Research on Humanities and Social Sciences www.iiste.

org
ISSN 2224-5766 (Paper) ISSN 2225-0484 (Online)
Vol.8, No.1, 2018

Attitude of Senior High School Teachers Toward Test

Construction: Developing and Validating a Standardised
Instrument
Frank Quansah Isaac Amoako
Department of Education and Psychology, University of Cape Coast

Abstract
Test construction is an essential part of teachers’ responsibility. Teachers are therefore supposed to craft well-
functioning items in ensuring effective teaching and learning. This study seeks to develop and validate a
standardised instrument in measuring teachers’ attitude towards test construction. The study further explores the
attitude of teachers towards test construction. The instrument was developed based on literature as well as personal
experiences of the researchers. The developed instrument was administered to 432 Senior High School teachers in
the Cape Coast Metropolis. Through an exploratory factor analysis, four dimensions were obtained which include:
planning, item construction, item review and assembling. A confirmatory factor analysis was then conducted to
examine the factor loadings of the items. After critical evaluation, the items on the instrument remained 32 which
was on a four point Likert scale. Further analysis revealed an overall negative attitude of SHS teachers towards
test construction. It is recommended that Ghana Education Service (GES) together with headteachers of various
SHS should ensure effective supervision of teachers in constructing test for students.
Keywords: Item construction, Item review, Testing, Assembling, Test construction

1. Introduction
The competency in test construction is an essential tool needed by every teacher if learning and instructional
objectives are to be effectively attained. The importance of tests in the educational system is enormous. Test
provides a platform by which any significant educational objectives can be achieved (Hamafyelto, Hamman-Tukur
& Hamafyelto, 2015). The effectiveness of learning goals, entrenched in the curricula of a school continues to be
the most fundamental sign pole for institutional superiority, educational development and individual goals.
Teachers are therefore required to have adequate knowledge in achieving these learning objectives in an accurate
and precise manner. Teachers must, thus, have the capability in the science and art of test constructing (D’Agostino,
2007).
A number of studies have explored teacher’s classroom test construction skills (Hamafyelto et al., 2015;
Kazuko, 2010; Onyechere, 2000). Ololube (2008) also evaluated test construction skills of professional and non-
professional teachers in Nigeria and reported that professional teachers tend to construct effective evaluative
instruments more than the non-professional teachers. It was also found in Ololube’s study that professional
teachers have the propensity to employ the various assessment techniques correctly, which is unlikely to happen
in the case of non-professional teachers. Onyechere (2000) found that some teachers craft poor tests while others
continue to use replica of test items because they seem to have inadequate skills in test construction. Hamafyelto
et al. (2015) discovered that Senior High School (SHS) teachers in Borno State, Nigeria, constructed items which
focused on lower cognitive operations. In Ebinye’s (2001) view, test construction has been found to be a major
source of anxiety among many teachers in Nigerian schools, especially, less experienced ones. This anxiety stems
greatly from lack of skills in test construction among these teachers. The problem of test construction was made
clear in a typical example:
A classroom teacher taught her pupils in second grade a lesson on ‘magnet’ and asked them, on
the following day, to write a six letter worded object which picks things. She expected almost the
whole class to return the word –‘magnet’ as their response. To her chagrin, the answer given
by more than 50% of the class was ‘mother’ (Daily Bread, 2011, p.23).
The teacher must have wondered what actually went wrong. Was it that she did not teach well or that the
pupils did not understand what was taught? The problem stems from neither the teaching nor the pupils’ learning
but from the way the test item was written. The question given by the teacher was not perfectly clear, thus, giving
room for more than one possible correct response.
In Ghana, a number of studies have indicated that teachers do not follow testing principles and consequently,
have poor testing practices (e.g., Anhwere, 2009; Amedahe, 1989). Amedahe (1989) revealed that SHS teachers
in the central region of Ghana have inadequate skills in testing. In a similar study among Junior High School
teachers in Ghana, teachers were found to have limited competencies in the management in the assessment
practices (Curriculum, Research & Development Division [CRDD] of Ghana Education Service, 1999). A critical
examination of literature indicates poor test construction skills of most teachers in all levels of education across
diverse subjects globally, and in Ghana to be specific (Anhwere, 2009; Amedahe, 1989; Ebinye, 2001; Hamafyelto

25
Research on Humanities and Social Sciences www.iiste.org
ISSN 2224-5766 (Paper) ISSN 2225-0484 (Online)
Vol.8, No.1, 2018

et al., 2015; Kazuko, 2010; Onyechere, 2000). This is really a great problem as students achievement are likely to
be reported with errors because poor items are used to measure achievement. Is it that teachers are not well trained
in test construction? Is it that teachers are trained well but feel reluctant in using what has been taught them? It is
important to state that these previous studies examined teachers test construction skills by asking them what they
actually do when crafting test items for students. However, these studies do not provide a comprehensive picture
of what teachers know. This is because teachers might have the knowledge but would be reluctant in practising
what he/she know. This is seen in the study of Ebinye (2001) who found that crafting test items appeared to be a
burden on teachers. Therefore, irrespective of the knowledge the teacher has, it is likely to construct poor questions,
or perhaps, repeat already existing questions (Onyechere, 2000).
In Ghana, teachers are trained in assessment of which test construction is an important component. In the
Colleges of Education, for instance, students are taken through a full course in educational assessment. The course
content allows these students to have a practical knowledge on test construction and assessment, in general.
Similarly, universities in Ghana who train teachers (e.g., University of Cape Coast, University of Education,
Winneba and Valley View University) also have a course in assessment for potential teachers to be trained in
assessment. This course also enlightens students on the construction of test items. Our personal observations have
confirmed earlier studies (Anhwere, 2009; Amedahe, 1989; Ebinye, 2001) that even though teachers are trained in
school assessment which includes test construction, most of them do not adhere to the rules governing these
practices which leads to poorly crafted questions. From our interaction with some teachers in some SHS (in the
Cape Coast Metropolis) during an educational out-programme, it appears that teachers attitude towards test
construction is nothing to boost of and this, to a greater extent, contributes to the construction of poor items. This
study seeks to empirically examine the attitude of SHS teachers through the development and validation of a
standardised scale in order to provide a standard measure of attitude towards constructing tests.

2. Development of the Instrument

The instrument was developed based on the behaviours exhibited by teachers in various schools. These behaviours
were observed by the authors of the instrument. Literature was, further, reviewed to obtain information on the test
construction behaviours of teachers (e.g., Allen & Yen, 2002; Nitko, 2001). Items were then cautiously crafted
based on literature and observations made by the researchers. Initially, 41-items were crafted but only 32-items
remained after the instrument had gone through several review and factor analysis. The items were on a four point
Likert scale of agreement (SD- strongly disagree, D-disagree, A-agree, SA- strongly agree). After the items were
crafted and reviewed, a pilot testing was conducted among 100 teachers from some selected SHS in the Sekondi-
Takoradi Metropolis. This was done to establish the validity and reliability of the responses which will be elicited
by the instrument. Some items were modified after the pilot testing of the instrument. Items like “Learners decide
item format to be used” was changed to “I prefer the item format of a classroom test to be decided by the learners”.
In all 4 items were reworded after the pilot testing. The instrument was then administered to 432 teachers in some
selected SHS in the Cape Coast Metropolis.

2.1 Ensuring Validity

The development of the ATC scale was carefully done to ensure the validity of responses solicited. Efforts were
made to ensure that the questions crafted represented attitudinal behaviours of teachers (Nitko, 2001). After the
items were crafted, they were also reviewed by experts (PhD students) in the Measurement and Evaluation field
to validate the instrument. This was done in line with Anim’s (2005) assertion that content and construct validity
is determined by expert judgement. Results from the factor analysis revealed that the Kaiser-Meyer-Olkin (KMO)
test of sampling adequacy and Bartlett;s test were not violated (See Table 1) based on Crocker and Algina’s (2008)
criteria. An exploratory factor analysis using Principal Component Analysis Method was, then, conducted to
determine the factors involved in the scale. The scree plot was used to determine the factors. The exploratory
analysis revealed four factors (See Figure 1).
Table 1: KMO and Bartlett’s Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy .602
Approx. Chi-Square 7119.354
Bartlett’s Test of Sphericity: df 628
sig. .000

26
Research on Humanities and Social Sciences www.iiste.org
ISSN 2224-5766 (Paper) ISSN 2225-0484 (Online)
Vol.8, No.1, 2018

Figure 1: Result on Scree Plot

A confirmatory analysis using Minimum Likelihood Method was further conducted to explore the factor
loadings of each item. Items with factor loading of .3 and below were rejected (See Table 2). After the factor
analysis, 32-items remained. Item like “I believe good items cannot be crafted without considering the learning
objectives” had factor loading of .235 and thus, was rejected. Based on the results from the confirmatory factor
analysis, the four factors were labelled as: planning, item construction, item review and item assembling.
Table 2: Factor Rotation
Items 1 2 3 4
To be honest, it is a waste of time trying to outline the purpose of a test when .633
planning for the test.
I just need my textbook to start writing test items. .532
I believe good items cannot be crafted without considering the learning objectives. .235*
I mostly do not prefer using test specification table in crafting questions. .632
I prefer to finish crafting the test before considering the thinking skills those items .642
measure.
Since I am the classroom teacher, I do not need to specify the content area I want .692
to test.
Planning a test is needless if I am the teacher .772
I prefer writing items based on what learners are expected to know whether taught .610
or not.
As a teacher there is nothing wrong with crafting items without considering the .767
learning objectives.
I prefer the item format of a classroom test to be decided by the learners. .451
It is not possible to always craft new questions for learners. .492
Crafted items do not necessarily have to match learning objectives. .463
I like to write tricky questions to test my students understanding. .589
Arranging of the options to multiple-choice items alphabetically is not .491
compulsory
I always refer to test specification table when constructing items. .475
I like to always write items with the same difficulty level. .739
There is the need to take items verbatim from textbooks used in teaching. .403
I usually construct test items few days for the paper to be written. .496
It is optional to review constructed items before it is administered .620
Checking for the item difficulty and discrimination after the test has been .714
constructed is not too necessary
It is essential to present more difficult items before less difficult items in .654
assembling crafted items
It is optional to number all the items on a test .443

27
Research on Humanities and Social Sciences www.iiste.org
ISSN 2224-5766 (Paper) ISSN 2225-0484 (Online)
Vol.8, No.1, 2018

Items 1 2 3 4
It is optional to provide clear directions for examinees on the test instrument .630
It is right to arrange options of test items horizontally .621
It is better to rely on past questions when constructing a test. .677
I like to prepare marking scheme after the test have been administered. .513
It is necessary to check for the clarity of crafted items .416
I prefer preparing marking scheme two or more days after constructing the test .438
I always like to arrange questions into sections based on their nature or type .486
I select questions from topics I think students have understood. .414
I think test specification table should be prepared by test experts and not the .553
classroom teachers.
It is essential to identify behaviours to represent a construct when crafting test .654
items.
I do not think it’s necessary to craft more items than actually needed. .654
1-Planning; 2- Item construction; 3- Item Review; 4-Assembling
*Item rejected

2.2 Estimating Reliability

Estimating reliability of items cannot be overlooked because every investigator consider it necessary in gathering
objective and accurate information. There is the need, therefore, to estimate the reliability of responses of the
construct of interest (Quansah, 2017). The reliability of the instrument was esyimated using the Cronbach’s Alpha
Reliability Method. The reliability estimate for each sub-scale as well as the whole instrument were investigated
(See Table 3). The overall reliability estimate of the instrument was .85. This reliability co-efficient is sufficient
enough to ensure reliable responses as indicated by Pallant (2010) that a reliability coefficient (alpha) of .70 or
higher is considered appropriate.
Table 3: Reliability Estimate for Sub-scales
Sub-scale No. of Items Coefficient
Planning 11 .81
Item Construction 11 .79
Item Review 3 .70
Item Assembling 7 .71

3. The Use of the Instrument and its Administration

The instrument, after its development and validation was named “Attitude towards Test Construction (ATC)
Scale”. The ATC scale is designed to provide much knowledge to stakeholders in education on the attitude of
teachers towards test construction. Specifically, ATC has been developed to assist headmasters/mistresses, Ghana
Education Service (GES), school counsellors and test experts in finding out the attitude of teachers towards test
construction in a particular school. This will provide relevant clues of the test construction practices of the teachers.
Teachers who have been found to be performing poorly can be administered the ATC scale to find out his/her
attitude towards test construction. This is because test construction practices has been found to be significantly
related to teacher effectiveness (Hamafyelto et al., 2015). The ATC scale can also be used as a research instrument
for students and other researchers who have interest in the area of test construction. These researchers can, thus,
adopt or adapt the instrument for their study.
The ATC scale can be administered to individuals or groups. For individual administration, the respondent
needs to be educated on the need to respond to the instrument. Effort must be made to establish good rapport with
respondent(s) so that accurate responses would be given willingly. The individual should be allowed to
independently respond to the instrument. On group basis, the investigator should ensure serene environment for
the respondents. Regardless of the individual or group of people who will be given the instrument to respond to,
their consent must be sought. It is important to ensure that ethical considerations are followed in the administration
of the instrument. In all, 25-30 minutes is appropriate for respondent(s) to respond to the instrument.

4. Scoring and Interpretation

The ATC scale has both positive and negative questions of which responses are measured on 4-point scale. In
scoring the items on the instrument, negative items are scored on point score from 1-4. That is, strongly agree is
valued 1-point, agree for 2-point, disagree is 3-point, and strongly disagree for 4-point. For positive items, strongly
agree is 4-point, agree 3-point, disagree 2-point and strongly disagree 1-points. Apart from items 14, 16, 23, and
32, the rest of the items are negative questions. For the overall attitude, the responses from all the items are added
and divided by the number of questions. The same computational method is applicable to the sub-scales (i.e,

28
Research on Humanities and Social Sciences www.iiste.org
ISSN 2224-5766 (Paper) ISSN 2225-0484 (Online)
Vol.8, No.1, 2018

calculating the composite score for the responses for a particular respondent or group of respondents). In
calculating for the attitude of respondents the mean of their responses is computed for and interpreted. In
interpreting the attitude of respondents of a particular item (e.g., item 5), the mean score of the responses is
compared with 2.5 ([1+2+3+4]/4=2.5). Mean scores less than 2.5 shows a negative attitude whereas mean scores
above 2.5 shows a positive attitude to that particular item. For the interpretation of individual scores, the mean of
the obtained scores is also compared with 2.5.

5. Exploring Teacher Attitude towards Test Construction

After the instrument has been validated, the attitude of the teachers were examined based on the validated items.
Table 4: Attitude of SHS teachers towards Test Construction
Sub-scales (Attitudes) No. of Items Mean SD
Planning 11 2.43 .74
Item Construction 11 1.90 .89
Items Review 3 2.03 .66
Assembling 7 2.14 .63
Overall Attitude 32 2.13 .72
Results (in Table 4) indicate that SHS teachers have negative attitude towards the planning of classroom test
(M=2.43, SD=.74), item construction (M=1.90, SD=.89), Items review (M=2.03, SD=.66), and Assembling
(M=2.14, SD=.63). Generally, teachers in SHS in the Cape Coast Metropolis were found to have an overall
negative attitude towards test construction (M=2.13, SD=.72).

6. Discussion
The need for teachers to construct good test in assessing their students have been underscored in literature
(Hamafyelto, 2015). While some teachers are found constructing poor items, others are found to be repeating
already existing questions (Onyechere, 2000). Some authors have attributed this to teachers’ limited knowledge
and skills in the area of test construction (e.g., Anhwere, 2009; Amedahe, 1989; Ebinye, 2001; Hamafyelto et al.,
2015; Kazuko, 2010; Onyechere, 2000). Others have attributed poor questions of teachers to the fact that teachers
see test constructions a major source of anxiety and burden (e.g., Ebinye’s, 2001). This present study revealed
another factor which also accounts for the poor construction of test items among teachers. Teachers were found to
have a negative attitude towards test construction. This may contribute to the construction of poor questions among
these teachers as indicated in previous studies. It is likely that teachers have the knowledge about test construction
but their attitude prevent them from utilizing the knowledge they have. Test construction, we might say, is a
difficult and rigorous task if teachers are supposed to do it effectively (Nitko, 2001). This explains the reason why
some teachers see test construction as a burden. The findings of this present study implies that even when teachers
are given adequate training in the area of test construction, it is unlikely that their skill attained might be put to use
if these teachers have negative attitude towards crafting the questions. This presupposes that the attitude of teachers
towards test construction is likely to act as a moderator in the relationship between knowledge and practice of test
construction.

7. Conclusions and Recommendations

Testing in education cannot be under emphasized because teaching and learning can never be complete without it.
Teachers would, thus, be ignorant of how well they are doing as well as how well the students are grasping the
concepts being taught. Nevertheless, these measure of teacher effectiveness and students performance can never
be seen if test are poorly constructed. Even though, teachers are taken through courses in their training, test
construct seem to be a nightmare (Nitko, 2001). It is believed that if attitude influences practice (Ebinye, 2001),
then there is the need for attitude of teachers to be explored. This is the foundation for the development of ATC
(Attitude towards Test Construction) scale. The instrument, therefore, provides a standardized measure of the
attitude of teachers towards test construction. It is important for stakeholders to re-orient teachers on the need to
follow test construction procedures and to put to use their skills attained from various training they have had. As
more training programmes through seminars and workshops are organised for teachers, stakeholders should be
aware of the fact that the training alone do not bring about the application of competencies gained but also their
attitude towards constructing the test. It is recommended that teachers should not only be trained constructing test
items but should also be enlightened on the need to adhere strictly to testing procedures. Ghana Education Service
(GES) together with headteachers of various SHS should ensure effective supervision of teachers in constructing
test for students.

References
Amedahe, F. K. (1989). Testing practices in secondary schools in the Central Region of Ghana. Unpublished
Master’s thesis, University of Cape Coast, Cape Coast.

29
Research on Humanities and Social Sciences www.iiste.org
ISSN 2224-5766 (Paper) ISSN 2225-0484 (Online)
Vol.8, No.1, 2018

Anhwere, Y. M. (2009). Assessment practices of teacher training college tutors in Ghana. Unpublished Master’s
thesis, University of Cape Coast, Cape Coast.
Anim, M. E. (2005). Social science research: Conception, methodology and analysis. Kampala: Makerese
University Press.
Curriculum Research and Development Division (CRDD). (1999). Investigation into student assessment
procedures in public Junior Secondary Schools in Ghana. Accra: Ghana Education Service.
D’Agostino, J. V. (2007). Quantitative research, evaluation, and management section. Ohio State University. 29
Woodruff Avenue, Columbus, OH 42310-1177. USA.
Daily Bread. RBC Ministries (2011). Our daily bread: For personal and family devotions. Michigan, USA:
Discovery House Publishers.
Ebinye, P. O. (2001). Problems of testing under the continuous assessment programme. J. Qual. Educ., 4(1), 12-
19.
Hamafyelto, R. S., Hamman-Tukur, A., & Hamafyelto, S. S. (2015). Assessing teacher competence in test
construction and content validity of teacher made examination questions in commerce in Borno State, Nigeria.
Journal of Education, 5(5), 123-128.
Kazuko, J. W. (2010). Japanese high school mathematics teachers’ competence in real world problem solving.
Keto Academy of New York and Teachers College Columbia University.
Nitko, J. A. (2001). Educational assessment of students. New Jersey: Prentice Hall.
Onyechere, I. (2000). New face of examination malpractice among Nigerian youths. The Guardian Newspaper
July 16.
Pallant, J. (2010). SPSS survival manual (4th ed.). New York, NY, McGraw Hill.
Quansah, F. (2017). The use of Cronbach Alpha Reliability estimate in research among students in public
universities in Ghana. Africa Journal of Teacher Education, 6(1), 56-64.

30