Ary Setya,Developing Writing Test 81

DEVELOPING WRITING TEST FOR EFL LEARNERS

Ary Setya Budhi Ningrum
STAIN Kediri, East Java

Abstract: This paper tries to present a general review of testing writing in

English as a foreign language. This paper covers some of the important and
relevant aspects of developing writing test such as definition of writing,
writing in ESL/EFL classroom, writing assessment, and validating writing
test. Special review is addressed to give a brief note of a comprehensive and
step-by-step procedure of developing test instrument proposed by Prof.
Soenardi Djiwandono.

Keywords: writing test, English as a foreign language

Testing is one of the most influential areas related to teaching as it can be

used to measure the effectiveness of teaching learning progress. In fact, if the
test used is good it can be used to check the effectiveness of the whole process.
Heaton (1988:5) insists on these aspects mentioning that both teaching and
testing are so closely interrelated that is virtually impossible to work in either
field without being concerned with the other. Hence, learning to teach also
requires learning to develop a good test.
There are two major parts of object of the test if we speak about language
testing. First part is in the sense of language skills namely speaking, listening,
writing, and reading. Secondly are linguistic components such as grammar,
vocabulary, and phonetics. In this paper, I will focus on testing language skill
namely writing test.

DEFINITION OF WRITING
Writing is a communicative act and a way of sharing observation,
information, thoughts, and ideas with others through written language (Cohen
et. al., 1989; Troyka 1987). As a communicative act, writing involves both
physical and mental process. Cohen et al. (1989) state further that as a physical
process, writing means producing graphemes and orthographic symbols in the
form of letters or combination of letters that relates to the sounds in spoken
language using hands while the eye movements follow over the words or
sentences. While physical activity, according to Richards (1990:101),can be
captured visually, mental activity, on the other hand, cannot be seen or observed
directly. It is the process in the writer’s mind that includes making connection
between ideas and processing thoughts to be expressed in a meaningful written
text by employing linguistics organization.
82 Lingua Scientia, Volume 3, Nomor 2, Nopember 2011

More specifically to the point, Troyka (1987: 3-4) characterizes writing

from its elements. Writing is seen as “a way of communicating a message to a
reader for a purpose.” In that case, writing entails four key elements namely,
communicating, message, reader, and purpose. Communicating means the act of
sending a message from the writer to readers. Message refers to the content,
ideas or information carried in the piece of writing. Reader is the receiver of the
message. Purpose is the intention of writing. All of elements attain a way of
transferring information from the writer, as the encoder of the information, to
the reader, as the decoder.
To make the message understandable the text produced in writing
process has to be communicatively effective. It means that a writer must have a
communicative competence, which includes grammatical, discourse,
sociolinguistics, and strategic competences (Savignon, 1983; Eanes, 1997).
Sauvignon (1983); Eanes (1997) further explain that grammatical competence
focuses on sentence-level grammar and requires knowledge of lexical items and
rules of morphology, syntax, sentence-grammar semantics, and phonology. While
discourse competence means the capability to connect sentences in stretches of
discourse and to form a meaningful whole out of a series of utterances. In other
words, it is concerned with intersentencial relationships. Sociolinguistic
competence is the knowledge of socio-cultural rules of language and of discourse.
It requires an understanding of social context in which the language is used.
Finally, strategic competence means the way of manipulating language in order
to meet communicative goals.
Furthermore, Brown (2001:343) points out that in order to make the
readers, as the addressees, able to interpret the linguistic message produced
through the process of writing, a writer needs to acquire the sub skills needed, as
shown below:
1) Produce graphemes and orthographic patterns of English,
2) Produce writing at an efficient rate of speed to suit the purpose,
3) Produce an acceptable core of words and use appropriate word order
patterns,
4) Use acceptable grammatical system (e.g. tense, agreement, pluralization),
patterns, and rules,
5) Express a particular meaning in different grammatical forms,
6) Use cohesive devices in written discourse,
7) Use the rhetorical forms and conventions of written discourse,
8) Appropriately accomplished the communicative functions of written texts
according to form and purpose,
9) Convey links and connections between events and communicate such
relations as main idea, new information, given information, generalization,
and exemplification,
10) Distinguish between literal and implied meaning when writing,
11) Correctly convey culturally specific references in the context of the written
text,
12) Develop and use a battery of writing strategies, such as accuracy assessing
the audience’s interpretation, using prewriting devices, writing with fluency
 Ary Setya,Developing Writing Test 83

in the first drafts, using paraphrases and synonyms, soliciting peers and
instructor feedback, and using feedback for revising and editing.

WRITING IN EFL CLASSROOM

As one of the four language skills, writing has occupied a place in most
English syllabuses. Even so, arguments are sometimes put forward for not
teaching students to write because it is felt that a command of the spoken
language and of reading is more important. For many students this may be true,
but today, given the importance of English as an international language, more
and more people need to learn to write in English for occupational or academic
purposes. Another thing that makes the ability to write in English important is
the fact that now English is the world’s major language for the communication of
research findings (Swales, 1990). Thus, the ability to write a scientific essay is a
major goal of tertiary education. Because of the reason above, instead of being
the last skill taught, writing has now become a much more important subject to
be taught in second/foreign language curriculum (Leki, 1994).
According to Rivers (1987), writing in language classroom can be
classified into four. The first type is writing or noting down which means
imitating or copying. An example of this is to copy English sentences or
paragraphs. The second is writing in the language or writing practice. The
activity might be in the form of writing sample dialogues, uncomplicated
translation exercises, dictation, and cloze procedures. The third type is
translation which refers to transferring passages or sentences from the native to
the target language. The fourth and the highest level of writing is called
expressive writing. It deals with the expression of original ideas in the target
language. Based on the types of writing above, Rivers (1987) summarizes writing
as the expression of ideas in a consecutive way, according to the graphic
conventions of the language—the ability to express himself in a polished literary
form which requires the utilization of a special vocabulary and certain
refinements of structure.
When students start to learn to write, at the beginning they will simply
write down English words and might be sentences. Slowly then, the students will
learn how to produce a sequence of sentences that are arranged in a particular
order and linked together in certain ways. It is not easy to produce such a well
organized piece of writing. Learning to write in either a first or a second
language is one of the most difficult tasks a learner encounters and one that few
people can fully master. Learning to write well is a difficult and lengthy process,
one that induces anxiety and frustration in many learners (Richard, 1990).
In EFL context, it is believed that developing writing skills is more
complicated than developing other language skills. Unlike speaking, for instance,
writing means producing a written message without the presence of the intended
readers. In this case the writer has to imagine the situational context and the
roles of his readers (Hughes and Porter, 1983). In order to be able to write well,
the foreign language learners need to be equipped with early and continued
writing experiences. As Byrne (1988:48) suggested that the writing program
must be planned carefully in order to develop the mastery of the writing skill
which the learner can use for a continually expanding range of tasks. Since the
84 Lingua Scientia, Volume 3, Nomor 2, Nopember 2011

writing skill will still be guided to large extent, the English teachers are
supposed to vary the activities given to the students in order to avoid
monotonous approach. Thus, the teacher’s task to develop the students’ writing
skill is more complicated than the other skills.

WRITING TEST
There are so many experts proposed what kinds of aspects should be
judged in writing test. First is Inman et al. (1979) who asserted that there are
five aspects should be assessed in a composition. They are logic, organization,
development, style, and mechanics. While Jacobs (1981) proposed five aspects as
well for ESL context, namely content, organization, vocabulary, language use,
and mechanics. Harris, (1974), in addition, proposed five general components:
content, form, grammar, style, and mechanics. Moreover, Heaton (1988) states
five general components or main areas for writing good prose such as language
use, mechanics, content, style, and a judgment skill that is ability to write in an
appropriate manner for a particular purpose with a particular audience in mind,
together with an ability to select, organize and order relevant information. It is
in line with Burgess & Head (2005) statement that an answer in writing test
that has some errors but achieves its communicative purpose will get a higher
mark than an answer that is grammatically accurate but does not meet the task
requirement. Other experts supported this mark by saying that effective writing
ability can be reached through a combination of sociocultural competence,
involving appropriate conventions, register, and (rhetorical) style; discourse
competence, involving ideas and their structuring, coherence, and cohesion with
an intended audience in the mind; and linguistic competence, involving
appropriate and broad lexis, fluent and accuracy syntax, and accurate
mechanics.
Furthermore, Langan (1985) has something different in evaluating essay
writing. He proposed five areas to be based for assessment in writing. They are
unity, support, coherence, and sentence skills. However, in fact, those elements
of writing have something in common; that is they are more or less the same.
Language, for example, has two elements. They are sentence structure and
diction. Mechanics has four points. They are paraphrasing, punctuation, spelling,
and capitalization. Style has five aspects. They are economy, simplicity, clarity,
congruity, and courtesy. Organization is the rhetorical form, while logic is
something to do with content. In short, among several features of composition to
be assessed, they have many things in common. They can be grouped into four:
content, organization, language, and style.

APPROACHES TO WRITING ASSESSMENT

Experts in writing assessment share something in common that there are
two basic approaches in the assessment of writing, namely direct and indirect
assessment. The former requires the students to write the actual composition by
organizing their ideas into a unified text. While the indirect assessment, on the
other hand, does not ask the students to write but it asks them to respond to
question about writing or to do a particular task related to writing skill. In other
words, they have to perform their knowledge about writing.
 Ary Setya,Developing Writing Test 85

Direct Writing Assessment

The most direct way of measuring students’ writing ability is to have
them write (Harris, 1974). That is why when assessing their writing ability they
must be asked to write. In a composition assessment, the students should be
presented with a clearly defined problem which motivates them to write. Heaton
(1975) says whenever possible, meaningful situation should be given to provide
necessary information for writing because: (1) situation determines the register
and style to be used in the composition, (2) it gives the possibility to obtain
greater degree of reliability in scoring, and (3) it has an excellent backwash effect
on the teaching and learning preparatory to the examination.

Indirect Writing Assessment

Writing can be divided into more specific ‘discrete’ elements, e.g
grammar, spelling, vocabulary, punctuation, and orthography and attempts can
be made to test these formal elements separately by the use of objective tests.
These tests would be indirect ones in that they would only be measuring parts of
what we take to be the construct of writing ability. What are tested may be
related to proficient writing, but they cannot represent what proficiency writers
can do (Weir, 1995). It would be very difficult to generalize from these types of
test as to how students might perform on more productive tasks which require
construction of a complete text. It would be difficult from these discrete item
tests to make direct statements about how good a writer or what he or she can do
in writing. Such indirect assessments have an extremely negative washback
effect on the teaching that precedes them.
The formats that can be used to assess writing skill can be classified into
three, i.e., open-ended, close-ended, and restricted response (Carol, 1980). In an
open-ended writing assessment, the students are given an opportunity to
initiate, elaborate or modify and to adopt the attitude they consider appropriate
in the response to the writing task. The nature of this format is closer to the type
of the test that is based on real-life performance and setting, that is writing for
academic and social purpose in the real world (Weir, 1995).

PROCEDURES OF SCORING
As mentioned previously, there are two approaches that can be used to
measure the students’ writing ability, namely the direct and indirect
measurement. Considering that each approach result in different tasks, a
different procedure in scoring is applied accordingly. In a direct measurement,
the score that the students get is derived from the rater’s judgment on the basis
of the pre-determined criteria stated in the scoring guide. On the other hand, in
an indirect measurement, the score is obtained from an objective scoring
procedure which is based on an answer key.

Scoring in Direct measurement

There are some proposed scoring guides by some experts deal with how to
evaluate writing in direct measurement of writing. First is coming from Jacobs
et al (1981) who classified the direct measurement of writing into two major
types. They are the holistic scoring and the frequency count marking. In the
86 Lingua Scientia, Volume 3, Nomor 2, Nopember 2011

holistic scoring, a piece of writing is viewed as a whole and complete idea rather
than as a separate element. The rater bases his judgment on his impression of
the composition and he might be guided by a holistic scoring guide in scoring the
composition. While the frequency count marking, on the other hand, is a
procedure in evaluating a piece of writing by tallying and/or enumerating certain
elements of the composition, like the number of cohesive devices, spelling errors,
grammatical errors, punctuations errors and things like that.
Another classification of direct measurement in writing is introduced by
Spandel and Stiggins (1990). They classify writing assessment into three types:
primary trait, holistic, and analytic scoring procedure. Primary trait scoring is a
procedure in scoring a piece of writing by focusing on the domain trait of the
piece, such as on descriptive, narrative, and argumentative writing. Holistic
scoring means scoring a piece of writing as a whole where each paper receives
only one score. The final score is not the total of sub scores. The difference
between primary trait scoring and holistic scoring is on the emphasis. In the first
procedure, a different mode of writing has a different scoring guide depending on
the types of discourse whereas in holistic scoring, there is no specific emphasis.
That is why the holistic scoring can be applied to all types of discourse. Analytic
scoring, unlike the first two procedures, scores a piece of writing by referring to a
list of features or sub skills on which a rater bases his judgment. The writing
quality is shown by the total of the sub scores.
Still another classification of direct measurement in writing is given by
White (1985), and Brown (2005), who classify evaluation on writing into two
basic scoring procedures. The first is holistic scoring and the second is analytic
scoring. In holistic scoring, a rater judges a piece of writing as a whole without
any separable aspects and their sub-scores. The holistic evaluation must come up
with a single score which does not result from summing up the sub-scores.
Analytic scoring, in contrast, come up with a single score resulting from
summing up of the sub-scores which are derived from the scoring of the features,
or aspects of the piece.
In short, there are two types in common for writing scoring. If the
procedure of scoring is based on the analysis of features, it is called analytic.
When the scoring is based on the judgment of rater(s) as a whole without
separating features or aspects, it is called holistic.

Scoring in Indirect Measurement

Indirect measurement of writing ability as stated earlier is basically a
measurement of knowledge about writing or sometimes it is called a
measurement of editorial skills. This is due to the fact that the students are
required to respond to questions about writing in an objective-type test, such as
multiple choices, or a completion test depending on the purpose or on what
aspects of writing to be measured. The type of task varies such as choosing the
most suitable topic sentences from a given option, identifying irrelevant
sentences in a paragraph, filling in the missing transitional signals, and many
others.
Actually in term of reliability, indirect measurement has higher
reliability coefficient that its counterpart as it is an objective-type test. This is
 Ary Setya,Developing Writing Test 87

due to the fact that indirect measurement uses an objective scoring system with
definite answer, so there will be no subjective judgment. A direct measurement,
in contrast, is usually difficult to obtain high reliability coefficient for the scores
depend on the raters’ judgment. Avoiding subjectivity is extremely difficult. To
overcome this problem, at least two raters, even more, are needed. A third rater
is required in case the scores from the two raters are more than pre-determined
maximum acceptable difference of scores.

DEVELOPING PROMPTS FOR WRITING TEST

Developing prompts for writing test should be done in carefully way since
the prompts used in the test will determine the successfully of the test itself. In
selecting the appropriate writing prompts, there are some aspects to be taken
into consideration. California Assessment Program (in O’Malley and pierce,
1996) proposed that each prompt used in writing assessment should meet the
following criteria. First, the writing prompt invites the desired type of writing or
genre. Second, it engages the thinking, problem solving, composing, and text-
making processes central to the type of writing. Third, it is challenging for many
students and accessible to all. Fourth, it provides equitable opportunities for all
students to respond. Fifth, it produces interesting, not just proficient writing.
Sixth, it is liked by many students.
Weir (1995) emphasized the importance of considering the students’
background knowledge when developing the test. This is due to the fact that
students write best when they find something they can write on. Concerning the
topics presented in a writing test, Reid (1993) asserted that effective writing
topics are the ones that (1) interest both the students and the readers, (2) are
accessible to all students, (3) involve the students in the topic, (4) are
comprehensible to the students, (5) are unbiased with regard to the culture of the
students, (6) are not culture bound (that is, they do not require intimate
knowledge of a particular culture, especially if the test is designed for the
students who come from different cultural backgrounds), and (7) allow the
students to learn while they write.
Based on the experts’ argument above, the prompt used in this test
writing are developed in such way that they have the main factors as following;
(1) the prompt do not require an ambiguous task so that the students do not
waste their time trying to figure out what is called for, (2) the prompt specify a
particular purpose mode of writing that is required only argumentation, (3) the
prompt offers interest to all the students so that they wrote with some genuine
concern, (4) the prompt specify the amount of time the students will have to
complete the writing, (5) the prompt specify the length of the text produced by
the students, (6) the last part of prompt tell the students what will be valued in
the writing.

VALIDATING WRITING TEST

A test can be used as a valuable device in giving overall impression about
teaching learning process has been done. By administering testing, the teacher
can evaluate the effectiveness of the syllabus as well as the methods and
materials (Heaton, 1988:7). However, to do so the test should fulfill some criteria
88 Lingua Scientia, Volume 3, Nomor 2, Nopember 2011

of being a good test. A good language test should possess three qualities i.e.
validity, reliability, and practicality (Harris, 1974).

Validity
Every test, whether it be a short, informal classroom test or a public
examination should be as valid as the constructor can make it (Heaton,
1988:159). Validity, according to Ebel & Frisbie, 1986 in Latief, 2000:98) refers to
the appropriateness of making specific inferences or of making certain decision
on the basis of scores from a test. In other words, the test must aim to provide a
true measure of the particular skill which is intended to measure: to the extent
that measures external knowledge and other skills at the same time, it will not
be a valid test (Heaton, 1988:159).
Differ with conventional concept that claims there are so many kinds of
validity; Djiwandono (2008:165) point out that validity is unitary concept. It
means that validity conceptually is only one kind, the variousness are lies on
how to prove the validity. They are mainly three ways to give a support or
evidence to validity, and we can choose one of the most appropriate one to get
evidence of our test validity. They are content validity, criterion-related validity,
and construct validity.
Content validity is a kind of validity that depends on a careful analysis of
the language being tested and of the particular course objectives (Heaton,
1988:160). It means that a test should contain a representative sample of the
course, the relationship between the test items and the course objectives always
being apparent. Furthermore, Gronlund (1985) in Latief (2000:1) states that
content validation is the process of determining the extent to which a set of test
tasks provides a relevant and representative sample of the domain of tasks
under consideration. Heaton (1988:160) stated that if we want to use content
validity to give evidence to our test, the test writer should first draw up a table of
test specifications, describing in very clear and precise terms the particular
language skills and areas to be included in the test. Then the important point is
that the test writer has attempted to quantify and balance the test components,
assigning a certain value to indicate the importance of each component in
relation to the other components in the test. By so doing, the test will achieve
content validity and reflect the component skills and areas which test writer
wishes to include in the assessment.
As cited from Djiwandono (2008:165), criterion-related validity can be
proved by comparing between the students’ score obtained and the scores from
the similar test which has been marked as a good test at the same time, we
speak of ‘concurrent validity’. For instance, we compare the students’
achievement in English course and the students’ scores in TOEFL. When the
correlation between the students’ achievement in the course and their
achievement in TOEFL test is high, the result of the test/measurement in the
English course has strong criterion validity evidence. Therefore, the result of the
English test is believed to have high concurrent validity. On the contrary, when
the comparison of the two test results makes low correlation, the result of the
test in the English course is said to have weak or low concurrent validity
evidence. If it concerns the degree to which a test can predict the test takers’
 Ary Setya,Developing Writing Test 89

performance in the future, it is called ‘predictive validity’. For instance, with

placement test: once courses are under way, we can check validation by
establishing the proportion of the students who were thought to be misplaced
(Hughes, 1989). This type of validity has also been called ‘empirical’, ‘pragmatic’,
or ‘statistical’ (Bell, 1981:198).
Whereas a test will be said has construct validity if it is capable of
measuring certain specific characteristics in accordance with a theory of
language behavior and learning (Heaton, 1988:161). This type of validity
assumes the existence of certain learning theories or constructs underlying the
acquisition of abilities and skills. For example, if the assumption is held that
systematic language habits are best acquired at the elementary level by means
of the structural approach, then a test which emphasizes the communicative
aspects of the language will have low construct validity. Conversely, if a
communicative approach to language teaching and learning has been adopted
throughout a course, a test comprising chiefly multiple choice items will lack
construct validity.
In Latief (2000:100), the aspects of the test which can be used as the
construct validity evidence are the tasks or activities required by the test that
the students should perform. A test should always ask the test takers (students)
to perform particular tasks or activities. Based on the tasks or activities that the
students should perform in the test is the basis of scoring. They are the construct
validity evidence. When the tasks or activity that the students should perform
and they are the basis of scoring reflect the students’ behavior which are being
measured, the result of the measurement or the test has a strong construct
validity evidence. Therefore, the result of the test is believed to have a high
validity. On the contrary, when the tasks or activities that the students should
perform in the test and they become the basis of scoring do not reflect the
students’ behaviors which are being measured, the result of the measurement or
the test does not have strong construct validity evidence. Therefore, the result of
the test is believed to have weak construct validity.

Reliability
Reliability of the result of language test refers to the preciseness of the
result in presenting the actual level of the language proficiency of the examinees
(students) (Latief, 2001:214). If the test is administered to the same candidates
on different occasions, then, to the extent that it produces differing results, it is
not reliable (Heaton, 1988:162). Reliability measured in this way is commonly
referred to as test/re-retest reliability to distinguish it from mark/re-mark
reliability. Another common reliability denotes the extent to which the same
marks or grades are awarded if the same test papers are marked by (i) two or
more different examiners or (ii) the same examiner on different occasion. In
short, in order to be reliable, a test must be consistent in its measurements.
Since there are so many kinds of test for language proficiency, so there
are many ways as well in calculating the level of reliability (see Djiwandono,
2008:171). The calculation of reliability level always required two sets of scores
to measure the consistency of the test. Correlation coefficient as a measurement
of the consistency of the test can be got by calculation method. There are so
90 Lingua Scientia, Volume 3, Nomor 2, Nopember 2011

many methods of estimating the reliability test. Based on the score obtained to
calculate correlation coefficient, there are eight kinds of reliability. They are (1)
test-retest reliability, (2) equivalent-forms reliability or alternate-forms
reliability, (3) split-half reliability, (4) cronbach alpha reliability, (5) cronbach
alpha for writing ability, (6) Kuder-Richardson (KR) reliability, (7) scorer
reliability or rater reliability, and (8) estimated reliability.
The approval of reliability is completely an empirical matter in which it
involves statistical analysis. The statistical analysis is used to show the
correlation in a various level. It is expressed in the form of correlation coefficient.
Since the reliability is a correlation in a various level, therefore, reliability is
actually a spectrum of level and is not dicotomically reliable and unreliable. The
reliability is spread from the highest to the lowest with some levels in between.
It is a form of a continuum of coefficient. The highest reliability is statistically
figured as 1.00. The reliability with value 1.00 means the score has absolute
consistency without any deviation at all. This kind of reliability is theoretical,
because in the reality almost there is no result of measurement which is
absolutely consistent without any difference at all, moreover in the measurement
of a multi-aspect subject such as language teaching. In practicing, the level of
reliability is usually found to be lower than the absolute correlation coefficient
(1.00) that is 0.95, 0.90, 0.70, and so forth.

Practicality
There are so many experts in language testing discuss about the
practicality of the test. According to Djiwandono (2008: 190), practicality of a test
does not have relationship with something abstract or theoretical, but it is
something done with its application, mainly in 1) the practicality of
administering the test, and 2) the financial aspect. It is line with Bell (1981:200)
that involved two parameters in determining practicality that is economy (in
terms of money and time) and ease. In addition, Harris (1974) asserted that a
test is said to be practical if it is economical in terms of cost and time, easy to
administer, score, and interpret.

TEST DEVELOPMENT PROJECT (TDP)

Since testing or assessment can not be separated from the language
teaching practice, we are as a practitioner (a lecturer) must be able to make a
good test. However, it is not as easy as we can imagine. To make a test
everybody can, but to make a good one? No one can guarantee that every
teacher/lecturer is able to make a good one. Fortunately, Prof. Dr. M. Soenardi
Djiwandono, a professor of language assessment of State University of Malang,
provides us systematic, comprehensive step by step procedures to develop a good
test called Test Development Project (TDP).
I can say that this Test Development Project (TDP) really works as it
guides us through systematic, comprehensive, step by step procedures. It is
systematic as we can not go on to the next step before the first step is truly done.
It is so, as the first step is a foundation of the next step. While it is said
comprehensive since the TDP comprises every aspect or component of any
language test. In the TDP proposed and developed by Prof. Dr. M. Soenardi
 Ary Setya,Developing Writing Test 91

Djiwandono, there are five stages of developing any language test—language

skills or language component as well. They are (1) preparation, (2) test plan, (3)
test development, (4) test validation, and (5) test try-out.
The first stage in TDP is preparation. In this stage, we must identify the
prospective test takers such as the institution, the level, number of the students,
and other related information to test writing. It is very important as different
tests are intended for different institutions or schools, or groups of test takes,
levels, and the like.
The second step is test plan. Finishing first step, we go to plan the test.
Starting from the available general and specific course objectives, we collect the
data. If the available course objectives are not appropriate, we proposed the
appropriate one and of course with justification. Then we determine the general
and specific test objectives. The general test objective must be in line with the
general course objective and specific test objectives must be in line with the
specific course objectives. Then we determine the kind and format of the test we
developed. Determining those two aspects are very crucial as it is a reflection
whether the test we developed is appropriate. Moreover, every kind of language
testing has its format. For example, you can not use multiple objective tests for
speaking test. After we decide the most suitable format for our test, we also
have to make Table of Specification (TOS). TOS covers specification of what we
will test and its proportion. Then we need to identify the source of the test
materials, the text used in the test, and the test item.
After completing the second stage, the next step is test development. Here
we develop the test draft, determine the answer key or scoring guide we will use
to check the students’ answer, and determine how the way to score and how
scores are converted to grades.
Finishing completing the third stage, the next is test validation. To validate
our test, Prof. Soenardi proposed three steps. First, we give our test draft to the
expert. It means we seek for the expert validation. The expert will review of the
entire test we made and we will get a feedback from the expert concerning to our
test draft’s content. Then we revise our test draft based on his/her feedback.
Next, we determine what kind of validity we will use and how to apply it. As
what Prof. Soenardi stated that validity is unitary concept that means we can
choose the appropriate one to validate our test. For reliability, the last step in
TDP is conducting the try-out. After we do the try out for our test, we analyzed
the result of our test. We identify the problems we found when conducting the
try out and the weaknesses of our test. Then we improve and revise our test to be
a better one.

CONCLUSION
Every teaching and learning process needs an evaluation. In this paper I
have briefly touched upon several issues related to assessment that writing
teacher should be aware of. A solid understanding of assessment issues should
be part of every teacher’s knowledge base, and teacher should be encouraged to
equip themselves with this knowledge as part of their ongoing professional
development.
92 Lingua Scientia, Volume 3, Nomor 2, Nopember 2011

REFERENCES
Bachman, L.F. 1990. Fundamental Considerations in Language Testing. Oxford:
O.U.P.
Bell, R.T. 1981. An Introduction to Applied Linguistics. London: Batsford
Academic and Educational Ltd. (Appendix C, “Language Testing”).
Brown, H. D. 2001. Teaching by Principles. An Interactive Approach to Language
Pedagogy (2nd ed). White Plains, New York: Pearson Education.
Brown, J.D. 2005. Testing in Language Programs: A Comprehensive Guide to
English Language Assessment. New York: McGraw-Hill.
Burgess, Sally & Head, Katie. 2005. How to Teach for Exams. Jeremy Harmer
(Ed). Harlow: Pearson Education Limited.
Byrne, D. 1988. Teaching Writing Skills: Handbook for Language Teachers.
London: Longman Group UK Limited Company.
Cohen, M., and Margaret, R. 1989. The Effect of distance on Students’ Writing.
American Educational Research Journal Vol.26, no.2, pp.143-159.
Das, K Birkam (Ed). 1989. Principles of Language Learning and Teaching.
Singapore: SEAMO Report Language Center.
Djiwandono, M. Soenardi. 2008. Tes Bahasa: Pegangan Bagi Pengajar Bahasa.
Jakarta: PT. Indeks.
Eanes, R. 1983. Content Area Literacy: Teaching for Today and Tommorow.
Albany: Delimar publisher.
Harris, D.D. 1974. Testing English as a Second Language. New York: McGraw-
Hill.
Heaton, J.B. 1988. Writing Engish Language Test. New York: Longman, Inc.
Hughes, A. 1989. Testing for Language Teachers. Cambridge: C.U.P.
Inman, B.A & Ruth, G. 1979. Aspects of Composition. Second edition. New York:
Harcourt Brace Jovanovich, Inc.
Jacobs, H.L., Zinkgraf, S.A., Wormuth, D.R., Hartfiel, V, F. & Hughey, J.B. 1981.
Testing ESL Composition: A Practical Approach. Massachusetts: Newbury
House Publishers.
Langan, J. 1985. College Writing Skills with Reading. New York: McGraw-Hill.
Latief, M.A. 2000. Validitas Hasil Pengukuran. Bahasa dan Seni, 28 (1): 95-104.
Leki, I. 1994. Teaching Second-Language Writing: Where We Seem to Be. In Karl,
T. (Ed.), Teacher Development: Making the Right Moves (pp. 170-178).
Washinggton, Dc: USIA.
Reid, J. 1993. Testing ESL Writing. New York: Prentice Hall.
Richards, J.C. 1990. The Language Teaching Matrix. New York: Cambridge
University Press.
Rivers, W. M. 1987. Interactive Language Teaching. Cambridge: Cambridge
University Press.
Savignon, Sandra J. 1983. Communicative Competence: Theory and Classroom
Practice. Massachusetts:Addison-Wesley Publishing Company.
Swales, J.M. 1990. Genre Analysis: English an Academic and Research Setting.
Cambridge: Cambridge University Press.
Troika, L.Q. 1987. Handbook for Writers. New Jersey: Prentice Hall Inc.
Weir, C.J. 1995. Understanding and Developing Language Test. Singapore:
Phoenix.
Ary Setya,Developing Writing Test 93