TESTING READING
The teaching and testing of reading has a history of research effort that
goes backfor decades. While L2 reading research has a more limited
history (as opposed tothe L1 research base), there is also a very large
database to draw on. This chapterwill outline briefly major themes
from research that, in combination, form theconstruct of reading
abilities (for both L1 and L2 reading). Determining the con-struct then
provides rationales for various instructional and assessment
practices.The focus of this chapter will not be an extensive review of
the reading construct.That has been developed in more detail in other
sources (e.g., Bowey, 2005; Koda,2005, 2007; Perfetti, Landi, &
Oakhill, 2005). Instead, the goal is to draw con-nections from the
reading construct to effective ways to teach reading and
testreading. The chapter will then briefly outline guidelines that should
be effectivefor teaching L2 reading abilities across a range of
curricular settings. It will alsooutline testing tasks that can be
effective means for assessing L2 reading abilities.
Fluent reading comprehension requires a number of processing
subskills andlinguistic knowledge bases. These processes and
knowledge resources allow thereader to comprehend texts to the level
required. The identification of these skillsand resources has been the
outcome of many research studies, and it remains thesource of much
ongoing research. In this section, research is reviewed that supportsthe
relationship between reading skills and reading comprehension. Much
of theresearch has been conducted in English L1 reading contexts,
though increasingamounts of L1 reading research in other languages
have also emerged in the past15 years (Cook & Bassetti, 2005b;
Frost, 2005; Joshi & Aaron, 2006; Koda, 2005).
Word recognition efficiencyEnglish L1 research on eye-movement
tracking has shown that good readersrecognize words on average in
about 200–250 milliseconds, they move their eyesahead approximately
eight letter spaces per focus, they make regressive eye-movements
about 12 percent of the time (often for slight adjustments),
andthey actively focus on more than 80 percent of the content
words and about35 percent of function words. In short, reading is
a process of very rapid wordrecognition carried out through fairly
consistent eye behaviors. Automaticity isa key to this rapid word
recognition process. The observable eye-movementprocesses of
fluent readers are quite similar in all languages, with variation
dueto differing amounts of linguistic information provided by
individual graphicforms (see Rayner, Juhasz, & Pollatsek, 2005). Word
reading efficiency is going tovary somewhat among different
orthographic systems (Frost et al., 2005; McBride-Chang et al., 2005).
L2 word reading efficiency, in some situations, can be a strong
predictor of L2 reading comprehension abilities (Kahn-Horwitz,
Shimron,& Sparks, 2005); in other settings, it will not be predictive
for multiple complex reasons.
Vocabulary knowledge Research on English L1 vocabulary
knowledge has demonstrated that fluentreaders have very large
recognition-vocabulary knowledge resources and thatvocabulary
knowledge is highly correlated with reading ability (see Bowey,
2005;Stahl & Nagy, 2006; Tannenbaum, Torgesen, & Wagner, 2006).
While estimates ofword knowledge vary greatly (from 19,000 to
200,000 words; Anglin, 1993; Na-tion, 2001), the most widely
accepted figure is that high school graduates knowon average
40,000 words as fluent L1 readers (Nagy & Anderson, 1984; Stahl
&Nagy, 2006). This is a very large number of words to learn and
most accountssuggest that many of these words are learned by
exposure to new words throughcontinual reading practice. Stanovich
(2000) has argued that extended exposureto print over years leads
to major differences not only in vocabulary knowledgebut also in
increasing comprehension and a range of measures of
conceptualknowledge.Research on L2 vocabulary knowledge has also
shown that vocabulary is cor-related with L2 reading comprehension.
Droop and Verhoeven (2003) reporteda strong relationship
between third and fourth grade L2 students’ vocabularyknowledge
and their reading abilities. Schoonen, Hulstijn, and Bossers
(1998)also reported very strong relationships between vocabulary and
reading, report-ing an r2 of 0.71 for eighth grade EFL students in
Holland. This relationship hasalso been clearly demonstrated in
research involving L2 reading assessment studies(Pike, 1979; Qian,
2002).Early experimental studies on vocabulary instruction have
demonstrated thatvocabulary learning can lead to reading
comprehension improvement (Beck,Perfetti, & McKeown, 1982;
McKeown et al., 1985), though the impact of vocabu-lary on
comprehension improvement is complex and requires intensive
instruc-tional effort (Nagy, 2005). In the past 20 years, there have
been relatively fewstudies of efforts to teach vocabulary explicitly and
then compare the experimentalgroup to a control group for reading
comprehension gains. However, Carlo et al.(2004) have demonstrated
that intense explicit vocabulary instruction with L2English fourth
graders leads to significant improvement over control groups, notonly
in greater vocabulary knowledge but also on a measure of reading
compre-hension abilities.
Morphology, syntax, and discourse knowledge Research on L1
morphological, syntactic, and discourse knowledge shows thatthey
all have an impact on reading comprehension. A number of
studies haveshown that morphological knowledge contributes to
reading comprehension:research by Anglin (1993), Carlisle (2003),
and Nagy et al. (2003) all argue thatmorphological knowledge
(knowledge of word parts) is very important to moreadvanced word
recognition and reading development (see Stahl & Nagy, 2006).The
contribution of syntax to reading is less well examined in L1 reading
contextsbecause L1 students develop implicit knowledge of most
grammatical structures.
(For this reason, L1 students are not commonly assessed for their
grammar know-ledge.) However, there is evidence that grammatical
knowledge (syntactic pars-ing) plays a role in L1 reading
comprehension (and it is intuitively obvious onreflection) (Bowey,
2005; Lesaux, Lipka, & Siegal, 2006; Perfetti et al., 2005). Thereis
extensive evidence that discourse knowledge contributes in important
ways toreading comprehension. Syntheses by Duke and Pearson
(2002) and Trabassoand Bouchard (2002) point to the importance of
discourse signaling mechanisms,organization patterns in texts, logical
relations across clauses and sentences, andtext structures that can be
recognized and learned (see also Hudson, 2007; Koda,2005).Research
on L2 syntax and discourse knowledge have both shown that
thereare strong relationships between these language knowledge
bases and readingcomprehension. This relationship also appears in
reading assessment research.Research studies with Dutch students
have shown that syntax is a powerfulpredictor of reading
comprehension abilities. Schoonen et al. (1998) showed thatsyntax
was a very strong predictor of reading ability in a multiple
regressionstudy. More recently, Van Gelderen et al. (2004) reported a
very strong relation-ship between syntactic knowledge and
reading comprehension. In readingassessment research, both
Alderson (2000) and ETS researchers (Enright et al.,2002) have
presented very high correlations showing that syntactic knowledge
isstrongly related to reading comprehension. In research on the
role of discourseknowledge, Carrell (1984, 1985) has shown that
discourse structure knowledgeis strongly related to reading
comprehension. Similarly, Horiba (1993) reportedthat Japanese L2
students at different proficiency levels used discourse know-ledge
differently in their recall of text information. Focusing more specifically
onthe role of discourse-based graphic organizers, Tang (1992) showed
that studentstrained to recognize the discourse structure of a text,
performed better on a com-prehension measure.
Strategic processing .L1 research on strategic processing during
reading (e.g., inferencing, compre-hension monitoring, and goal
setting) demonstrates that strategic processesand metacognition
influence reading comprehension. Discourse
comprehensionresearchers have shown that inferencing that arises
from “learning from texts”has an important impact on
comprehension (Goldman & Rakestraw, 2000; Nation,2005; Perfetti et
al., 2005). Similarly, comprehension monitoring (as in monitoringfor
problems in text comprehension) appears to be a predictor of
comprehensionabilities (Cain, Oakhill, & Bryant, 2004). At the same
time, these abilities, beingmetacognitive in nature, are not simple
reading strategies. Rather, they constitutea range of skills and abilities,
and represent a range of strategic responses to
textdifficulties.Experimental research on comprehension instruction
and strategy trainingis extensive (see Pressley, 2006; Trabasso &
comprehension. Important evidence supports answering main idea
questionsas a post-reading task, using semantic mapping of ideas from
a text, previewingspecific information from the text, asking student to
formulate questions about atext, filling in and generating graphic
organizers that reflect the organization ofthe text, visualizing
information from the text, raising awareness of
discourseorganization of the text, among others. Overall, a number of
reasonably effectivestrategies have been identified in instructional
research, though combinations ofstrategic responses to texts
appear to be more effective in supporting comprehen-sion
development (Trabasso & Bouchard, 2002; see Grabe, 2004). The
best strat-egic approaches to reading instruction involve reciprocal
teaching, transactionalstrategies instruction, and concept-oriented
reading instruction (Block & Pressley,2002; Guthrie, Wigfield, &
Perencevich, 2004; Pressley, 2006).Research on L2 strategic
processing is far more limited. There are relatively fewstudies that
demonstrate a relationship between reading strategies and
readingcomprehension. Chen and Graves (1995) showed that
previewing a text was apre-reading strategy that improved student
comprehension. Klingner and Vaughn(2000) drew on reciprocal
teaching concepts and developed a four-strategy pro-gram for
teaching strategic reading, Cooperative Strategic Reading. Results
showedsome improvements in reading strategy use and in vocabulary
growth based onthe approach. In a recent meta-analysis of L2
reading strategy research, Taylor,Stevens, and Asher (2006) reviewed
the existing empirical research in L2 readingstrategy training (10
published studies and 12 dissertations) and concluded thata low to
moderate effect exists between strategy training and L2 reading
com-prehension improvement. The analysis is encouraging, but it
should be treatedcautiously due to the limited database available for
the analysis.
Extended exposure to printL1 research on extended exposure to
print has demonstrated a strong relation-ship between amount of
reading (over long periods of time) and improved read-ing
comprehension (Guthrie, Wigfield, & Von Secker, 2000; Guthrie,
Wigfield, &Perencevich, 2004; Stanovich, 2000). Stanovich and
colleagues, in a series of stud-ies, showed that exposure to print
(amount of reading) was an important inde-pendent predictor of
reading ability (see Stanovich, 2000 for overview). Sénéchal(2006) also
showed that reading exposure was a significant predictor of
readingcomprehension among 90 fourth grade French-speaking
students. Interestingly,despite many claims about Sustained Silent
Reading (SSR) and Free VoluntaryReading (FVR), there are no
rigorously controlled experimental studies that showa strong
relationship between either of these instructional approaches and
read-ing comprehension abilities.Research on extensive reading is
relatively unexplored in L2 reading. The oneset of studies that has
indicated the positive effects of extensive reading on read-ing
comprehension was that carried out by Elley over a period of 20 years
(Elley,2000). In these studies, he has shown that getting students
to read extensivelyover a long period of time consistently improved
reading comprehension abilities, Bouchard, 2002). Many
studiesdemonstrate a causal impact of instructional skills and
strategies on reading
Motivation L1 reading motivation is an area with only limited
research focused explicitlyon reading abilities. However, a few
studies have shown that more motivatedreaders both read more
and have better reading comprehension abilities. Guthrieet al. (2000)
showed that third grade and fifth grade readers who were more
motivated read more and that eighth and tenth grade readers
who were more motivated also were better at reading
comprehension. Guthrie, Wigfield, andPerencevich (2004) also
showed that specific instructional contexts can improvereading
motivation and, as a consequence, improve reading comprehension.
Thereis very little research on L2 reading motivation, a construct that is
quite differentfrom general L2 language learning motivation (cf. Mori,
2002).All of the above components contribute directly to reading
comprehensionand represent aspects of the construct of reading
comprehension. (There areadditional components not covered in
this review that also impact L2 readingcomprehension – e.g.,
working memory, background knowledge – though thesecomponents
require more complex descriptions and evaluations.) In almost
allcases, L1 research demonstrates that training in these
components leads to im-proved reading comprehension. In
addition, the experimental training studiesindicate that these
component skills are likely to be useful components of acurriculum
designed to improve students’ L2 reading abilities. Comparable
L2research on training impacts of component skills on reading
comprehension is still needed (see Grabe, 2004).
1 In the L2, learners have a much smaller linguistic knowledge
base of the L2when they begin reading. Their knowledge of
vocabulary is much more lim-ited; their knowledge of syntax is
similarly limited and there are no nativeintuitions about structure;
their knowledge of markers of discourse structureand their awareness
with text organization in the L2 will also be limited.2 L2 students,
overall, will have much less experience with reading exposure inthe L2.
They simply will have had much less practice in L2 reading.3 L2
students will experience L2 reading differently because they have
experi-ences reading in two different languages and because
cognitive processingwill involve two language systems (e.g., accessing
the bilingual lexicon, usinga joint strategy system – Garcia, 1998;
Kern, 1994) (see also Cook & Bassetti,2005a: Multi-competence
Hypothesis).4Aside from the possibilities of developing somewhat
distinct cognitive process-ing, students engaged in L2 reading will
also experience a range of transfereffects (cognitive skills, strategies,
and goals and expectations). Some transfereffects will involve
interference from the L1; others will facilitate L2 readingprocesses.
(See Dressler & Kamil, 2006; Geva & Siegal, 2000; Koda, 2005
onthe Interdependence Hypothesis and the Underlying Cognitive
Abilities, orCentral Processing Hypothesis.) Specific issues related to
transfer includethe Linguistic Threshold Hypothesis and unique
aspects of L2 reading. TheLinguistic Threshold Hypothesis argues
persuasively that a certain level of L2linguistic knowledge is needed
to support more fluent reading comprehen-sion processes. Unique
aspects of L2 reading include the extensive use ofglosses while
reading, the effort to carry out mental translations while reading,and
the extensive use of bilingual dictionaries and guessing word
meanings.5 L2 reading is also distinct in that readers rely on a
different combination ofgeneral background knowledge. Drawing on
information about “how theworld works” sometimes varies between
L1 and L2 reading experiences.6 Moreover, L2 readers will encounter
distinct social and cultural assumptionsin L2 texts that they may not
be familiar with or find somewhat hard toaccept. Certain types of
inferencing that might be routine in L1 reading maynot support
comprehension processing in the L2, particularly in cases
ofengaging in reading for purposes of “reading to learn.”
L2 Reading Assessment
Discussions of language assessment of all types start with
considerations abouttest validity. This chapter will assume that the
concept of validity, or constructvalidity, is available for review
through other sources. It is sufficient to note thatvalidity is an
extended argument from multiple perspectives (construct repres-
entation, reliability, comparative assessment, consequential impact,
and usability)that persuasively argues for the appropriate and fair
use of a test in a givencontext. In this section, I will assume that
these principles should guide assess-ment activities.In discussing
reading assessment, one must decide if the discussion is to
focusprimarily on classroom assessment, informal assessment, and
alternative assess-ment practices, or on standardized assessments.
Classroom assessment of read-ing development has a much wider
scope than standardized assessment options.In situations of formal
comprehension assessment in the classroom, often as
anachievement test, comprehension gains are assessed on a
specific text or set oftexts that has been recently taught.
Classroom settings for reading assessmentalso include informal
reading inventories or miscue analysis (reading aloud one-on-one with
an evaluator who notes errors on a record sheet and then
determineswhat progress a student has made or what instructional
support is needed by the student).
The classroom context also allows for various types of alternative assessmentoptions for determining student progress. In the classroom, one has the option ofcontinuous
ongoing assessment (quizzes, observations, record keeping of home-
work, interviews, progress charts, amount of reading, etc.). In such
settings,almost any language task that is a teaching task can also
be used as an assess-ment task. What might be lost in the way
of relatively weak validity or consist-ency for any given reading
task or measurement in the classroom setting iscountered by the
continual nature of assessment practices of all types. At thesame
time, teachers and administrators have a responsibility to develop
appro-priate tasks and appropriate interpretations of task
outcomes so that studentsare not evaluated unfairly. For this reason,
it is important to look at the types oftasks developed for more
standardized reading tests and consider how thesemajor tests
incorporate and reflect the reading construct, and how they
engageL2 learners in fair and appropriate assessment tasks (see
Appendix).Unfortunately, there are not many relevant, easily
accessible, and usefulclassroom-based tests of English L2 reading
abilities. This problem is clearlydemonstrated by a recent review of
L2 literacy development in US K-12 settings(August & Shanahan,
2006a). As the editors state, “The assessments cited in
theresearch to gauge language-minority students’ language proficiency
and to makeplacement decisions are inadequate in most respects”
(August & Shanahan, 2006b,p. 17). There is certainly a need for good,
well-developed low- to medium-stakesreading tests that can be used in
a variety of classroom contexts, that are gradedfor multiple proficiency
levels, that are affordable, and that can support instruc-tional decision-
making.In contrast to the more open classroom settings, standardized
assessment prac-tices are far more constrained by concerns of validity,
reliability, time, cost, use-ability, and consequence. Most standardized
tests attempt to establish a student’slevel of reading comprehension
ability, either in relation to some set of criteria orin relation to a
wider population. The time available for such an assessment
islimited and the test must be fair and useful. These concerns
strongly limit thetypes of reading assessment tasks that can be
used. Until fairly recently, stand-ardized L2 reading assessment has
not been overly concerned with the develop-ment of reading
assessment in terms of an evidence-based construct of
readingabilities tied to the group of students being assessed. However,
efforts to developstandardized reading tests in the past 15 years
have focused much more expli-citly on the construct of reading
and claims that can be made for reading pro-ficiency based on
evidence from the test. There are a number of good examples
ofstandardized assessments being developed from an initial set of
claims about thenature of L2 reading ability and a set of tasks that
would measure the relevantcomponent skills.The development of the
IELTS (International English Language Testing System)represents one
example of building a standardized test from construct assumptionsand
the gathering of appropriate evidence (Clapham, 1996). Similarly, the
effortsto redesign the TOEFL have only recently (since 1995) been
driven by the priorestablishment of an appropriate L2 reading
construct and evidence to support
assessment tasks that would measure this construct (see Chapelle,
Jamieson,& Enright, 2008). Additional approaches to L2
standardized assessment that arebuilt from claims about reading
abilities include the suite of Cambridge Englishproficiency exams (Weir
& Milanovich, 2003) and the Advanced English ReadingTest in China
(Weir, Huizhong, & Yan, 2000). These approaches to L2
readingassessment strongly document arguments for an L2 reading
construct, the importance of specific components of reading ability,
the types of tasks that can assessthese component abilities, and the
creation of overall tests that generate evidencefor the claims made
(thus building a validity argument for the appropriateness ofthe test
that has been developed).Drawing on evidence from research on
reading abilities to argue for a readingconstruct is one way in which
reading assessment practices have improved. How-ever, the
relationship is reciprocal; it is also the case that careful reading
assess-ment research has helped provide evidence for the
component abilities centralto L2 reading, as well. That is, the
evidence provided from assessment researchhas influenced
conceptualizations of component abilities underlying L2
readingcomprehension.For example, it is now clear from assessment
research that L2 vocabulary know-ledge is a powerful component of
L2 reading abilities (Pike, 1979; Qian, 2002).Similarly, and perhaps
more surprisingly for some, L2 grammar knowledge is amajor
component ability for L2 reading comprehension (Alderson, 2000;
Enrightet al., 2002; Pike, 1979). Appropriate reading strategies (as
opposed to test-takingstrategies) used in testing contexts also
appear to be an important componentof L2 reading abilities. While
there is other compelling evidence for these com-ponents of L2
reading ability (as well as other components of reading ability), itis
important to recognize that these component skills have also
emerged fromresearch on L2 reading tests.L2 Implications for
Reading Instruction andAssessmentOverall, the combination of
research on L1 and L2 reading abilities suggests thatthere is a
reasonably good set of implications for L2 reading instruction
andassessment suggested by research results. Reading
comprehension requires thefollowing skills and knowledge
resources:1The ability to decode graphic forms for efficient word
recognition2The ability to access the meaning of a large number of
words automatically3The ability to draw meaning from phrase- and
clause-level grammaticalinformation4The ability to combine clause-
level meanings to build a larger network ofmeaning relations
(comprehend the text)5The ability to recognize discourse-level
relationships and use this informa-tion to build and support
comprehension.
6 The ability to use reading strategies with more difficult text and for
a rangeof academic reading tasks.
7. The ability to set goals for reading and adjust them as needed.
8. The ability to use inferences of various types and to monitor
comprehensionin line with reading goals9 The ability to draw on prior
knowledge, as appropriate10Abilities to evaluate, integrate, and
synthesize information from a text toform a situation model of
comprehension (essentially what the reader learnsfrom the text)11The
ability to maintain these processes fluently for an extended period of
time12The motivation to persist in reading and to use the text
information appro-priately in line with reader goalsIn an ideal world,
each of these implications from research would be subject toa set of
instructional training studies and longitudinal development studies
todetermine the potential for turning implications into effective
applications in theclassroom. Once interesting applications are
developed, it is important to deter-mine the effectiveness of those
applications more generally for the developmentof L2 reading abilities.
Of course, we cannot wait for this ideal to be carried outbecause we
need to improve L2 students’ reading abilities at the present momentas
best we can. Instructional practices, based on current evidence,
need to beused in classrooms while additional evidence is
gathered. The best that we canoffer are practices that have been
examined and found useful, and then teachersshould draw on their
expertise and experience to build the larger curriculumframework
for effective teaching.
Teaching L2 Reading
The major argument of the chapter to this point is that a number of key
readingsubskills can be taught successfully, and further, that the
learning of these subskillswill contribute to a learner’s reading
comprehension abilities. How these skillsshould be taught most
effectively is indicated to some extent by the researchreviewed
above. However, there are many instructional approaches that
canpotentially contribute to the development of reading abilities. While
there is notyet extensive empirical research on the effectiveness of
many practices, there isteaching expertise and experience which
support these approaches until controlledevidence is collected and
assessed. Quite a lot is already known about promisinginstructional
practices. In some cases, we know that instructional activities car-
ried out consistently have been useful with some groups of students
and shouldbe useful with a number of other student groups. We
know that instructionalactivities which receive enough instructional
time, intensity of effort, and priorityin the curriculum can lead to
significantly improved reading skills development. We know that
students respond well to a number of instructional activities
thatimprove reading skills. We can build on these starting points
while additional research is being carried out.
The goal for reading instruction, at a general level, is to
incorporate key com-ponent skills and knowledge into a reading
curriculum. Specific instructionalactivities included in the curriculum
follow from the major themes developedearlier in the article and the
resulting implications. To describe how to carry outeach suggestion
would amount to multiple teaching-instruction handbooks, atask
that goes far beyond the scope of the present chapter. More generally,
goodsuggestions and examples for many of the issues described here
can be found ina number of good reading textbook sets (e.g., Anderson,
2003; Blanchard & Root,2007; Silberstein, Dobson, & Clarke, 2002; for
many good instructional examples).What can be offered in this chapter
is a set of more general curricular principleswhen building a reading
curriculum and rethinking instructional practices. Theseprinciples
include:1 A curricular framework for conceptualizing L2 reading
instruction that shouldintegrate major skills instruction with extensive
practice and exposure to print(building upon a needs analysis, goals
and objectives for teaching and testing,attractive and plentiful
resources, appropriate curriculum framework, effect-ive teacher
support, effective teaching materials and resources).2 Reading
materials and resources that need to be interesting, varied, good-
looking, abundant, accessible, and well-used.3 Some degree of
student choice along the way in selecting major readingsources.4
Reading skills that are introduced and taught by examining the primary
textsused in the reading course. There should not be a need for special
materials tointroduce reading skills (though additional activities for
further practice arenecessary). If skills are meant to help
comprehension, they should help withcomprehension of the major
texts being read in a class. This link betweenskills and instructional
texts also raises metalinguistic awareness of how textsare put together
linguistically.5 Lessons that are structured around pre-reading,
during-reading, and post-reading activities, and these activities should
be varied from one major readingto the next.6 Instruction that is built
on an integrated curriculum framework and can sup-port the following
developmental goals:(a)Promote word-recognition skills(b)Build a
large recognition vocabulary(c)Practice comprehension skills that
combine awareness of grammar, mainidea identification, and
comprehension strategies: strategy instruction isnot separate from
text comprehension instruction(d) Build awareness of discourse
structure (recognize main ideas, recognizemajor organizing patterns,
recognize how the information is organizedin parts of the text,
recognize overt signals of text structure, recognizeanaphoric
relations in texts, recognize other cohesive markers in texts)
(e)Promote strategic reading(f )Practice reading fluency (build
reading rate, build text passage readingfluency, read and reread at
home with parent or tape or self )
Testing L2 Reading
Carrying out appropriate reading assessments also requires a
translation from“implications from research,” as well as an effort to
consider useful applicationsdirectly from assessment research to
realistic classroom situations. Again, a thor-ough set of practical
recommendations and associated example activities is be-yond the
scope of this chapter. However, a number of good examples of
readingassessment tasks can be found in Alderson (2000), Hughes
(2004), and Weir andMilanovic (2003). It is also important to
highlight two key concepts for readingassessment. First, reading
assessment tasks are a restricted purpose for reading.The context
for assessment itself precludes any strong assumption of a match
toauthentic reading in the “real world.” One consequence is that
assessment tasksdo not need to be avoided or radically distorted
because they are not the same asreading in the real world. Realistic
reading assessment tasks, as opposed to realreading tasks, may be
the better benchmark (see Alderson, 2000). Second, read-ing
assessments need to take into account both students’ proficiency
levels andstudents’ ages. Tasks need to change to fit a given
proficiency range and studentmaturity level as part of an
appropriate reading assessment battery.In closing the discussion of
reading assessment, 10 recommendations forgood assessment
practices are offered (though the list could easily include
20recommendations):1 Students should be tested on a range of
relevant skills.2 Students should be encouraged to read longer
texts (for advanced assess-ment, 700 –1,200 words, assuming 120 –
150 wpm).3 Background knowledge influences all comprehension
and needs to be ac-counted for in a positive way (multiple topics,
multiple tasks, general topics,limited interdependence of items
within some subset of tasks).4 Group tasks might be used to engage
discussions of reader interpretations oftexts.5 Extensive reading
should not be discouraged by assessment procedures.6 The
importance of identification and fluency skills needs to be
explored(reading word lists, oral reading for one minute, silent reading
on computer,timed reading, assessment of rereading).7 Tests might
explore ways to assess synthesis skills, evaluation skills, strateg-ies,
metacognitive knowledge, and skills monitoring (text monitoring
whilereading).
8Reading might be tested within a content-focused battery (but
item inter-dependence has to be a concern).9Tests might consider
item types that take advantage of computer interfaces(e.g., allow
a text to disappear after reading, use a few hypertext links in
atest passage, combine information from multiple texts to complete a
task).10Many skills might be measured usefully through informal
assessment optionsin classroom contexts. What one loses in
reliability and objective controlscould be countered by the many
formal and informal assessments that canbe made in the
classroom. (But informal assessment is not a substitute formore
formal testing.)
Concluding CommentsThere are a number of additional
recommendations that can be made for build-ing L2 reading
instruction, planning effective multi-level reading curricula, devel-oping
appropriate assessment practices, and providing feedback on
learningprogress (assessment for learning). This article has sought to
develop the founda-tion that leads to useful implications for
reading instruction and assessment. Ithas also outlined a simple
array of curricular guidelines for reading instructionand assessment
practices that can be developed or adapted to a fairly wide rangeof L2
reading contexts. At the same time, a short article of this type
can onlybegin to scratch the surface of the potential instructional and
assessment optionsand variations that can help make a difference
in reading success with L2students. The key to these ongoing efforts
is to continue exploring effective prac-tices for reading instruction
and tasks for reading assessment that are based onimportant and
relevant reading research and persuasive instructional research.
