Various
Approaches to Language Teaching:Language
testing involves both linguistics and psychology as it is concerned with languages
and learning. The process may be called an experiment because the objectives or
the learning tasks are defined in order to study the learner's behaviour through
instruction and it is aimed at using the statistical techniques to assess or study
the learner's behaviour and hence it is evaluative. Thus language testing may
be said to be consisting of three main factors, viz., language, learning and evaluation.
For purposes of learning, language is indispensable and hence it is extremely
difficult to consider language as a separate entity from learning. However, it
can be said that there is a very close interrelation between language and learning
and water tight compartmentalisation between the two is not possible. Language
Learning means not only the learning of a second language or a foreign language
but also the learning of mother-tongue or the first language.
There have been in the past different schools of thought in the context of language
as strictly compartmentalised entities. Such scholars do not take congnisance
of the learner's needs. Their view point is that language is a thing in itself
and it exsits in the written texts and there is nothing like Spoken Language.
This view point supports the grammar-translation method of language exists only
in written texts and mostly in literary ones, emphasis in language teaching is
laid down upon the teaching of literature. In other words, the idea is that of
mind training and of transfer of training. Accordingly testing in such situations
is mainly concerned with the framing of translation tasks. As already mentioned
earlier, this school of thought gives very little importance to the learner's
needs and the content of teaching material is derived from the written literary
texts which is considered to be the Real Language. The only aspect taken into
consideration by proponents of this view point is that some attempt is made to
arrange the instructional material in such a way that the learner begins from
the easiest material and proceeds towards more difficult material. Such sequencing
does not have any scientific basis and it is mostly subjective.
The second school of thought is that language is a kind of machine which acts
as the stimulus-response mechanism. According to the proponents of this view point,
learning is the extreme form and it is said to be consisting of conditioned responses
to various situations in the environment. Language is argued to be not possessing
any direct connection with the environment but it is said to be a part of human
behaviour. Therefore, scholars supporting this view consider that there is no
need to link in any meaningful way with the world to which it refers. Harris and
some others argue that linguistic analysis proceeds theoretically from sound to
sentence and it is concluded that learning itself proceeds in the same way. According
to them teaching of language necessitates the analysis of language into its components
viz., the structures. This analysis is called the immediate constitute analysis.
Under this system instruction starts from the smallest units and proceeds from
sound to sentence making use of different teaching aids and machinery This is
where the use of language laboratory comes into picture and the method of instruction
may be said to be the Audio-Lingual Method. The assumption is that learning takes
place by generalisation and by analogies. In this method drills and exercise are
considered to be important and therefore, testing invariably consists of items
parallel to drills and exercises. To a very large extent the tests of the type
mentioned will at best indicate whether the learner is able to use the language
in real life situations. The latter aspect could be taken care of if and only
if open ended items like translation, guided and free compositions and free response
items are included in the test.
The third school of thought that is worth mentioning in this context is that language
is creative and simultaneously it is also rule based. Scholars like Chomsky who
hold this opinion argue that language and learning are interdependent. The language
described by them is an abstraction which is far away from learning in its formalization
as opposed to the other schools of thought mentioned above. As yet the influences
or the applications of this approach to language teaching are not crystal clear.
The three major assumptions that are made in this context are:
1.
The aim of language teaching is to develop the grammatical and communicative competences
among the language learners. According to Chomsky underlying of surface realisations
there exists a deep structure and in the context of language learning the learner
has to master the deep structure and the operation of various transformations
employed.
2.
A second language and foreign language learner has to adopt the same kind of strategy
as one does in learning the mother-tongue or the first language learner has to
adopt is not yet known clearly as mentioned by Lyons and Wales. But the transformation
of the generative approach is directed towards the way native speaker learns his
own language, its formalization will be in L1 terms and language teaching making
use of it must also approach the second and foreign language learners in the same
terms.
3.
The third assumption is related to the learning of mother-tongue as it accepts
the demands of the situations or the environment in which the language is used.
This point of view emphasizes tasks and problems rather than patterns and repetitions
in language teaching and at the same time strategies are emphasized as opposed
to simple memory. Therefore, the testing strategy related to the third school
of thought will be much more situational and will be less concerned with pure
language features such as the segments, stress, etc. Tests related to this view
point set problems which may not be too far away from the translation problems
but are directly related with the corpus of instruction. Therefore, comprehension
and composition which need resolutions are included in such tests. The test items
in such tests will also aim at assessing the creative ability of the learners.
The present day language teaching is mainly based upon the belief that spoken
language must be taught first before proceeding to the teaching of written language.
While the teaching method employed is mainly Audio-Lingual, the instructional
material is normally prepared considering of conversations or dialogues employed
in various real life situations. Therefore, the method is a conglomeration of
various existing methods and therefore, it may be called an eclectic method.
The strategies and techniques for the construction of good language tests have
been discussed in the present volume in addition to the discussion on evaluation
of methods, materials and media and provision of a set of instruments meant for
such evaluation.
The need for measurement and evaluation has been discussed providing a set of
instruments for the evaluation of Language Instructional Material.Later emphasis
has been laid down upon the measurement of performance of the language learners
and discussion of various factors involved in the development of instruments for
testing the achievement made. The techniques of testing of the major important
language skills, viz, listening, speaking, reading and writing have been discussed
in detail providing examples from Indian language wherever necessary with a hope
that they will aid the test maker in developing good language tests. In addition
to this, some light has been thrown on the aspects of testing of Culture and Literature.
Needless to say that not much literature in the area of language evaluation and
language testing is available and the present volume is aimed at providing basic
guidelines for the test makers in Indian languages. It is hoped that this volume
will prove to be useful for all those who are engaged in language teaching and
language testing with reference to Indian languages in particular.
Some
Basic Linguistic Factors and Language Testing:As the theory of linguistics
has been applied to the task of language teaching, the application of linguistics
to language testing cannot be ruled out. Language teaching and testing are so
intensely interrelated that the basis for teaching as well as testing is the same.
Therefore it would be necessary to consider those basic linguistic factors which
play an important role in language testing. Broadly speaking, the following five
factors may be considered as important in the context of language testing:
1.
A language is a set of habits.
2.
Language is primarily speech and secondarily writing.
3.
A language is what its native speakers say and not what some one thinks ought
to say (language is descriptive and not prescriptive).
4.
It is the language that should be taught and not facts about the language.
5.
Language learning is not simply the mastery of the 'code' but also the 'use of
code'.
Let us consider each of the above factors and see to what extent they play a role
in the context of language testing.
1.
A Language is a set of Habits :A second or foreign language learner can be
said to know a language or to have attained proficiency in the target language
if and only if he can respond quickly and automatically n natural language situations.
It is therefore necessary for the test maker to decide to what extent speed should
be a factor in the test. A speed text may be defined in its strict sense as the
one which contains such easy items that all the testees can answer them correctly
given enough time. Though this is the precise and strict definition of speed test,
in practice very few tests are pure speed tests. On the other hand there is another
category to tests viz., power tests which can be timed in such a way that only
certain group of examinees are able to complete answering before the time is called
and the writer of language tests can be set to measure the extent to which the
learners have mastered the set of habits. Incidentally, the difference of power
test as opposed to the speed test is that the difficulty of the items is steeply
graded, some items being too difficult for most or any of the subjects to answer
correctly though they will probably have time to reach them1. Regarding the general
language tests, it may be said that most of them are not highly speeded, but aim
at the automatic response from the examinees. The objective types of items possess
an advantage over the free-response types in the sense that the responses for
such items can be made more quickly.Needless to say that in a second or foreign
language learning situation, the learner generally have a tendency to transfer
their mother tongue habits to the target language. It is in this context that
the use of contrastive analysis has been made the results of which form the basis
for selection and gradation of language instructional material for teaching a
second/foreign language. Particularly in the context of second or foreign language
testing, the examines of which possess the same native language background, it
is always useful to use the results of contrastive analysis and to include those
language patterns which are likely to pose problems for the learner due to the
interference of his mother tongue habits. In constructing the test items involving
the problem areas due to interference of habits from the mother tongue to the
target language, the examiner must make use of the mother tongue habits of the
learners in constructing the distractors for the test items. For example, a test
item of the following type in Telugu is most likely to pose the structural problems
for the Hindi speakers. In Telugu, the first person plural haas two different
words viz.,
1.
memu (exclusive of the listener) and
2.
manamu (inclusive of the listener)
miru ikkada undandi
gudiki vel?l?i
vastam
(You stay here. We will go to temple and come back)
The blank in the above item should be filled in by using 'memu' and not 'manamu'.
But in Hindi, irrespective of whether the first person plural is inclusive or
exclusive, the blank can be filled in by using ham (we).
From the practical point of view, it may not be possible to be more dependent
on contrastive analysis in the testing of a language as a second or foreign language.
If a language test has to be designed for a wider audience who may possess varied
language backgrounds, the procedures of contrastive analysis become more inapplicable.
It is therefore, necessary that the test maker should find out some means of sampling
the total repertoire of the target language pattern and construct the test consisting
of items representing different aspects of the target language and its use. The
general argument that some items in the test will be easier for the students of
particular language backgrounds than others cannot stand valid and need not be
considered as a defect because the test is focused upon various factors of the
target language and reflects the actual language learning situation.
2.
Language is Primarily Speech and Secondarily Writing : In the context of second
and foreign language teaching in the present day times, more emphasis on developing
proficiency in the spoken language is given and it is only after ensuring that
the learners have attained an optimum level of achievement in the spoken skill,
they are led to the written skills. However, by the end of the instructional programme
the learners are expected to achieve proficiency not only in the spoken skills
but also in the written skills. It is for this reason that the first levels of
language instruction and language testing are mainly based upon the spoken language
and the measurement of listening comprehension and production are given importance.Adequate
and appropriate methods of testing the listening comprehension, auditory perception
and the testing of speaking have been developed and each have been discussed in
detail in this volume. But it should be noted here that for he testing of the
spoken skills, special facilities and amenities like the language laboratory,
the tape recorder etc., are required for efficiently testing these skills and
such facilities are not generally available everywhere. Therefore, for this reason
the efficient testing of spoken skills may not be possible everywhere.
3.
A Language is what its native speakers say, not what some one thinks they ought
to say: The belief that a language is what its native speakers say and not
what some one thinks they ought to say, implies that the language instructional
materials should be descriptive. In other words such instructional material should
have the basis on the actual speech of the native speakers who use different social
and regional contexts. It is such actual speech of the native speakers who use
different varieties of the language successfully in different social and regional
contexts. It is such actual usage tat has to be taken as the basis for language
teaching and language testing and not the ideal artificial variety of language
which does not have any existence perhaps outside te classroom. The test maker
also has to keep in mind the features of the spoken language as opposed to the
highly structural conservative grammatically of the written language.In the early
years of language testing, most of the tests required the examinees to make decisions
about the grammatical accuracy of language. An English example of such tests is
as follows :
Directions :- Insert 'shall' or 'will' in the following.
Don't worry, I ---
have plenty of time.
Along with the progress of developments in the area of
language testing a clear cut distinction has been made between the problems involving
the colloquial language and those calling for the sensitivity to written style.
The instructions for the former type of question are simply to ask the testee
to select the alternative that sounds exactly like what a native speaker of the
target language would say.
4.
It is language that should be taught and not facts about the language :In
the name and guise of language teaching, most language instructional programmes
are concentrating on teaching more about a language and not the language itself.
The basic theory of language testing is that, what is tested in the examinations
is what exactly is taught. Therefore, the so called language examinations or tests
concentrate on assessing the learner's knowledge about the language and not his
knowledge of the language. It may be said that the language teaching existing
in many educational organization is not skill oriented. Pattanayak has rightly
pointed out that "all the current examinations test the student's knowledge
about a language rather than his performance in it."2In the context of language
instruction, grammar should be used as a means to an end and not as an end in
itself. Teaching the student to make statements about the language is a wasteful
actively that deprives him of the precious time during which he could try to develop
the language skills or he could be made to learn to imitate the forms of the target
language correctly and practice them until they become automatic. In the initial
stages of language instruction, perhaps some grammatical explanation may be occasionally
useful is enabling the learner to understand the new patterns of the target language.
Thereafter the learners should be asked only to manipulate the patterns and not
to explain them.
5.
Language learning is not simply the mastery of the 'code' but also the 'use of
code' :Learning of a second or foreign language involves the mastery of not
only the structure of language, the vocabulary etc., but also the various kinds
of language use in different social and cultural contexts. In other words it involves
the mastery of code in addition to the use of the code. If a learner of a second
or foreign language is said to have attained proficiency in the target language,
it implies that he has attained adequate control of the sentence patterns, the
vocabulary etc., in addition to the mastery of various kinds of language use in
different contexts. Thus learning a language involves acquiring knowledge of the
code together with the ability to use that knowledge in producing appropriate
utterances and in understanding what is said by other native speakers.In order
to be able to effectively communicate with the native speakers of the target language,
the learner should be able to control the language structures and patterns apart
from making use of appropriate styles and registers depending upon the context
of situation. This is what has been labeled as communicative competence. Thus
one should keep in mind very clearly that mastery of a second or foreign language
is attaining proficiency both in grammatical competence as well as the communicative
competence.From this point of view, it would be necessary to test the examinee's
ability in both the area of grammatical and communicative competence if one has
to test the language proficiency of a learner. It is in this context that tests
should be focused on both these aspects. The aspects of testing language proficiency
discussed in this volume will highlight these factors in detail. These arguments
hold good in the context of language testing also. Therefore, the language tests
must be constructed in such a way that specimens of the language or other direct
responses to language are evoked. They should not simply not be asked to identify
the parts of speech or sentence types but should have a chance to exhibit his
ability or competence to use the 'code'.
Basic
Ideas in Measurement and Evaluation
Measurement
and Evaluation :
The terms Measurement and Evaluation though possess distinctly
different meanings are quite often confused and are frequently used interchangeably.
The term Measurement refers to quantitative descriptions behaviour, things or
events i.e;., behaviour described in terms of numbers.The term Evaluation provides
a broader concept than Measurement. While it involves quantitative descriptive
i.e., description expressed in words. In other words, Evaluation involves the
interpretation of what is measured. In addition to the description of behaviour
in terms of numbers and words, Evaluation also includes value judgements about
the thing described. In the context of evaluating the achievements of a student
in an instructional programme, apart from the performance evaluation, the evaluator
considers the effectiveness of instruction in terms of methods, materials and
media. Therefore the primary or the basic criteria to be used for Evaluation are
the course objectives or instructional objectives or criteria by which worth is
determined. Thus defining the instructional objectives (major and minor) clearly
and preferably and first step in the process of Evaluation.
Following are procedural
steps typically followed in the evaluation of student's achievement :
1.
Identification of course objectives (the expected or desired learning outcome).
2.
Defining the objectives in behavioural terms (in terms of the learner's terminal
behaviour).
3.
Constructing appropriate tools or instruments for measuring the behaviour.
4. Applying or administering the tools/instruments and analyzing he results
to determine the degree of learner's achievement in the instructional programme.
Thus above four steps are basically the same in the evaluation of instruction,
curriculum or the programme as a whole.Both Measurement and Evaluation require
a broad variety of tools or instruments such as tests, rating scales, inventories,
checklists, questionnaires etc.
Basic
Factors in the process of Measurement :
The process of measurement is
generally thought of in terms of quantitative descriptions of the measured phenomena
or in other words the process is thought of in terms of numerical test scores.Measurement
in Education broadly consists of arranging or ordering of individual learners
in accordance with their responses to specific tests connected with the learning
that important elements in this process are :
1.
The Test Situation (to which the individual learners respond)
2.
The Individual Learner's Responses to such situations.
3.
Ordering of Individual Learner's based on the assessment judgement of their responses
expressed in terms of scores/grades.
1.
The Test Situation :The first and foremost element in the process of Measurement
is that there will be a number of well-defined test situation top which the individual
learners respond. They present stimuli (or tasks) which require appropriate responses
on the part of the individuals. Such situations may take a broad range or variety
of forms such as :
(a) Essay questions
(b)
Paragraph questions
(c)
Short-answer questions
(d)
True/False questions
(e)
Multiple choice questions etc.
Apart
from these usual types of test items, certain situation may involve role-play,
construction of material objects or manipulation of apparatus etc.These situations
may be administered through various types of media viz., printed page, tape reorders,
films and film strips, oral instructions, physical gestures or combination of
two or more of these. The mode of responses elicited by the test situations may
be verbal, non-verbal, cognitive, affective or manipulative.
There are two restriction on any test situations to be observed in the process
of measurement :
(i)
Each situation must be exactly repeatable from one occasion to another, to the
extent possible. It should be identical for all students in the group of learners
being tested.Although this restriction could be observation in most of the test
situations, it is relatively difficult in the case of oral tests. In the case
of oral tests, each situation is normally unique to the particular occasion as
the questioning may proceed in any kind of unexpected and unpredictable directions
depending upon the examiner's interests and the sequences of responses given by
the examinees.
(ii)
The test situations will be identical for all individual in the group of learners,
as far as possible.This restriction, does not mean that the test situations
will be perceptually identical but only aim at their being able to meet a set
of objective specifications.
2.
Learner's Responses to Test Situation :It is the learner's responses to test
situations that are most crucial and important in the process of measurement,
although the first important step in this process is the test situation. The scores
or numbers which form the quantitative measurement are based upon what the learner
does in response to the test situations. These responses could either be verbal,
non-verbal, written manipulation of physical objects etc., depending upon the
type of the test.
There are two factors concerned with learner's responses viz.
(a)
Direct measurement and
(b) Indirect measurement
Direct
measurement deals with the responses as products. It is concerned with the product
in and for itself, its qualities and its desirable charactistics.Indirect measurement
deals with a symbolic response. One that stands for some kind of process that
has gone on behind the scene. The interest of the examiner is not in the symbol
itself but in the which is symbolized ; in other words, the probable series of
mental operations which the learner has performed and which he ahs indicated by
means of a symbol.
3.
Ordering of Responses :The basis and fundamental operation in the process
of educational measurement involves the judgement by the examiner(s) of the quality
or appropriateness of the response with respect to a given situation. The responses
of the students will be arranged by awarding scores which permit ordering in an
ascending order. In other words, the examiner compares the responses of the examiners
and arrives at a decision that one examinee exhibits more ability than the other(s).
Such judgements on the part of the examiners are in a way subjective as their
decisions generally depend upon what the examiners look for on the student responses,
their backgrounds, habits of perception and frames of references. The minimization
of such subjectivity is the most essential characteristic of Good Measurement.
Dyer3 suggests the following three ways through which the subjectivity in examiners'
judgements can be minimized :
(a)
The examiner makes his judgements following the actual productions of the responses
by students (Essay. question typifies this case).
(b)
The examiner anticipates what the responses to the test situation will be and
makes his judgement. (Multiple choice question typifies this case).
(c)
The examiner attempts attempts in advance to develop models of likely responses
that he considers to be good, bad and indifferent, and then, when the actual responses
become available, he orders them by matching each one to the appropriate model.
(Short answer question typifies this case).
Types
of Evaluation:
Evaluation in the context of language may be divided into
two main varieties :
(a) On-going evaluation or continuous evaluation and
(b) Terminal evaluation
The
first kind of evaluation viz. ongoing evaluation is meant to keep on getting regular
feed back at every stage of h programme during its process viz. planning, preparation,
production and application, This would enable the programme to see the success
or otherwise of it.
Again two kinds of evaluation could be thought of from another point of view and
they are
(a)
Brief Evaluation and
(b) Extensive Evaluation
(a)
Brief Evaluation :Extensive Evaluation involves the analysis of a programme
in all its main and sub aspects. The evaluator has to rate and weigh each of them
individually and consolidate the total rating based on which he makes his value
judgement. This is more objective and valid.The kinds of evaluation have been
further classified into two categories viz., Formative and Summative Evaluation.
(b)Formative
Evaluation :Formative Evaluation is that process of evaluation which is done
from time to time in the case of an instructional programme and from quality either
of the instructional programme, the techniques and methods, materials or media.
(c)Summative
Evaluation :Summative Evaluation is that kind of evaluation which takes into
consideration the periodic evaluations that have been done and in addition, a
total evaluation of the programme, process or product is made and the conclusions
are arrived at keeping in view the outcomes of the periodic evaluations in addition
to the final evaluation.
Testing
or Measurement in the Context of Language Teaching/Learning:
Programme
evaluation involves the evaluation of the teaching methods, media of instruction,
language instructional material in language instr4uctional material in language
education in addition to the learners' performance. Language tests are the measuring
tools to assess the learners' achievement and therefore, they are applied to the
learners and not to the materials or the methods or the teachers. They are designed
to measure the learners' knowledge of the language being learnt or his competence-both
grammatical & communicative-in the target language at a particular time during
the course of language instruction. Such knowledge of the others or with a standard
norm that may be fixed. Measurement is what the results of the tests show which
in itself does not have much meaning. But the inference or conclusions that can
be drawn from the measurement is more crucial and important and this is what is
called the Evaluation.As already mentioned earlier measurement and evaluation
are distinct from each other, but they are logically related. Measuring the knowledge
of a learner is a means of evaluation not only in respect of the learner but also
the teacher, the teaching material and the medium of teaching. In other words,
the test results are neutral and are to be drawn which fall under evaluation.
An experiment scientifically is the process of testing a hypothesis. In the context
of language teaching the term experiment may be defined as the means of relationship
between the teaching material, teaching methods and the learners' achievement.
In this context, hypotheses are framed such as if method 'A' is used with a set
of selected instructional materials, a particular group of learners will achieve
a certain amount of mastery of the target language in a fixed period of time.
In other words this is nothing but the investigations of the effect of selected
teaching material in a specific learning situation on the learners. In order to
assess which material is more effective and which method is more advantageous,
it would be essential to evaluate the effectiveness of various methods and materials.
Aims
and Objectives of Language Testing :The first and foremost aim of language
testing is that of research. Whether in the context of mother tongue teaching
or second/ foreign language teaching, tests will be essential to assess the quantum
of learning that has taken place from time to time. Such testing would be necessary
for the evaluation of either language teaching methods, materials or media and
it becomes essential to compare the experimental and control conditions to test
the research hypotheses.Language tests are generally understood to be used as
tools for measuring of the learner's competence or knowledge of the language.
Broadly speaking language tests may be said to be of four major types:
1.
Achievement
2. Proficiency
3. Aptitude
4. Diagnostic
This classifications is based upon the use of the tests.
Various
Types of Tests:
1. Achievements Tests :Achievements tests are aimed
at the assessment of what has been learnt by the learner from the language instructional
programme. In other words they are aimed at finding out the quantum of language
skills acquired by a learner during the course of instruction (or the use of such
tests is made to measure and find out how much has been learnt of what has been
taught i.e., of the syllabus). The results of such tests are made use of not only
for declaring the success or failure of a learner in the examination, but at times
they are also made use of in taking decision about the learner's future. The test
results are also made use of in making alterations or changes in the syllabus
or the teaching method or the presentation of material. Thus the uses of achievement
tests are manifold the main use being the testing of learner's achievement. It
should be noted here that the only capability of the achievement tests are manifold
the main use being the testing of learner's achievement. It should be noted here
that the only capability of the achievement test is to indicate how much of the
syllabus has been learnt, but it cannot make predictions about the learner's future
performance unless the syllabus is designed for this purpose.
2.
Proficiency Tests :Proficiency tests are used for assessing what has been
learnt may be from a known or an unknown syllabus. In other words such tests are
used to find out the knowledge of a learner that is already existing. The different
between achievement and proficiency testing is that the example, TOEFL (Testing
of English as a Foreign Language), English Proficiency Test Battery, Cambridge
Proficiency Examinations and Michigan Test of English Proficiency may be cited
as examples of Proficiency Tests.
3.
Aptitude Tests :Aptitude tests are made use of in assessing one's proficiency
in language for language use. While a proficiency test assesses the adequacy of
control in the target language, aptitude test is said to be assessing the amount
of linguistic skill required for learning of languages. The distinction between
Proficiency and Aptitude Test is very subtle and unclear. Aptitude test may be
differentiate from the Achievement Tests in the sense that the former does not
necessarily require the examinee to possess any knowledge of the language aimed
at teaching and the latter does. In the context of language testing, a Proficiency
Test may be said to be taking place at a certain point after the instruction has
started and has relation to future non-language performance. An aptitude test
is concerned with the inherent aptitude for language learning.
4.
Diagnostic Tests :Diagnostic Test differs from the other type of tests in
the sense that it relates to the use of information obtained and to the absence
of a specific skill in the learner. Achievement, Proficiency and Aptitude are
concerned with both use and skills of language, whereas a Diagnostic Test is made
use of by the teacher for the information provided from the presence or absence
of a part or whole of one of he skills. In other words Diagnostic Test is made
use of by the teacher for the information provided from the presence or absence
of a part or whole of one of the skills. In other words Diagnostic Test helps
in discovering thelearner's deficiencies in specified areas of language learning.
Diagnostic Tests generally yield a profile which is of a greater interest than
a single total score.
Some scholars have made distinctions between tests and examinations. It may not
be out of place to see how the examinations and tests differ from each other if
at all they do. Davies notes "The notion of test conjures up vague ideas
of psychology and of intelligence, whereas an examination suggests the end of
term multi-subject ordeal"4 Some scholars have an advantage in accepting
its influence over the curriculum. It may be noted here that the examinations
has an advantage in accepting its influence over the characteristics of tests
and in addition it has the feature of being able to influence the curriculum.
Normally tests and examinations are interchangeably used just as the terms measurement
and evaluation are. However a broad distinction between 'Examination' and 'Test'
could be made in the sense that the former is normally meant for assessing overall
achievement which is not followed up with remedial efforts.
In
simple terms it may be said that examination refers to the total area of language
measurement and the term 'Test' refers to a specialised part within it and therefore,
a test may be conceived as a kind of examination.Before proceeding further with
the discussion about the language tests and their construction, it would be necessary
to consider some of the criteria that good language tests are supposed to be possessing.
A good test is said to be
satisfying the following three characteristics :
(a)
The test should be simple.
(b)
The syllables should be teachable (It should reflect the syllabus that it covers
adequately and without distortion of relative weights)
(
c) The effects should be beneficial
(a) By simple what is meant is that the test should be easy to administer and
that the examiner is clear as to what is being tested and it should contain all
the necessary characteristics of an objectives test.
(b)
By teachable what is that the syllabus that forms the basis for the test should
be fairly detailed so that the instruction could be effective. It also implies
that the instruction should be possible for an average language teacher.
(c
) By beneficial, what is meant is that the test would provide the basis for the
improvement of the teaching material or the teaching techniques or the syllabus,
while it helps in placing the learners on a comparative scale.
In
addition to the above features, good tests are required to be reliable and valid.
Reliability
: The examiners would expect their tests to measure as accurately as possible
what they intend to measure. In language teaching, reliability can be achieved
only by making the tests as objectives as possible. By objectively, it is meant
that an impartial speaker of the target language would agree upon the correct
and incorrect responses of the language learners. This necessitates the construction
of test items in such away that there could be one and only one acceptable correct
response to each item.
As all the examinees cannot be said to be possessing the same intelligence, knowledge
of the target language etc., the subject unreliability cannot be done away with.
Therefore, the concern of the test makers should be to see the reliability of
the instrument leaving aside the subject unreliability. Whether the language test
is meant to assess the learner's communicative competence or his grammatical competence,
we should be clear in our minds that the testing of the total knowledge of the
language learner is impossible. Therefore, only a representative sample of the
total knowledge that a learner possesses can be tested for which a selection has
to be made out f the syllabus based on which are considered to be fair and representative
ones. Such selection or the process of sampling depends on the construct that
the examiners have. This aspects will be discussed in the next section on validity.
Language tests may be unreliable for various other reasons that the wrong or defective
sampling and they are concerned with the language data and the test instructions.
For instance a syntactic test item may contain vocabulary that the examinee does
not know. Similarly the instructions may be defective in the sense tat they maynot
be clear and may be ambiguous or contain technical words that the examinee is
not aware of. In order to avoid to avoid these factors which make the test unreliable,
the instructions must be precise, clear and unambiguous. The range of vocabulary
used in the test must be within the pupil's knowledge. Such confirmations need
the testing of tests. This is by trying them out on a sample of the same category
as the learners belong to and thereafter eliminating the items hat are found to
be defective in the try out. This is what precisely is called the item analysis.
Validity
:According to Pilliner5, the validity of any examination or test procedure
may be broadly defined as the extent to which it does and what it is intended
to do. Four kinds of validity are talked about in the context of tests viz.,
1.
Content Validity
2. Predictive Validity
3. Concurrent Validity
4.
Construct Validity
1.
Content Validity :If the test contains questions that require the subject
to perform all the activities that have been taught in the course of instruction
and accounts for the strong resemblance between the forms and exercise and the
test items and procedures, it is said to possess 'Content Validity'. Establishing
the content validity of a test does not require the comparison of results from
other tests, but it is a matter of expert judgement. This is therefore, to some
extent, subjective and becomes unreliable.
2.
Predictive Validity :If the results of a test can be used to predict the success
of the examinees in the performances of some other related task, it is said to
possess the "Predictive Validity".
3.
Construct Validity :If the results of one test are confirmed by a parallel
test of the same kind aimed at measuring the same thing they are said to possess
what is called "Concurrent Validity".
4.
Construct Validity :Needless to say that language test or any test or experiment
has a basis. In the context of language testing, it is the theory of language
that forms the basis. While making a selective sampling of the linguistic and
communicative aspects of language instruction for inclusion in the tests, the
selected items are based upon various aspects viz., the linguistic and communicative
aspects. If the test represents the items with appropriate weightages on different
aspects of language instruction and the distribution of the items is found to
be appropriate, then such a test is said to possess what is called the "Construct
Validity".
Cronbach6 in his "Essentials of Psychological Testing" has provided
the following
table
which illustrates these four kinds of validity :
| |
Question
asked |
Procedure
|
Principle
used |
Examples
|
|
Predictive
|
Do test
scores predict a certain important future performance |
Give test
anduse it to pre-dict the outco-me. Sometimelater obtain ameasure of the outcome.
Com pare prediction with outcome. |
Selection
and classification. |
Admission
testfor Medical students comp ared with later marks.
|
|
Concurrent |
Do test
scorespermit an estimate of a certain present performance?
|
Give test,
obtaindirect measure of the other per- formance.Compare
the two. |
Testsintendedas
a substitutefor a less convenient procedure. |
Group
mentaltest comparedto individual test |
|
Contents
|
Does this
test givea fair measure of performance onsome important set of tasks? |
Compare
the items logically to the content supposed to be measured. |
Achievementtests. |
A
test of shorthand ability isexamined to |
|
Construct
|
How can
scores ontest be explained psychologically? |
Set up
hypotheses. Test them experimentally by by any
suitable procedure. |
Tests
used for description oror in scientific research
|
A test
of artaptitude isstudiedto determine howlargely scor-es depend on art training,experience
inWestern Cul-ture etc. |
CRITERION-REFERENCED VERSUS
NORM-REFERENCED TESTS
The tests whether in language or in other discipline may be broadly classified
into two categories viz., Criterion-referenced and Norm-referenced Tests.In the
context of the validity of language tests we have already talked of the fact that
a test can be considered to be valid if and only if it successfully measures what
it is supposed to measure and nothing else. Normally in the second language testing
the usual aim of the tests can be attributed to what may be called "Fluency"
or "Command of the basic grammatical structures" and/or "ability
to function in the target language in real life situations". While these
attributes to second or foreign language testing sound concrete, but still more
narrowly specified and operational definitions would be necessary. In other words
the pre-requisite of the making of a test is that the objectives are clearly defined
in what may be called the behavioural terms. Such objectives should to the extent
possible be specified in terms of the learner's terminal behaviour. These objectives
stated in behavioural terms form the basis or the criteria for the construction
of a test. Thereby, it can be said that the bases of the bases of the test are
the criteria set forth as mentioned here. Such tests are usually called the Criterion-Referenced
Tests.
Norm-Referenced
Tests :
The
Norm-Referenced Tests are those which aim at placing the examinees based on their
performance in the tests against a pre-defined degree of achievement or norm.
The main consideration in such tests is the norm that is predetermined against
which the examinees are placed at different points in the scale with reference
to the standard scores or percentiles set up. The key concept of such tests is
comparing an individual with others of his references group. Therefore, such tests
are called Norm-Referenced Tests.It is needless to say that a test can be both
Criterion-Referenced and Norm-Referenced. From the point of view of the types
of test items and the skills that are aimed at testing, it may be called the Criterion-Referenced
Test and from the point of view of placing the examinees on a scale with reference
to a set norm that might determine the passing or failure, it may be called the
Norm-Referenced Test.