1.
Introduction
Entrance into institutions of tertiary education
in Japan has been described as a hierarchical
system of 'exam hell' by more than one critic
(see, e.g., Cutts 1997; Yoneyama 1999). Families
invest enormous resources, both in time and
money, to cram for the tests. In some cases,
if initially unsuccessful at entering the college
of choice, students even devote a year (or more!)
after high school to further prepare to sit
the exam yet again. To what extent this 'hell'
affects in the same way the entire college-bound
student population is open to debate. Nonetheless,
in a competitive atmosphere these tests for
entrance into colleges are held with great import
by students, parents, institutions, and the
general public. Considering the authority these
college exams hold in Japanese society, a commensurate
assessment of the quality of the tests themselves
seems lacking. Although they may have privately
held beliefs, for numerous reasons (beyond the
scope of this paper) the entrance test developers
themselves, as well as the institutions where
they are employed, are especially hesitant to
offer publicly data that would objectively evaluate
these numerous admissions exams.
This short paper will attempt to begin to address
this shortcoming in research on Japanese entrance
exams not by offering any hardcore statistical
analyses, but rather by simply calling attention
to the need for more professional responsibility
and research at the institutional level. Specifically,
the L2 proficiency portion of the entrance exam
at a small university of commerce in Tokyo will
be considered. The main question that will be
addressed is whether either a standardized or
performance-based language assessment would
be more appropriate as part of the institutional
entrance exam.
2.
Standardized & Performance Proficiency Testing:
Norm-referenced vs. Criterion-referenced
Numerous books on constructing and using language
tests have been written (see, e.g., Lado 1961;
Valette 1977; Hughes 1989; Bachman 1990; Alderson,
Clapham et al. 1995; Brown 1996). In most of
these discussions the authors classify a measurement
strategy as either norm-referenced (NRM) (i.e.,
standardized) or criterion-referenced (CRM)
(i.e., domain-referenced or "performance"
tests) (Fulcher, personal communication). Brown
(1995) clearly delineates the differences of
these two types by focusing on the categories
of "test characteristics" and "logistical
dimensions."
To elaborate briefly, NRM's are general tests
intended to be used to classify students by
percentile for measuring either aptitude or
proficiency for admissions into or placement
within a program. CRM's, on the other hand,
are more specific, achievement or diagnostic
tests intended to be used for motivating students
by measuring to what percent they have achieved
mastery of the taught/learned material. Griffee
(1995, p. 21) points out how the type of interpretation
that a NRM, such as the Test of English as a
Foreign Language (TOEFL), offers is the relative
performance of the students compared with that
of all the others resulting in, ideally, a bell
curve distribution. Contrastingly, the CRM,
such as a locally produced achievement test,
measures absolute performance that is compared
only with the learning objective, hence a perfect
score is theoretically obtainable by all students
who have a mastery of the pre-specified material,
or conversely, all students may fail the test.
Closely related to the distinction between NRM
and CRM tests is a theme that is repeated in
much of the language testing literature (see,
e.g., Brown 1996, p. v). This is the concern
that since program-level decisions and classroom-level
decisions are two entirely different categories
of testing, different assessment instruments
must be implemented. Admissions and placement
decisions are questions of proficiency and students
are ideally spread out in a continuum for which
a NRM is the test of choice. Language skills
are tested generally and students can then be
grouped accordingly into ability levels for
decisions of either admission into a program
or streaming into different classes within a
program. Comparisons of average proficiency
levels within a program, or across institutions
on a state, national, or international scale,
are other program-level concerns that are best
addressed with a NRM.
In classroom-level decision-making, on the other
hand, diagnostic or achievement assessment is
most helpful. For this end, CRMs are most accurate
in helping teachers (and administrators) to
assess the strengths and weaknesses of individual
students with regard to the curriculum goals,
as well as checking progress and achievement
within such a program. Using data from CRM testing,
teachers can be better equipped to assign final
grades for a course and administrators can make
more informed decisions about curriculum changes
or adjustments. This concern with matching informational
needs with the appropriate testing instrument
is critical to "sound testing." "Bad
or mediocre testing is common, yet most language
professionals recognize that such practices
are irresponsible and eventually lead to bad
or mediocre decisions being made about their
students' lives." (Brown 1996, p. vii).
First, a very brief review of the relevant literature
surrounding second language proficiency testing
will be undertaken. What theoretical and logistical
concerns must be addressed by the administrators
and developers of such a test? To contextualize
the discussion, the milieu in which the Japanese
college entrance exams take place must be considered,
and specifically, the situation at this college
in Tokyo will be described. Finally, based on
these discussions, an argument will be made
for implementing a standardized, norm-referenced
EFL proficiency test as part of the entrance
exam to this institution.
2.1
Theoretical & Conceptual Considerations
Beyond this basic division of NRM and CRM test
types are other theoretical considerations that
must be made when developing a language test.
Skehan (1988) points out how theory must be
the driving force behind any language test.
He mentions (p.211) that the oft heard proposal
of solving the problem of language testing by
following around one's students and observing
their communicative performance in a variety
of situations is not only impractical, biased,
unethical, and unreliable, but, above all, the
data simply may not be representative. The underlying
challenge of language testing, then, is to achieve
representativeness. Deciding what and how to
test requires recourse to a specific linguistic
theory (Henning 1987). The theoretical buck
stops with testing since "testers, by researching
into the structure of language proficiency,
are attacking fundamental problems in language
teaching and applied linguistics." (Skehan
1988, p. 211).
Davies (1978, p. 149) identified such a tension
between assessment and pedagogy when he spoke
of a testing continuum that ranges from the
analytic to the integrative. Oller's (1979)
theory of Unitary Competence Hypothesis (UCH)
purported that integrative tests would be the
best measurement tool, conflicting with previous,
analytic assumptions (see Lado 1961; Davies
1973) which assumed skills-by-levels matrices.
The Communicative Language Teaching (CLT) paradigm
for language testing holds that performance,
rather than standardization, should be the goal
of measurement. Real life tasks, face validity,
authenticity, and performance were four important
key words surrounding the CLT model for testing
(Fulcher 2000). This communicative approach,
a psycholinguistic-sociolinguistic "stage"
in the history of language testing (Weir 1990),
was what Morrow (1981) coined "the Promised
Land." Recently Fulcher (2000) reiterated
Davies' emphasis on the similarity between the
analytic-integrative and reliability-validity
spectrums and pointed out how this debate on
performance vs. discrete-item testing had actually
taken place decades before.
In fact, the competence/performance issue is
a concern in linguistics that has long interested
language-testing experts, ever since the distinction
was originally proposed by Chomsky (1965). Linguistic
competence (a person's knowledge of an L1 or
L2) is, obviously, more perfect than actual
linguistic performance (real language use) even
in 'native speakers'. When testing L2 learners,
then, an assessment of competence can only be
guessed at indirectly by estimating a student's
performance. Even a successful test of the underlying
psychological construct of competence such as
the TOEFL "only provides an estimate of
the student's performance, which is only a reflection
of the underlying construct, or competence."
(Brown 1996, p. 29)
If measuring language proficiency of learners
for admissions into university, then, administrators
must be aware that if the assessment is, by
definition, not entirely accurate, at the very
least high standards must be upheld to ensure
fairness in the selection process. In language
testing for admissions, being objective necessarily
implies addressing the concerns of both validity
and, even more important in the eyes of this
author, reliability. Does a test actually measure
what it claims to be measuring-- for example,
in a college admissions exam, general L2 proficiency
(i.e., Is the test "valid"?)? More
importantly, does the language portion of an
entrance exam give the same objective results
every time it is administered (i.e., Is the
test "reliable"?)? Understanding the
inherent inaccuracy of testing to begin with,
most standardized NRM test developers rigorously
analyze the validity and reliability of the
tools they develop to ensure that the measurement
of learner performance is at least as statistically
accurate as possible.
2.2 Logistical Concerns
In addition to these lofty concerns of theoretical
and statistical ideals, harsh reality forces
administrators and teachers to compromise in
addressing practical and logistical issues.
How to allocate limited resources is a constant
concern for any institution, especially universities.
It is rarely true that infinite time and money
exists to both teach and test languages, or
any other subject for that matter. Test cost,
construction, administration, and scoring are
logistical concerns that all involve allocation
of time and money. Most experienced administrators
will admit (e.g., Brown 1996, p. 33) that language
testing decisions are, more often than not,
determined more by these practical concerns
than by theories of language learning and assessment.
Nevertheless, within the time/cost parameters
that exist in a given entrance test environment,
test developers and examiners have a professional
responsibility to properly address issues of
theory, validity, and reliability.
3.
English Entrance Exams in Japan: cultural/situational
considerations
Japan has one of the highest rates of post-secondary
school attendance among all industrialized nations,
with 2.5 million undergraduates enrolled at
over 600 national, public, and private four-year
universities (Hirowatari 2000). Over half of
all Japanese teenagers, then, apply to take
a college entrance exam for admission into a
tertiary institution. Most such admissions exams
include a compulsory English proficiency sub-test
although EFL is not a state-required subject
at primary, secondary, or tertiary schools in
Japan. Partly because of this university entrance
exam focus on English, while only a handful
of students are exposed to language classes
in primary school, over 10 million 12 to 18
year olds, and another million or so university
students, 'elect' to study English.
3.1 "Exam hell":
Hensachi
In Japan there is a commonly held belief in
"the educationally credentialized society,"
or gakureki shakai. In many cases, the
extraordinary emphasis on ranking colleges and
universities has led to a brand-name sensitivity
that may affect a person for their entire life.
One effect of a gakureki shakai is a
phenomenon that has been labeled "exam
hell."
As was mentioned in the introduction, the so-called
"exam hell" is pressure felt by many
young adults in Japan (as well as South Korea,
Taiwan, Singapore, and other Asian countries
for that matter). Most teenagers are expected
to prove their intellectual mettle (or exam-taking
talent) on these fact-oriented exams, even though
they are rarely pushed to excel once they have
matriculated at a college or university (see
McVeigh 1997). Entrance into a university is
often equated with passing the test, and in
actuality this is often the case. Though admissions
procedures are becoming more creative in recent
years, the majority of colleges have resisted
any change in a system that has been in place,
arguably, since the Meiji Era in the late 1800s
(Amano 1990). Indeed the university entrance,
and overall education, system itself is inherently
immobile (Frost 1991; Schoppa 1991), and has
been described as a societal 'filtering' mechanism
to create a class structure where otherwise
none purportedly exists (see, eg, Cutts 1997;
McVeigh 1997; McVeigh 1998).
Students are strictly ranked according to hensachi,
the "abstract notion of a national norm-referenced
person-indexed score." (Brown 1995, p.25).
Using this score, high school and prep school
teachers advise their students about which university
entrance exams they should take based on their
probability of acceptance (a high school teacher's
reputation is on the line if their students
shoot too high and miss their mark-- conservatism
that is a necessity). In fact, the largest cram
school syndicates in the Tokyo and Osaka area
publish hensachi ranking lists of two
and four-year colleges which students and teachers
use to make application decisions.
3.2 Changing demographics
Japanese society is now faced with two demographic
challenges that have been termed shoushika
(low birthrate syndrome) and koureika
(aging syndrome). These changes, of course,
have repercussions throughout society, and schools
are already witnessing the effects. Most universities
in Japan have seen, first, a slowing in the
rising rate of applicants, and, now, an overall
decrease in the number of students sitting the
yearly exams. This has forced a normally conservative
sector of society to move in relatively innovative
ways in an attempt to counteract their growing
inability to attract students. Even top name
schools in the higher echelons of the rankings
have had to consider the ramifications of less
and less applicants each year. Not least of
their concerns is financial, of course, since
entrance exam fees are a substantial source
of revenue (in the $ millions) even for the
prestigious, but inexpensive, national universities
($150 per student). No school in Japan can afford
to sit on the laurels of past achievement and
national prestige, least of all the colleges
occupying the lower rankings. Recently college
prep and cram schools have instituted a new
"F" rank, designating those colleges
where the entrance exam is a mere formality
since any student that applies is automatically
accepted, given a "free pass."
3.3 Analyses of Japanese
university entrance exams
With that being said, the conservative world
of university entrance exams is slow to change.
Though the doors are slowly opening since schools
are acknowledging that they must lower the bar
to keep the freshman class (and, ergo, the coffers)
filled, the content of entrance exams, especially
the English tests, have changed little. Although
the administrators and board of directors may
be anxious for progressive change, the test
developers themselves control, for the most,
the content of the entrance exams.
In fact, Brown & Yamashita (1995; 1995)
conducted a longitudinal study of the content
of English language entrance exams at Japanese
universities and found insignificant change
over a period of years. This would tend to indicate
that university test developers are content
to examine the students as always. After analyzing
the content of 21 English entrance exams, Brown
& Yamashita (1995; 1995) made many important
findings, of which the following are most pertinent
to this discussion:
1. Though the Japanese Ministry of Education
guidelines implemented in 1993 strongly urged
greater emphasis on listening skills, much of
the content of the entrance exams still involve
discrete-point grammar and translation items.
2. Considering that most teachers use simplified
texts for classroom pedagogy, the readability
indexes indicated that the text passages in
the exams were quite difficult, ranging from
a 'native'-speaker eighth grade (junior high)
level to thirteenth grade (university) level.
3. The considerable variety of test items on
an English entrance exam indicates that the
ability to take exams may be more important
than actual language proficiency.
It
would appear to this author, that while the
Ministry of Education has been encouraging a
more communicative approach to language education,
most university examiners must be either:
1. unsympathetic to the aims of the state education
planners or
2. ignorant of proper testing methodology or
3. unaware of the negative effect of grammar-translation
testing "washback."
Most probably, all three reasons are partly
to blame for the present state of affairs in
entrance testing. First, Marshall (1994) adeptly
delineates the processes by which university
professors have historically tended to buck
bureaucratic shido ("guidance").
Secondly, Terauchi (2000) summarizes a Japan
Association of College English Teachers (JACET)
study that revealed an appalling lack of graduate
training in applied linguistics and English
Language Teaching (ELT) at university and college
foreign language faculties. Finally, Mulvey
(1999) has pointed out that even the possible
positive effects of "washback" from
testing grammar-translation, i.e., greater reading
ability, have not occurred in Japan.
4.
Edo College of Commerce Entrance Exams
Edo College of Commerce (ECC) is a pseudonym
for a four-year business university in Tokyo
with an undergraduate enrollment of 2800, as
well as a very small graduate program. There
is at the moment only one faculty-- commerce--
with two majors-- commerce and management. In
2001 the school expanded by adding a faculty
of management with four new majors. The overall
student enrollment has not been affected by
this curriculum change, however. The school
has a solid reputation as a school of business
and commerce, but it is clearly a third, or
fourth, tier institution. Although it has not
yet been classified as an "F" rank
school by the prep and cram school syndicates,
students with high hensachi do not normally
apply to or enroll at this school.
4.1 Entrance pathways at ECC
There are a total of nine different "pathways"
for undergraduates students to enter ECC: 3
different "Recommendation" interview
paths, a newly established "AO" (i.e.,
"Admissions Office") interview path
for students with unique qualifications, a "Sports"
scholarship pathway, a "Center" exam
that is developed and administered by a central
agency (much like the SAT in the U.S.), and,
the "Regular" path by which three-quarters
of the students enter ECC, two different locally-developed
entrance exams administered at the beginning
and end of February (n.b., the Japanese school
year is from April to March). These "Regular"
exams administered by faculty members of ECC
are entirely developed by a team of ECC full-time
professors of kokugo (Japanese), history,
math, business, and foreign languages.
4.2 English proficiency
testing at ECC
The English portion of these exams is developed
from scratch in-house every year by the entire
staff of six foreign language teachers. The
content of the language exam is for the most
part determined practically. Individual teachers
are very conscious of completing the onerous
task of test writing as swiftly and painlessly
as possible. For this reason, the previous years
exams are used as models and only slight modifications
are made year-to-year. It is no surprise that
since there is no resident testing "expert"
(among the six faculty, only one holds a PhD
in ELT, and less than half have masters level
training in applied linguistics) and nothing
but very rudimentary knowledge of language assessment,
the full-time professors among the staff at
ECC hold very little discussion regarding either
appropriate testing methodology or exam quality.
There is, however, much discussion and concern
for "saving face," i.e., avoiding
mistakes in clarity, grammar, spelling or misprints.
There is also a lot of care taken with maintaining
the security of the tests; they are kept under
lock and key as they go through the various
proof-reading processes over the course of three
months (October- December).
Interestingly, the English exam has always been
a required sub-test but became an elected test
in 2001. The top-down decision to make this
portion of the exam elective was made by the
Admissions Committee seemingly in response to
the falling number of applicants mentioned above.
The reasoning is that by not requiring English
more prospective students will be likely to
sit the entrance test, given the general aversion
to English study among high school students
in general and especially among the ECC applicants.
5.
Conclusion: Rationale for Adopting a Standardized,
Norm-referenced Entrance Exam at ECC
It is common knowledge in the fields of applied
linguistics and ELT, and education in general,
that there is no "perfect" test for
any given purpose. What is necessary, however,
is constant monitoring and improvement of any
given testing program in order to ensure both
language teachers and learners that the inevitable
compromises are both conscious and rational
decisions, not randomly based on 'blind' tradition.
This current study shows that there is clearly
room for improvement in language proficiency
testing on the ECC entrance exam. Especially
given the high face validity of all university
entrance exams in Japan, or any other society,
the test should be a rigorously analyzed instrument.
"Evidence of validity should be presented
for the major types of inferences for which
the use of a test is recommended. A rationale
should be provided to support the particular
mix of evidence presented for the intended uses…
For each total score… estimates of relevant
reliabilities and standard errors of measurement
should be provided in adequate detail to enable
the test user to judge whether scores are sufficiently
accurate for the intended use of the test."
(American Psychological Association., et al.
1985, pp. 13-20)
In conclusion, after careful consideration of
both the abovementioned theoretical and logistical
concerns of both standardized and performance
testing, as well as the state of affairs vis
a vis entrance exams in Japan in general and
at ECC specifically, rationale for adopting
a norm-referenced language proficiency exam
will be discussed. The arguments for choosing
a NRM tool rather than a more performance-oriented
CRM test will center around five issues of authenticity,
measuring proficiency and not achievement, washback,
practicality, and test evaluation and adoption.
The argument that performance-based tests are
more 'authentic' than NRM tools is tenuous.
By definition a language test is unnatural,
no matter how 'communicative' the design and
intent. "Any language test is by its very
nature inauthentic, abnormal language behavior,
for the test taker is being asked not to answer
a question giving information but to display
knowledge or skill."(Spolsky 1985, p. 39)
This being said, standardized test developers
are aware of the need for as much 'authenticity'
as possible, witnessed by the TOEFL 2000 project,
for example. Simply because a test is performance-based
does not necessarily guarantee authenticity.
Furthermore, communicative, or performance,
language tests (such as the CUEFL project) are
characteristically criterion-referenced (Morrow
1991, p.112)-- excellent for diagnostic measurements
but poor indicators of general language proficiency
of learners in relation to other learners. If
an institution is to make informed entrance
decisions being fair and accountable for its
admissions decisions, a norm-referenced test
is essential.
Test 'washback' is often cited as an argument
against standardized testing. Most informed
test developers are conscious of the possibility
of an exam's influence on language pedagogy.
This phenomenon, while not to be dismissed,
is not very well understood even by experts.
In fact, the complexities of washback, though
implicating Japanese exams in a negative phenomenon,
have not demonstrably shown with empirical evidence
the alleged connection of NRM testing with outdated
classroom pedagogy. Rather, could 'tradition'
(or, more critically, lack of proper faculty
development-- 'FD' is a popular catchphrase
at Japanese universities recently) be the 'culprit'
in a CLT gap in theory and practice in the Japanese
EFL classroom? Nevertheless, if washback does
exist in the entrance exam environment of Japan,
it is certainly not because of 'standardized'
testing, a term that is an oxymoron in the world
of Japanese university admissions. Must not
educators provide a more 'standard' assessment
tool with which to decide the future lives of
millions of Japanese youth?
At ECC, as at nearly every college in Japan,
an admissions office, or even secretary, does
not exist. There is no such title, or official
role, of "admissions counselor." The
burden of admissions policies, college recruiting,
test development, and actual administration
of these exams is all carried out by a rotating
team of faculty and administrative staff, further
supported by ad hoc committees. None of the
teachers or administrative staff involved was
hired specifically for any of these admissions
related jobs. Those 'unlucky' faculty members
and staff chosen for administrating the admissions
process do an admirable job, given the inherent
limitations and incredible logistical complexities
of handling, and testing, thousands of applicants
yearly. Quite simply, a performance type of
entrance exam would be totally unrealistic (and
probably vetoed by the entire faculty!) given
the inevitable logistical complications of testing
and marking an open-ended, interactive language
exam.
Finally, an in-house entrance testing program
developed, modified, and ostensibly improved,
yearly to meet the specific needs of the ECC
applicant population and language program would
be ideal. Such a project, however, is problematic
(if not impossible) given the (wo)man-hour resources
and expertise available for the task. There
is a clear need for considering the adoption
of a commercially available, standardized NRM
as a tool for making appropriate admissions
decisions. Though not ideal, such a test would
be a vast improvement over the unscientific
testing approach being implemented at present.
Such a test would have the added benefit of
situating ECC applicants in a national and international
macrocosm, aiding school executives, administrators,
and faculty to make more informed intelligent
admissions, language program, and classroom
level decisions.
6. References
Alderson,
J. C., C. Clapham, et al. (1995). Language
test construction and evaluation. Cambridge,
Cambridge University Press.
Amano, I. (1990). Education and examination
in modern Japan. Tokyo, Tokyo University
Press.
American Psychological Association., American
Educational Research Association., et al. (1985).
Standards for educational and psychological
testing. Washington, DC, American Psychological
Association.
Bachman, L. F. (1990). Fundamental
considerations in language testing. Oxford [England]
; New York, Oxford University Press.
Brown, J. D. (1995). Differences between
norm-referenced and criterion-referenced tests.
Language Testing in Japan. J. D. Brown
and S. O. Yamashita. Tokyo, The Japan Association
for Language Teaching: 12-19.
Brown, J. D. (1995). "A gaijin
teacher's guide to the vocabulary of entrance
exams." The Langage Teacher(December):
25.
Brown, J. D. (1996). Testing in language
programs. Upper Saddle River, N.J., Prentice
Hall Regents.
Brown, J. D. and S. O. Yamashita (1995).
"English language entrance examinations
at Japanese universities: what do we know about
them?" JALT Journal 17(1): 7-30.
Brown, J. D. and S. O. Yamashita, Eds. (1995).
Language Testing in Japan. JALT Applied
Materials. Tokyo, The Japan Association for
Language Teaching.
Chomsky, N. (1965). Aspects of the
theory of syntax. Cambridge, M.I.T. Press.
Cutts, R. L. (1997). An empire of
schools : Japan's universities and the molding
of a national power elite. Armonk, N.Y.,
M.E. Sharpe.
Davies, A. (1973). The construction
of language tests. The Edinburgh course in applied
linguistics: vol.4. testing and experimental
methods. J. P. B. A. a. A. Davies. London, Oxford
University Press.
Davies, A. (1978). "Language
testing: survey article 1." Language teaching
and linguistics abstracts 11: 145-159.
Frost, P. (1991). Examination hell. Windows
on Japanese education. E. R. Beauchamp.
New York, Greenwood Press: 291-305.
Fulcher, G. (2000). "The 'communicative'
legacy in language testing." System 28(4):
1-15.
Griffee, D. T. (1995). Criterion-referenced
test construction and evaluation. Language
Testing in Japan. J. D. Brown and S. O.
Yamashita. Tokyo, The Japan Association for
Language Teaching: 20-28.
Henning, G. (1987). A guide to language
testing : development, evaluation, research.
Cambridge, Mass., Newberry House Publishers.
Hirowatari, S. (2000). "Japan's
national universities and dokuritsu gyosei hojin-ka."
Social Science Japan 19(September): 3-7.
Hughes, A. (1989). Testing for language
teachers. Cambridge [Cambridgeshire] ; New York,
Cambridge University Press.
Lado, R. (1961). Language testing
: the construction and use of foreign language
tests ; a teacher's book. London, Longman.
Marshall, B. K. (1994). Learning to
be modern : Japanese political discourse on
education. Boulder, Westview Press.
McVeigh, B. J. (1997). Life in a Japanese
women's college : learning to be ladylike. London
; New York, Routledge.
McVeigh, B. J. (1998). The nature
of the Japanese state : rationality and rituality.
London ; New York, Routledge.
Morrow, K. (1981). Communicative language
testing: revolution or evolution? The communicative
approach to language teaching. C. J. Brumfit
and K. Johnson. Oxford, Oxford University Press:
143-157.
Morrow, K. (1991). Evaluating communicative
tests. Current developments in language testing.
A. Sarinee and Regional Language Centre. Singapore,
SEAMEO Regional Language Centre: 111-118.
Mulvey, B. (1999). "A myth of
influence: Japanese university entrance exams
and their effect on junior and senior high school
reading pedagogy." JALT Journal 21(1):
125-142.
Oller, J. W. (1979). Language tests
at school : a pragmatic approach. London,
Longman.
Schoppa, L. J. (1991). Education reform
in Japan : a case of immobilist politics. London
; New York, Routledge.
Skehan, P. (1988). "Language
testing. part 1." Language Testing 21:
211-221.
Spolsky, B. (1985). "The limits
of authenticity in language testing." Language
Testing 2: 31-40.
Terauchi, H. (2000). English for Academic
Legal Purposes in Japan. Tokyo, Liber Press.
Valette, R. M. (1977). Modern language
testing. New York, Harcourt Brace Jovanovich.
Weir, C. J. (1990). Communicative
language testing. New York, Prentice Hall.
Yoneyama, S. (1999). The Japanese
high school : silence and resistance. London,
Routledge. |
