognitive abilities are measured by teacher-made tests. Again
the quality of the test is crucial. Tremendous pressure is on a
student whenever he is going to take a test. Sometimes it means his
entire future depends upon one or two hours of his performance on
the test. What will happen if one is not feeling well when the time
comes to take the test?

In 1962 the Department of Teacher Training implemented a
new curriculum in both two and four year programs of teacher education.
The new curriculum added one course in educational measurement and
evaluation in the four year program. This course covered how to
construct an objective test from planning to interpretation of the
test score. It was 1966 when the first graduates from the new
curriculum were assigned to schools. It was hoped that these new
teachers would help other teachers in the same school learn how to
construct objective tests. However seniority is a very real barrier
in communication between old teachers and new teachers. If the
seniority barrier assumption is true, there will be a difference in
the new teachers' needs and the older teachers' needs for an in-
service program in testing. It is interesting to investigate
whether the older teachers need more testing services.

Because of the centralized system of education in Thailand,

information and policy are diffused from the Ministry of Education to

the Provincial educational officers and then to the teachers.

Within one province the central amphur will receive the information at
first-hand. Other amphurs in the province will get them later
depending on how far they are from the central amphur. In Khon Kaen,
there is a university which is located in the central amphur. These
two facts lead to the conclusion that teachers in the central amphur
should have had more help and can have acquired knowledge about
educational measurement easier than those in other amphurs.
Therefore, they should have less need for in-service courses than
teachers who are not in the university area.

The level of the teachers' certificate furnishes another
way to classify teacher groups. The holders of higher certificates
took more courses in their four-year program and should have more
knowledge about educational measurement. Teachers who hold the
certificate gained in a two-year program, it is hypothesized, may
have more need for help than do the holders of higher teacher
certificates.

As a result of this study, one should be able to conclude
a) which group of teachers in Khon Kaen Province has the most urgent
need, and b) what area of service is most needed. It is hypothesized
that teachers who have Teacher Certificates or lower, who are not
in the university area and who are older teachers (teach more than
8 years), have the highest need, and that writing the objective items

will be the most needed area of service.

Purpose of the Stud -
The purpose of this study is to identify the areas of

instruction in measurement which are needed most by the teachers in

the Thailand province of Khon Kaen, and also to compare teachers

who are grouped according to the amount of their experience in teaching,
to the location of their schools, and to the kind of teaching certificate
they hold. Actual and perceived needs in four fundamental areas will

be measured: planning test construction or curriculum analysis,

test item editing, item analysis, and reporting test scores.

Hypotheses

The major hypotheses to be examined are as follows:

1. Teachers who are in the university area may have

fewer actual and perceived needs than teachers who are

farther from the university in Khon Kaen Province.

2. Teachers who hold higher teacher certificates may

have less actual and perceived needs than teachers

who hold the teacher certificate in Khon Kaen Province.

3. Older teachers (more teaching experience) are

expected to have more actual and perceived needs than

newer teachers in Khon Kaen Province.

4. No difference is expected in actual and perceived

needs among the four subject matters covered in the study.

Besides these four hypotheses about main effects, all
interactions will be tested. These interaction effects are determined
in two parts: the interactions between the three variables of
location, level of teacher education and level of experience; and
the interactions between the three variables and the repeated
measure variable, the four subject matter areas covered in this

study.

Definition of Terms

 

Actual need is defined as a lack of knowiedge in a sub-

 

area of tests and measurement as is disclosed by responses to groups
of related test items that are designed to test desirable knowledge
of measurement and evaluation.

Perceived need is defined as the feeling of lacking

 

desirable knowledge in tests and measurement as measured by a
Likert Type format questionnaire about tasks in measurement and
evaluation.

Amphur is a political subdivision of a province. There
are 13 amphurs in Khon Kaen Province.

University Area is the Central Amphur of Khon Kaen
Province. The university is Khon Raen University.

Far From University Area describes all amphurs in Khon
Kaen Province except the Central Amphur. It is also called
outside university area.

Certificate teacher is a teacher who graduated from a

 

two-year teacher training curriculum, or its equivalent.

Higher certificate teacher is a teacher who graduated

 

from a four-year teacher training curriculum, or its equivalent.

New teacher is a teacher who has been teaching for

 

less than eight academic years.

Old teacher is a teacher who has been teaching for eight

 

or more academic years.

10

Overview
The study is organized into five chapters, followed by
appendices.

In Chapter II, the literature relevant to the general
problem and related areas is reviewed. The design of the study,
the pilot study, the instrumentation, and the methods of analysis
are discussed in Chapter III. The research data and results of the
study are presented in Chapter IV. The final chapter contains a
summary of the study, the conclusions, a discussion of the findings

and implications for future research.

CHAPTER II

REVIEW OF THE LITERATURE

The intent of this chapter is to review studies that are
related to the problem as described in Chapter I. The review is
divided into three sections: the importance of classroom testing;
the qualifications of teachers as evaluators; and in—service programs

in testing and evaluation.

The Importance of Classroom Testing

 

It is generally agreed that evaluation of pupil progress is
a major aspect of the teacher's job. A good picture of where the
pupil is and of how he is progressing is fundamental both to
effective teaching and to effective learning.

Classroom testing may be used in the form of quizzes,
midterm and final examinations, and as general guideposts for
teaching. Tests may be given at various times during the school
year in order to stimulate learning through review and to determine
status and progress in learning. As Mehrens and Lehmann (1973, p. 167)
state: "Teacher-made tests are frequently the major basis for
evaluating the students' progress in school."

Teacher-made tests, by their definition, are constructed by
a classroom teacher for use in his particular classes under conditions
of his choosing. Classroom achievement tests, if properly prepared,
are tailor—made for a certain class of pupils taught in a certain

manner. They can be related to the kind of educational objectives

ll

12

involved in the class and also to the relative emphasis each received.
They also can be shaped to fit the particular characteristics of the
pupils who compose the class. The teacher is in the best position
to know the exact nature of the characteristics of his class. Further,
since the apprOpriateness of any achievement test item is determined
by the educational objectives of the class, the teacher is in an
excellent position to build test items that measure suitable areas of
the pupils' achievements and that are of suitable difficulty.

Commercially prepared achievement tests may be used to obtain
some of the information needed by teachers, and they may even be used
to motivate students. But it is unusual for standardized tests to be
administered more than once a year. Teacher-made tests are better in
the sense that they are more relevant to a teacher's particular objectives
and to his class. As Mehrens and Lehmann (1973, p. 202) say, "Classroom
tests, despite some of their limitations, will never be replaced because
they (a) tend to be more relevant, (b) can be tailored to fit a teacher's
particular instructional objectives, and (c) can be adapted better to
fit the needs and abilities of the students than can commercially
published tests."

Also Ebel (1962, p. 24) states:

Even if good ready-made tests were generally

available, a case could still be made for teacher-

prepared tests; the chief reason being that the

process of test development can help the teacher

define his objectives. This process can result

in tests that are more highly relevant than any

external tests are likely to be. It can make the

process of measuring educational achievement an

integral part of the whole process of instruction

as it should be.

Schwartz and Tiedeman (1957, p. 110) say that teacher-made

tests are valuable because they

l3

1. Assess the extent and degree of student
progress with.reference to specific classroom
activities.
2. Motivate students.
3. Permit the teacher to ascertain individual
strengths and weaknesses while the pupil is studying
a particular subject—matter area.
4. Provide information for reporting purposes.
5. Provide immediate feedback for the teacher,
who can then decide whether a particular unit or
concept needs reteaching and/or whether individual
pupils are in need for remedial instruction.
6. Provide for continuous evaluation.
A good classroom test must be valid, reliable, and objective.
It must possess good discriminative power, and it must be characterized
by questions with a range of difficulty suitable to the abilities of
all the students in the class. Bell (1972, p. 8) says: "A teacher
who gives tests and uses them in the classroom without ever knowing
whether and to what degree such qualities are present is flying blind."
Teacher-constructed tests should be designed to measure
areas of strengths and weaknesses, and thus enable the teacher to
plan for a better program of organization, preparation, or presenta-
tion, for the purpose of reteaching. Furthermore, Thorndike and
Hagen (1969, pp. 30-33) suggest that functions of teacher-made tests are:
a) Motivation.
"Tests that are well constructed and effectively
used can motivate students to develop good study
habits, to correct errors, and to direct their activities
toward the achievement of desired goals." (p. 31)
b) Diagnosis and instruction.
"The items on which an individual fails or on

which many members of a class group fail can serve
to identify points needing further study." (p. 31)

14

c) Defining teaching objectives.

"We may know a teacher by the tests he makes.

They tell what he is truly valuing in his pupils,

even tbough.he himself does not know it, and they

influence profoundly what his students will learn." (p. 32)

d) Differentiation and certification of pupils.

"The teacher inevitably has a responsibility for
certifying pupils' accomplishments to higher levels

of the educational enterprise and to the world outside

the school." (p. 33)

The authorities seem to agree that in constructing a test,
the teacher must be concerned with the same factors as are the author
and publisher of a standardized achievement test-validity, reliability,
and objectivity.

Evaluation of pupil achievement, in their view, is one of
the teacher's important responsibilities. It is important that the
teacher's device be well thought out and well made. For any type of
written test, it is desirable to have a definite plan in advance of
preparing the test items. The development of such a plan requires
an analysis of the outcomes one is trying to achieve in the teaching
of a particular course or unit and of the significant segments of
content through which those objectives are to be realized. In
addition, the plan should include the allocation of test items among
the content areas and objectives, the types of items to be used, the
total numbers of items in the test, and specifications for the spread
of item difficulties.

In Thailand, more emphasis is on the writteniexaminations in which
teachers have to depend on the-tests they have made individually.

Evaluation at the end of the school year is based almost entirely

on the teacher-made tests (Paeratakul, 1963, in Thai edition). Only

15

the performance during the five days of the last week of the school
year is the basis for determining which.students will go to the next
grade or repeat the same grade in the next school year. So there is
much more stress on teacher-made tests in Thailand than in the

United States.

The Qualifications of Teachers As Evaluators

Teachers occupy a central role in the testing and evaluation
process. Teacher-made tests are frequently the major basis for
evaluating the progress of students. Many teacher-made tests do not
serve their function optimally because of time pressures on teachers
and, more important, because test-writing is not generally a part of
the professional training of teachers (Grobman, 1971).

In 1955 Victor Noll reported a study of requirements in
measurement by state departments of education for teaching certificates
and by 80 institutions preparing teachers in all parts of the country.
The sampling included large publicly supported institutions, large
privately supported institutions, teachers colleges, and liberal
art colleges. A similar survey was reported in 1956 by Allen. The
results in her study were based on questionnaire returns from 288
higher institutions with programs;for the preparation of teachers, an
88 percent return. The findings of the two surveys are in
substantial agreement. They showed that an introductory course in
measurement was not generally required by state departments of
education for a teaching certificate. Most institutions preparing
teachers, they found, offer an introductory course in educational

‘measurement but comparatively few require it for a recommendation for

16

a teaching certificate.

Another study done under the Russell Sage Foundation in
1967 by Goslin states that

...with respect to formal training in test and

measurement techniques (courses taken in college,

attendance at clinics on standardized testing, and

the like), the data indicate that less than 40 percent

of all teachers have had more than minimal exposure

(one course) to such training,Jand that a sizable

prOportion of teachers have never had a course in

measurement techniques or attended a clinic at which
testing was discussed. Data also indicate that

teacher involvement in testing is both high and of

potentially great influence on the educational system

and the children within it. If we accept testing

and the use of test results as a part of the teacher

role, efforts need to be made to make sure that

this part of their role is not carried out by

chance. (1967, p. 127)

A more recent study was conducted by Roeder (1972) who
surveyed the qualifications or skills of the teacher as an evaluator.
The 940 elementary teacher-training institutions located in every
state and the District of Columbia, were mailed a one-page questionnaire.
A total of 916 responses (97.4 percent) were received. The data
indicate that 57.7 percent of the institutions which were surveyed, or
496 institutions, did not require their prospective elementary
teachers to complete a course in evaluation; 12.1 percent (104)
required nothing more than a one or two semester hour course; 17.8
percent (158) required a three semester hour course and only 1.4
percent (12) required four or more semester hours or course work
in evaluation. Roeder concludes that the majority of teachers receive

only a minimal exposure to the complex world of evaluation. That

is, most of today's elementary teachers are not prepared to make

or use tests 0

17

Grobman (1967, 1971) sees testing as an integral part
of teaching. She found that many teachers testified that the
improvement of their test building skills has resulted in the improve-
ment of their teaching, since it has made them.more sensitive to their
teaching aims and the implementation of these. She claims that few
teachers have had an opportunity to acquire a sufficient background
in evaluation and testing to enable them to measure adequately the
skills and understanding they wish to encourage in their students.
Ebel (1961, p. 68) states that a teacher who is competent
in educational measurement should:

(1) Know the educational uses as well as the limitation,
of educational tests.

(2) Know the criteria by which the quality of a test
should be judged and how to secure evidence relating
to these criteria.

(3) Know how to plan a test and write the test questions
to be included in it.

(4) Know how to select a standardized test that will
be effective in a particular situation.

(5) Know how to administer a test properly, efficiently,
and fairly.

(6) Know how to interpret test scores correctly and
fully, but with recognition of their limitations.

Further, in the 62nd Yearbook of the NationaljSocietygfor
the Study of Education (1963) Hagen and Lindberg express a belief that
classroom teachers: (1) must have a basic understanding of the
evaluative process, including an understanding of the basic concepts
of reliability, standard error of measurement, and validity as essential;
(2) should know the general types of tests available and be aware of

their uses, strengths, and limitations; and (3) should be able to

18

COmbine available test data with other records and to interpret them
as they relate to individual children in their classes.

Chaval Paeratakul (1963), the first person from Thailand who
graduated in Tests and Measurements in the United States, along with
other professionals in this area agreed that there was an urgent need
for teachers in Thailand in tests and measurements. Other evidence to
support this idea was the change in the curricula of teacher training
education programs in 1962 and the nationdwide in—service program

from 1965—1966. Teachers in Thailand have just begun to know what
the objective tests are all about. They definitely need more help

in getting acquainted with these subject matters.

 

In-Service Prggrams in Testing_and Evaluation

From studies reviewed in the last section, it is clear that
the problem of competence of teachers in educational measurement and
evaluation is real and substantial. They lead to the conclusion that
many teachers do not seem to be adequately prepared in this respect
and apparently they recognize their inadequacy.

It is of the utmost importance to educational progress that
the competence of teachers to measure educational attainments be
improved. Ebel (1961) says that far more harm, perhaps ten times as
much harm, is currently being done to student learning as a result
of the shortcomings of the classroom tests by which a student's
educational efforts are largely stimulated, directed, and evaluated,

than is being done by all the faults of external testing programs.

“'He suggests in-service programs as a means for improving the competence

of teachers in measurement on the basis of experience with a variety

19

of in-service programs intended to help teachers solve measurement
problems. These programs, variously referred to as conferences, seminars,
institutes, or workshops, have ranged from an afternoon lecture to

a three-day preschool program with several followbup meetings later

in the year. Some of these programs were sponsored by a single school
system and involved the teachers of that school in all subject areas

and at all levels. Others reflected the interest of a single professional
group, such as engineers or nurses.

Goslin (1967, p. 129) suggests the publication of a short
booklet‘ designed explicitly to deal with the problems faced by teachers
in their role as testers as a supplement to a "more formal training
in testing" to prospective teachers.

Hagen and Lindberg (1963) claim that each school system
must take responsibility for providing orientation and in-service
training in testing because it is unrealistic to assume that teachers
can learn all that they need to know about testing during their pre-
service courses.

There was a panel discussion about the improvement of the
competence of classroom teachers in measurement and evaluation held
by the National Council on Measurements Used in Education, at the 1959
annual meeting. The discussion centered on the questions of why in-
service training is a problem, and what can be done to improve in-
service training at the local level.

As an example one local in-service program was presented

in specific steps:

20

1) Preparation of local bulletins. A series of
bulletins, supplementing the published materials
concerning the tests in use in the county, have
been written tying infithe testing program with
the local program of instruction.

2) Use of school test coordinators. These test
coordinators met regularly, especially before and
after a scheduled testing program, to discuss
problems involved in administering, scoring, and
interpreting the test results.

3) Extension courses.

4) Faculty workshops. A considerable number
of faculty workshops, varying in length from

one to four or five sessions, have been held.

5) Development of community interest in the
measurement program, through a judicious use

of local newspaper publicity, talks to
parent-teacher associations.

6) Use of local norms.

7) Provision of adequate physical facilities.

8) Use of demonstrations, lectures, etc., in
the district's In-Service Training Center.

9) TV Workshop on testing.
(Durost, pp. 32-33)

Rather than the in-service program, McSwain (1959) focused
on the role of the college faculty in the teacher education program.
He said that during the period of student teaching each student-
teacher should learn, from teacher and faculty advisor, about these
areas: (1) a functional concept of evaluation and measurement in
creative teaching and purposeful self-education; (2) techniques used
in evaluating the behavioral responses of pupils; and (3) criteria
to be applied in selecting and using educational measures to improve
teacher diagnosis and effective procedures in preparing and

interpreting cumulative child-accounting records. And, he added,

21

tzhe degree of competence in applied measurement developed by each student
during this professional internship depends basically upon the
leadership of the faculty counselor.

In the same panel discussion, Charles Stevenson (1959)
talked about measurement in the teaching-learning situation. He
stated that the in-service program should cover: (1) common under-
standing of what is to be measured among teachers, supervisors,
coordinators, and superintendents; (2) the type of evaluative
instruments which will be adequate in a particular situation;
(3) clarification of how the results of measurement can be used to
improve the teaching-learning situation; and (4) methods and procedures
to be employed in improving these techniques of measurement.

The last presentation in this panel, by Margaret Stevenson,
was a discussion of the role of the classroom teacher in school
testing programs. She suggested that after the school system has
formulated its philosophy on testing; after representatives of all
levels of school personnel have c00perated in developing the program;-
after the testing program has been given a definite place in the school
curriculum--then the major hurdles have been passed, and the school
system is ready to address itself to the more direct problem of
improving.classroom teacher competence in measurement and evaluation.
That is, improving classroom teacher competence in measurement and
evaluation must not be treated as an isolated question. It must be
viewed in context, against the background of the numerous problems
involved in establishing an adequate, democratically structured
testing program in the school curriculum.

The importance of in-service education for all educational

22

personnel is recognized throughout the literature of the teaching
profession (Ebel, Ed., 1969, p. 645). The rapid expansion of knowledge,
which has been reported extensively over the past several years, and
its effect on changing methods and in develOping technology utilized
in the classroom are major factors in making programs of in-service
education necessary.

A statewide research study in Tennessee about in-service
education was conducted by Brimm and Tollett (1974). The conclusions
of the study can be summarized as the following:

(1) The primary purpose of in-service education
is to upgrade the teacher's classroom performance.

(2) If continuing professional growth is to be

taken seriously, administrators and teachers must

pool their knowledge and resources and seek to

make in-service programs more responsive to the

needs and interests of practicing classroom teachers.

The two chief weaknesses of the in-service program as
Ebel (1961) sees them are: (1) they are too brief. While an hour
or two a year spent in considering measurement problems under the
guidance of a specialist is far better than nothing at all, it is
unreasonable to suppose that satisfying, enduring progress in solving
the manifold problems of educational measurement, or in deve10ping
the requisite knowledge, understanding, and skills, can be made in
so short time; and (2) they involve too much talking and too little
doing. For the cultivation of a practical art like educational
measurement, sound pedagogy requires a mingling of theory and practice.

Ebel presents the ideal program of in-service training for improving

the competence of teachers in measurement as the following:

23

Suppose that a school administrator and his staff
have decided to focus attention for a year on the
improvement of classroom testing. Suppose they engage
a specialist in educational testing to meet with them
five times during the year, at intervals of six weeks or
so, for a day or two. Participation in the initial
program might well be limited to five, six, or seven
groups of four to six teachers each.

The goal of each group would be to make, to use
and to analyze a quality test in a subject which all
members of the particular group were teaching. Examples
of subject area in which these tests might be developed
are: fourth grade mathematics; sixth grade geography;
eighth grade English; or high school history; chemistry;
or economics.

The first meeting of each participating group
would be devoted to a description of the entire project,
with special consideration of the first step - the
preparation of specifications for the test to be
developed. Sample specifications would be presented
for study and analysis. Between the first and second
sessions each teacher group would work out the
specifications for its test. These could be reviewed
at the second meeting, and work on item writing would
be launched. The third meeting could be devoted to
item review and test assembly, the fourth to test
administration and analysis, and the fifth to a review
of the test developed and of the entire project as a
learning experience. (Ebel, 1961, pp. 70-71)

A program like this, Ebel says, would produce not only a
handful of excellent tests but also a sizable group of teachers whose
competence in measurement was vastly improved and highly respectable.

One of the most effective in-service programs in educational
measurement in Thailand was carried out by Chaval Paeratakul in 1965-66.

v//The purpose of the program was to improve the quality of the teacher-
made tests. It was a five-day workshop. The material covered.
included: purposes of measurement; curriculum analysis as a blue-
print for test construction; types of test items; item analysis;
scores and norms; and reporting the test result to parents and

authorities. The morning sessions were lectures by authorities.

24

The afternoon sessions were practicums. Teachers were assigned to

one of the small groups working on blueprints, writing test items,
performing item analysis, and reporting standard scores. The training
had emphasized test construction and writing test items according to

Bloom's Taxonomy.

Summary,

In most of the literature, educational measurement and
evaluation are viewed as essential and constructive (Ebel, 1972;
Schultz, 1971; Sharp, 1966; Jessel, 1972; and Mehrens, 1967). 'Occasionally
some suggest that education could go on perfectly well, perhaps much
better than it has in the past, if tests and testing were abolished.
This may imply that the tests these critics have seen or have used
themselves are not good. No doubt some bad tests have actually been
worse than no tests at all.

Test scores are powerful educational tools, but they are
only tools. They will not give direct, complete answers to the
practical educational questions. They will not point unequivocally to
a specific course of action in a given set of circumstances. They
may indicate that something needs to be done. They may provide data
that will help in deciding what to do. But they will not make the
decision.

Several writers (Mehrens, 1967; Trump, 1967; Holt, 1971;
McMorris et a1, 1972; and Dyer, 1962) suggest that the limitations
and underlying principles of tests must be recognized if one wants to
use tests as good measuring tools in education.

The qualifications of teachers as evaluators are widely

25

reported as less than satisfactory for most teachers. Studies on
competence of teachers in measurement and evaluation conclude that

most teachers do not have adequate knowledge in this area (Goodman, 1971;
Noll, 1955; Allen, 1956; Goslin, 1967; and Roeder, 1972). These

studies also suggest that more publication and service programs need

to be Operated to improve the situation.

The authorities conclude that in-service or workshop programs
are needed to improve the competence of teachers in measurement. The
purpose of the program is primarily to help teachers solve measurement
problems. Much time, effort, and money needs to be spent to initiate
or improve programs of services and assistance to teachers. The justifi-
cation is that this is an important method of raising the standards
of education.

These help programs may be as little as a short booklet
(Goslin, 1967), or they may be conferences, seminars, or workshops
.which range from one afternoon to a five-session workshop (Huddleston,
1959; Ebel, 1961). The contents to be covered are the basic knowledge
in educational measurement and evaluation that teachers should know,
such as the limitations of educational tests, the properties of tests,
ways of planning and constructing tests to use for one's own purposes,
test administering, and how to interpret test results.

Teacher~made tests in Thailand are much more important to
effective teaching and to effective learning than in the United States
since standardized tests are not available yet. And, also, more
stress is put on the final examination than in the United States.

The teacher-made final examination has at least seventy-five

percent weight on evaluating "pass" or "fail" for the

26

students. The qualifications of Thai teachers as evaluators are much
worse than in the United States. Also, the need for improving the

competence of teachers in Thailand is very much greater than in the

United States.

CHAPTER III

PROCEDURE

This study can be classified as a comparative study.
Questionnaires, sent through the mail or delivered personally, were
used to collect the data. This chapter provides a description of the
population, sampling procedure, instrument, data collection, and plan
for data analysis. Findings in the study are presented in Chapter IV

“and conclusions are given in Chapter V.

Population

Georgraphically, Khon Kaen province is in the northeastern
part of Thailand, about 200 miles from Bangkok by the "Friendship"
(Mitraparp) highway. In area, the province is 14,076 square-
kilometers. The province is divided into thirteen amphurs and these
thirteen amphurs have 116 tambons (subdivisions of amphur). The
population is 1,136,939 people, the majority of whom are farmers.

Educationally, there are 1047 schools, one technical
institute, and one university. 0f the 1047 schools, thirty-five offer
a high.school program either alone or in combination with the elementary
and secondary levels in the same school. Forty-three of 1047 schools
are schools for adults, which have both elementary and secondary
levels.

The population of interest is teachers in these 1047
schools. There are 5,913 teachers and 198,246 students. The average

ratio of teacher to student is 1:33. The breakdown of teachers into

27

28

eight groups is presented in Table 3.1. The largest group (1534) is
composed of teachers who are not in the central amphur, who have been
teaching more than eight years and who hold the teacher certificate
or lower. It is estimated that more than ninety percent of these
teachers are in elementary schools or teach in elementary grades. Most
of the big schools (more than 1000 students) are in the university
area, so the number of teachers in the central amphur is the largest
among the thirteen amphurs. The smallest group (210) is composed of
teachers who are in the university area, who hold at least the higher
teacher certificate, and who have been teaching more than eight years.
Of the 5,913 teachers, sixteen hold master's degrees and
349 have at least the baccalaureate degree. Six percent of all teachers
(365) hold at least baccalaureate degrees and most of them are teaching
in the university area. Eighty-nine percent of the teachers (5249)
teach at the elementary level (grades 1 to 7) and seventy-one
percent (4212) in lower elementary level (grades 1 to 4). 0f the
totals, 844 teachers are in fifty-eight private schools, mostly in
the university area. All others are in public schools, either in
provincial municipal sections or under the Departments of Elementary

and Secondary Education.

Sampling;grocedure

A list of all the teachers in Khon Kaen province was
provided by the‘Ministry of Education. Along with the names, the
location of schools, ages of teachers, years of teaching experience,
and levels of teacher education were given. The teachers were

classified into eight groups, according to location of schools,

29

 

 

 

 

 

 

 

 

2am , an: :3 3L
sou
once new web «mna MQNH muumuo>wn
on» ma uo
. mop
33 SN man as a: 5335a
was sH
ooaofiuomxm ouooauonnm A ounmauunww . moomwuumwu
one: uo..muw w .ouw w none moon one: so .mu» w .muw a done soon a
Hmuos oumoumwuuou umnwwm . uoaog no oumoauauuou unnomoa soaumoog

 

 

 

 

muzmHMmmxu qumode mo mudmw 92¢
MH<UHMHHMMU MHMU<HH mo mm¢AU .Aoomom mo onH<UOA Mm QmHhHmm<Au

zmdm 20mm 2H mmumu<ua ho MHQZDZ A<HOH

H.m mqde

 

30

years of teaching, and levels of teacher education. A tentative number was
assigned to every teacher for the purpose of random selection. A table

of random numbers was used. A systematic random sampling procedure was
used to get a sample of 40 teachers in each of the 8 groups for the

total of 320 teachers.

Because of the difficulty of transportation and uncertainty of
the mail, it was expected that the percentage of return would be lower for
the teachers who are outside than those inside the university area. There-
fore, 80 more teachers were randomly selected, 20 from each of the 4 groups
outside the university area. The first 40 returns of each of these 4 groups
was used in the analysis. If less than 40 were returned, in any of the
groups, additional numbers of teachers were selected and personal delivery
questionnaires were distributed. However, only a very few additions were
necessary. Sincemost teachers in Thailand are government officers and
their jobs are alike because of the very strong centralized educational
system, the biasedness contributed by substitution of the sampling is
minimum according to the homogeneity of the population.

No problem of data collection was expected from those in the
university area because most of the teachers received their questionnaires
by personal delivery. Since the agent collecting the questionnaires lived
in the university area, the followbup was done weekly. Because of the
personal relationship and with.some help from the Head Masters, a

hundred percent return was received in the University area.

Instrument
The instrument used was composed of two parts:
1) a questionnaire on perceived needs using a Likert

type answer format.

31

2) test items about basic knowledge in measurement

and evaluation.

There were four subject matters covered in the test and
in the questionnaire. Each part was composed of seven items for the
perceived needs and five items for the test questions. The items
in the questionnaire portion were descriptions of tasks in the four
subject matters and are called "perceived needs." The items in the
test portions covered basic knowledge in measurement and evaluation
corresponding to the four subject matters that teachers should know,
according to the curriculum of the Measurement and Evaluation workshop
program during 1966-1968. These test items are called "actual needs."
The four areas are:

l) planning test construction or curriculum analysis;

2) test item editing;

3) item analysis;

4) reporting test scores.

Both test and questionnaire were constructed with eight
items in each part, with a total of sixty-four items (see Appendix A)- The
instrument was then translated into Thai. The translation was checked
with five Thai students who were enrolled in the College of Education,
Michigan State University.

The resulting instrument was tried out on thirty teachers
in Khon Kaen, who were then omitted in selecting the sample. The
questionnaire and test were distributed to this group in the last
week of February, 1974, and all returns were collected in the first
week of March, 1974. All responses were coded and data cards were

punched. The computer programs were run to compute reliabilities,

32

item analysis, inter-item correlations, and correlations between
actual and perceived needs.

For the perceived needs scale, the reliability of .87
by Hoyt's procedure was obtained. The inter-item correlation was ev-
aluated. One item from each part of the perceived needs scale was
deleted. The final form was composed of twenty-eight items, seven
for each area of the tasks in measurement and evaluation (see
Appendix B.). Both inter-item correlations and variance of items
were used as criteria for deletion. Among eight items in Test Planning,
items 3 and 4 had negative correlations with some of the other items.
Item 3 correlations ranged from -.236 to .451 and item 4 from —.246
to .444 (the highest one with item 3). The variance of item 3 was
higher than the variance of item 4 (.913 and .907 respectively).
Therefore, item 4 was deleted. Among the items in Item Editing,
item 16 was taken out. Its correlations ranged from -.252 to .334
with four out of the seven correlations negative. In the Item Analysis
area, item 18 was deleted because of its two negative correlations
with item 21 and 22 (-.007 and -.036 respectively). The lowest inter-
item correlations within the last group of items, Reporting Test
Scores, were found on item 29 (.175 and .212). All of the correlations
in this group were positive and most of them were higher than .40.

In addition to statistical procedures, the psychometric
procedure was also considered. Item 3 and 4 (see Appendix A) in Test
Planning were very much alike, i.e., "Deciding what ability should be emphasized
in the test" and "Determining the ratio among memory, application,
and reasoning." Item 16 of the second group was too general. It

was "writing better items". Therefore, this item was deleted from the

33

group Item Editing. The next two items were also selected out
because they were too general in idea or very broad, i.e., item 18
in Item Analysis group was "Selecting good items", item 29 in
Reporting Test Scores group was "Assigning grades to students."

Most of the test items (actual needs) were difficult. The
difficulty indices ranged from .06 to .70. Only three items had an
index of difficulty above .50. The Hyt's reliability was .41.

Three items were deleted from each of the four parts. The deletions
were based on the item discrimination index (the point biserial
correlation). The deletion was considered separately in each group of
measurement areas of the in-service program. Within each of the four
subject matters, the item discrimination indices and item difficulties
were considered in deleting the items. In the first subject matter,
the three lowest discrimination indices were .001, .066, and .153.
Though item 3 had a difficulty index of .467 its discrimination index
was .001. Item 8 had the highest discrimination index among these
three items but had the lowest difficulty index (.100).

In the Item Editing group item 9, 12, and 15 were deleted
according to low discrimination indices and low levels of difficulty.
Item 9 had discrimination and difficulty indices of -.112 and .267
respectively. Item 12 had .153 and .200. Item 15 had -.137 and .267.
The three lowest discrimination indices in the next group, Item
Analysis, were .009, .124, and .254 with.the levels of difficulty .300,
.333, and .167 respectively. In the last group, Reporting Test

Scores, the deleted items were items 25, 26, and 29 with discrimination

34

:indices of -.098, -.126, and -.080 and levels of difficulty were
.167, .033, and .100 respectiVely.

Those items retained were then revised. All alternative
choices were reordered and reworded if necessary. One alternative
among the five-choice items was deleted. The deletion was of the
alternative that was the least attractive, i.e., no one chose it or
the discrimination index was high and positive. All items then
presented four choices, in order to reduce the difficulty of the test
items for this particular group of teachers.

The final actual needs test contained twenty items, five
for each part (see Appendixzn). The correlation between the scores
for total perceived needs and for total actual needs was very low

(I - .126), and not significant.

Data Collection

The revised edition of the questionnaire along with the
demographic items and letter (see Appendixli) was sent to teachers
in the sample in the last week of May, 1974, both by mail and personal
delivery. All teachers in the university area received personal
delivery, school by school. Most teachers outside the university
area received mail questionnaires. However, some of those who were
closest to the university area got theirs by personal delivery. To
insure a high percentage of response, the follow-up was done weekly
either by postcards or personal contact. By the second week of June,
1974, all returns from the four groups in the university area had been
received. The number of returns required from the outside the

university area was reached, with the help of some personal contact,

35

by the last week of June, 1974._ As was expected, most of the first
forty returns from each.group in the outside university area were from
people nearby amphurs. The farther the school from.the amphurs, the
less likely that the responses would be returned.

Since the structure of society and the school system in
Thailand are very much the same, teachers who graduated in the same
level of teacher certificate must have the same program of study.
Because of a very strong centralized educational system, their
teaching jobs and routine work are also alike. It is reasonable to
assume that teachers who hold the same teacher certificate have the
same characteristics. That is the distortion of data is less likely
to occur\when additional number of teachers is drawn in order to reach

the desirable number of responses.

Design
The design of this study is described as a 2x2x2 factorial

design with.four repeated measures each on actual needs and perceived
needs. The four areas of knowledge of educational measurement were:
planning test construction or curriculum analysis (PL); test item
editing (IE); item analysis (IA); and reporting test scores (RE).
Table 3.2 presents the lay-out of the three independent variables:
locations of schools; levels of teacher education; and teaching

experience. The design is crossed and balanced.

 

36

TABLE 3.2

DESIGN OF THE THREE INDEPENDENT VARIABLES:
LOCATION OF SCHOOLS, LEVELS OF TEACHER EDUCATION,

AND TEACHING EXPERIENCE

 

 

 

 

 

 

 

 

 

 

Locations Levels of Experience M M M
Education M1 2 3 4
Old
n - 4O
Cert.
University New
. . n - 40
Area (Central f‘
Old
Amphur) High n - 40
Cert. *—
New
n - 40
Old
n - 40
Cert. ‘“
Far From New
n - 40
University
Old
(Other High n a 40
Amphurs) Cert. I
New
n - 40

 

 

 

 

 

 

 

Old - teachers who had taught at least eight years

New - teachers who had taught less than eight years

 

37

On each scale of needs, the following null hypotheses were
tested.

1) There is no difference in needs between teachers located
in the area near the university and those farther from the university
in Khon Kaen province.

2) There is no difference in needs between teachers who
hold the teacher certificate or lower and those who hold higher certificates
in Khon Kaen province.

3) There is no difference in the needs of the experienced and
of the newer teachers in Khon Kaen province.

4) There is no difference in needs among the four subject
matters in measurement covered in the study.

5) There is no location by level of teacher education
interaction.

6) There is no location by experience interaction.

7) There is no experience by level of teacher education
interaction.

8) There is no location by experience by level of teacher
education interaction.

9) There is no subject matter by location interaction.

10) There is no subject matter by level of teacher education
interaction.

11); There is no subject matter by experience interaction.

12) There is no subject matter by location by level of
teacher education interaction.

13) There is no subject matter by location by experience

interaction.

38

14) There is no subject matter by experience by level of
teacher interaction.
15) There is no subject matter by location by experience

by level of teacher education interaction.

Analysis

First, the correlation between actual needs and perceived
needs will be evaluated to discover whether there was any relationship
between them. If there is no significant correlation at all, the
perceived needs and actual needs will be analyzed separately using the
2x2x2 factorial design with four repeated measures. If there is a
relationship, the multiple repeated measures, both actual needs and
perceived needs would have been analyzed in the same lay-out.

Before the analysis of repeated measure design can be performed,
the variance-covariance matrices of both actual and perceived needs will
be tested for the compound symmetry assumption. Box's chi square test
of the homogeneity for variance-covariance matrix will be computed
on each of the variance-covariance matrices of actual and perceived
needs. If there is a relationship between actual and perceived
needs, and the homogeneity of variance-covariance for both matrices
are the same either accepted or rejected, the multiple repeated measure
analysis will employ the multivariate test procedure. If the outcomes
of the Box's chi square tests are not the same, separate analysis
will be performed on each of the needs scales. All of the hypotheses
will be tested at the significant level of .05 each.

Beside these hypotheses testings, descriptive data were

provided to serve as a detailed investigation of some relevant

39

information such as average number of teachers, average number of
students, sex, opinion on national testing, grade taught and subject

taught.

Summary
A list of all 5,913 teachers in Khon Kaen province was

obtained from the Department of Education, Rhon Kaen Education Office.
The teachers were classified into eight groups according to two
categories each, of school location, level of teacher education, and
teaching experience. Forty teachers were randomly selected within

each group for the total of 320 plus an additional eighty teachers.

A questionnaire containing the demographic questions, perceived need
scales, and actual need test items was constructed and tried out in
March, 1974. The revised edition of the questionnaire was then sent

to those 400 teachers in May, 1974. The required sample size of forty
in each group for the total of 320 was reached by the end of June,
1974. All responses were coded and data punch cards were produced.

The 2x2x2 factorial design with four repeated measures was used to
analyze the data on both actual needs and perceived needs. The decision
whether univariate or multivariate analysis will be employed depended
upon the results of Box's chi square test of the compound symmetry

of the variance-covariance matrices on each need. Graphic presentation
will be considered when there is the signficiance of interaction.

Descriptive data along with null hypotheses will also be presented.

CHAPTER IV

ANALYSES AND RESULTS

At the end of the first week of July, 1974, all data were
coded. Responses to the perceived needs were coded 4,332,132
corresponding to the rating from highest need to no need at all on each
item. Responses to each item of the actual needs were coded 1 to 5,
with 1:4 assigned to responses §:d and 5 assigned where the response
was omitted. Coding systems for all demographic data were also
created and used to record the information. Next, all data cards
were punched, and were ready for use on July 16, 1974. A FORTRAN
program was written to obtain the 4 subscale scores on each of the
needs scales as well as the total scores for both needs. The accuracy
of the program was checked against a sampling calculated by hand. Both
yielded the same scores. The program also produced a punched card for
each individual respondent. Each card contained information on group
identification, demographic data, four subscale scores for perceived
needs, total score for perceived needs, four subscale scores for
actual needs and total score for actual needs.

All programs were executed by the CDC 6500 at the
Computer Center, Michigan State University. The canned programs used
were PROFILE to get Box's chi square test for homogeneity of variance-
covariance matrix; ACT routine in the CISSR system:to get the cross-
tabulation or contingency table; FORTAP program to compute Hoyt's
reliability and item analysis; and FINN program to perform the

correlation, cell means and cell standard deviations, and the analysis

40

41

of repeated measure design.

Before any analysis was done, a re—evaluation of the instrument
was made. The reliabilities of both the perceived and actual needs scales
by Hoyt's procedure were .91 and .42 respectively. The low reliability
on the actual needs was probably caused by two factors. First, the test
items were rather difficult. Positively skewed distribution was found.
[Many of the low scores were probably obtained through.guessing. Another
factor was that the test was quite short (twenty items). The split-
half (odd-even split) procedure was employed as another method of
estimating reliability. This estimated reliability was .47. If the
test was doubled in length (assuming parallel content) the reliability
would have been .64.

In this chapter, the results of the data analyses are presented.
The descriptive data and some demographic results are discussed first.

The correlations between actual and perceived needs are given next. The
final section deals with the hypotheses of interest as tested by the
repeated measure analyses, graphs of interactions and some follow-up or

post hoc tests.

Descriptive Data

 

The description of the sample is provided to compare the come
position of the sampled teacher groups with.the population that was
discussed in the previous chapter. The average number of teachers and
average number of students in the schools in the sample are presented
in Table 4.1 along with their standard deviations. The distribution of
school size is skewed. Most of the big schools are in the central amphur
With thirty to seventy teachers and 1100 to 1500 students. The over-all

average number of teachers within a school is twenty-nine with an average

42

TABLE 4.1

AVERAGES AND STANDARD DEVIATIONS OF THE NUMBER OF TEACHERS
AND NUMBER OF STUDENTS IN SAMPLED SCHOOLS CLASSIFIED

ACCORDING TO EIGHT GROUPS OF TEACHERS

 

 

 

 

 

 

 

 

 

Location Teacher Experience Number of Teachers Number of Students
r Educati” 1 if S.D. § ii 3.1).
1 1
New 21.85 11.10 730 254.60
Cert. ’ ' 4
University
Old 33.45 14.29 1120 430.98
Area Higher New 42.65 18.65 1255 496.11
Cert. —-
01d 46.05 10.89 1457 217.08
outside New 18.95 4.65 672 95.37
University Cert. ——
01d 18.35 5.58 570 92.54
Higher
Cert.
Old 20.75 5.92 674 157.70

 

 

 

 

 

43

TABLE 4.2

INTER-CORRELATIONS AMONG NUMBER 012' TEACHERS,

NUMBER OF STUDENTS, PERCEIVED NEEDS, AND ACTUAL NEEDS

 

NTE.. ..NST PN AN
NTE 1.000
NST .520 1.000
PN -.069 -.145 1.000
AN .160 .152 .004 1.000

 

 

NTE - Number of teachers in a school

NST - Number of students in a school

PN - Perceived needs

Actual needs

E

of 892 students, yielding the ratio of 1:31 teacher to students which
is lower than 1:33, the population value.

Table 4.2 presents the inter-correlations among four variables:
number of teachers in schools, number of students in schools, perceived
needs; and actual needs. The number of teachers as well as number of
students in schools serve as the index of size of school the teachers
are in. The correlation between the number of teachers and number of
students in schools is .520 which is very high. All other correlation
coefficients are low, especially the correlation of interest, between

perceived needs and actual needs, which is .004. The correlation

44

between size of school and perceived needs is low and negative, between
size of school and actual needs is low and positive.

The responses on the descriptive questions are counted
according to category of answer choices. Only the frequency
count is of interest. Of all 320 teachers, only fifty-one teachers
(16 percent) had attended one of the in-service programs on measure-
ment and evaluation given between 1961 and 1973. Of these fifty—
one teachers who had participated in the in-service program, twenty-
three are from the central amphur, thirty-six hold the higher
certificate, and thirty-three have been teaching for at least
eight years.

The majority of teachers in the sample do not like the
idea of national testing. Sixty-four percent of the teachers
(205) responded no to that question. Of the 160 teachers who
have taught more than eight years, seventy-seven percent (123 teachers)
disagree with the issue and fifty-one percent (eighty-two teachers)
from the less experienced group responded no.

The other breakdowns of the total 320 teachers are by
sex, grade taught, and subject taught. There are 141 male and
179 female teachers (44% and 56% respectively). Most of the
teachers teach in the elementary grades (grades 1 to 4) and
teach all subjects.

One hundred nineteen teachers (37%) teach in grade 1 to
4, ninety teachers (28%) in grades 5-7, one hundred three teachers (32%)
in grade 8 to 10, and eight teachers (32) in grades 11-12. Sixty-
three percent of 320 teachers teach a particular subject. Twenty-

four teachers teach Thai, thirty-one Sciences, thirty-six Mathematics,

45

forty-four Social Sciences and Humanities, thirty-seven English, and
thirty teach Home Economics, Physical Education, Art and Handicrafts,

or Agriculture.

Homogeneity of Variance-Covariance Matrices

As was reported earlier, the correlation between the perceived
needs and actual needs was very low. More detailed investigation
was done to evaluate the correlation between the subscales of the two
need scores. For planning and curriculum analysis, the correlation
was -.O70, Test Item Editing .065, Item Analysis .053, and Reporting
Test Scores .018. All correlations were low and not significant at
the 95 percent level of confidence (an r of .098 is required for
significance at .05 for df - 400). Since there is no correlation
between actual and perceived needs the testing of hypotheses of interest
on a multiple repeated measures design yields the same results as the
combination of the tests on each variable, providing that the over-all
Type 1 errors are the same.

The test of the compound symmetry assumption was performed
twice, once for the four subscales on the actual needs and again on
the perceived needs. Table 4.3 provides the variance-covariance
matrix as well as the correlation among the four subscales on actual
needs. Again all correlations were low, and only two of them above .10

- .174 and r - .175). The Box's chi square test of

(rPL,IE IE,IA
homogeneity for the pooled variance-covariance matrix is 11.41 at

8 degrees of freedom.which is not significant at the 95 percent level
of confidence. Since the hypothesis of the compound symmetry is not

rejected, the analysis of variance on repeated measures (Mixed Model

46

repeated measure) is used to test all hypotheses of interest concerning
the actual needs.

Table 4.4 presents the variance-covariance and correlation
matrices among the four subscales on perceived needs. All correlations
are high, ranging from .43 between Test Planning and Reporting Test
Scores, to .591 between Item Analysis and Reporting Test Scores. The
variances of Test Planning, Item Editing, and Item Analysis are about
the same (22.974, 20.816, and 21.980 respectively), but the variance Of
Reporting Test Score variable was the highest at 32.414. The Box's
chi-square test of homogeneity for the pooled variance-covariance
matrix on the perceived needs was 48.59 at eight degrees of freedom
which was significant at the .05 level of confidence. (The critical
value is 15.51). Therefore, the hypothesis of the compound symmetry
was rejected. The multivariate test was employed to test the hypotheses

concerning the perceived needs.

Repeated Measure Analysis on Actual Needs

 

Summary data, cell means and cell standard deviations of the
four subscales of the actual needs are presented in Table 4.5. To
get a general profile of the four dependent variables, the means
and standard deviations for all subjects are also presented. The
highest mean was on the Item Editing variable (2.328), and Test
Planning (2.175) was the next highest. Third was Item Analysis
with a mean of 1.706 and Reporting Test Score was the lowest of all at
1.272 mean. Profiles of all four groups of teachers outside the central
amphur were similar to the general profile except for a slight

difference for the new higher teacher certificate holders in which

47

TABLE 4.3

VARIANCE-COVARIANCE AND CORRELATION MATRICES OF FOUR TEST SCORES

 

 

TESTS

PLVF‘ IE IA RE
PL 1.092
IE .173 1.268
IA .120 .237 1.447
RE .070 .117 .099 1.088
PL 1.000
IE .147 1.000
IA .096 .175 1.000
RE .065 .099 .079 1.000
PL = Test Planning
IE 8 Item Editing
IA - Item Analysis
RE - Reporting Test Scores

Box's chi square test of homogeneity for the pooled variance-

covariance matrix - 11.41, df = 8.

(Not significant at .05 level).

48

TABLE 4.4

VARIANCE-COVARIANCE AND CORRELATION MATRICES

OF FOUR PERCEIVED NEEDS SCORES

PERCEIVED NEEDS

 

 

PL IE ' IA RE
PL 22.974
IE 12.319 20.816
IA 11.236 11.571 21.980
RE 11.726 14.265 15.784 32.414
PL 1.000
IE .563 1.000
IA .500 .541 1.000
RE .430 .549 .591 1.000
PL - Test Planning
IE - Item Editing

IA - Item Analysis

53

Reporting Test Scores

Box's chi square test of homogeneity for the pooled variance-
covariance matrix - 48.59, df = 8.

(Significant at .05 level).

49

TABLE 4.5

MEANS AND STANDARD DEVIATIONS OF FOUR TEST SCORES

(ACTUAL NEEDS)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Location Teacher Experience Tests Tota1__
Education PL IE IA RE
New 2.20 1.95 1.63 1.05 6.825
n.= 40 0.99 1.11 1 1.10.+\0.99. 2.745
University 111
In = 40 1.05 1.03 1.28 0.98 2.230
New 1.85 2I48 1.75 1.85 7.925
“18“ . n =.40 1.10 1.22 1.17 1.14 2.422
Area
01d 2.45 2.43 2.13 0.93 7.925
Cert. '
n = 40 1.04: 1.28 . 1.24 0.83 2.740
New 2.15 2.58 1.58 1.40 7.700 _‘
, n = 40 1.03 1.11 1.13 1.15 2.345
Out81de C t ' ___
er - Old 2.00 2.18 1.10 0.98 6.250
n = 40 0.99 1.08 0.90 1.05 2 488
”“1V8r31ty New 2.45, 2.43 1,63 1.48 7.975
“18“ . n = 40 1.18 0.96 1.37 1.13 2.824
Area Cert Old 2.15. 2.58 1.98: 1.10 7.800
n = 40 0.98: 1.20 1.35 1.03 2.584
2.1 2. 2 .70 .2 2
All Groups 75 3 8 1 6 1 7
1.045 1.126 1.203 1.043

 

 

 

 

 

New a Teachers who have been teaching for less than eight years.

Old = Teachers who have been teaching at least eight years.

PL ' Test Planning

IE

Item Editing

51

Item Analysis

i3

Reporting Test Scores

50

group the means on Test Planning and Item Editing were about the same,
i.e., 2.45 and 2.43 respectively. All four profiles for the four
groups within the central amphur were quite different from each
other and from the general profile. All teacher certificate holders
showed higher means on Test Planning (2.20 and 2.15) than means on
Item Editing (1.95 and 2.03). Large discrepancies were found among
the higher teacher certificate holders within the central amphur. For
the new teachers, the mean for Test Planning was much lower than the
mean on Item Editing (1.85 versus 2.48) and the reversed order between
Item Analysis and Reporting Test Score was observed with the values
of 1.75 and 1.85. For more experienced higher teacher certificate
holders, means for Test Planning and Item Editing were about the same,
with the mean for Item Analysis rather close to them (2.45, 2.43,
and 2.13 respectively). These three means were quite different from
the mean of 0.93 for the Reporting Test Score variable. The cell
standard deviations varied from .83 to 1.37.

The tests of all hypotheses concerning the actual needs are
presented in Table 4.6. The univariate repeated measure analysis
was used. Fifteen hypotheses were tested. Four of them were on the
main effects, the other eleven concerned the interactions.

From Table 4.6 seven of the null hypotheses were rejected

at a = .05. They were:

the main effect of teacher education (F i 8.44, p < .003);

the interaction between school location and teacher

experience (F - 3.878, p < .0499);

the subject matter main effect (F - 67.28, p < .0001);

the subject matter by location interaction (F - 3.021,

51

TABLE 4.6

UNIVARIATE REPEATED MEASURED ANALYSIS 0N TEST SCORES

 

 

 

 

 

Sources DF MS F P
Location (L) 1 0.200 0.123 .727
Teacher Education (T) 1 14.450 8.855 .003
Experience (E) 1 1.250 0.766 .382
LT 1 0.078 0.048 .827
LE 1 6.328 3.878 .0499
TE 1 0.528 0.324 .5691
LTE 1 4.513 2.765 .097
R:LTE 312 1.632
Subject Matter (M) 3 73.133 67.280 .0001
ML 3 3.284 3.021 <.05
MT 3 1.029 0.946 >.05
ME 3 3.252 2.991 <.05
MLT 3 1.679 1.545 >.05
MLE 3 0.704 0.648 >.05
MTE 3 4.545 4.181 <.01
MLTE 3 4.069 3.743 <.05
RM:LTE 936 1.087

(.05) a 2.61 F 0

F3,1000

3 1000 (.01) - 3.8
D

52

p 5 .05);
- the subject matter by experience interaction CF - 2.991,
p < .05):
- the subject matter by teacher education by experience interaction
(F I 4.181, p < .01);
- and, the subject matter by location by teacher education
by experience interaction (F - 3.743, p < .05).

The other nine hypotheses were not rejected, at a 95 percent
level of confidence.

The significance of the four-way interaction supported the
difference of the profiles among the eight groups of teachers,
discussed earlier (under table of means and standard deviations on
the actual needs). A followhup analysis was performed using the
same computer program, but the variables were transformed to pairwise
comparisons between Test Planning and Reporting Test Scores, Item
Editing and Reporting Test Scores, and Item Analysis and Reporting
Test Scores. The three-way interaction among location, teacher
education, and experience was computed for each transformed variable.
The univariate F ratio was significant only for the pairwise
comparison of the Test Planning and Reporting Test Scores (F - 11.0774
with 1 and 312 degree of freedom, p < .001, see Appendix B). The
other two comparisons were not significant.

On examining all eight profiles, the most contrasting
profiles were from the less experienced higher teacher certificate holders
within the central amphur (university area) and the more experienced
and higher teacher certificate holders within the central amphur

(university area). The first group had the lowest mean on Test Planning

53

but the highest mean on Reporting Test Scores. On the contrary, the
latter group tied for the highest mean on Test Planning but was the

lowest on Reporting Test Scores. These reversed profiles suggested

that the interaction was not ordinal.

Another followhup was done to test the three—way interaction
among location, teacher education and experience on the Test Planning
and the Reporting Test Scores variable as well as the other two
variables. The multivariate F ratio was significant at a = .05
(F - 3.618 with 4 and 309 degrees of freedom, p < .0068, see
Appendix D) and the univariate F ratio indicated that only the three-
way interaction on the Reporting Test Score variable was significant
at a = .0049 . Although there was no significant difference in
scores for Test Planning, Item Editing, and Item Analysis among the
eight groups of teachers, there were some differences among them
on the Reporting Test Score variable, which seems to account for
the four-way interaction results.

Since the four-way interaction was significant, a general
profile analysis was not made and the following discussion is made
on the condition of the presence of these interactions.

The interaction among the subject matter. 18V€110f teacher
education and experience was significant at a - .05 (F = 4.181,

p < .01). It was concluded that the profiles of the four groups of
teachers, regardless of the location of their schools, were different,
as can be seen in Figure 1. The difference on each of the

four subscales on the actual needs between the less experienced

and more experienced teachers across teacher certificate holders

and higher teacher certificate holders were quite substantial. On

54

Test Planning, Item Editing and Item Analysis variables, the less
experienced groups were higher than those with.more experience within
the groups of teacher certificate holders. This relationship was
reversed within the groups of higher teacher certificate holders.
The pattern of differences was not reversed for the Reporting Test
Score variable. The differences, rather, were a matter of size.
The difference between less experienced and more experienced teachers
on the Reporting Test Scores variable was .035 for the teacher
certificate holders, and .650 for the higher certificate holders
The differing size of the differences between less experienced and
more experienced teachers as well as the direction of differences
' was also found in the Item Analysis variable, (i.e., .115 within the
teacher certificate holders and -.365 within the higher teacher
certificate holders). Again the analysis of individual variables
revealed that the interaction between teacher education and experience
was significant only on the Reporting Test Scores variable at
o - .0125 (F = 6.895 with 1 and 312 degrees of freedom, p < .0091).
Figure 2 presents the profiles of less experienced teachers
and more experienced teachers to explain the interaction between
subject matter and experience. The differences between less experienced
teachers and the more experienced teachers on the four sub-scales
were -0.025, 0.055, -0.125 and 0.342 for the Test Planning, Item
Editing, Item Analysis and Reporting Test Scores Respectively.
The significance of the interaction between the subject
matter and location of school is illustrated by the graph in
Figure 3. The differences between teachers in the central amphur

and those outside the central amphur on Test Planning, Item Editing,

55

FIGURE 1

GRAPH PRESENTATION OF CELL MEANS FOR SUBJECT MATTER,

TEACHER EDUCATION, AND EXPERIENCE (ACTUAL NEEDS)

NI

2.50 H

2.00 ..

 

1.50 .,

 

 

1.00 ..

 

 

 

 

 

 

PL IE IA RE
New teacher with teacher certificate
--——- 01d teacher with teacher certificate
._.... .._.. New teacher with higher teacher certificate

. . . . . . . Old teacher with higher teacher certificate

56

FIGURE 1

(CONTINUED)

Test Planning Item Editing Item Analysis Reporting

Test Scores

 

 

 

 

 

 

 

(PL) (IE) (IA) (RE)
New 22175 2.265 1.605 1.225
Cert.
Old 2.075 2.105 1.490 1.190
High New 2.150 2.455 1.690 1.665
Cert. Old 2.300 2.505 2.055 1.015

 

 

 

57

FIGURE 2

GRAPH.PRESENTATION OF CELL MEANS FOR SUBJECT MATTER

AND TEACHING EXPERIENCE (ACTUAL NEEDS)

 

 

 

 

 

 

 

 

 

 

 

 

3'6
2.5
2.00
1.5
1.00

PL IE IA RE
PL a Test Planning pL IE 1A RE
IE - Item Editing New 2.163 2.360 1.648 J1.445
IA-Item Analysis 01d 2.188 2.305 1.773 l1.103
RE - Reporting Test Scores

New teachers

—————— Old teachers

58

FIGURE 3

GRAPH PRESENTATION OF CELL MEANS FOR SUBJECT MATTER

AND LOCATION OF SCHOOLS (ACTUAL NEEDS)

NI

2.50

2.00

1.50

1.00

 

T

 

 

 

 

 

PL
PL 8 Test Planning
IE I Item Editing

IA = Item Analysis

E3

Reporting Test Scores

 

University area

 

Non-University
area

 

IE RE
University area
Not in the
University area
PL IE IA RE
2.163 I 2.223 1.848 1.308
2.188 | 2.443 1.573 1.240

 

 

 

 

59

FIGURE 4

GRAPH PRESENTATION OF CELL MEANS FOR LOCATIONS OF SCHOOLS

AND TEACHING EXPERIENCE (ACTUAL NEEDS)

 

 

 

ii
8.00 .1
«L
a rUniversity area
7.50 x
+ 1
4
7.00 T l bNon-University area
,1.

 

New ' Old

New - Teachers who have been teaching for less than 8 years

01d - Teachers who have been teaching 8 years or more

 

 

New Old
University
area 7.375 7.688
Non-
University
area 7.838 7.025

 

 

 

 

60

Item Analysis and Reporting Test Scores were -0,025, -0.220, 0.093
and.0.068 respectively. The interaction was disordinal with.respect
to location of schools.

On the between-subject part, a significant interaction
between location of school and experience was observed. This inter-
action was explained by the reversed differences in the total actual
needs (Fig. 4). The difference between less experienced and more
experienced teachers within the central amphur was -0.313 (means of
7.375 and 7.688) while the same difference for teachers outside the
central amphur was 0.813 (means of 7.838 and 7.025). The
significance of the teacher education main effect was also observed
(F - 8.855, p < .003). The mean for teacher certificate holders
was smaller than the mean for higher teacher certificate holders

(7.056 and 7.906 respectively).

Repeated Measure Analysis on Perceived Needs

The results of the data analysis on perceived needs are
presented in the same pattern as in the previous section. First
the summary data are discussed, then the testing hypotheses are
presented followed by the graphic presentation of any interactions
that were significant.

All summary data, means and standard deviations, both of
subscale scores and total perceived needs are presented in Table 4.7.
The general profile illustrates all group means on Test Planning
(20.816), Item Editing (20.906), Item Analysis (22.606), and Reporting
Test Scores (20.897). The maximum perceived needs on Item Analysis

with the three other means were about the same. Within the central

61

TABLE 4.7

MEANS AND STANDARD DEVIATIONS OF FOUR PERCEIVED NEEDS SCORES

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Location Teacher Experience Needs Total
Education PL IE IA RE
New 20.90 20.53 23.05 21.08 85.550
n a 40 5.07‘ 5.43 4.27 J 5.00 15.671
University Cert.
Old 21.93‘ 21.40 23.68 21.73 88.725
1
New 20.23 19.73 22.78 21.10 83.825
Cert. Old 21.60 21.20 23.03 21.28 87.100
n - 40 3.59 4.18 4.22 5.82 14.604
n a 40 ' 5.23 4.73 4.33 4.49 15.100
Outside Cert.
, Old 19.83 19.40 20.93 18.95 79.100
New 21.30 21.68 23.18 21.45 87.600
Cert. 01d 20.10 21.93 22.25 19.95 84.225
n - 40 04.484 4:22 4.96 5.28 14.591
20.810 20.906 22.606 20.897
All Groups
4.793 4.563 4.688 5.693

 

 

 

PL - Test Planning

IE - Item Editing

IA - Item Analysis

RE - RepOrting Test Scores

New 8 Teachers who have been

Old - Teachers who have been

teaching less than eight years

teaching eight years or more

62

amphur (university area), the four groups had almost the same profile.
There were small decreases from Test Planning to Item Editing and
increases from Item Editing to Item Analysis. The means for
Reporting Test Scores dropped to about the same as Test Planning. 0n
the other hand, for those teachers outside the central amphur (outside
university area), the means for three groups (except experienced
teacher certificate holders) increased from Test Planning to Item
Editing and increased to Item Analysis. Finally, for the Reporting
Test Scores variable, the means decreased to about the same as the
first two subscales. For experienced teacher certificate holders
outside the central amphur the profile was the same as for those
groups within the central amphur (university area).

Contradictory information between the total actual needs
and the total perceived needs was found on the more experienced teacher
certificate holders outside the central amphur (outside university area).
These teachers had the lowest means of the perceived needs of all
eight groups, but they had the lowest actual needs score which
indicates that they had the highest need. The group that reported
the highest perceived needs was the more experienced teacher certificate
holders in the central amphur (mean of 88.725).

The tests of fifteen hypotheses concerning the perceived
needs were performed by the multivariate repeated measure analysis.
The between-subject part, i.e., hypotheses about location, teacher
education, experience, and interactions among them, was tested by
the univariate F ratio while the within-subject part, i.e., hypotheses
about the subject matter and interactions with it,.by the multivariate

F ratio.

UNIVARIATE AND MULTIVARIATE REPEATED MEASURED ANALYSIS ON PERCEIVED NEEDS

63

TABLE 4.8

 

 

Sources DF Univariate Multivariate
' MS F P F P
Location (L) 1 92.450 1.468 .227
Teacher Education (T) 1 17.112 0.272 .603
Experience (E) 1 15.313 0.243 .622
LT 1 135.200 2.146 .144
LE 1 336.200 5.337 .022
TE 1 13.613 0.216 .642
LTE 1 12.013 0.191 .663
R:LTE 312 62.997
Subject Matter CM) 3-310 24.349 .0001
ML 3-310 3.351 .0194
MT 3-310 0.458 .7120
ME 3-310 1.115 .3433
MLT 3-310 0.763 .5155
MLE 3-310 0.414 .7434
MTE 3—310 1.082 .3571
MLTE 3-310 0.552 .6474

 

 

 

64

FIGURE 5

GRAPH PRESENTATION OF CELL MEANS FOR LOCATION OF SCHOOLS

AND SUBJECT MATTER (PERCEIVED NEEDS)

>fl

23.00 ‘,

22.00 .

21.00

IAJLJ
V

 

 

20.00
1

 

 

 

 

 

PL IE IA RE

PL - Test Planning University area

IE - Item Editing -—--—- Outside university

IA - Item Analysis

 

 

RE - Reporting Test Scores

PL IE . IA . . RE
University area 21.165 20.715 .23.135 21.298
outsideagzivemity 20.470 21.103 . 22.085 20.500

 

 

 

 

 

 

65

FIGURE 6

GRAPH PRESENTATION OF CELL MEANS FOR DOCATION OF SCHOOLS

AND EXPERIENCE OF TEACHERS (PERCEIVED NEEDS)

 

 

 

 

 

 

3(-
87.00),
86.00),
85.00),
84.00V
83.00 L
82.001.
81.00,:L
'1:
* New Old
University area
------- Outside university area
New Old
University area 84.688 87.913

 

Outside university
area 86.638 81.663

 

 

 

 

66

All univariate F and multivariate F ratios are presented
in Table 4.8. A rejection of the null hypotheses was indicated, at
a - .05, on three of the fifteen hypotheses: the interaction between
location of schools by experience (F = 5.377, with 1 and 312 degrees
of freedom, p < .022); the subject matter main effect; and the
subject matter by location of school interaction (multivariate F of
3.35 and 24.349 respectively with 3 and 310 degrees of freedom,

p < .0001).

Since the interaction between location and subject
matter was significant, the general profile was not interpreted.

One profile was made for all teachers in the central amphur (university
area) and another for those outside the central amphur (outside university)
area). Figure 5 presents these two profiles. For teachers in the

central amphur (university area), the means vary from 21.165 for

Test Planning to 20.715 for Item Editing to 23.135 for Item Analysis

and finally to 21.298 for Reporting Test Scores. For teachers outside

the university area the mean scores were 20.470, 21.103, 22.083

and 20.50. The largest difference between the groups within and those
outside the university area was found on Item Analysis subscale with

the value of 1.05.

The interaction between locations of schools and experience
was explained by the difference between less experienced teachers and
more experienced teachers within the central amphur and the difference
for teachers outside the central amphur on the total perceived needs
score. The mean for less experienced teachers in the university area
was 84.688 and more experienced teachers 87.913 with the difference

of -3.225. For teachers outside the university area the mean for the

67

less experienced group was 86.638 and the more experienced group
81.663 with the difference of 4.975. The type of interaction was

disordinal (Fig. 6).

Additional Analysis

 

Besides testing the hypothesesvof interest, further analyses
were done to compare actual and perceived needs among various groups
of teachers, defined by the following independent variables: having
attended the in-service program; sex; favoring a national testing
program; and grade taught. The analyses were done individually
on both total scores and subscale scores. Means of the groups, F
ratios testing the difference between groups, and the probabilities
of the F ratios are presented in Tables 4.9 to 4.12.

In Table 4.9, the comparison between teachers who had
attended the in-service training program and those who had not attended
was made. For total perceived needs, teachers who attended the in-
service training had perceived themselves as having a greater need
than those who had not attended the training program but the amount
of difference was not large enough to be significant at a B .05 level.
On all four subscales within the perceived needs, no significant
differences were found. In the test of actual needs, the teachers
who had not attended the in-service training program had higher need
(indicated by lower scores from the test) than those who had attended
the training program. Again, the amount of the difference was not
large enough to show any significance. However, by examining the
subscales separately, the difference in Item Editing was significant

at o .05 in favor of those who had not attended the in-service training

68

TABLE 4.9

COMPARISON ON PERCEIVED AND ACTUAL NEEDS BETWEEN TEACHERS

WHO ATTENDED AND DID NOT ATTEND THE IN-SERVICE PROGRAM

 

 

 

 

In-Service
Variables Attended Did Not Attend F P less than
n :_51 n e_269
(X) (X)
Perceived Needi
PL 20.71 20.84 0.03 .8588
IE 21.39 20.81 0.68 .4110
IA 22.35 22.65 0.18 .6758
RE 21.94 20.70 2.04 .1541
Total 86.39 85.00 0.32 .5694
Actual Needs
PL 2.25 2.15 0.35 .5545
IE 2.63 2.27 4.25 .0402
IA 1.78 1.69 0.25 .6203
RE 1.33 1.26 0.20 .6563
Total 8.00 7.38 2.44 .1195

 

PL - Test Planning

IE - Item Editing

IA - Item Analysis
RE - Reporting Test Scores

69

TABLE 4.10

COMPARISON ON PERCEIVED AND ACTUAL NEEDS BETWEEN TEACHERS

WHO WERE FAVORABLE AND NOT FAVORABLE TO THE NATIONAL TESTING POLICY

 

 

 

 

 

National Testing
Variables Favored Not Favored F P less than
n f_115 n d 205
(X) (X)
Perceived Needs
PL‘ 21.68 20.33 5.91 .0157
IE 21.33 20.67 1.53 .2167
IA 23.00 22.39 1.26 .2630
RE 21.28 20.68 0.80 .3710
Total 87.29 84.07 3.02 .0832
Actual Needs
PL 2.05 2.24 2.46 .1175
IE 2.43 2.27 1.58 .2095
IA 1.68 1.72 0.09 .7600
RE 1.45 1.07 5.13 .0242
Total 7.62 7.40 0.49 .4828
PL - Test Planning
IE - Item Editing

IA - Item Analysis

(3

Reporting Test Scores

70

TABLE 4.11

COMPARISON ON PERCEIVED AND ACTUAL NEEDS

BETWEEN MALE AND FEMALE TEACHERS

 

 

 

 

 

Sex
Variables Male Female F P less than
n e_141 n {_179
(X) (X)
Perceived Needs
PL 20.30 21.22 2.88 .0908
IE 20.81 20.98 0.11 .7362
IA 22.63 22.59 0.71 .9331
RE 20.90 20.89 0.00 .9916
Total 84.65 85.68 0.33 .5647
Actual Needs
PL 2.14 2.20 0.25 .6172
IE 2.37 2.30 0.32 .5710
IA 1.84 1.60 3.21 .0742
RE 1.33 1.22 0.83 .3641
Total 7.69 7.32 1.60 .2063
PL - Test Planning
IE I Item Editing
IA - Item Analysis
RE - Reporting Test Scores

71

TABLE 4.12

COMPARISON ON PERCEIVED AND ACTUAL NEEDS AMONG TEACHERS

WHO TAUGHT DIFFERENT GRADE LEVELS

 

 

 

 

 

Grade Taught
Variables G1 G2 G3 G4 F P less than.
n e_119 n :_90 n f_103 n :_8
(X) (X) (X) (X)
Perceived Needs
PL 21.08 21.59 19.91 19.87 2.25 .0830
IE 20.85 21.56 20.61 18.25 1.65 .1784
IA 22.50 23.19 22.17 23.25 0.83 .4789
RE 20.27 22.11 20.58 20.62 1.97 .1185
Total 84.70 88.44 83.27 82.00 1.91 .1284
Actual Needs
PL 2.11 2.29 2.16 2.12 0.52 .6665
IE 2.13 2.21 2.57 3.37 5.54 .0011
IA 1.47 1.87 1.83 1.87 2.39 .0689
RE 1.22 1.22 1.31 2.12 1.91 .1281
Total 6.93 7.59 7.86 9.50 4.32 .0053
PL - Test Planning G1 . Taught grades 1-4
IE 8 Item Editing G2 - Taught grades 5-7
IA - Item Analysis 03 - Taught grades 8-10
RE - Reporting Test Scores G - Taught grade 11 and higher

72

program. The mean for teachers who had attended the program was

2.63 and for those who had not attended it was 2.24. wThe F ratio was
4.25 with a probability less than .04. The other three subscales
showed no significant differences.

The comparison between teachers who favored a national
testing policy and those who did not is presented in Table 4.10.

In both the total perceived needs and total actual needs, there were
no significant differences between the two groups. However, when the
four subscales within each need were analyzed, significant
differences were found in the scores for Test Planning in the perceived
needs and for Reporting Test Scores in the actual needs. The
perceived needs in Test Planning for teachers who favored the
national testing policy was higher than for those who did not

(means were 21.68 and 20.33 respectively, F - 5.91 and p < .0157).
Conversely, the actual needs in Reporting Test Scores were higher

for teachers who did not favor the national testing policy than
those teachers who did (means were 1.07 and 1.45 respectively,

P - 5.13 and p < .0242).

A similar comparison was made between male and female
teachers. No significant differences were reported in any of the
subscales or in the total scores for both perceived and actual needs.
A slight difference was found on the total perceived needs with the
higher scores made by female teachers. Female teachers had lower
scores in the actual needs, but these were not low enough to be
significant. Group means on all needs, along with their associated

F ratios and probabilities are presented in Table 4.11.

73

When teachers were compared according to the grades taught,
a highly significant difference was observed in the Item Editing
actual needs. Those who taught in grades 11 and higher had the lowest
actual needs (means of 3.37) and next to the lowest were teachers
who taught in grades 8 to 10. The actual needs in Item Editing
increased as the level of grade taught decreased. Teachers who taught
in grades 1 to 4 had maximum needs (mean 2.13). The F ratio was 5.54
which was significant at a 3 .01 (see Table 4.12). This finding
agrees with the previous hypothesis which indicated that teachers
who had higher teacher education had less actual need. Since most of
the teacher certificate holders or lower were teaching in the lower
levels, it is to be expected that teachers in grades 1 to 4 should show
the highest actual needs. The pattern of differences among the four
groups of teachers, classified by the grade level they taught, was
also reflected in the scores for the total actual needs. Means were
6.93, 7.59, 7.86, and 9.50 for those who taught in grades 1 to 4,
5 to 7, 8 to 10, and 11 and higher, respectively. The F ratio was

4.32 with probability less than .0053.

Summary

The correlation between the actual and perceived needs
was very low and homogeneity of the variance-covariance matrix
for the actual needs was accepted but rejected for the perceived needs.
The decision was made that the analyses should be separated between
actual and perceived needs. The first descriptive discussion was
on information from the demographic items. The data obtained tended

to support that the sample represented the population. The univariate

74

repeated measure analysis was employed to test hypotheses concerning
the actual needs. The four-way interaction was significant. The
general profile was not interpreted. Summary data indicated that the
more experienced teacher certificate holders outside the university area
had the highest actual needs scores. Among the subject matter Reporting
Test Scores was the most needed area. Item Analysis was next. The
teacher education main effect was also significant. The means indicated
that the teacher certificate holders needed help more than the higher
teacher certificate holders.

The multivariate repeated measures analysis was employed
to test hypotheses concerning the perceived needs. The subject matter
by location of school interaction was significant. Interpretations
of profiles were made separately for the central amphur teachers
and for those outside. Both groups had the most need in Item Analysis.
Teachers in the central amphur were higher than those outside the
university area on their perceived needs on the Item Analysis
variable. In the university area, Item Editing was the least needed
but in the outside group Test Planning was the least needed.
Disordinal type interaction was found on the location of school
by experience interaction. Means of the total perceived needs
indicated that the more experienced teacher certificate holders
outside the university area had the least perceived needs, which
was in contradiction with the findings on actual needs.

The hypotheses testing results were as follow:

- Actual needs

 

1) There was no difference in actual needs between teachers

in and outside of the university area.

75

2) Teacher certificate holders had more actual needs than
the higher teacher certificate holders.

3) There was no difference in actual needs between more
and less experienced teachers.

4) There were some difference in actual needs among the
four subject matters; Reporting Test Scores seemed to be the most
needed (lowest mean).

5) There was no interaction between location of schools
and level of teacher education.

6) There was a disordinal interaction between location of
schools and teaching experience. Less experienced teachers in the
university area had more actual needs than less experienced teachers
outside the university area but the more experienced teachers in the
university area had less actual needs than the more experienced
teachers outside the university area (Fig. 4).

7) There was no interaction between teaching experience
and level of teacher education.

8) There was no three-way interaction among location of
schools, teaching experience, and level of teacher education.

9) There was a disordinal interaction between subject
matter and location of schools. Test Planning and Item Editing
variables both in and outside the university areas had about the
same needs. For Item Analysis, the university area group had
more need than the outside university area group but the effect was
reversed in the Reporting Test Score variable.

10) There was no interaction between subject matter and

level of teacher education.

76

11) There was a disordinal interaction between subject matter
and teaching experience. The less experienced teachers seemed to have
less need in. Reporting Test Score.

12) There was no interaction among subject matter, location
of schools and level of teacher education.

13) There was no interaction among the subject matter,
location of schools and teaching experience.

14) There was a three-way interaction among the subject
matter, teaching experience, and levels of teacher education (Fig. l).

15) There was a four-way interaction among the subject
matter, location of schools, teaching experience, and level of
teacher certificate. The follow-up test observed that there was
no difference in scores for Test Planning, Item Editing, and Item
Analysis among eight groups of teachers, but there were some differences
among them on the Reporting Test Scores variable.

- Perceived needs

 

1) There was no difference in perceived needs between
teachers in and outside the university area.

2) There was no difference in perceived needs between
the two groups of teachers classified by teacher education levels.

3) There was no difference in perceived needs between
more and less experienced teachers.

4) There were some differences in perceived needs in the
four subject matters. The Item Analysis variable seemed to be the
most needed while the other three were about the same.

5) There was no interaction between location of schools

and level of teacher education.

77

6) There was a disordinal interaction between location of
schools and teaching experience. Among the less experienced teachers,
those who were in the university area had fewer needs than the outside
university area, but for more experienced teachers, outside university
area had fewer needs than the university area.

7) There was no interaction between teaching experience
and level of teacher education.

8) There was no interaction among location of schools,
teaching experience, and levels of teacher education.

9) There was an interaction between the subject matter
and location of schools as presented in Figure 5. The university
area group had more needs in the Test Planning, Item Analysis, and
Reporting Test Score variables but had less need in the Item Editing
variable.

10) There was no interaction between the subject matter
and level of teacher education.

11) There was no interaction between the subject matter
and teaching experience.

12) There was no interaction among the subject matter,
location of schools, and levels of teacher education.

13) There was no interaction among the subject matter,
location of schools, and teaching experience.

14) There was no interaction among the subject matter,
teaching experience, and level of teacher education.

15) There was no four-way interaction among the subject
matter, location of schools, teaching experience, and levels of

teacher education.

78

Analyses were also done to compare actual and perceived needs
among various groups of teachers by the following independent variables:
having attended the training program; sex; favoring a national testing
program; and grade taught.

Among teachers who attended and did not attend the in-
service program, the former group had perceived themselves as having a
greater need but had lower actual needs (indicated by higher scores
from the test) than those who had not attended the training program.
There were no significant differences on needs among teachers who
favored and did not favor the national testing policy. Neither were
there any differences between the sexes.

There were differences among teachers who taught a different
grade. Teachers who taught in higher grades had lower needs than

those who taught in lower grade levels.

0

CHAPTER V
SUMMARX AND CONCLUSIONS

In—service programs in educational measurement, with emphasis
on test construction procedures, i.e., test planning and curriculum
analysis, test item editing, item analysis, and reporting test scores,
have been given to teachers in various parts of Thailand by the Ministry
of Education and the Test Bureau at the College of Education, Prasarnmitr.
Undoubtedly, the cost of this in-service program.was high and the teachers
who joined the program might not be those who actually needed the most
help. It was the purpose of this study to find out which groups of
teachers actually need in-service training and in which areas of measure-
ment the need is the greatest. This study also yields some follow-up
information on the effects of the earlier in-service programs. Two types

of needs, actual and perceived, were investigated.

Summary_and Discussion

The data of this study were based on two factors: a) the
perceived needs which were measures of the value judgment of teachers in
the sample; and b) the actual needs which were measures of their
knowledge in measurement and evaluation. A Likert type questionnaire was
constructed to measure the perceived needs. The items were a random
sample of the tasks usually involved in the process of test construction.
The respondent was asked to rate each task from 1 to 5 where l meant most
needed and 5 not needed at all. The first edition of the questionnaires

had eight items for each of the areas of in-service training for a total

79

80

of thirty-two items. The actual needs were measured by a random sampling

of test items measuring knowledge on educational measurement similar

to those items used in basic measurement courses at Michigan State University.
There were thirty-two items, eight each under the four areas of the in-
service program. Both questionnaire and test were tried out and the
reliabilities were found to be .87 and .41 respectively.

The revised editions of the questionnaire and test had twenty-
eight items and twenty items respectively. They were sent together with
a demographic data sheet and introduction letter to 400 teachers who were
randomly selected from eight strata in Khon Kaen province. The strati-
fication was based on the three variables: location of school-university
area or out of university area; levels of teacher education—teacher
certificate holders or higher teacher certificate holders; and teacher
experience-less experience or more experience using eight years as the
dividing line. Responses were received from teachers forming an equal
sample size for each of the eight strata.

The correlation between total actual needs and total perceived
needs was very low, i.e., .004. It was indicated that there was no
relationship at all between the actual and the perceived needs. Since
most of the teachers had little knowledge about the subject matter, they
felt a great need in every task. At a particular level of actual needs,
there were some teachers who would like to know more about the subject
matter and some who reported lower perceived needs. For the latter
group, the perceived needs were compared with routine work in school and
teachers seemed to decide that they had no time to pay attention to
measurement problems. At the level of their present knowledge on the
subject matter, they seemed to think that they could carry on test

construction and measurement in general very well. The variation in their

81

perception of their needs among those teachers who had the same level of
actual needs caused the low correlation between the two needs. Another

way to explain the low correlation between perceived and actual needs is

by the observation that marginal distributions of both.variables were skewed.
Most of the teachers reported high perceived needs and low actual needs.

The scatter diagram formed a tendency of curvilinear relationship between

the two variables. Therefore, the Pearson product moment correlation yielded
a low value.

From this study, the perceived needs were the teachers' value
judgments about their needs in measurement and evaluation while the actual
needs were what they knew in the subject matter areas. Since value judgments
were different among those teachers who had the same levels of actual needs,
this also contributed to the low correlation coefficient between the two needs.

The correlation between the number of teachers in the respective
respondents' school with their perceived needs was low and negative at -.069
and with their actual needs was low at .160. Since there was a high relation—
ship between the number of teachers in school and the size of schools, the
low correlation observed indicated that there was no relationship between the
size of schools and the actual needs and between the size of schools and
the perceived needs. The size of school was also related to the number of
grades in school. Big schools generally were those that had grades 1 to 12.
No matter how big the schools the teachers were in, their jobs were still
determined by their levels of teacher education. Big schools included all
levels of teachers, from those who had no teacher education at all to those
who had master's degrees. The variation of teacher certificate in the big
schools was greater than the variation in small schools.

The test of homogeneity of the variance-covariance matrix was not

significant for the four subscales of the actual needs but was significant for

82

the perceived needs. The univariate repeated measures analysis was applied to
the actual needs and the multivariate test to the perceived needs. The
three way factorial design (2x2x2) with.four repeated measures were
analyzed by the FINN program. The design was crossed and balanced.

The analysis of the actual needs showed that the interaction
among location of schools, level of teacher education, teaching experience,
and the subject matter was significant. Descriptive followbups showed that
the less experienced teachers holding only the teacher certificate who worked
within the university area and the more experienced holders of the higher
teacher certificate who also worked within the university area had the most
contrasting profiles. The four way interaction was explained by the
significance of the location by teacher education by teacher experience
interaction on the contrast variable of Test Planning and Reporting Test
Scores. The three—way interaction was not significant when Test Planning
was analyzed individually, but the interaction was significant for Reporting
Test Scores alone. In the presence of the fourdway interaction, the
subject matter by the level of teacher education by experience interaction,
the subject matter by experience interaction, the subject matter by location
interaction, the subject matter main effect, the location by experience
interaction, and the teacher education main effects were significant
at a - .05. All interactions were presented in graph form.

The F ratio for testing two hypotheses concerning the main
effects were significant. They were the subject matter main effect and
the teacher education main effect. On the general profile actual needs
in Reporting Test Scores was the highest need (lowest scores from the
test) followed by Item Analysis and by Test Planning with Item Editing

the lowest. However the general profile could not be made applicable to

83

all groups of teachers because the fourdway interaction was significant.
The significance of the teacher education main effect was followed by

the notion that the teacher certificate holders had more actual needs than
the higher teacher certificate holders.

0n perceived needs, only three null hypotheses were rejected.
The subject matter by location interaction was studied along with the
subject matter main effect. For teachers in the university area, the
more experienced teachers had perceived needs ranging from highest to
lowest in the following order: Item Analysis; Reporting Test Scores;

Test Planning; and Item Editing. But for those outside the university
area, the perceived needs on Item Editing was second, followed by Reporting
Test Scores and Test Planning. In general, the factor of Item Analysis
had the highest scores in percieved needs. The other null hypothesis
rejected was the location of schools by teaching experience interaction.
The interaction was disordinal which could be explained by the observation
that less experienced teachers within the university area had lower
perceived needs than the more experienced teachers within the university
area but less experienced teachers outside the university area had higher
perceived needs than more experienced teachers outside university area.
The difference between perceived needs of all (both inside and outside

the university area) of the less experienced teachers was smaller than

all of the more experienced teachers.

Further comparisons were made to find whether there was any
difference among levels of the following independent variables: having
attended the in-service program; sex; favoring a national testing program;
and grade taught. On total perceived and actual needs, a difference

was found on the grade taught variable for total actual needs only.

84

It was indicated that the higher the grade taught, the higher the test
scores and hence the lower their actual needs. The same observation was
true for actual needs in Item Editing but not for the other three subject
matters. The score in actual needs for Item Editing also differed between
teachers who had attended the in-service program and those who had not,
even though no difference was found in the total needs. The means
indicated that those who had attended had lower actual needs.

Between those who favored a national testing program and those
who did not, the differences were found on Test Planning perceived needs
and Reporting Test Scores actual needs. The means showed that teachers
who favored a national testing program had higher Test Planning perceived
needs than those who did not. For Reporting Test Scores actual needs,
teachers who favored national testing needed less. There were no

significant difference at all between male and female teachers.

Conclusions and Implications

The descriptive information supported that the random sample
was very similar to the total population. This suggests that the bias due
to the substitution of the samples in the outside-university area is
minimum. The smaller teacher-student ratio for the sample as a whole was
contributed by the larger proportion of central amphur schools (university
area) than in the total population. Most of the large schools, all of
which.were in the university area, had lower teacher—student ratios.
Most of the schools outside the university area had much higher
ratios because of the difficulty of getting teachers in the remote
areas. Also, some of these remote schools are one-room schools.

Sixteen percent of the 320 teachers in the sample had experience

85

in the in-service training program between 1961 and 1973. .Seventy-

one percent of these fifty-one teachers held at least the higher teacher
certificate, and sixty-five percent were more experienced teachers,

who had been teaching for at least eight years. The conclusion could

be drawn that in those years the program of in-service training was
given to those who had already had more teacher education. Because

of the nature of the subject matter, with some mathematics involved,

the holders of only the teacher certificate and those teachers who

had not had any teacher education at all may have turned down the
invitation to join the program.

A cross-tabulation between responses on attended/did not
attend the in-service program and favored/did not favor a national
testing policy revealed that of fifty-one teachers who did attend the
in-service program, thirteen teachers favored a national testing policy.
Thirty-eight teachers who attended the in-service program were unfavorable
to a national testing policy. Twenty-six percent (thirteen out of
fifty-one) of the teachers who attended the in-service program
favored a national testing policy. On the other hand, thirty-eight percent
(102 out of 269) of the teachers who did not attend the in-service
program reported favoring such a policy.

In the past, the in—service training program emphasized
item editing in an attempt to improve the quality of the teacher-made
test. The results of this attempt were reflected in the lower needs
on both the actual and perceived needs tests than were found in other
subject matter, i.e., lower than Item Analysis within percieved

needs and lower than both Reporting Test Scores and Item Analysis in

86

actual needs. Because of the interaction of school location, teacher
education, teacher experience and subject matters, the needs for each
group were different. Therefore, the subject matter should be arranged
according to the needs of a majority of teachers in each training
session. The results of this study seem to indicate that it would be
appropriate to emphasize item analysis procedures in future in-service
programs because it was the area with the highest perceived needs score
and was next to the highest in the test of actual needs. Further, item
analysis is the next step in improving tests after the items have been
written and tried out.

In the Item Editing subject matter, teachers in the higher
grades had less actual need than those who taught in lower grades.

This may suggest that in the past, both in the in-service training
program and in teacher-training curriculum, the emphasis was on item
editing, i.e., writing the test questions more than any other area
of subject matter in measurement and evaluation.

In testing the difference between teachers who had attended
and had not attended the in-service program, no significant difference
was found between these two groups. But the data showed that there
were some slight differences between them. In perceived needs,
teachers who had attended the in-service training program felt more
need than those who had not attended the program. This result might
suggest that those who attended the program seemed to realize that
there were several things that they must know about tests and
measurement. The data, also, indicated that teachers who had

attended the training program had higher scores on the actual needs test

87

(meaning that they had lower needs) than who had not attended the program.
And, again, the significant difference was found in the Item Editing
subject matter.

One of the nationdwide programs conducted by Chaval
Paeratakul during 1965 included spending two days out of the five-
days sessions on test item editing alone. The activities included
theory in writing good test items with examples of good and less
good items. Then each teacher had to write some test questions and
criticize them item by item. Another piece of evidence
to support the position that there is less need in Item Editing subject
matter was that in Chaval Paeratakul's measurement textbook, the first
book in Thai language about measurement, the discussion about item
editing subject matter takes four out of the nine chapters.

The study also suggests that there should be a compulsory
program requiring the holders of teacher certificates and teachers
with no formal teacher education to attend the training sessions,
especially those who have been teaching for at least eight years
outside the university area and for new teachers within the university
area. Special attempts should be made to give service to these two
groups of teachers. The high actual needs for these two groups may
be the result of a lack of interest and or of heavy loads or of routine
teaching. Most of the teachers studied taught in the lower elementary
grades (1 to 4), taught all subjects and for about 5 hours a day.

They apparently found no time to study or to pay much attention to

other professional activities.

88

Recommendations for Further Study_

Since the previous in-service training seemed to yield
little benefit to the teacher certificate holders, perhaps because
the work may have been too hard and conceptually the subject matter
too difficult for them, there should be a study of any future in-service
training to find out which group of teachers benefits the most from
the training. Another attempt should be made to find suitable ways
of presentation and interpretations of subject matter for those who
have had no teacher education at all. An experimental study should be
done on the effect of the in-service training on actual measurement
in the schools. Each year the Department of Education in Khon Kaen
has about 120 positions Open for graduates from the teacher
training institutions. These new teachers might be randomly assigned
to training and no-training groups. The no-training group would
then be sent directly to the assigned schools. The training group
would first go through a training program and then be sent to assigned
schools. The followbup study could then compare the quality of the
tests these new teachers made. The efficiency of the training program
might be evaluated from the proposed study.

Another experiment might be done to investigate the benefit
from the training service. The gain of knowledge after the training
should be studied to compare the gain made by holders of the teacher
certificate with the gain made by the holders of the higher teacher
certificate. The quality of the tests they use should also
be investigated to discover the extent to which the new
knowledge they have acquired is put into practice, and under

what circumstances the application is facilitated.

89

If there is no improvement in the quality of the tests the in-service
training is in vain. The effort and money that was put on the program
were wasted.

The application of the test results has been highly important
in the Thai educational system. One test result might determine the
students' whole future. If the test was not valid and reliable,

a capable student might be forced to drop out of school. In order to
evaluate students' abilities based on valid and reliable measures,
sound national testing programs should be developed and put into
practice. Standardized tests should be constructed and be available
to all teachers. Funds should be provided by the Ministry of
Education to construct national norms. A National Test Bureau should
be established, to be a center of testing services and to carry

on all educational testing jobs. Programs of in-service training
could be part of the function of this National Test Bureau. Different
programs should be established for different groups of teachers,
depending on their needs and the levels of their education. For
example, teachers with minimum educational background should be
provided with basic information about how to construct a test, to
write good items, and to interpret the test scores, since these
subject matters do not require much knowledge of statistical procedures.
Item analysis and transforming scores should be tasks done for them by
personnel at the local educational offices.

Since the four-way interaction (subject matter by location
of schools by level of teacher education by teaching experience) was
found, yet needs of teachers in each group were different in some

way, different programs of in-service training must be planned for

90

particular groups of teachers and these programs must be flexible.

That is, it is desirable that before planning an in-service program,

a pilot study he conducted to find out the needs of teachers in the
particular location and the level of their knowledge about the subject.

Besides the standardization of tests and administration
of the in-service program, the National Test Bureau should provide
an information service, either through the preparation and distribution
of a teachers' handbook on test construction, or through a quarterly
journal on educational measurement and evaluation, or desirably through
both.

The College of Education at Khon Kaen University should add
another dimension to respond to teachers' testing needs. As the closest
resource, the university can provide an intermediate help to all
teachers in the province. The services can be as individualized as mail
corresponding, providing a practical suggestion to individual teacher's
problems. The university test and research center should offer services
such as consultation and information services. Graduate students
majoring in Measurement and Evaluation should acquire most valuable
experience through participation in this program.

A summer work-shop is another service which the university can
provide to teachers. Of course content and structure should be drawn
carefully based on participants' characteristics and needs. The session
should be about one to one and a half months long. Practical considera-
tions would be more important than a theoretical orientation. Before
the participant teachers leave the campus, they should be convinced that
they can get help from the staff any time on any measurement problem that

they may have in the future. They should leave with the feeling that

91
the university is trying to help all teachers, both participants and non-
participants. The university is the place they can turn to when they

would like to have their problems solved.

APPENDIX A

Directions:

PMT I 0

APPENDIX A

THE INSTRUMENT (FIRST EDITION)

You need not identify yourself, but please answer each of

the questions as accurately and as honestly as you can.

Please mark X in the apprOpriate columns for each item.

Do you think you need to know more about these tasks?

Choosing the type of test to use
Determining the length of the test

Deciding what ability should be emphasized
in the test

Determing the ratio among questions that
call for memory, application, and
reasoning.

Determining the number of items
for each topic

Administering the test
Setting the difficulty level
Arranging the items into the test

writing good item stems (in multiple-
choice item)

Writing reasonable alternatives
(in multiple-choice item)

92

Urgent Need

Very Much Needed

Some Need

Little Need

Don't Need

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

ll.
12.
13.
14.
15.
16.

17.

18.

19.

20.

21.

22.
23.
24.

25.

26.
27.
28.

29.

93

Writing good true-false items

Scoring essay tests

How to measure in the cognitive domain
How to measure in the affective domain
How to measure in the psycho—motor domain
Writing better items

Understanding the major uses and
limitations of item analysis data

Selecting good items.

Computing and understanding the item
difficulty index

Computing and understanding the item
discrimination index

Understanding the difference and
relationship between difficulty and
discrimination

Finding the validity of a test
Finding the reliability of a test

Revising items

Reporting and writing recommendations
to the principal and superintendent

Choosing the type of test scores
Computing and understanding T-scores
Computing and understanding percentiles

Assigning grades to students

Urgent Need

Very Much Needed

Some Need

Little Need

Don't Need

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

30.

31.

32.

94

Drawing and interpreting profiles

Using norms

Distinguishing between norms and
standard

Urgent Need

Very Much Needed

Some Need

Little Need

Don't Need

 

 

 

 

 

 

 

 

95

PART II Mark.X in front of the best answer of each.item.
1. In evaluating an achievement test for possible use in your school,
the most important issue is whether the test

a) is based upon a thorough survey of teaching practices
over the whole country.

b) matches in content the objectives of instruction in
the school.

c) has sufficiently high predictive validity.
d) is convenient to administer and score.
e) has sufficiently high reliability.

2. The most important concern in judging how well a test samples
a content area is its

a) completeness.
b) correlation with a criterion.
c) representativeness.
d) variety.
3. What is the first decision made by the test constructor?
a) Characteristics of students taking the test.
b) Number of test items to use.
c) Types of test items to use.
d) Difficulty of the test.
e) Purpose of the test.

4. In constructing a table of specifications for a classroom test,
the number of test items allotted to each cell should be

a) varied each term, so that students are unable to
detect the amount of emphasis to be given to each
area in the test.

b) approximately equal, so that no particular objective
will receive undue emphasis.

c) balanced between knowledge and higher level learning
outcomes.

d) determined by the emphasis to be given during instruction.

96

5. Which of the following will not be accomplished by constructing
a test blueprint?

a) help to clarify instructional objectives.
b) improve the validity of the test.

c) relate evaluation to teaching.

d) help to "balance" the test.

e) save construction time.

6. Which one of the test attributes is most likely to be use
of a table of specifications?

a) Difficulty
b) Objectivity
c) Reliability
d) Sampling
e) Usability

7. Which type of item format provides greatest content sampling in
the shortest possible testing time?

a) essay

b) matching

c) multiple-choice
d) short-answer

e) true-false

8. Which of the following is not a purpose of the classroom
achievement test?

a) Measuring attainment of instructional objectives.
b) Diagnosting individual learning difficulties.
c) Judging relative mastery of basic knowledge.
d) Assessing relative curricular effectiveness.

e) Assessing instructional effectiveness.

97

9. The best source of possible alternatives for a multiple-choice
item is the

a) explanations of poor students in the class.
b) teacher's knowledge of the subject matter.
c) questions of bright students in the class.
d) textual material assigned to the class.

10. Multiple-choice items written in the form of questions
a) assume complex reading skills.

b) require no more detailed alternatives than do
incomplete statements.

c) have greater clarity than items phrased in
incomplete statements.

d) lend themselves to easier detection of
construction flaws.

11. It is good practice in multiple—choice, single-correct
answer test items to

a) randomize the order of the answers, putting the
correct alternative in each position an approximately
equal number of times.

b) put the correct answer last or next to last more
often than in any other position.

c) let chance take care of the order of alternatives
within the item.

d) arrange placement of the correct answers according
to some predetermined scheme.

12. "None of the above" is an appropriate distractor for a multiple-
choice item in cases where

a) the other distractors are partially but not
absolutely correct.

b) the best answer is an absolutely correct answer.

c) the number of possible responses is limited to
two or three.

d) a large variety of possible responses might be given.

e) guessing is apt to be serious problem.

98
13. Which, if any, of the following should characterize the stem of
a good multiple-choice test item?
a) It should be a complete direct question.

b) It should be the subject portion of the proposition
the item is intended to test.

c) It should express or imply a specific question.
d) It need not have any of the foregoing characteristics.

14. How should the question and the responses of a multiple-
choice test item be changed to make the item easier?

a) question more specific, responses more alike.
b) question more specific, responses more diverse.
c) question more general, responses more alike.
d) question more general, responses more diverse.
15. With respect to the multiple-choice test item a good distracter is
a) only an untrue statement.
b) only irrelevant statement.
c) both untrue and irrelevant.
d) either untrue or irrelevant.

16. Within a given item format, the items of a test should be
arranged by

a) subject-matter content.

b) increasing level of complexity.
c) level of reading difficulty.

d) a random pattern.

17. If experts agree on the correct answer to a test item, the
item is

a) fair.
b) objective.
c) reliable.

d) specific.

99

18. Fundamentally, item discrimination indicies attempt to indicate

a) the ability of items to distinguish between persons
having different levels of that characteristics
which.the test measures.

b) how well the items can be distinguished from one
another within the test.

c) the degree to which items from different tests
can be discriminated.

d) the ability of the item writer to develop items
which are readable.

19. If three good students out of four answer a question correctly,
what is its index of discrimination?

a) 1.00
b) .75
c) .25

d) Insufficient information is given to answer the question.

20. It is likely that item discrimination will be at a maximum
when the item difficulty index approaches

a) 1.00
b) .75
c) .50
d) .25

e) 0.00

100

Items 21 through 24 are based on the following data:
A class of 200 students was divided at the median into a high
and a low group. Their responses to five 4 choice multiple-choice

items are tabulated below. Correct responses are underlined.

Number of students marking

Item Number Grogp Response 1 Response 2 Response 3 Response 4

1 High 89_ 20 0 0
Low ‘45 3O 20 5
2 High 0 0 199 0
Low 0 15 '85 O
3 High 35- 10 10 45
Low 55_ 20 0 25
4 High 20 5 62_ 10
Low 20 25 35 20
5 High 30 30 0 '49
Low 5 55 5 32.

Use the following KEY to answer each of the next four items:

KEY: a) Item 1
b) Item 2
c) Item 3
d) Item 4
e) Item 5

21. Which item discriminated best between high and low groups?
22. Which item has the highest index of difficulty?
23. Which item has difficulty index of .50?

24. Which item shows negative discrimination?

101

24. The most easily defensible use of tests in schools is for

a)
b)
e)

d)

accreditation of schools.
motivation of study.
teacher evaluation.

pupil guidance.

26. If a pupil had a score at the median of his class, we know he

a)

b)
e)
d)
27. Norms are
a)
b)
e)
d)

is at a score halfdway between the lowest and
highest score.

has a score higher than those of half his classmates.
is doing acceptable work.

has a score at the mean.

lgag£_likely to remain current for which type of test?
Achievement

Aptitude

Interest

Projective

28. On test XYZ pupil A scored at the 80th percentile and pupil B

scored at

the 40th percentile. What statement may we make

relative to their abilities?

a)

b)

c}

d)

e)

Pupil A has twice the measured ability of Pupil B.
Pupil A's test score will be twice that of Pupil B.

If Pupil A's score is superior to those of 70 pupils,
B's score is superior to those of 35.

Percentile ranks are meaningless in comparing the
scores of A and B.

We need to know more about the distribution of scores
before we can compare the two percentiles.

102

29. A distribution of linear s-scores is always

a)
b)
e)
d)

e)

similar to the raw score distribution.
a rectangular distribution.

a normal distribution.

a peaked distribution.

a skewed distribution.

30. A person whose z-score was zero would have a raw score equal to

a)
b)
e)
d)

zero a

the mean.

the median.

none of these.

31. If one desires to give equal weight to several tests when
combining scores, he will obtain the best results if he uses

a)
b)
C)
d)

e)

percentages.
percentile scores.
raw scores.
standard scores.

any of the above--it makes no difference.

32. Is it wise to base a pupil's grade on his achievement in
relation to his ability? Why?

a)

b)

C)

d)

Yes, because achievement depends on ability.

Yes, because it gives every pupil an equal chance
to make high.marks.

No, because the ability required in a particular course
is difficult to assess properly.

No, because to do so simply intensifies the competition
for high marks.

APPENDIX B

THE INSTRUMENT (FINAL EDITION)

To my colleagues who teach in Khon Kaen Province schools,

I am studying for my doctor's degree at Michigan State University,

East Lansing, Michigan. The last requirement is that I do a research
project. Because when I have finished I expect to return to the Faculty
of Education at Khon Kaen University, and because Khon Kaen is my home
city, I am studying whether the teachers in our province need and want

help in making and using classroom tests.

The Changwad Education Officer and the Amphur Education Officers are
interested in helping in this study, and I am very grateful to them.
Your Amphur Education Officer is distributing the questionnaires and
will return them to me. But nobody except me will know just how you
answered. When you have made your answers, put your questionnaire in
the enclosed envelope, £331 it, and hand it to your principal, who
will start it on its way back to me. Do not sign your name on the
envelope. Answer each question as accurately and as honestly as

you can. Your responses to this questionnaire will not be identified

to you individually, therefore do not worry about being wrong.
If my study is to be acceptable and later become useful, it is important

that you take this task seriously and I get back all the questionnaires.

It will take about thirty minutes for you to fill out the form.

103

104

Thank you very much for your help. The findings will probably make

the basis for designing a course or workshop to be offered to

all teachers.

Cordially Yours,

Sor Wasna Pravalpruk

105

Directions:

You need not identify yourself, but please fill in all
information on this page. This information is for research purposes
only.

Answer each of the questions on this and the following pages

as accurately and as honestly as you can.

PART I Please fill in the blanks and check the appropriate categories.

Name of school
Amphur
What grade are you teaching?

What subjects are you teaching?

 

How many teachers are there in your school?

How many students are there in your school? (approximate)

 

Your sex B Male B Female
Level of teacher education:
[:J Cert. and lower
[:1 Higher cert. and higher
Years of teaching experience:
[:J Less than 8 years

[:1 Eight years and more

106

Did you attend the Measurement and Evaluation
Workshop program during 1966-1968?

D Yes D No

Are you in favor of national testing?

[I Yes D No

107

PART II Please mark X in the apprOpriate columns for each item.

Do you think you need to know more about these tasks?

8
'o
o
o
'o :z 'o
o o 'o
o .o -o o o
z: o o :z o
n o :z
u :8 :z o
8 63.“
SE83“
semaa
1. Choosing the type of test to use
2. Determining the length of the test
3. Deciding what ability should be emphasized
in the test
4. Determining the number of items
for each topic
5. Administering the test
6. Setting the difficulty level
7. Arranging the items into the test
8. writing good items stems (in multiple-
choice item)
9. Writing reasonable alternatives
(in multiple-choice item)
10. Writing good true-false items
11. Scoring essay tests
12. How to measure in the cognitive domain
13. How to measure in the affective domain
14. How to measure in the psycho-motor domain
15. Understanding the major uses and
limitations of item analysis data

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

16.

17.

18.

19.
20.
21.

22.

23.
24.
25.
26.
27.

28.

108

Computing and understanding the item
difficulty index

Computing and understanding the item
discrimination index

Understanding the difference and _
relationship between difficulty and
discrimination

Finding the validity of a test
Finding the reliability of a test

Revising items

Reporting and writing recommendations
to the principal and superintendent

Choosing the type of test scores
Computing and understanding T-scores
Computing and understanding percentiles
Drawing and interpreting profiles

Using norms

Distinguishing between norms and
standard

Urgent Need

Very Much Needed

Some Need

Little Need

Don't Need

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

109

PART III Mark X in front of the best answer of each.item.

1. In evaluating an achievement test for possible use in your school,
the most important issue is whether the test

a) is based upon a thorough survey of teaching practices
over the whole country.

b) matches in content the objectives of instruction in
the school.

c) has sufficiently high predictive validity.
d) has sufficiently high reliability.

2. The most important concern in judging how well a test samples
a content area is its

a) length.

b) correlation with a criterion.
c) representativeness.

d) variety.

3. In constructing a table of specifications for a classroom test,
the number of test items allotted to each cell should be

a) varied each term, so that students are unable to
detect the amount of emphasis to be given to each
area in the test.

b) approximately equal, so that no particular objective
will receive undue emphasis.

c) balanced between knowledge and higher level learning
outcomes.

d) determined by the emphasis to be given during
instruction.

4. Which of the following will not be accomplished by constructing
a test blueprint?

a) improve the validity of the test.
b) relate evaluation to teaching.
c) help to "balance" the test.

d) save construction time.

110

5. Which type of item format provides greatest content sampling in
the shortest possible testing time?

a) matching.
b) multiple-choice.
c) short-answer.
d) true-false
6. Multiple-choice items written in the form of questions
a) assume complex reading skills.

b) require no more detailed alternatives than do
incomplete statements.

c) have greater clarity than items phrased in
incomplete statements.

d) lend themselves to easier detection of
construction flaws.

7. It is good practice in multiple-choice, single-correct
answer test items to

a) randomize the order of the answers, putting the
correct alternative in each position an approximately
equal number of times.

b) put the correct answer last or next to last more
often than in any other position.

c) put the correct answer first more often than in
any positions.

d) arrange placement of the correct answers according
to some predetermined scheme.

8. Which, if any, of the following should characterize the stem of
a good multiple-choice test item?

a) It should be the subject portion of the preposition
the item is intended to test.

b) It should express or imply a specific question.

c) It need not have any of the foregoing characteristics.

d) Both (a) and (b).

111

9. To make a multiple-choice item easier, we should make the

a) question more specific, responses more alike.
b) question more specific, responses more diverse.
c) question more general, responses more alike.
d) question more general, responses more diverse.
10. Within a given item format, the items of a test should be
arranged by
a) decreasing level of complexity.
b) increasing level of complexity.
c) level of reading difficulty.
d) a random pattern.
11. Fundamentally, item discrimination indicies attempt to indicate
a) the ability of items to distinguish between persons
having different levels of that characteristics
which the test measures.
b) how well the items can be distinguished from one
another within the test.
c) the degree to which items from different tests
can be discriminated.
d) the ability of the item writer to develop items
which are readable.
12. If three good students out of four answer a question correctly,
what is its index of discrimination?
a) 1.00
b) .75
c) .25
d) Insufficient information is given to answer the question.

112

Items 13 through 15 are based on the following data:
A class of 200 students was divided at the median into a
high and a low groups. Their responses to five 4 choice multiple-

choice items are tabulated below. Correct responses are underlined.

Number of students marking

Item Number Group Response 1. Response 2. Regponse 3. Response 4.

 

1 High 99_ 20 0 0
Low .3; 30 20 5
2 High 0 0 $99 0
Low 0 15 99_ O
3 High 99. 10 10 45
Low .99 20 0 25
4 High 20 5 99' 10
Low 20 25 35 20

Use the following KEY to answer each of the next four items:

KEY: a) Item 1

b) Item 2
c) Item 3
d) Item 4

13. Which item discriminated best between high and low groups?
14. Which item has difficulty index of .50?

15. Which item shows negative discrimination?

16.
a)
b)
C)
d)

17.

113

The norms for which type of test need to be revised more often?

Achievement
Aptitude
Interest

Projective

On test XYZ pupil A scored at the 80th percentile and pupil B
scored at the 40th percentile.

What statement may we make

relative to their abilities?

a)

b)

C)

d)

18.
a)
b)
C)
d)

19.

Pupil A has twice the measured ability of Pupil B.
Pupil A's test score will be twice that of Pupil B.

If Pupil A's score is superior to those of 70 pupils,
B's score is superior to those of 35.

Percentile ranks are meaningless in comparing the
scores of A and B.

A person whose z-score was zero would have a raw score equal to

zero a

the mean.

the mode.

none of these.

If one desires to give equal weight to several tests when

combining scores, he will obtain the best results if he uses

a)
b)
C)
d)

percentile scores.
raw scores.
standard scores.

any of the above--it makes no difference.

114

20. Is it wise to base a pupil's grade on his achievement in relation
to his ability? Why?

a) Yes, because achievement depends on ability.

b) Yes, because it gives every pupil an equal chance to
make high marks.

c) No, because the ability required in a particular course
is difficult to assess properly.

d) No, because to do so simply intensifies the competition
for high marks.

115

4 s v '
quu Lwauh$1u30u3nsauunu
V O V V A '4
93011033080143 LNOTUUTiyfiyfi Lanap‘m Michigan State University
v A é V ¢ V a 4 a t d A V

1N8N7§aLuin3 33u uqnm38 3zmsqm3saflimm3flwu3 L 80330LN03003130N3083
QV o d G v ' ( ' V Q Qt]
00u3zlflm3a3u uu33m038008uunfilunmzflns3fi3887 unzisuunutflun3utnnmsqn0u
4‘ o J V v v 4 V V a/
0 u$083n3ifins3170303300040311838g1u3au30883133108300031080108000883008
fithuu

av V ' 4 ' a v u o é ' V V ' 4

00u1008833n13nua33083ufins3003330N3nuazs3Lns 0W3u1018033073uus

I y d v I a o V v V
1u18730330 mamanwizqmm3ul3m 5030 N3ufins3fin3783tnsuaz33u3hszflupuvnuuu
V t I O . U V t v
80003nfllun3nququuinuu3ungluy Ldsw3umsun303uw0uu01873083 020383180;
' a a u v A V ' V v '
lupuiaflns3fin3ia310007030N301w03:10:3ui3uaqlmfisunslfl
V ' I
unnaau03unnfi3zlm17003ifinu300:39000000uLM3fih 08:3:1003330303-
I I iv A I l v 1 V v
fiOUflBQMfiUlEuTWUUhhfl N3ulnmsqLfluu30080M3u 108010070090833N83031uniuwnma
' e
Lm3uu
yd a é Ru 4 V
unnaan03ufl m man man m 8333:10u333031u3nlfl 000u010w0383u
4 v vv V' v ‘v
NfihLaualiluuaufiqn 301NM3uWU383nnaflluuunnn08
A (I w v v I

0808303130831700538flflizluaun0L0038Nu30038833N13nfislu031080000

V A 7 V 4 V“ q;
80003Nfiwnnannw0um3u033nfinnm083w3uuazaquuuasnu3nnulwhen 10800w1:q0flu

I 4 ' V
n33n73un0300M3un3m10038003u

v v 4
OfiflflfifilJulJO’fl

(
( 8.338u3 07:338quu )

116

033030
. v vv V I ‘ ‘v‘
lfl7000u0303nw0001030130 Nhn33un3031uuuuasnu3nflvz13100313083

V

a ' ‘4
VVULMWUN

.v ' '4
OBUH o n§m3tmu00033u1u300330 “$003 OQIHTBQMLHNWZHM

031n07h3umhn3700u ........... ....
ibdasu ..................
333280u .................
33u3u0§10130178u ........

3hu3ufih170u (fli:N3m) ........

g 1
twfiNBQM3u I 338 003
' 1.1.

308; l flnfi. 87003033

 

 

 

 

 

I
flnfl.§0 flfqg3 07080n33

1|

 

 

 

V I

I
33u3u07m3u80u u08033 a 0

 

 

 

I
a fluazu3nn33 a 0

 

 

 

I V

V I
N3u188LN3001N0373008030N781083

 

L80 1300 ..........

 

 

 

 

I
Tutnu

 

I y K ‘ I
M3u3001ufin3180073um007:Lmfiufaiu

I
D 31011 1113811

 

 

 

 

117

I I Q 4 v A I v] I y I I V
fiflflfi b W1UflU1fi953fiBfilfiMLWULflflfi UL?BQWB1U§M§ULMLWUQ1fi fiWMWUQWUWWBGfifli

V I V I
uwndqfi Tunw ‘fk‘aalumaqfi 9 (WBQfifiTufififififi) uazd b,m,€, :aqaalflvufiqmaqfi a

4 "
haiumaqnwiLau

v 4 IV
maqnniuwnmqm TumaqnwiLau

 

9 b 1 m C C

 

y I
9. fifiTlaflfiflleflMfifldflflfianLfiu fliuuu?aamuu

 

V
b. naififiuunhdwnu11%BQ%aaau

 

V V V
m. meaBUfidiluuaniinnwwmwaauaaaziiunq

 

. I V V V 4V
¢. n1inwuufidquflaTflhaiv:aanwaaaunwa

 

 

 

 

a4 -‘1
a. fifififiiflufifln
u y «4
b. nninfiuufifiqfinuwnqwuwaamaaau
‘ V I V V an .1
d. n1?Liuqmaaauumazmatfluwaaaumquzun
d. unitfiuufifififim1u1uflaaanuuuLfianmau
-

 

v 4 V 4
@. nnitfluumuLaaanmaaauuUULaafimau

 

V
90. nwitfiuuflaaanunugnaw

 

V V
99. nwiquhzuuumaafluunnamuu

 

V I 6’
9b. nfiiaflnmaaandhflmqw LmuWuamwamuwfiwaw:

.--...._. ..qb

 

v V 4 v a f v 4
9m. finiaflmuwugfinuiammuflwflnqunLiuu

 

I
9;. fifiifiIfihB: Lfiu1u3%1Wfi:fifiB1 ummfihufi

 

(V

V I
9;. manwuufingLmuwflun1i1Lminzumaaau
F...w--_r--‘. _-. -.M... .- ...-___....-..-_.

I
vb. fifiTHHUQWHfiSfiIWMMNfiflflfldflfifiuflfifidfiflfiflI

flfififln

‘p. ‘n-w-q‘

 

9w. hwififiudmuazmuquuunuwaqafiuwvafiuuumaa

V
flaflflu

 

9d. fiQWNUWfiWWITSMQWIfiQWNUWflIWUUflZBWUfi?

V
WWQUfiflBGflflfiBU

 

 

 

 

 

 

 

62$.

b0.

b9.

biz).

lob.

5x

118

V 4
maqn1$n1nmqm

V
nwimqhuwnLfiuamiqwaqmaaan
L v
nwrunhdwuLéauuwaqmaaau
v V V
nqrflinflgaunimflaaau
I
nwinranagmflizvfifififihL?uuua:iwuawuma
I
afluq
4 V 4 .
finitaafi1fififlmfiaamzuuuLwaLauauanwiaau
fin:fifiuqmuazmawnuuwumaafizuuutr
0’ G’
nwififiudmuazhawuunwufiaqtUa:tmuima

( Percentile )
I

V
n17LflfuuLfiunhawnafiuninluhwunWQflwaa
fih1?uuTnuTmnwiwq Profile

v I ‘
iwwflmtnmmuynm ( Norm )
V I

I I a
fiaumnmfiqizudwoanmflifimuaznwmigwu

( Standard )

Iv
Tumaqn1$Lau

 

9

19

€

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

WEUé m

I 4 I
M1UlMU7WQfi

. 4' V
fiamfiflBUfiWfififlfififififiIfififihMMfiUMTWU WUWUWMWBUWfififl

119

V I

v d v 4' 4 I
9. maaan w:1aa11nfihqmau61u7:qLiuuwaqmwumaiv:

I
whiufiafiflgwnmutfiuauhai
u 4. '1" I Q Id
dfinwuwnwuun aquthemnwumaaufiwuuq
flflawntfivamiqga

I
fifl??fll%flfifi§d

I V

InnaifivwimnawmaaBUhiaumanuLfiauwfimwLfluqlfififi Liwgfi

fl.

V
mawnuwamaqmaaflu
V( V v (45 V
mdwufihwufiflEQfiaaaunungtnmmmm011
v V I“ I”,
fidwuLfluwaumumflawaaanmadmwuu

I I v

V I
flaaauawumazflaflfidwuumnmwqfifiquLflauw

a v4"
1UfifiTfifiVUfififi1uum83flflufiBQfififiuddflflfif?tflflfififlfiflnfififl fi???:

H.

I V VI V I V
tfléuunnfiffififlfiniaautfifliulufihL?uugdwt§34fl9zfiuwn1ufiaaaanwfima
I I
flLmqfififiTuumazmauflaqLflau13mw
I V V
maiw:flfihmenfifiiuflwufiawuafiuaznwiqmLufiua

V
haivzfifiuunThUflggaau

. a r V t v I v v
fiWTMWWWTWQQLfiTfitMMfifiQflTfi38H$1UfiWUWBTUfi UfilIUIBTMU

I yv I
maqumaaauflhdwuLfiuqmiafloéu

I v (
fiauiufidwnfihfihfiWufiqiL?uuua:nwiaau
I V V
fidunfiuumvfiuduwaquwaaan

I
mqufliznfithaw

V V V V V
maaanfififiiuudaaniwuwnmafiqufiUTfiLdawqumwimauuaufiqfi

H.

I
unnfihg
4
uUULaahmau
V
unnTutfiuufifimanfihq

uuughafl

90.

120

v v». A
InfinitfiuuflaaauuUULfianmau WfiflizlnhmnuwuLwaa:1I

fl.

vv ‘ v v I 4 v

quufiLiuumaqqmmqwuaIMIIHWunqInfiumfiIIMMMWUWuizmugq
V IV

mILfiamszauuTfififi iumaqafinqnnwn

v 4 4V v I

m1LaafivzLfihizwqunulfluawuuquaqugn

I V

. V e "V
nwuumuquaqmauuimmmLvuuwmaqnwimwnazii

V 4 VIII 0 I V 4 5 ° '
InmaaanunnLaammau fifi?1flfi1flfififlfl§fifl¥1Ufifl fi.fi.fi.d. M539. uumdinwuumauqqii

fl.

V° I V v I v V A v
TufiwmaugmaQIuma fi.%.fi.d.9. fiazmuTflmeflnfiIumaaauuuqaznu
v V. IV IV V IV a A
IflWmemanagmefl Lmuma I. WuuwhndwmuLaanaun

Iv IV

v V 0 V I v 4 A
IfiWuanmanagmufl Imuma n. Tuuwnmfiwmataanauq

v. V4 v I v #0 V
ImmwmangmiuLInIfiuuuuimunuuflquuuannqi n1num11

0 V 4 4 v '
mauquumaaanunuLaantUfiaiuanumzauwaT:

I.

V I
an?LfluflizfiwufleaflizTUhLIBWumfimanqumfluLfianLfluauuwunu
martfluflizTUhfimwnLaww:Lawzvq

V
LfiuflizTuaunnlmfilh

3
MI 0. “fit I.

4 VV 4 I I A . v
lWfl?:quflflfifluuUULfifififlflndfiflgufiifiLfin fiQTMWWM

OJ o ”1 v 4‘19“"
mauun.1quwa-aq fl uavmauLaan fifilfifldfiuufiflfl
I
fiBuUHD1NWHLW12fi01fl mazmautfianqunuuwnq
. VV 4 ”4.;
flauUWfifionfififid uazmauLaanTnaLmuonuuwnq

V V I
mauflfiuwulundwq uazmauLfiafimnqfiunqnq

V V

4 v I v v v 4 I
fifiTtTUIflBfiBDuflfitfiflLflfilfluflflfiBUMQfiSUUfifiiLTUQBUWIT?

4 V ‘1 I
IiuovqfiflanwnLflqwu
4 V I
Liuqaqmmflawuifluwfl
V
L?uqmwudfiWUIauw1—fih

I V
I?uqflquii51w

121

V
99. dfiquvfiuunmaqwaaaufiaa:17
It! I v v
n. manuawuqimmwzunqumL?uuaanmwui:fiumawufi1nnim

4W 'v "JVBV
fl. fififiufififlfiTQM?tunIflflfiflflflfifitflflflBHWWfiflflBUflqUfiflfiflUMQQ5UU

I V V

I v
fl. fififlflflfifififihumfimfiQfiadflflfiflhuflfi:flflflflfivfififlflflflflufififlflfithu

V v 4 v v Ivvvv“
a. hdqnanufiinmaqgaanmaaaumv:UiuflngaaauumaszWfiTanfi

I V V

V V d V v 4 I V I5 I4 ‘ .
9b. 1Hfiflfiflflflflu Ivfiflfiliflulfid m 1U é fiumaumauugn ufiMIWflBfiBUIBHUMflfiuW?

. I
IwuunLMIWI

v '4" V ' 4 4 4
I. usaanInsfi LWTfizflfldMTWU3810flulwuBfi
vv v I v 4
Immafiawuqufinq mauwfl 9m uq @é
a 4" 4 I I 4 I I I,
Tumuudquuntiuu boo flu uuqaanLfiu b nquwanquhzuuugquaznquhzuuumw

I v I

V V
nqua: 900 mu “WTWIIWIfifidfilflflfififlflflflfldfiflfiflnfififlLaflfiflflu 6 $3 (flfiflflflgfigflfififlah

 

 

4 v v
Mmfiu‘lm)
o v 4 i I v d
VWUUUUHLTUU NBUUWatfifilfiafi
V I v 4 v 4 v 4 v 4 v 4
fig fifiMUfiLiUu WILEBH fl. fidlfiflfl fl. “Ilfiflfi fl. Wfilfiflfi Q.
9. ad £2 b0 0 o
I
fifi gg mo be a
E. @G o o 900 o
I
fifi o 95 fig 0
m. 9Q m6 90 90 (K
I
fifi £5 b0 0 m:
6. I b0 C bf 90

2L-mm
_)

5

5‘

a

a I

be

122

VV 4 V I V 4
Ifimdtaanvwoawqfinanma 9m an 9c

90').

9‘0

9&0

9b.

0.

fieluufiafiuwvvfiuungamafi
V V I
waiuuflrthfiIIMUInqwu .60

V
maqufifiHuqvafiuunLfluau

(a v 4vvv'4
anmUInm (Norms ) flaqwaanuflizanTuummaqfliuflaquniwuaqun

H.

V

$3 9
V
fifl b
V
flfl m
V
flfl C
A

1

Jul-“if \ .1

 

 

V &

weaaue11n5MQMfilunwififiuw *
V

flflfiflnfififlfifiufifih

V v

Iaaau1mm11nauqa

V
fiafiflndhfihuha

V v d 4 V 4 f v. 0’ 4
9a. Tufiaaanuxnnu I Lfififififlfi. Tfiflzuuum do Lflaitmuuwa ( Percentihe ) Lmnmwum.

..V d K ( IV V I 6V V
Lflfizuuu do LUBTLIHIMfi nfiuififlafififiufifidfifid flflqfigfifiafl

fl.

%.

fl.

0.

I
fi.flfififiufifinfiifilflu b LMfiflBQ m.
I
fizuuumaq fi.Lfiu b tMfififiI %.
V V I A V I A
01 fi.1mh:uuu§qnfiwfiuau do flu %.I:Tfifi:uuu§qnawhuau ma an

I r I I I
fifiuuuqLflflitmu1me1ufifi11nuuwumanwiLfl?untfiunh:uuumaqunt?uu b mu

{1* ‘~ I ~ d1"T
9d. UWUfiLiUufiUM qufizuuu Z Lfiuguu LIIthuuumn ( hzuuu mvwnmaaau WUfiTG)

I
meTI

%.

H.

Q.

(
HUU
1|
' 4
tMfiTfiULQfiU ( Mean )
' 4 V 4
twwhzuuuwflmuihuwanfi ( Mode )

I
Infififimangn

 

123

I v V V I V V
95. hgfiuuflqmaafifiiiduhruuuvnnflaaauuawuazuuTfi81mamiwaduu?aufiuunmaqmaaau

u ' v V :1
flfifizunLMWfiu Lflfifi17931fifizuuufiflfiqfi$dvraM§fi
" 6'
n. tflaftmuima ( Percentile )
‘Zl. muuufiu ( Raw Score )
fl. muuummgqu( Standard Score )
V 4 v I. v I v
a. WmnzuuuunnquunTh InMWWufiqwuunfiuumnqunu

V 4 uAT 4u4.4u
b0. figfififvtlufltuuuufflLfflTlGrade UHLTUU fiUfi??fiUfidfiUWUfiLiUuMWLWUUfiUflfifiM

_ 4 v v A U 4 fl'
anuwimuiahdnnmumwaqunLiuumuuuuiflLu
I
fl. m3: LWTWSUflIWU%UB§fiDfl31Mfi1u1Em
V V I
fl. fiQT LWiqszTuunL?uuflTan1a¢v:Thinh:uuu§qmeflnu

v] A I 4 0’ 4 I I
Lflfi77 LWTfi:flififlfiquqTfiqu1fifiufidqufiflfiflTZHBUWIMHUUEU

I . ya IVA v
Q. 1Mh1? LWTWSMWWHLfifififiiflfldflUM?:1fifltuuugd

 

 

APPENDIX C

 

APPENDIX C

PAIRWISE PROFILE ANALYSIS 0N ACTUAL NEEDS

 

 

Sources df MS F P less than
All Groups

PL-RE 1 130.46 128.05 .0001

IE-RE 1 178.45 168.16 .0001

IA—RE 1 30.18 25.83 .0001
Location

PL-RE 1 0.35 0.35 .557

IE-RE 1 3.31 3.11 .079

IAeRE 1 1.70 1.46 .229

Teacher Education

PL-RE 1 0.04 0.04 .845
IA-RE l 1.50 1.28 .258

Teaching Experience

PL-RE 1 5.44 5.44 .022
IE-RE l 3.31 3.11 .079
IA-RE l 8.79 7.52 .007

Location x Education

IE-RE l 0.76 0.71 .399
IA-RE l 1.14 0.97 .324

124

125

PARWISE PROFILE ANALYSIS 0N ACTUAL NEEDS

continued
Sources df MS F Y 1888 than
Location x Experience
PL-RE 1 1.50 1 .47 .226 ’
IE—RE 1 0.01 0.01 .939 f____fi%
IA—RE 1 0.69 0.59 .443 a“ if.
Education x Experience 2
PL—RE 1 7 .44 7 . 30 . 007 ‘
IE—RE 1 6.80 6.41 .012
IArRE 1 11.82 10.12 .002
Location x Education x Experience
PL—RE 1 11.29 11.07 .001
IE—RE 1 1.06 1.00 .319
IArRE 1 0.98 0.84 .361
Within
PL—RE 312 1.019
IE—RE 312 1.061

IArRE 312 1.169

 

 

APPENDIX D

APPENDIX D

MULTIVARIATE ANALYSIS OF VARIANCE

ON FOUR SUBSCALES ON ACTUAL NEEDS

 

 

Sources df MS F P less than
Location 4 & 309 2.258 .063
Teacher Education 4 & 309 2.554 .039
PL 1 0.80 0.73 .393
IE 1 6.90 5.44 .020
IA 1 8.45 5.84 .016
RE 1 1.38 1.27 .261

Experience 4 & 309 2.549 .040
PL 1 0.05 0.05 .831
IE 1 0.25 0.20 .655
IA 1 1.25 0.86 .353
RE 1 9.45 8.69 .004

Location x

Teacher Education 4 & 309 1.219 .303

Location x

Experience 4 & 309 1.504 .201

Education x

Experience 4 & 309 3.065 .017
PL 1 1.25 1.15 .286
IE 1 0.90 0.71 .400
IA 1 4.51 3.12 .078
RE 1 7.50 6.89 .009

“"55.- u: m3,

 

127

MMLTIVARIAIE ANALYSIS OF VARIANCE

ON FOUR SUBSCALES ON ACTUAL NEEDS

 

 

continued
Sources df MS F P less than
Location x
Teacher Education
x Experience 4 & 309 3.618 .007
PL 1 3.20 2.93 .088
IE 1 2.28 1.80 .181
IA 1 2.45 1.67 .194
RE 1 8.78 8.07 .005
Within
PL , 312 1.092
IE 312 1.268
IA 312 1.447

RE 312 1.088

BIBLIOGRAPHY

BIBLIOGRAPHY

Ahmann, J.S., and Clock, MQD. Evaluating Pupil GrOwth. Boston,
Massachusetts: Allyn and Bacon, Inc., 1971.

Allen, Margaret E. "Status of Measurement Courses for Undergraduates
in Teacher-Training Institutions." Thirteenth Yearbook
National Council on Measurements Used in'Education, 1956: 67-73.

 

Bell, R.D. "The Risky Business of Classroom Testing." Business
Education Forum. 27 (October 1970): 8-10.

 

Black, H. They Shall Not Pass. New York: William.MOrrow & Co., Inc., 1963.

 

Bowman, A. "In-Service Training of Teachers in Measurement." The 18th
Yearbook of the National Council on Measurement in Education.
1961: 43-49.

Brimm, L. and Tollett, D.J. "How do Teachers Feel About In-Service
Education?" Educational Leadership. 31 (March 1974): 521-525.

 

Durost, N. "Problems in In-Service Training of Teachers in the Use of
Measurement and Evaluation Techniques." The 16th Yearbook of
the National Council on Measurement in Education. 1959: 31-33.

 

Dyer, S. "Testing's Nine Major Pitfalls." Senior Scholastic. 80
(March 14, 1962): 21T-23T.

 

Ebel, R.L. (Ed.) Encyclopedia of Educational Research. 4th ed., 1969.

Ebel, R.L. Essentials of Educational Measurement. Englewood Cliffs,
New Jersey: Prentice-Hall, Inc., 1972.

Ebel, R.L. Essentials of Educational Measurement: Instructor's Manual.
Englewood Cliffs, New Jersey: Prentice-Hall, Inc., 1972.

Ebel, R.L. "Improving the Competence of Teachers in Educational
Measurement." Clearing House. 36 (October 1961): 67-71.

 

Ebel, R.L. "Measurement and the Teacher." Educational Leadership.
20 (October 1962): 20-24.

 

Ebel, R.L. "The Social Consequences of Educational Testing." School
and SOCiety. 92 (1964): 331-334.

Findley, W.G. (Ed.) "The Impact and Improvement of School Testing
Program." (Article by Elizabeth Hagen and Lucille Lindberg)
The 62nd Yearbook of the National Society for the Study of
Education Part II, 1963: 232-253.

128

129

Findley, W.G., and Bryan, Miriam M. "Tests Like Handguns..." (From
Education Daily, January 20, 1972). 'Phi Delta Rappan
53 (March 1972): 442.

 

Gerberich, J.R. "The Development of Educational Testing." Theory
Into Practice. 2 (October, 1963): 184-191.

 

Goslin, D.A. Teachers and Testing. Russell Sage Foundation, 1967.

 

Grobman, Hulda, "Classroom Testing and the Biology Teacher: An
Annotated Bibliography." American Biology Teacher. 29
(April, 1967): 282—285.

Grobman, Hulda "Clansroom Testing in Biology: An Annotated Bibliography
(11)?" American Biology Teacher. 33 (February 1971): 86-90.

Heenan, D.K. and Chaval Paeratakul Education in Thailand: Evaluation
of Instruction in Thailand. East Lansing, Michigan State
University, 1966.

Holt, J. "I Oppose Testing, Marking and Grading." Today's Education.
60 (March 1971): 28-31.

Jessel, J.C. "Focus on Issues: Should we Declare a Moratorium on
Mental Testing." Contempgpy Education. 44 (November 1972):
127.

 

LaSage, W. "The Hassle Over Standardized Testing." Instructor. 82
(March 1973): 45-56.

Link, F.R. "Teach-Made Tests." National Education Association
Journal. 52 (October 1963): 23-25.

 

McMorris, R.F. et al "Effects of Violating Item Construction Principles."
Journal of Educational Measurement. 9 (Winter 1972): 287-295.

 

McSwain, E.T. "The Role of the College Faculty in the Teacher Education
Program." The 16th Yearbook of the National Council on
Measurement in Education. 1959: 34~38.

 

Mehrens, W.A. "The Consequences of Misusing Test Results." National
Elementapy Principal. 47 (September 1967): 62-64.

 

Mehrens, W.A. and Lehmann, I.J. Measurement and Evaluation in Education
and ngcholpgy, New York: Holt, Rinehart and Winston, Inc.,
1973.

 

Mehrens, W.A. and Lehmann, I.J. Measurement and Evaluation in Education
and Psychology: Instructor‘s Manual. New York: Holt,
Rinehart and Winston, Inc., 1973.

 

Noll, V.H. "Problems in the Pre-Service Preparation of Teachers in
Measurement." 'The 18th Yearbook of the National Council on
Measurement in Education. 1961: 35-42.

 

130

Noll, V.H. "Requirements in Educational Measurement for Prospective
Teachers." School and Socie_y. 82 (September 17, 1955):
88-90 a

 

Roeder, H.H. "Are Today's Teachers Prepared to Use Tests?" Peabody
Journal of Education. 49 (April 1972): 2394240.

Rose, H.A. and Elton, C.F. "Testing, on the Bias." Journal of College
Student Personnel. 12 (September 1971): 362-364.

 

Rosner, B. "Substance of Training in Measurement and Evaluation in
Pre-Service Preparation of Teachers." The 18th Yearbook of
the National Council on Measurement in Education. 1961: 51-54.

Schutz, R.E. "The Role of Measurement in Education: Servant, Soulmate,
Stoolpigeon, Statesman, Scapegoat, All of the Above, and/or
None of the Above." Journal of Educational Measurement. 8
(Fall 1971): 141-146.

 

Schwartz, A.A., and Tiedeman, D. Evaluating Student Progress in the
Secondary School. New York: Longmans, 1957.

 

Sharp, Evelyn W. "How Teachers See Testing." Educational Leadership
24 (November 1966): 141-143.

 

Stevens, C.A. "Measurement in the Teaching-Learning Situation."
The 16th Yearbook of the National Council on Measurement in
Education. 1959: 39-42.

Stevens, Margaret "The Role of the Classroom Teacher in School Testing
Programs." The 16th Yearbook of the National Council on
Measurement in Education. 1959: 43-46.

 

Thorndike, R.L. and Hagen, Elizabeth Measurement and Evaluation in
Psychology and Education. New York: John Wiley and Sons, Inc.,

1969.

 

 

Trump, J.L. "What's Wrong with Testing." Theory Into Practice.
2 (1963): 235-242.

 

Tylor, Jean L. "Of Tools and Fools." Phi Delta Kappan. 55
(December 1973): 285-286.

 

Wronski, S.P. and Raw Sawagdie-Panich Education in Thailand:
Secondary Education, Manppwer and Education Planning in
Thailand. East Lansing: Michigan State University, 1966.

‘D’J’l’d Lwa’ma mafiamfi’ma Tmud’nmwwfl‘n w$:um,ec:o.<

 

 

(3". . J..- .

 

MICHIGAN smTE UNIV. LIBRARIES
1111111111111111W111111111111111
31293100191877