You are on page 1of 9

See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/308749113

Discrete-point vs. integrative testing

Article · January 2017

CITATIONS READS
5 14,132

1 author:

Sahbi Hidri
Université de Tunis
29 PUBLICATIONS   180 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Linking a Suite of Examinations to the CEFR View project

Measuring the construct of writing in a static vs. dynamic assessment exam among ESP learners View project

All content following this page was uploaded by Sahbi Hidri on 01 April 2021.

The user has requested enhancement of the downloaded file.


Discrete Point and Integrative Testing 5001

Geyser, G. (2010). Progress testing in TESOL. Retrieved from http://www.ESL Article.com


Shohamy, E. (1992). Beyond proficiency testing: A diagnostic feedback testing model for
assessing foreign language learning. The Modern Language Journal, 76, 513–21.
Shohamy, E. (2001). The power of tests: A critical perspective on the uses of language tests. Harlow,
England: Pearson.
State of California, Department of Education. (No date). Test performance descriptors
(p. 11). Retrieved from http://www.celdt.org/resources/im/

Suggested Readings

Coombe, C., Davidson, P., & Lloyd, D. (Eds.). (2005). Proceedings of the 7th and 8th Current
Trends in English Language Testing (CTELT) Conferences, Vol. 4. Dubai, UAE: TESOL Arabia
Publications.
Gu, L., & So, Y. (2015). Voices from stakeholders: What makes an academic English test
international? Journal of English for Academic Purposes, 18, 9–24.
Milanovic, E. (2002). Common European Framework of Reference for Languages: Learning,
teaching, assessment: Language examining and test development. Strasbourg, France: Council
of Europe.

Discrete Point and Integrative Testing


SAHBI HIDRI

Framing the Issue

Language testing trends are dictated by the views of language and language learn-
ing. Teaching methods and/or approaches, such as the grammar-translation
method (GTM), the audio-lingual method (ALM), pragmatic expectancy grammar
(PEG), total physical response (TPR), communicative language teaching (CLT)
require different learning skills, strategies, methodologies, teaching styles and phi-
losophies all of which have a deep impact on testing, leading to many testing tech-
niques (Spolsky, 1985). There have been major disparities between discrete-point
(1960s) vs. integrative testing (1970s). In order to grasp the unremitting develop-
ments in teaching and testing, an investigation of the relationship between lan-
guage testing and applied linguistics (AL), and whether the latter has impacted
language testing paradigms, has an overriding consideration. AL tries to find
remedies to language problems by probing into the learning aspects on the one
hand and knowing about the language on the other and it is at this level that

v8_lbp-D.indd 5001 10/4/2017 2:58:06 PM


5002 Discrete Point and Integrative Testing

designing tests resides in converting AL theoretical considerations on the notion of


construct (McNamara, 2004) into practical tips. Therefore, any reshaping in lin-
guistic theory necessitates a change in language testing.
With the appearance of the ALM in an era that was also called “Structuralist”
(Carroll, 1961/1972), learning was perceived to rest on habit formation where oral
language is prioritized over the written one. The testing paradigm of the ALM was
labeled as “discrete-point” (Carroll, 1961/1972) and it overwhelmingly praised the
use of multiple-choice (MC) items, such as the TOEFL whose initial major purpose
was to measure the test takers’ knowledge of the various grammar structures. The
theoretical views underlying discrete-point assessment indicated that different lan-
guage elements and structures are isolated and that both kinds of language, written
and spoken, are similar. For instance, drilling and memorization of linguistic struc-
tures are delivered in short bits of isolated utterances in which learners have to
manifest knowledge of language rather than its use. The teacher, who is the only
source of information, is in charge of developing writing assignments for her class.
What is typical about discrete-point testing is the discrete nature of bits of language.
The role of test developers is to measure samples of language structures for test tak-
ers who are supposed to generalize them to other language forms. For instance, in
testing listening, test takers are administered short oral utterances to measure their
oral comprehension. Test items, such as true/false questions or prompts, MC and
picture tasks, can be employed to test segmental features of minimal pairs, stress,
intonation, grammar, and vocabulary. In addition, discrete-point tasks hinge on the
use of phonemic discrimination, response paraphrase and evaluation.
Given the limitations that impinge on context, the PEG approach emerged. The
testing paradigm is called integrative which means the ability to predict or guess
words/items in an utterance (Oller, 1979) where C-tests were thought to be a good
measure of overall proficiency. The PEG approach, also called “functional,” grap-
ples with uses of elements of sentences into a larger context of utterances character-
ized by a list of all possible functions that users may encounter. Integrative testing
presupposes an inductive mastery of grammar rules both in terms of knowledge
and use in which the test takers are asked to use language in context, such as gap-
filling (cloze tests), summarizing, and dictation. In so doing, different abilities come
into play, such as knowledge about vocabulary, grammar, discourse, reading and
writing skills, and strategies. Hence, all skills become important. Dictation is the
second most heavily used testing technique that requires short-term memory
(Farhady, 1979). Thus, a corollary to this trend is the notion of subjectivity that was
very common in scoring such tests in the absence of validity, reliability, and statisti-
cal tests that would lead to fair judgment on test takers’ ability. As a case in point,
the theoretical view of listening implicit in integrative testing requires a blend of all
the bits of sentences. For Heaton (1988), it is better to target three types of knowl-
edge in integrative testing. For instance, in testing reading, these types are linguis-
tic, textual, and background where the latter includes knowledge of the world.
Advocates of integrative testing (e.g., Oller, 1979) maintain that dictation, cloze
tests, and writing are a good sample of integrative test items. This has to be used in
longer texts and probably in a more extended discourse.

v8_lbp-D.indd 5002 10/4/2017 2:58:06 PM


Discrete Point and Integrative Testing 5003

Making the Case

Many studies have investigated the divergences between discrete-point and


integrative testing. For instance, in a study on a comparison between recall and
summary techniques in measuring reading comprehension ability for beginners
in French, Riley and Lee (1996) stressed the major differences between both test-
ing techniques. They further maintained that, unlike integrative testing, such
testing methods did not offer much regarding context of utterances. Shohamy
(1984) found that in MC and open-ended tasks, test takers performed differently
and that they preferred MC test items. In using three testing techniques, cloze
items, free written recall and open-ended items, Lee (1987) pointed out that there
was a significant correlation between the nature of test items and language expe-
rience among test takers in Spanish. Wolf (1993) underlined significant effects
between MC, open-ended, and deletion cloze tasks, with test takers accomplish-
ing better results in MC items. Spolsky (1985) stressed the relevance of using two
approaches in language testing: structural and notional-functional where the
former is dependent on knowledge of grammar structures administered in a
discrete-point format while the latter is based on the use of cloze tests. In a study
on the use of discrete-point and integrative partial dictation in listening, Cai
(2012) found that partial dictation has a high value of internal consistency, as it
matched test takers’ language ability level. According to Oller (1979), discrete-
point tests sentences where the test takers should show the ability to use one bit
at a time while integrative testing measures the ability to use many elements and
that it necessitates language use and not only knowledge of it.
Despite its narrow scope in reaching high levels of face validity and good item
discrimination (Grotjahn, 1986), integrative testing has been praised for its objec-
tive scoring and high reliability and validity indices (Eckes & Grotjahn, 2006). In
addition, in integrative testing, the different elements of language are intertwined
together to test not only knowledge of the language but also its use. Many research-
ers have explored the limited scope of using discrete-point or integrative testing.
In integrative testing, several skills are blended together and they are meant to
measure test takers’ ability to comprehend and process information. Tests some-
times assess fuzzy constructs. It is, therefore, the test developers’ responsibility to
design tests that measure what they are intended to measure.

Pedagogical Implications

In order to further probe discrete-point vs. integrative testing, a special focus


will address the receptive and productive nature of questions and prompts ger-
mane to both paradigms. The productive mode touches on the act of bringing
forth input, such as what crops up in speaking or writing. The receptive mode,
however, entails receiving input and then processing it for comprehension pur-
poses, such as reading and listening. Channel refers to the medium of

v8_lbp-D.indd 5003 10/4/2017 2:58:06 PM


5004 Discrete Point and Integrative Testing

communication in a test, whether be it oral or written/audio or visual. Reading


and writing can be looked on as written channel tests where test takers observe
forms and then read or write. Oral channels are listening and speaking requiring
the test takers to hear and react to the input (e.g., listening to directions).
However, mode brings up the receptive (reading and listening) and productive
(speaking and writing) nature of skills. For instance, in the listening skill, the
channel is audio, while its mode is receptive. Whatever the complexity of the
language skills might be, test developers and raters should not mix up both con-
structs. They should be persuaded to mediate on objective scoring, practicality
of test administration and fairness (e.g., no bias toward any facet). One kind of
test item is the use of close-ended formats, such as MC, yes/no, and true/false
questions. Such test items may be inauthentic, for instance, in testing a very com-
municative task, such as discussion or role play. To counterbalance this pitfall,
test designers should target other test modes and channels to set up a compre-
hensive picture of the construct being measured.
In order to further address these differences, Table 1 highlights the substantial
differences of discrete-point vs. integrative testing. From the paired comparison
samples, it could be argued that integrative testing offers better and possibly more
chances of contextualizing test items.
However, when compared to communicative testing, both discrete-point and
integrative testing do not proffer ample opportunities for test takers to be exposed
to a larger frame of context, as the two trends are biased and overlook the com-
municative nature of authentic and real-life testing situations. In Table 2, the meth-
ods and approaches in both paradigms are different, with the ALM rooted in
behaviorism and the PEG in cognition. The table describes the nature of tasks and
test items that test developers may consider in test design. For instance, discrete-
point testing focuses on targeting the aural skills, like listening, while integrative
testing highlights the use of test items, for instance, in writing. This shift paved the
way for a new perception of language and language learning that was being
shaped, moving from discrete parts in the early days of language teaching to the
idea that language would be weaved into wider structures that call for a bigger
involvement on the part of test takers.
The following item samples are associated with both paradigms. The first tech-
nique in discrete-point testing is phonemic discrimination. A test item such as mini-
mal pairs consists of administering isolated sentences where test takers are asked
to select the word they hear. This technique is called discrete-point because no clue
on context of the utterance is provided. The examples provided here deal with
testing the listening skill.

Example 1
Test takers hear:
Mary got dizzy and hit her head on the wall.
Then they read:
Mary got dizzy and hit/hid her head on the wall.

v8_lbp-D.indd 5004 10/4/2017 2:58:06 PM


Discrete Point and Integrative Testing 5005

Table 1 Differences between discrete-point vs. integrative testing.

Discrete-point testing Integrative testing

Language embodies bits of linguistic structures. Language typifies an amalgamation of


components.
Language items are fragmented and Language items are integrated and
decontextualized. contextualized.
Test items cannot predict test takers’ It is possible to predict the test takers’
performance. performance.
Test takers should manifest their knowledge of Test takers should show not only their
grammar structures. knowledge of structures, but also their
ability to use them in context.
Language rests on habit formation. Language is grounded in guessing.
Skills and subskills (listening, speaking, Skills and subskills are tested in a
grammar, reading, writing, phonology, combined way.
morphology) are tested in a discrete way i.e.,
separately.

Table 2 Nature of discrete-point vs. integrative test items.

Testing paradigms

Discrete-point Integrative

Methods and approaches ALM: behaviorism PEG: cognition


Nature of tasks and test items True/False Cloze tasks
Yes/No Summarizing
MC Dictation
Phonemic discrimination Translation
Minimal pairs Essays
Paraphrase recognition Oral interviews
Response evaluation Interview
Phoneme recognition Role play
Spelling Gap-filling
Grammar Summarizing

The guessing effect for this item would be 50%. This task can be criticized for
being inauthentic, since it is very rare for test takers to be exposed to similar situ-
ations. The second testing technique is paraphrase and recognition. For instance, test
takers hear an utterance then they are instructed to select one option from a range
of options, preferably four, that best reflects the main idea of the sentence or
utterance.

v8_lbp-D.indd 5005 10/4/2017 2:58:06 PM


5006 Discrete Point and Integrative Testing

Example 2
Test takers hear:
Three men broke into the house
They read

a. Three men furnished the house.


b. Three men left the house.
c. Three men stole the house.
d. Three men painted the house.

The purpose of the task is to check whether test takers are capable of discerning
the meaning of “broke into.” The four options should focus on the meaning of one/
two items of the original sentence. It is a paraphrase in a sense that the test takers
should be equipped with the ability to select the most paraphrased item that best
reflects the word(s) in the main sentence. The third testing technique is called
response evaluation. Test takers hear a question then select the most appropriate
option from a four-option test item. The tasks are administered in short decontex-
tualized sentences or utterances.

Example 3
Test takers hear
How often do you go to the cinema?
They read

a. Yes. I do.
b. Once a month.
c. No. I don’t.
d. Around 200 dollars.

In integrative testing, different techniques are utilized. For instance, dictation


integrates more than one skill, such as memory span, spelling, sound segments,
grammar, lexis and comprehension (Heaton, 1988).
Example 4
In his dream, the (1) …………………… boy saw many people (2) ……………………
a large and green (3) …………………… decorated in flowers and (4) ……………………
trees. He did not (5) …………………… anyone, as they all (6) …………………… new
faces to him. (7) …………………… kept wandering, trying to (8) ……………………
his girlfriend. She promised (9) …………………… be here on time. (10)
…………………, he heard a voice (11) …………………… his name: “What are (11)
…………………… doing here?” No one (12) …………………… invited you. I think
(13) …………………… came to the wrong (14) …………………….

The problem with dictation, unlike communicative dictation, is that test takers are
instructed to provide the exact spelling of the correct word; thus being tested on a very

v8_lbp-D.indd 5006 10/4/2017 2:58:06 PM


Discrete Point and Integrative Testing 5007

narrow range of skills. Therefore, it would be safer if they are provided with longer
pauses to allow them to insert the right word. Test developers should avoid deleting
words from the very first sentences of the input. They should allow some time for test
takers to get used to the text and build some background knowledge on it. Reading or
playing longer dictation passages in one shot may impose a very heavy memory bur-
den on test takers. Rather, it should be done in meaningful units. Sometimes, there are
problems with a deleted word, as it may turn out to be a high frequency word that is
overused in the passage, such as words 12 and 13 in example 4. It remains a matter of
common sense then for the test designer to decide on what to delete.
Cloze tests are often administered in reading tests where every nth word (e.g.,
fifth, seventh, or tenth word) is deleted and test takers are asked to complete the
blanks with the missing word. Cloze tests measure a wider range of skills of
language ability than dictation.

Example 5
I have always kept wondering about cloning and its side effects on people. I have read a
lot about the reasons behind cloning and I still cannot come to grips with the real reasons
that push people to (1) ………………….. to cloning: why do people (2) …………………..
to have a replica of (3) …………………..? Why do they pay all (4) …………………..
money to be cloned or (5) ………………….. clone other people? Do they (6)
………………….. want to live forever? And (7) ………………….. do scientists invest
a lot (8) ………………….. such an enterprise? Is cloning (9) …………………..
accepted by all nations and (10) …………………..?

Apart from constituting low percentages of face validity, cloze tests depend on the
nature of deleted items, especially when they are functional words that could be
easily discerned without even being exposed to the listening or reading input. To
achieve a higher reliability index, cloze tests should consist of as many deleted
words as possible. For Heaton (1988), the number of blanks for advanced-level test
takers should be from 40 to 50. Test takers are encouraged to provide either the most
acceptable or exact word and they should not be penalized for spelling mistakes;
however, they should be penalized for grammar mistakes if the provided word
distorts meaning. No matter how discrete and/or integrative tests might be, it is
salient to take into account both approaches to test design. As invaluable classroom
testing and teaching techniques and even though they seem different, both testing
paradigms cannot circumvent the pitfalls of measuring a very narrow range of skills.

SEE ALSO: Behaviorist Learning Theory; Cloze Tests; Mentalist Learning Theory;
Reliability; Validity

References

Cai, H. (2012). Partial dictation as a measure of EFL listening proficiency: Evidence from
confirmatory factor analysis. Language Testing, 30(2), 177–99.

v8_lbp-D.indd 5007 10/4/2017 2:58:06 PM


5008 Discrete Point and Integrative Testing

Carroll, J. B. (1961/1972). Fundamental considerations in testing for English language


proficiency of foreign students. In H. B. Allen & R. K. Campbell (Eds.), Teaching English as
a second language, New York: McGraw-Hill. (Reprinted from Testing the English proficiency
of foreign students [1961]. Washington, DC: Center for Applied Linguistics.)
Eckes, T., & Grotjahn, T. (2006). A closer look at the construct validity of C-Tests. Language
Testing, 23, 290–325.
Farhady, H. (1979). The disjunctive fallacy between discrete-point and integrative tests.
TESOL Quarterly, 13, 347–57.
Grotjahn, R. (1986). Test validation and cognitive psychology: Some methodological
considerations. Language Testing, 3, 159–85.
Heaton, J. B. (1988). Writing English language testing. London, England: Longman.
Lee, J. F. (1987). Comprehending the Spanish subjunctive: An information processing
perspective. Modern Language Journal, 71, 50–7.
McNamara, T. (2004). Language testing. In A. Davies & C. Elder (Eds.), The Handbook of
Applied Linguistics (pp. 763–83). Oxford, England: Blackwell.
Oller, J. W. Jr. (1979). Language tests at school: A pragmatic approach. London, England: Longman.
Riley, G. L., & Lee, J. F. (1996). A comparison of recall and summary protocols as measures
of second language reading comprehension. Language Testing, 13, 173–89.
Shohamy, E. (1984). Does the testing method make a difference? The case of reading
comprehension. Language Testing, 1, 147–70.
Spolsky, B. (1985). What does it mean to know how to use a language? An essay on the
theoretical basis of language testing. Language Testing, 2, 180–91.
Wolf, D. (1993). A comparison of assessment tasks used to measure FL reading
comprehension. Modern Language Journal, 77, 473–89.

Suggested Readings

Brown, H. D. (2004). Language assessment: Principles and classroom practices. New York, NY:
Pearson.
Davies, A. (2003). Three heresies of language testing research. Language Testing, 20(4), 355–68.

v8_lbp-D.indd 5008 10/4/2017 2:58:06 PM

View publication stats

You might also like