423339

2012
LTR16110.1177/1362168811423339AkakuraLanguage Teaching Research

LANGUAGE
TEACHING
RESEARCH

Language Teaching Research

Evaluating the effectiveness of

16(1) 9–37
© The Author(s) 2012
Reprints and permissions: sagepub.
explicit instruction on implicit co.uk/journalsPermissions.nav
DOI: 10.1177/1362168811423339
and explicit L2 knowledge ltr.sagepub.com

Motoko Akakura
University of Auckland, New Zealand

Abstract
This study examined the effectiveness of explicit instruction on second language (L2) learners’
implicit and explicit knowledge of English. Explicit instruction on the generic and non-generic use of
English articles was delivered by CALL activities. Four tasks assessed acquisition: elicited imitation,
oral production, grammaticality judgement, and metalinguistic knowledge tasks. A pretest and
two posttests were conducted immediately and six weeks after the treatment. Durable effects
for explicit instruction were found on measures of implicit knowledge and on ungrammatically
supplied items on measures of explicit knowledge. This study’s findings contribute towards our
understanding of the efficacy of explicit instruction on implicit and explicit knowledge at relatively
advanced stages of L2 acquisition.

Keywords
implicit knowledge, explicit knowledge, explicit instruction

I Introduction
Not all grammatical forms are acquired with ease in a second language (L2). Structures
such as English articles are non-salient or fragile features (N. Ellis, 2006) that are not
readily perceived by mere exposure to the language alone. As early as 1967, Corder
(1967) pointed out that a defining concern of L2 research is that there are some aspects
of language for which L2 users fail to make input become ‘intake’; that is, to acquire
implicit language knowledge that may be used for comprehension or production pur-
poses. Frequently, non-salient features such as English articles may not easily be ‘noticed’
at the stage of input (VanPatten, 1996). Noticing is argued as being a necessary step prior
to acquisition taking place (Schmidt, 2001). The difficulty of noticing non-salient forms
has led researchers to approach theories on grammar teaching in various ways (see

Corresponding author:
Motoko Akakura, University of Auckland, 74 Panapa Drive, St John’s, Auckland 1072, New Zealand
Email: [email protected]
10 Language Teaching Research 16(1)

Hinkel & Fotos, 2002), with the assumption that ‘hard to acquire’ structures may require
explicit intervention. Instruction is considered to be explicit if rule explanation com-
prises any part of the instruction (the direct rule-giving approach), or if attention is drawn
to underlying rules in the input.
For those ‘hard to acquire’ structures, various forms of intervention have been
researched. A decade after Norris and Ortega’s (2000) meta-analysis of studies investi-
gating the effects of instruction on L2 learning, studies dealing with L2 acquisition
through some means of intervention remain extensive. But the fundamental issue regard-
ing the role of instruction in developing non-salient L2 forms continues to be debated
(DeKeyser, 2005; R. Ellis, 2002, 2008; Krashen, 2003). Some evidence suggests that a
number of grammatical forms of the L2 seem to resist intervention (e.g. Macaro &
Masterman, 2006).
In part, the uncertainty over the efficacy of explicit instruction is due to the difficulty
of operationalizing implicit language knowledge. Implicit language knowledge refers to
knowledge of a language that may be accessed instantaneously during spontaneous com-
prehension or production. It differs from explicit language knowledge, which is knowl-
edge about language and how the language can be used. There is a paucity of work
directly evaluating the acquisition of communicative ability, such as the capacity for
fluent speech, considered to be evidence of implicit language knowledge (R. Ellis, 2008).
Much of the research evaluating L2 acquisition has employed measures of language abil-
ity that assess explicit knowledge rather than implicit knowledge (R. Ellis, 2009; Norris
& Ortega, 2000). The latter tends to require more labour intensive measures, such as
freely constructed responses, that involve more subjective scoring. This measurement
problem has been pointed out as having added to the controversy regarding the efficacy
of explicit instruction (Hulstijn, 2005). Until recently, few studies had addressed this
issue, largely due to methodological difficulties in differentiating between implicit and
explicit knowledge.
Recent work by Ellis and colleagues (R. Ellis, 2005; R. Ellis et al., 2009; Han & Ellis,
1998) suggest that it is possible to provide a relatively separate measure of both knowl-
edge types. Unplanned language use under speeded conditions that require a primary
focus on meaning is said to tap implicit knowledge. Language use with a primary focus
on form, conducted without time pressure, is said to tap explicit knowledge (R. Ellis,
2005). While pure measures of either implicit or explicit knowledge are non-existent,
these empirical developments in assessing language knowledge allow the possibility of
investigating whether explicit instruction impacts on implicit knowledge. Within a cog-
nitive approach to L2 acquisition, improving the ability to communicate fluently and
confidently in an L2 is considered to be the ultimate goal of instruction, so it is appropri-
ate for L2 pedagogy to focus on developing implicit knowledge in addition to explicit
knowledge (R. Ellis, 2008).

1 CALL as a method of intervention

While there are various ways for providing explicit instruction, a major advantage of
Computer Assisted Language Learning (CALL) is that language learners may play an
active role in the learning process with a greater degree of autonomy in how they learn,
Akakura 11

which in turn is expected to lead to higher motivation (Stepp-Greany, 2002). CALL may
also provide opportunities for advanced learners who are no longer in formal language
classrooms to study autonomously, independent of time and space (Felix, 2003).
Adopting a direct rule-giving approach is generally not considered to ‘actively engage
the learners in taking responsibility for their own learning’ (R. Ellis, 2008, p. 149).
However, this disadvantage may be augmented by the nature of CALL activities that
require dedicated commitment on the part of learners to complete the lessons by them-
selves. If successful, it may help ensure autonomous learning that is predicted to be a
necessary component for successful uptake.

2 L2 acquisition of English articles

English articles are ideal target forms for investigating the effects of instruction, as they
are considered to be an element of grammar that expresses ‘highly abstract notions that
are extremely hard to infer, implicitly or explicitly, from the input’ (DeKeyser, 2005,
p.5). Pedagogically, articles are frequently cited to be problematic even among advanced
learners (e.g. Butler, 2002; Huong, 2005). Some evidence suggests that the distinction
between the indefinite and definite articles resists intervention (Dulay et al., 1982), as
they are governed by complex rules. The teaching of generic articles has sometimes even
been neglected. Textbooks propose that learners ‘leave out articles before plural and
uncountable nouns when we are referring to something general’ but they omit stating that
articles are also used to refer to something general as well (e.g. Parrott, 2000, p. 47). The
use of articles to express generic reference in English is problematic. Linguists have
proposed that there is a continuum based on the distinction between generic and non-
generic articles (Jacobsson, 1998). Claims have been made for teaching articles at
advanced stages of acquisition (Master, 1997), yet this is not customary.
There have been a relatively large number of experimental studies since the 1990s
focusing on instruction of English articles. This research has tended to highlight article
difficulty. Instruction typically covers the use of non-generic articles among adult learn-
ers and most of these studies show some effect of instruction. Some have reported statis-
tically significant positive effects (Bitchener & Knoch, 2009; R. Ellis et al., 2008), some
limited gains (Huong, 2005; Shimamune & Jitsumori, 1999; Shiozawa & Simmons,
1995), while one study found no significant effect compared to the control (Brender,
2002). Studies on the acquisition of generic articles remain inadequate. Studies examin-
ing generic articles (Master, 1994; Muranoi, 2000) collapse the distinction between
generic and non-generic articles into total accuracy scores, so that the distinction is lost.
Evidence on implicit knowledge of English articles is also limited, as few studies have
employed oral measures in data collection (e.g. Muranoi, 2000; Rovers, 2002). Table 1
shows a summary of studies on L2 English article acquisition.
The present study contributes to the ongoing debate on the effectiveness of instruction
on the development in L2 knowledge of advanced learners. Outcomes are assessed on
both knowledge about language (explicit knowledge) and knowledge of language that is
available for fluent use (implicit knowledge), in the same theoretical and methodological
direction established by Ellis et al. (2009). This study investigates the following research
question:
 12
Table 1. Summary of research investigating the effectiveness of explicit instruction on English article acquisition, 1994–2009

Study Sample size Target Instruction Length of Tests Posttests Results

structure instruction

Master, 1994 ESL in the USA a, an, Explicit instruction 6 hours Written: 9 weeks Beneficial effect of
14 university generic a was combined (over 9 Gap-filling task instruction claimed,
students generic the with implicit input weeks) but uncertain in terms
+ 33 Control (inductive learning of implicit and explicit
e.g. underlining knowledge as a) inductive
article occurrences). instruction was also
offered, and b) the written
cloze test was a speeded
task.
Shiozawa and EFL in Japan a, an, the Experimental 1.5 hours Written: 20–32 Gains were seen on the
Simmons, 1995 132 university Group did (over 6 Journal extracts weeks gap-filling task but not
students gap-filling exercises classes) Gap-filling task in written production
+ 222 Control followed by (journal extracts).
discussion of
difficult items.
Shimamune and EFL in Japan a, an, the Asked to read a Unspecified Written: 4 weeks Explicit instruction
Jitsumori, 1999 24 university page of article rules (until Sentence- 6 weeks enhanced ability to
students with examples mastery) completion task 8 weeks describe rules but not
+ 24 Control before practicing Explaining in the long term. Claims
on CALL drill- article error fluency training using
type activities CALL activities superior
involving sentence- to explicit instruction in
completion tasks. the long term.
Muranoi, 2000 EFL in Japan a, an, Explanation given in 1.5 hours Oral: 0 weeks Interaction enhancement
61 university generic a L1 on why article Story 5 weeks (IE) with formal debriefing
students errors occur, and description task, had greater impact than
Language Teaching Research 16(1)
Table 1. (Continued)
Akakura

Study Sample size Target Instruction Length of Tests Posttests Results

structure instruction

+ 30 Control the role of indefinite Picture IE with meaning-focused

article and generic description debriefing.
a explained. task, Results uncertain in
Written: terms of the sole effect
Picture of explicit instruction as
description task, there were both implicit +
Grammaticality explicit input.
judgement
Brender, 2002 EFL in Japan a, the Brender’s flow Unspecified Written: 3 weeks Explicit instruction led to
56 university chart for article amount Cloze test improvement in cloze test
students selection was (over 14 Essay writing scores and less error in
+ 35 Control explained together weeks) task essay task although not
with various statistically significant.
consciousness Possible implicit input
raising tasks. in addition to explicit
instruction.
Rovers, 2002 ESL in New Zealand a, the Group 1: Unspecified Oral: 1–2 days No sig. effect on the oral
18 ESL classroom Deductively solving amount Picture task (implicit knowledge).
students the rule behind first (one class description task, Inconclusive results on
+ 7 Control and subsequent session) Written: measures of explicit
mention. Grammaticality knowledge due to
Group 2: Rule judgement dissimilar pretest scores
given explicitly + Error between groups, and use
production practice. correction task of same test as posttest.
(Continued)
13
 14

Table 1. (Continued)
Study Sample size Target Instruction Length of Tests Posttests Results
structure instruction

Tashima, 2004 ESL in the USA. a, an, the Experimental 4.5 hours Written: 4 weeks No statistically significant
6 Group received 3 (over 3 Cloze test gains after explicit
+ 6 Control tutoring sessions on weeks) instruction in choosing
grammar exercises. between definite and
indefinite articles.
Huong, 2005 ESL in Vietnam a, an, the  Explicit grammar 7.5 hours Written: 1 week Group receiving Cognitive
67 advanced instruction provided (over 5 Gap-filling task 3 weeks instruction outperformed
English teacher- via; Group 1: weeks) Error the traditionally trained
trainees commercially correction task group in the immediate
*no control group available functional posttest only.
approach vs. Group
2: a Cognitive
Grammar approach
Paradowski, 2007 EFL in Poland English Comparison of 4.5 hours Written: 2–8 Overall result (total
28 students (aged articles, relative structures in English (6 X 45 min Gap-filling task, months scores) show significant
16–17) pronouns and Polish with classes over Transformation gain over the Control
+ 116 Control and adverbs, a focus on form: a school task Groups. However, gains
expletives, e.g. finding correct year) for individual grammar
reported counterparts items are uncertain.
speech, in Polish, filling
conditionals Polish texts with
English equivalent,
developing grammar
algorithms.
Language Teaching Research 16(1)
 Akakura

Table 1. (Continued)

Study Sample size Target Instruction Length of Tests Posttests Results

structure instruction
Ellis et al., 2008 EFL in Japan a, an, the Exposure to written Unspecified Written: 6 weeks
Explicit written feedback
49 Japanese EFL corrective feedback time to read Picture 10 weeks
resulted in greater
university students on three narrative feedback description task accuracy in writing task
in ESOL classes writing tasks, on 3 story Error and consistency in error
written after first writing tasks. correction task correction in the delayed
reading the stories. posttest.
Bitchener and ESL in New Zealand a, the Explicit instruction Group 1: Written: 1 week Significance effect of
Knoch, 2009 39 university on first and 30 min rule Picture 8 weeks instruction reported
students subsequent mention presentation description task 6 months for written production.
+ 13 Control of articles given as + 5 min 10 months Treatment group
written feedback on to read consistently outperformed
picture description feedback. the Control on all
tasks. Group 1 Groups 2 posttests.
further received and 3: 5
30 min rule mins to read
explanation. feedback.
15
16 Language Teaching Research 16(1)

What are the effects of explicit instruction on the acquisition of generic and non-generic article
usages in L2 English, as measured by (a) tests of implicit knowledge, and (b) tests of explicit
knowledge?

II Method
Participants in the experimental group were exposed to instruction delivered through
CALL activities on a website that can be used by learners on their own. Controls did not
receive instructional treatment. Assessment tasks adapted criteria set by Ellis et al.
(2009). Measures of implicit knowledge consisted of an elicited imitation task and an
oral production task. These had a focus of attention on meaning, and were conducted
under time pressure (see Appendix 1). Measures of explicit knowledge were a grammati-
cality judgement task and a metalinguistic knowledge task, with a primary focus on form
and no time constraints (see Appendix 2). There were three testing times: a pretest, an
immediate posttest, and a delayed posttest six weeks after the treatment. The target gram-
matical forms were generic (indefinite a/an, and definite the) and non-generic use of
English articles (indefinite a, an and definite the).

1 Participants
The study was conducted among 94 participants in New Zealand for whom English is an
L2; 21 native speakers participated for test reliability purposes. A background question-
naire showed that the majority (78%) were of Asian origin, with almost half having
Chinese as their first language (L1), followed by Korean, Japanese and Malay. The aver-
age number of years they had spent in an English-speaking country was five years, rang-
ing from less than one month to 24 years. The mean age was 28. The mean age at which
they started learning English was 10 years old. The average number of years in formal
English education was nine years. Difficulty with written articles was reported by 44%
of the L2 participants, and for spoken articles 39%. The majority of participants’ profi-
ciency levels fell within the advanced-to-near-native level (46%) or upper-intermediate
(45%), while those in lower-intermediate comprised only 9%.1 Participants were ran-
domly assigned to the experimental group (n = 49) and control group (n = 45). An inde-
pendent-samples t-test did not show any significant difference between the two groups
on pretest scores, their proficiency levels, L1, age or gender, indicating that the Control
and Experimental Groups performed similarly at the start of the study, and without dif-
ferences in prior knowledge of English articles.

2 Instructional materials
Each lesson began by explaining the ‘form/function’ mappings of articles (Batstone &
Ellis, 2009), and concluded with 20 multiple-choice questions focusing on form. The
first 10 required selecting correct articles, and the final 10 required judging grammatical-
ity. Practice quizzes were inserted for each point made in the explanatory section so as to
reinforce the focus of each lesson. As English articles are, by nature, dependent on what
the speaker intends to focus on when using the noun, a descriptive approach was taken in
Akakura 17

Table 2. Description of the lessons

Lesson Target form Examples

1 Non-specific focus of the My girlfriend is a helicopter pilot.

indefinite articles a and an If you can make 20 pizzas in an hour, you’re hired.
2 Non-generic articles Robert: Well, love is a reason I want to marry
you Jasmine.
(There are also other reasons like your house,
your car, and your credit card.)
Jasmine: The reason I want to get married to you
is love.
(Love is the only thing that matters, honey.)
3 Generic articles A cat is much smaller than alion.
The kiwi bird is in danger of becoming extinct.
*A kiwi bird is in danger of becoming extinct.

explaining how articles are used and the reasons why they are used in that way. The les-
sons broadly incorporated aspects of DeKeyser’s (1998) output-based instruction
whereby explicit instruction is followed by form-focused activities providing the oppor-
tunity to use articles. The descriptions of the lessons are provided in Table 2. Each of the
three lessons required less than 60 minutes to complete and was completed over a 1-week
period.

3 Instruments and procedures

A pretest was administered a week prior to posttest 1. The experimental group completed
article lessons during the first week after the pretest. Posttest 2 was completed in week
seven. A split-block design was used for the two measures of implicit knowledge. Three
similar versions of each test were counterbalanced to rule out content familiarity as a
reason for improvement in posttest scores. Analysis of variance conducted on the pretest
scores showed no differences between the three versions. For the two measures of
explicit knowledge, test items were randomly scrambled to create three versions of the
test that became the pretest, posttest 1 and posttest 2.
All pretests and the oral posttests (elicited imitation and oral production tasks) were
conducted one-on-one with the researcher. Participants’ speech data were recorded digi-
tally. For the written portion of the posttests, participants had the choice between com-
pleting an online CALL or paper version. The CALL version was created in such a way
that the learners had to complete all the lesson materials. Email links were provided with
the instruction ‘do the task without looking at any other material and without help from
other people.’ It is not possible to know whether participants adhered to this instruction,
which was based on trust alone.

a Elicited imitation task: This task consisted of 14 true or false sentences in which gram-
matical (n = 10) and ungrammatical (n = 10) articles were tested. Participants were asked
to listen to a story while looking at a series of 20 pictures illustrating it. The pictures were
18 Language Teaching Research 16(1)

taken from children’s picture books with no text (DuPasquier, 1988; Mayer, 1969, 1974).
Based on these, the researcher created narratives embedding target articles. Whereas the
pictures depicted the true story, half of the recorded narrative contained sentences that
were contrary to the pictures. For example, participants heard; ‘Once in his room, * boy
put the frog inside his wooden box. Is this true or false?’ while the picture showed the frog
in a glass jar. Participants were given an answer sheet where they could tick their answers:
true, false or not sure. Nouns that created obligatory occasions for articles tested were
typed beneath the pictures. A bell sound signalled that the recorded narrative had reached
the sentence corresponding to the pictures so as to alert respondents to check whether
what they would hear next, matched the picture. The ‘True or False’ was intended to focus
attention on meaning, and is an adaptation of the ‘Belief’s questionnaire’ used in Ellis
et al. (2009) . Participants were then asked to describe the picture orally using all of the
words provided under the picture. The instruction said; ‘Sometimes you may have to
change the sentence from what you hear to match the pictures.’ Each task began with oral
instructions and practice items after which participants were given the opportunity to ask
questions. Reliability (Cronbach’s alpha) for the pretest was .804. Two examples of the
recorded target sentences and descriptions of the corresponding pictures together with the
words that had to be included in the elicited imitation are presented in (1).

(1) a. Recording:
‘One of the insects that attack in large numbers is the bee. Is this true or false?’

Picture of a dog being chased by a swarm of bees.

Words provided: One of the insects, attack, bee

b. Recording:
‘*During the night when no one was looking, unexpected thing happened. Is this
true or false?’

Picture of a frog stepping out of a glass jar at night, while the boy and dog are
fast asleep on the bed.

Words provided: During the night when…, unexpected thing…

A story-based elicited imitation task has not previously been used to measure implicit
knowledge and so a validation study was conducted to confirm that the task met the cri-
teria established by Erlam (2006) as a measure of implicit knowledge; for a detailed
account, see Akakura (2009).

b Oral production task: This task required narrating a story for the same sequence of
pictures as the elicited imitation task, but in their own words. The reason for using the
same pictures was because task familiarity is reported to (1) reduce cognitive load during
performance (Bygate, 1999), and (2) increase language complexity (Ortega, 1999; Skehan
& Foster, 1999). In order to discourage reliance on hearer knowledge and thus the over-
use of the definite article, participants were asked to pretend their audience were children
Akakura 19

who were listening to their story for the first time. Photos of children were used as props.
In order to encourage the use of generic articles, they were also asked to explain words
that might be difficult for the children. Correlation coefficients calculated to give an
estimate of inter-rater reliability of scoring for the pre- and posttests were: pretest,
r = .991, p = .000; posttest 1, r = .998, p = .000; posttest 2, r = .982, p = .000.

c Grammaticality judgement task: Participants were given unlimited time to judge 20

underlined portions of sentences for grammaticality, both grammatical (n = 10) and
ungrammatical (n = 10). They were given two practice items using non-article forms.
A confidence assessment measure was incorporated within the judgement scale with the
coding: 1. correct, 2. probably correct, 3. probably incorrect and 4. incorrect, forcing
participants to make a judgement. If the sentence was grammatical, correct answers
included both ‘correct’ and ‘probably correct’. Responses were scored as either correct
(1 point) or incorrect (0 points). Reliability (Cronbach’s alpha) for the pretest was .706.
Examples of sentences for judgement are given in (2).

(2)
a. One of the famous flightless birds of New Zealand is the kiwi.
b. ‘What are you drawing?’ asked mum.
  * ‘It’s supposed to be a elephant,’ said Hannah crossly.

d Metalinguistic knowledge task: Participants were asked to correct 10 sentences each

containing an article error that was underlined (n = 10). Next, the rule-stating section
required participants to provide written explanations for the ungrammatical articles (n = 5).
Again, participants were given as much time as they needed, and completed two practice
items before commencing. Responses were scored as either correct (1 point) or incorrect
(0 points). Reliability (Cronbach’s alpha) for the pretest was .644. An example sentence
for correction is given in (3), and for the grammatical explanation in (4).

(3) Among the animals protected by the wild life protection agency is panda.
a.  the panda
b. * a panda

(4) The thief stole a expensive vase from the living room.
a.  ‘an’ is the article that need [sic] to be use [sic] when the next word that comes after
it, start [sic] with vowels.
b. * an before ‘expensive’ [because] ‘expensive’ is adjective.

4 Analysis and task validity

Descriptive statistics for both groups on all four tasks were calculated. Average effect
sizes (Cohen’s d-index) were also calculated to measure the degree of instructional effec-
tiveness. For the grammaticality judgement task and elicited imitation task, (1) total
scores were calculated along with separate scores for (2) generic and non-generic arti-
cles, and (3) grammatical and ungrammatical items. For the metalinguistic knowledge
task, which only tested ungrammatical articles, only (1) and (2) were calculated.
Grammatical and ungrammatical items were examined separately based on previous
20 Language Teaching Research 16(1)

Table 3. Correlation matrix for the four tests (L2 learners, n = 94)

Test Metalinguistic Elicited Oral Speaking

knowledge imitation production proficiency
Grammaticality judgement .26* .11 .15 .04
Metalinguistic knowledge − .28* .35** .03
Elicited imitation − .48** .39**
Oral production − .29**
Notes: * p < .05, ** p < .001

Table 4. Loadings for principal component factor analysis

Component 1 Component 2
Eigenvalue 2.017 1.123
Percentage variance 40.339 22.450
Cumulative percentage 40.339 62.789
Test:
Oral production .687 .376
Elicited imitation .794 .205
Grammaticality judgement −.036 .754
Metalinguistic knowledge .214 .766
Speaking proficiency .783 −.197
Note: Extraction method: principal component analysis with no factors imposed

findings (R. Ellis et al., 2009), suggesting that ungrammatical items provide a better
measure of explicit knowledge. In the current study, scores were calculated in the same
way for the elicited imitation task in order to consider whether the same difference is
evident on a measure of implicit knowledge.
The validity of the task instruments is supported by correlations between tasks indi-
cating the strength of relationships (see Table 3) and a principal component factor analy-
sis revealing factors underlying the task scores (see Table 4). Speaking proficiency is a
measure independent of article use; it was rated on the oral production task using the
IELTS speaking band. It is a time-pressured task hypothesized to correlate more with
implicit knowledge, and was included for validation purposes.

III Results
1 Elicited imitation task
The descriptive statistics on the elicited imitation task (see Table 5) show around 50%
accuracy levels on the pretest, with higher scores on non-generic articles compared to
generic articles, and on grammatical items compared to ungrammatical items. These
accuracy scores generally increased over both posttests with the Experimental Group
outperforming the Control on all components.
Akakura 21

Table 5. Descriptive statistics and effect sizes for the elicited imitation task

Experimental group Control group

M SD n d M SD n
Non-generic articles (maximum = /12):
Pretest 7.45 2.363 49 −0.21 7.98 2.659 45
Posttest 1 9.29 2.131 49 0.49 8.18 2.367 45
Posttest 2 10.02 1.984 49 1.07 7.64 2.460 45
Generic articles (maximum = /8):
Pretest 2.84 1.650 49 −0.08 2.98 1.685 45
Posttest 1 4.51 1.850 49 0.85 2.96 1.796 45
Posttest 2 4.96 1.645 49 1.16 3.00 1.745 45
Grammatical articles (maximum = /10):
Pretest 5.35 1.843 49 −.020 5.76 2.278 45
Posttest 1 6.96 1.779 49 0.52 5.98 1.983 45
Posttest 2 7.80 1.779 49 0.94 6.02 1.983 45
Ungrammatical articles (maximum = /10):
Pretest 4.94 1.853 49 −0.14 5.20 1.902 45
Posttest 1 6.84 1.760 49 0.88 5.16 2.044 45
Posttest 2 7.18 1.752 49 2.39 4.62 1.800 45
Total (maximum = /20):
Pretest total 10.29 3.202 49 −0.20 10.96 3.567 45
Posttest 1 total 13.80 2.993 49 0.79 11.14 3.690 45
Posttest 2 total 14.98 3.058 49 1.37 10.64 3.276 45

Results of a one-way ANOVA reveal that there is an overall significant difference in

the effect of instruction for group on posttest 1 (F(1,92) = 14.866, p = .000), with a
medium effect size (d = 0.79). This is sustained with an increase in posttest 2 (F(1,92) =
44.023, p = .000), where the effect size is large (d = 1.37). The ANOVAs reveal group
differences on posttest 1 for both the grammatical items (F(1,92) = 6.398, p = .013, a
medium effect size d = 0.52), and ungrammatical items (F(1,92) = 18.339, p = .000, a
large effect size d = 0.88). These differences increase in posttest 2 for both the grammati-
cal (F(1,92) = 20.894, p = .000, a large effect size d = 0.94), and ungrammatical items
(F(1,92) = 48.838, p = .000, a large effect size d = 2.39). There were significant differ-
ences between groups for non-generic articles in posttest 1 (F(1,92) = 5.702, p = .019, a
small effect size d = 0.49) and in posttest 2 (F(1,92) = 26.758, p = .000, a large effect size
d = 1.07). There were also significant group differences and large effect sizes for the
generic articles, in both posttest 1 (F(1,92) = 17.037, p = .000; d = 0.85), and posttest 2
(F(1,92) = 31.387, p = .000; d = 1.16).

2 Oral production task

The descriptive statistics on the oral production task (see Table 6) reveal high
levels of accuracy in the pretest scores that increase for the Experimental Group, but
22 Language Teaching Research 16(1)

Table 6. Descriptive statistics and effect sizes for the oral production task

Experimental group Control group

M SD n d M SD n
Non-generic (percentage accuracy):
Pretest 76.92 19.348 49 0.02 76.42 20.713 45
Posttest 1 77.76 16.698 49 0.05 76.80 19.394 45
Posttest 2 86.59 17.503 49 0.41 78.69 20.883 45
Generic (percentage accuracy):
Pretest 5.78 21.387 49 −0.41 18.89 38.860 45
Posttest 1 13.27 24.528 49 −0.15 17.22 29.107 45
Posttest 2 13.78 24.506 49 −0.17 18.36 29.991 45
Total:
Pretest total 91.63 10.434 49 0.14 89.96 12.731 45
Posttest 1 total 93.86 8.183 49 0.26 90.98 13.407 45
Posttest 2 total 94.98 7.750 49 0.47 90.24 12.147 45

generally remain the same for the Control Group. Scores are calculated as percentage
accuracy for obligatory occasions created for each of the five article usages being
examined.
One-way ANOVAs show no overall difference between groups in posttest 1 (F(1,92) =
1.609, p = .208) but an overall significant difference on posttest 2 (F(1,92) = 5.161,
p = .025), with a small effect size (d = 0.47). The ANOVAs reveal statistical group differ-
ences on the production of non-generic articles which become apparent in posttest 2
(F(1,92) = 3.977, p = .049), with a small effect size (d = 0.41). The production of generic
articles initially shows statistical significance in favour of the Control Group in the pre-
test (F(1,92) = 4.198, p = .043, a small effect size in favour of the Control Group
d = –0.41) but becomes insignificant in both posttests.

3 Grammaticality judgement task

The descriptive statistics for the grammaticality judgement task (see Table 7) show
relatively high levels of accuracy on the pretest, with higher scores on judgements of
non-generic articles over generic articles, suggesting greater difficulty with generic
articles. Participants in both groups judged ungrammatical articles a little better than
grammatical articles. The Experimental Group marginally outperformed the Control
Group on all components of the posttests, and explicit instruction leads to greater
group differences in posttest 2 on judgements of ungrammatical articles. The effect
size is medium (d = 0.70). The effect sizes for total task scores, however, remain small
throughout: (d = 0.26, d = 0.39, d = 0.43).
Results of a one-way ANOVA reveal no overall difference for group on posttest 1
(F(1,92) = 3.496, p = .065). However, there is an overall group difference on posttest 2
(F(1,92) = 4.457, p = .037), with a small effect size (d = 0.43). The ANOVAs do not reveal
any group differences for the grammatical items in either posttest 1 (F(1,92) = .495,
Akakura 23

Table 7. Descriptive statistics and effect sizes for the grammaticality judgement task

Experimental group Control group

M SD n d M SD n
Non-generic (maximum = /12):
Pretest 11.10 0.941 49 0.45 10.53 1.517 45
Posttest 1 11.43 0.890 49 0.31 11.11 1.153 45
Posttest 2 11.55 0.843 49 0.44 11.11 1.153 45
Generic (maximum = /8):
Pretest 5.55 2.072 49 0.06 5.42 2.158 45
Posttest 1 6.69 1.686 49 0.31 6.16 1.770 45
Posttest 2 6.63 1.590 49 0.30 6.07 2.093 45
Grammatical (maximum = /10):
Pretest 7.94 1.701 49 0.11 7.73 1.993 45
Posttest 1 9.06 1.464 49 0.15 8.84 1.522 45
Posttest 2 8.78 1.662 49 0.07 8.67 1.552 45
Ungrammatical (maximum = /10):
Pretest 8.71 1.354 49 0.34 8.22 1.491 45
Posttest 1 9.06 1.314 49 0.45 8.42 1.373 45
Posttest 2 9.41 0.788 49 0.70 8.51 1.646 45
Total (maximum = /20):
Pretest total 16.65 2.394 49 0.26 15.95 2.946 45
Posttest 1 total 18.12 2.185 49 0.39 17.26 2.250 45
Posttest 2 total 18.18 2.038 49 0.43 17.18 2.570 45

p = .483) or posttest 2 (F(1,92) = .107, p = .744). However, there are differences

between groups in favour of the Experimental Group on both posttests for the ungram-
matical items; posttest 1 (F(1,92) = 5.315, p = .023), with a small effect size (d = 0.45),
which increases in posttest 2 (F(1,92) = 11.649, p = .001) to a medium effect size
(d = 0.70). The ANOVAs do not reveal any group differences for generic articles in
any of the three tasks (F(1,92) = .087, p = .769, F(1,92) = 2.280, p = .134, F(1,92) =
2.200, p = .141).

4 Metalinguistic knowledge task

The results on the metalinguistic knowledge task (see Table 8) show that both groups
performed similarly on the pretest with negligible effect sizes in favour of the Experimental
Group. The Experimental Group outperformed the Control Group on all components of the
posttests with mostly large effect sizes. The overall effect of explicit instruction is evident in
the total scores for both posttests, which were large (d = 1.10, d = 1.07).
The ANOVA shows an overall significant group difference in posttest 1, (F(1,92) =
28.787, p = .000), where the effect size is large (d = 1.10). This is sustained in posttest 2
(F(1,92) = 27.344, p = .000), again with a large effect size (d = 1.07). The sub-components
contributing to this result are scores on generic articles, non-generic articles, and in stating
24 Language Teaching Research 16(1)

Table 8. Descriptive statistics and effect sizes for the metalinguistic knowledge task

Experimental group Control group

M SD n d M SD n
Non-generic (maximum = /9):
Pretest 8.06 1.14 49 0.16 7.87 1.29 45
Posttest 1 8.80 0.50 49 0.87 7.87 1.42 45
Posttest 2 8.88 0.39 49 0.87 8.11 1.19 45
Generic (maximum = /6):
Pretest 3.39 1.58 49 0.07 3.27 1.62 45
Posttest 1 4.47 1.21 49 0.92 3.27 1.39 45
Posttest 2 4.65 1.32 49 0.86 3.56 1.22 45
Correction of articles (maximum = /10):
Pretest 8.31 1.52 49 0.01 8.29 1.47 45
Posttest 1 8.92 0.95 49 0.45 8.40 1.32 45
Posttest 2 9.12 1.05 49 0.50 8.60 1.01 45
Article rule-stating (maximum = /5):
Pretest 3.14 1.22 49 0.24 2.84 1.24 45
Posttest 1 4.35 0.72 49 1.52 2.73 1.32 45
Posttest 2 4.41 0.76 49 1.28 3.07 1.27 45
Total (max = /15):
Pretest total 11.45 2.292 49 0.13 11.13 2.455 45
Posttest 1 total 13.27 1.381 49 1.10 11.13 2.380 45
Posttest 2 total 13.53 1.501 49 1.07 11.67 1.942 45

the reason for ungrammatical articles, which are all significant with large effect sizes in
both posttests:

•• generic posttest 1, F(1,92) = 20.140, p = .000, d = 0.92;

•• generic posttest 2, F(1,92) = 17.539, p = .000, d = 0.86;
•• non-generic posttest 1, F(1,92) = 18.422, p = .000, d = 0.87;
•• non-generic posttest 2, F(1,92) = 18.186, p = .000, d = 0.87;
•• rule-stating posttest 1, F(1,92) = 55.142, p = .000, d = 1.52;
•• rule-stating posttest 2, F(1,92) = 39.374, p = .000 d =1.28.

There are also group differences for correction in posttest 1 (F(1,92) = 4.813, p = .031),
with a small effect size (d = 0.45). This difference increases slightly in posttest 2 (F(1,92) =
6.007, p = .016), with a medium effect size (d = 0.50).

IV Discussion
A summary of the significant effects found for instruction described in the previous sec-
tion is provided in Table 9. On measures of implicit knowledge, significant effects for
instruction were found for both generic and non-generic articles in the immediate posttest
Akakura 25

Table 9. Summary of significant effects found for explicit instruction

Measures of explicit Measures of implicit

knowledge knowledge

Grammaticality Metalinguistic Elicited Oral

judgement knowledge imitation production
Generic:
Posttest 1 – ** ** –
Posttest 2 – ** ** –
Non-generic:
Posttest 1 – ** * –
Posttest 2 * ** ** *
Grammatical:
Posttest1 – n/a * –
Posttest 2 – n/a ** *
Ungrammatical:
Posttest1 * *** ** n/a
Posttest 2 ** *** ** n/a
Total:
Posttest 1 – ** ** –
Posttest 2 * ** ** *
Notes: * p < .05, ** p < .001, *** error correction is p < .05 and rule stating is p < .001

of the elicited imitation task, and this was sustained with an increase on delayed posttest-
ing. Similar results were found when analysed in terms of grammatical vs. ungrammatical
items. The oral production task reflected the difficulty of generating target forms that is
characteristic of this type of production task. The average accuracy total in the pretest is
the highest of all four tasks at almost 91%, while at the same time production of generic
articles is very low, with only 11% of the participants producing generic indefinite articles
a/an, and just 3% of participants producing generic definite article the in their stories (see
Table 10). Production of generic articles remains low in both posttests. The beneficial
effect of explicit instruction on the oral production task became apparent only in the
delayed posttest. This may have been due to an underestimation of participants’ gains
caused by the high pretest scores; that is, there was a possible ceiling effect. Differences
in language performance may be influenced by the measures being used to elicit data
(Norris & Ortega, 2003), causing inherently less demanding tasks to result in higher
scores than complex tasks. This is evident between the two measures of implicit knowl-
edge, with the elicited imitation task being the more complex task of the two.
The durability and robustness of the effectiveness of explicit instruction after six
weeks on measures of implicit knowledge is a major finding. Previous general trends
reported by Norris and Ortega (2000) in their meta-analysis indicate that the effect of
instruction may last beyond immediate post-experimental observations. However, unlike
the current study, gains in previous studies have tended to gradually deteriorate over
time. The results of the current study corroborate previous findings showing effects for
instruction on implicit knowledge only after a delay (R. Ellis, 2009; Mackey, 1999).
26 Language Teaching Research 16(1)

Table 10. Number of obligatory occasions created in the oral production task

Non-generic Generic

indefinite a indefinite an definite the indefinite a/an definite the

Pretest: Control and 1,027 144 2,691 38 10
Experimental
Average per participant 10.93 1.53 28.63 0.40 0.11
(n = 94)
Pretest: Experimental 551 74 1,523 14 6
group
Average (n = 49) 11.24 1.51 31.08 0.29 0.12
Posttest 1: 570 91 1,431 17 7
Experimental group
Average (n = 49) 11.63 1.86 29.20 0.35 0.14
Posttest 2: 569 99 1,405 15 9
Experimental group
Average (n = 49) 11.61 2.02 28.67 0.31 0.18

R. Ellis (1990) notes that instruction may have a delayed effect when grammatical
concepts help learners attend to the features of the input and so acquire them procedur-
ally. The delayed effect may be due to the extra time required for internal processing,
which is considered necessary for input to be converted to implicit knowledge (Gass,
1997; Nassaji & Fotos, 2004; VanPatten, 1996).
The current results suggest that explicit instruction has an effect on implicit language
knowledge. This has been a subject of theoretical debate. Contrary to some article acqui-
sition studies (Huong, 2005; Shimamune & Jitsumori, 1999; Shiozawa & Simmons,
1995), this study indicates that instructed learners can develop implicit knowledge of
non-salient forms through self-paced autonomous study on CALL activities.
On measures of explicit knowledge, the beneficial effects of explicit instruction were
limited to ungrammatical stimuli only. The Experimental Group made significant improve-
ment on the metalinguistic knowledge posttests. This suggests that explicit input can trans-
late into explicit knowledge (if only for ungrammatically supplied items) among advanced
L2 learners. On the other hand, despite greater potential for improvement on generic arti-
cles due to low pretest scores on the grammaticality judgement task, instruction did not
significantly affect judgement of generic articles. The fact that generic articles pose prob-
lems for L2 learners is worth noting. Prior studies investigating the effect of instruction on
generic articles (Master, 1994; Muranoi, 2000) tended to only provide total accuracy
scores, so that differences between generic and non-generic articles were not apparent.
A reason for the differential effect of explicit instruction on measures of explicit
knowledge may be that judging article errors better reflects the use of explicit knowl-
edge. Current results support previous findings (R. Ellis et al., 2009; Han & Ellis,
1998) that suggest ungrammatical sentences are better indicators of explicit knowl-
edge. The difference, however, may also have been due to the nature of the target form
in the present study.
Akakura 27

Ungrammatical articles may have been easier to detect compared to grammatical arti-
cles in the current grammaticality judgement task. The complex nature of articles may
have contributed to the difficulty in judging the grammaticality of grammatical articles,
while not affecting judgement of ungrammatical articles which were mainly supplied as
omissions.2 For example, L2 participants unfamiliar with generic articles may consider
either one of the following to be wrong, because they are asked to judge different forms
of articles with overlapping meanings.

(5) a. A penguin uses its wings when swimming under water.

b. One of the famous flightless birds of New Zealand is the kiwi.

On the other hand, violations of any function through omission may have made the
error immediately obvious, allowing for a correct judgement.

(6) a. Among New Zealand’s rare birds is brown kiwi.

b. Pablo Picasso is painter and artist who changed the history of 20th Century art.

The fact that all permissible usages must be considered before correctly determining that
an article is ‘grammatical’ stems from articles having multiple functions for the same form.
Such complexity in the form–meaning relationship is claimed to influence grammatical
difficulty (DeKeyser, 2005) and may be the cause of difficulty with generic articles.

V Conclusions
One implication that can be drawn from this study is that explicit instruction can benefit
implicit L2 knowledge as well as explicit knowledge. Thus, these results corroborate
previous research that has reported a positive effect for explicit instruction (e.g. DeKeyser,
1998). The study also demonstrates that it is possible to obtain relatively separate meas-
ures of implicit and explicit knowledge of L2 grammar.
There are also pedagogical implications. In the context of the tasks and the target
structure chosen for the present investigation, the autonomous study on CALL activities
helped foster implicit knowledge. Time had almost no detrimental effect on the acquisi-
tion of articles as implicit knowledge; the Experimental Group performed best in the
delayed posttests. Results also suggest that explicit instruction assists understanding of
generic and non-generic article ‘rules’, although additional instruction may be necessary
to enable learners to decide whether sentences containing articles are well formed or
deviant. On one level, these findings support Master’s (1997) claim that articles should
be taught even at advanced stages of English acquisition. The results are encouraging for
those who are no longer in any language program but prefer to work autonomously in
improving their L2 abilities. On another level, it provides a case for a direct rule-giving
approach in CALL activities. The current study suggests that explicit instruction was
effective in improving oral production of articles and helped learners become better ana-
lysts of article errors.
This study is not without its limitations. It only shows that explicit instruction in a
computer-mediated environment is effective. Further research is needed to investigate
28 Language Teaching Research 16(1)

classroom-based instruction on articles. It is also possible that there were ceiling effects
in some of the measures, which may have under-evaluated the effect of the treatment.
There is also an issue of task complexity, evident between the oral production and elic-
ited imitation tasks. This study was designed with multiple tests so as to mitigate the
methodological limitations of each task.
A possible expansion of the present investigation would be to see whether the observed
patterns of performance are the same for different grammatical structures (particularly
other non-salient forms) and in different languages. Future research investigating effects
of instruction on implicit knowledge may benefit from conducting posttests over a longer
term to avoid making premature inferences about the role of instruction. Replications of
the current study using traditional face-to-face instruction may provide empirical results
applicable to classroom-based L2 development. Also, while no conclusive answer on the
relationship between implicit and explicit knowledge or on the interface hypothesis is
possible as L2 participants of this study had existing explicit/implicit knowledge, there
was evidence that instruction led to the development of implicit knowledge. Future stud-
ies may be able to investigate the relationship between the two knowledge types by
teaching a structure that L2 learners have not yet begun to acquire.

Notes
1. Participants’ proficiency levels were determined by rating their stories produced in the pretest
according to the IELTS speaking band descriptors (public version), i.e. (a) fluency and coher-
ence, (b) lexical resource, (c) grammatical range and accuracy and (d) pronunciation. An
average total was then used to determine proficiency in five categories ranging from beginner
to near native.
2. Ungrammatical items were created through omission of articles. In most cases, it is not
possible to create an ungrammatical item by replacing the indefinite with the definite article
(for example, the use of the definite article would not make ‘A cat has good vision in the dark’
ungrammatical), or vice versa.

References
Akakura, M. (2009). Effect of explicit instruction on implicit and explicit second language knowl-
edge: An empirical study on English article acquisition. Unpublished doctoral thesis, The
University of Auckland, New Zealand. [Electronic Version.] Digital Dissertations.
Batstone, R. & Ellis, R. (2009). Principled grammar teaching. System: An International Journal of
Educational Technology and Applied Linguistics, 37, 194–204.
Bitchener, J. & Knoch, U. (2009). The value of a focused approach to written corrective feedback.
ELT Journal, 63, 204–11.
Brender, A.S. (2002). The effectiveness of teaching articles to (-Art) ESL students in writ-
ing classes using consciousness-raising methods. Unpublished EdD, Temple University,
Philadelphia, PA.
Butler, Y.G. (2002). Second language learners' theories on the use of English articles: An analysis
of the metalinguistic knowledge used by Japanese students in acquiring the English article
system. Studies in Second Language Acquisition, 24, 451–80.
Bygate, M. (1999). Quality of language and purpose of task: Patterns of learners' language on two
oral communication tasks. Language Teaching Research, 3, 185–214.
Akakura 29

Corder, S.P. (1967). The significance of learner's errors. International Review of Applied
Linguistics in Language Teaching, 5, 161–69.
DeKeyser, R.M. (1998). Beyond focus on form: Cognitive perspectives on learning and practic-
ing second language grammar. In C.D.J. Williams (Ed.), Focus on form in classroom second
language acquisition (pp. 42–63). Cambridge: Cambridge University Press.
DeKeyser, R.M. (2005). What makes learning second-language grammar difficult? A review of
issues. Language Learning, 55, 1–25.
Dulay, H., Burt, M., & Krashen, S. (1982). Language two. New York: Oxford University Press.
DuPasquier, P. (1988). The great escape. Boston, MA: Houghton Mifflin.
Ellis, N. (2006). Selective attention and transfer phenomena in L2 acquisition: Contingency, cue
competition, salience, interference, overshadowing, blocking, and perceptual learning. Applied
Linguistics, 27, 164–94.
Ellis, R. (1990). Instructed second language acquisition: Learning in the classroom. Oxford/
Cambridge, MA: Blackwell.
Ellis, R. (2002). Does form-focused instruction affect the acquisition of implicit knowledge? A
review of the research. Studies in Second Language Acquisition, 24, 223–36.
Ellis, R. (2005). Measuring implicit and explicit knowledge of a second language: A psychometric
study. Studies in Second Language Acquisition, 27, 141–72.
Ellis, R. (2008). Explicit knowledge and second language learning and pedagogy. In J. Cenoz
& N.H. Hornberger (Eds.), Encyclopedia of language and education: Volume 6: Knowledge
about language (pp. 143–153). 2nd edition. New York: Springer.
Ellis, R. (2009). Measuring implicit and explicit knowledge of a second language. In R. Ellis,
S. Loewen, C. Elder, R. Erlam, J. Philp, & H. Reinders (Eds.), Implicit and explicit knowledge
in second language learning, testing and teaching (pp. 31–64). Bristol/Buffalo, NY/ Toronto:
Multilingual Matters.
Ellis, R., Loewen, S., Elder, C., Erlam, R., Philp, J., & Reinders, H. (2009). Implicit and explicit
knowledge in second language learning, testing and teaching. Bristol/Buffalo, NY/Toronto:
Multilingual Matters.
Ellis, R., Sheen, Y., Murakami, M., & Takashima, H. (2008). The effects of focused and unfo-
cused written corrective feedback in an English as a foreign language context. System: An
International Journal of Educational Technology and Applied Linguistics, 36, 353–71.
Erlam, R. (2006). Elicited imitation as a measure of L2 implicit knowledge: An empirical valida-
tion study. Applied Linguistics, 27, 464–91.
Felix, U. (2003). Pedagogy on the line: Identifying and closing the missing links. In U. Felix
(Ed.), Language learning online: Towards best practice. Netherlands: Swets & Zeitlinger
Publishers, 147–71.
Gass, S.M. (1997). Input, interaction, and the second language learner. Mahwah, NJ: Lawrence
Erlbaum.
Han, Y. & Ellis, R. (1998). Implicit knowledge, explicit knowledge, and general language profi-
ciency. Language Teaching Research, 2, 1–23.
Hinkel, E. & Fotos, S. (2002). New perspectives on grammar teaching in second language class-
rooms. Mahwah, NJ: Lawrence Erlbaum.
Hulstijn, J. (2005). Theoretical and empirical issues in the study of implicit and explicit second-
language learning. Studies in Second Language Acquisition, 27, 129–40.
Huong, N.T. (2005). Vietnamese learners mastering English articles. Unpublished PhD disserta-
tion, University of Groningen, Groningen, The Netherlands.
Jacobsson, B. (1998). Notes on genericity and article usage in English. Studia Neophilologica, 69,
139–53.
30 Language Teaching Research 16(1)

Krashen, S.D. (2003). Current issues and controversies: Does grammar teaching work? What
about ‘comprehensible output’? In S.D. Krashen (Ed.), Explorations in language acquisition
and use: The Taipei lectures (pp. 30–67). Portsmouth: Heinemann.
Macaro, E. & Masterman, L. (2006). Does intensive explicit grammar instruction make all the dif-
ference? Language Teaching Research, 10, 297–327.
Mackey, A. (1999). Input, interaction, and second language development: An empirical study of
question formation in ESL. Studies in Second Language Acquisition, 21, 557–87.
Master, P. (1994). The effect of systematic instruction on learning the English article system. In
T. Odlin (Ed.), Perspectives on pedagogical grammar (pp. 229–52). Cambridge: Cambridge
University Press.
Master, P. (1997). The English article system: Acquisition, function, and pedagogy. System, 25,
215–32.
Mayer, M. (1969). Frog, where are you? New York: The Dial Press.
Mayer, M. (1974). Frog goes to dinner. New York: The Dial Press.
Muranoi, H. (2000). Focus on form through interaction enhancement: Integrating formal instruc-
tion into a communicative task in EFL classrooms. Language Learning, 50, 617–73.
Nassaji, H. & Fotos, S. (2004). Current developments in research on the teaching of grammar.
Annual Review of Applied Linguistics, 24, 126–45.
Norris, J.M. & Ortega, L. (2000). Effectiveness of L2 instruction: A research synthesis and quan-
titative meta-analysis. Language Learning, 50, 417–528.
Norris, J.M. & Ortega, L. (2003). Defining and measuring SLA. In C. Doughty & M.H. Long
(Eds.), The handbook of SLA (pp. 716–61). Malden, MA: Blackwell.
Ortega, L. (1999). Planning and focus on form in L2 oral performance. Studies in Second Language
Acquisition, 21, 109–48.
Paradowski, M.B. (2007). Foreign-language grammar instruction via the mother tongue. In
F. Boers, J. Darquennes, & R. Temmerman (Eds.) Multilingualism and applied comparative
linguistics, Vol. 1: Comparative considerations in second and foreign language instruction
(pp. 151–67). Newcastle-upon-Tyne: Cambridge Scholars Press.
Parrott, M. (2000). Grammar for English language teachers. Cambridge: Cambridge University
Press.
Rovers, M. (2002). An investigation into the effects of two instructional treatments on the devel-
opment of ESL learners' explicit and implicit knowledge of articles. Unpublished Masters
Dissertation, The University of Auckland, Auckland, New Zealand.
Schmidt, R. (2001). Attention. In P. Robinson (Ed.), Cognition and second language instruction
(pp. 3–32). Cambridge: Cambridge University Press.
Shimamune, S. & Jitsumori, M. (1999). Effects of grammar instruction and fluency training on
the learning of the and a by native speakers of Japanese. The Analysis of Verbal Behavior, 16,
3–16.
Shiozawa, T. & Simmons, T. (1995). Do improved examination scores indicate acquired skills?
A study in teaching the use of articles in journal writing through a continuous testing regimen.
Journal of International Studies, 14, 133–49.
Skehan, P. & Foster, P. (1999). The influence of task structure and processing conditions on nar-
rative retellings. Language Learning, 49, 93–120.
Stepp-Greany, J. (2002). Student perceptions on language learning in a technological environment:
Implications for the new millennium. Language Learning and Technology, 6, 165–80.
Tashima, Y. (2004). Teaching English articles by the binary system. Unpublished Master’s
Dissertation, Portland State University, OR.
VanPatten, B. (1996). Input processing and grammar instruction in second language acquisition.
Norwood, NJ: Ablex.
Akakura 31

Appendix 1
Measures of implicit knowledge
Elicited imitation task (Version A: one of three versions)
Recorded instruction and test sentences

This is a True or False, and creative sentence-making task. The pictures you have in front
of you tell a story. You will hear a story to go with the pictures. The pictures always tell
the correct version of the story but the tape will sometimes not.

As you listen to the story, you will hear this bell sound.  The bell will be followed by
one sentence, and the question “is this true or false?” Then there will be a pause. During
this pause, you are asked to do two things.

First, indicate on your answer sheet, whether you think the sentence is true or false. The
pictures will give you a hint for some of the questions; otherwise, you may have to rely
on what you already know. If you think it is true, tick in the TRUE box. If you think it is
false, tick in the FALSE box. If you are not sure, please tick in the NOT SURE box.

After ticking one of the boxes on your answer sheet, then please say your sentence. The
words you must use are provided under the picture. Please use all of the words without
changing them in any way. Sometimes you may have to change the sentence from what
you hear to match the pictures. This task is not testing your memory. So you don’t have
to repeat what you hear, but you can use what you hear to give you an idea of how you
can create your sentence. Be as creative as you can, but please remember to use correct
English and to use all the words provided without changing them in any way. First here
is a practice.

{If participant summarizes, or repeats before ticking, point these out, re-do the practice
if necessary}

1. There was once a little boy called James. It was his sister’s birthday and his fam-
ily was going out for dinner, and so James was getting dressed.  His pet dog,
turtle and frog watch(ed) him. “Is this true or false?” [True]
2. James told his pets to stay at home and wait for him.  He did not see that one of
them had jumped into his pocket. “Is this true or false?” [True]
3. Everyone was dressed nicely and ready to leave. But no one seemed to notice that
one of his pets was hiding in his pocket.  James wave(d) good-bye to his pet
frog and turtle. “Is this true or false?” [False]
---------------------------------------------------------------------------------------------------------------------
Do you have any questions? If you do, please ask now. {Stop to answer questions}
Remember to tick True, False or Not Sure first, and then say your creative sentence using
all of the words under the picture without changing them. Now the story will begin.
---------------------------------------------------------------------------------------------------------------------
32 Language Teaching Research 16(1)

1. There was once a little boy called James who loved animals. He had a pet dog
named Oscar. One day, James found a frog and took it home. At night, they all
went upstairs.  Once in his room, (the) boy put the frog inside his wooden box.
“Is this true or false?” [False]
2. James and Oscar soon fell fast asleep.  During the* night, when no one was
looking, (an) unexpected thing happened. “Is this true or false?” [True]
3. The frog got out of the glass jar. Next morning, James couldn’t see his frog. He
searched his room. Oscar tried to help, but his head got stuck in the glass jar.
Have you noticed that so many dogs are having accidents recently?  I think that
(the) dog will be extinct in fifty years. “Is this true or false?” [False]
4. James didn’t notice what was happening to Oscar, but he did notice that the win-
dow had been left open.  Oscar was wondering if the* frog had escaped when
James really had an accident. “Is this true or false?” [False]
5. Luckily, Oscar wasn’t hurt. Now that they were outside, James continued search-
ing. Oscar found a strange object hanging from a tree. It was a beehive, but Oscar
didn’t know that.  You see, (a) dog will sing to anything suspicious. “Is this true
or false?” [False]
6. While Oscar was barking, James suddenly cried out “Ouch!”  James was badly
bitten on his left leg by (an) angry animal. “Is this true or false?” [False]
7. Oscar was also in trouble. Suddenly lots of small insects were chasing him. What
could they be?  One of the* insects that attack in large numbers is the bee. “Is
this true or false?” [True]
8. James was looking at a tree when something suddenly flew out and he fell back-
wards.  What flew out was an old kiwi bird that lived inside the oak tree. “Is
this true or false?” [False]
9. Now, this owl belongs to a family of owls that is protected by the wild life agency.
Why are they protected?  It could be because (the) owl will die out if they’re
not protected. “Is this true or false?” [True]
10. But James was the one who needed protection. The owl flapped its large wings and
started attacking James. James ran and ran, but the owl was always right above. 
This shows that it is the bird that rules the* sky. “Is this true or false?” [True]
11. Finally the owl gave up. But before James knew what was happening, he found
himself on top of a deer’s head. He tried to move, but couldn’t.  We all know
that (a) deer is a big, strong animal for (a) little boy. “Is this true or false?” [True]
12. The deer started running. Oscar saw this and started following.  Suddenly, (the)
deer stopped in front of a car. “Is this true or false?” [False]
13. James was thrown off the cliff. Oscar couldn’t stop and also started falling.
Splash, went James, splash went Oscar. Oscar quickly climbed onto James’ head.
 You know, a dog is usually (a) good swimmer. “Is this true or false?” [True]
14. But Oscar didn’t like getting wet. Well, as they sat in the pond, James thought he
heard something. When they looked behind a log, they found James’ frog together
with a whole family of frogs.  Did you know that a frog likes to live in wet
places? “Is this true or false?” [True]
15. James promised he would keep the frog where there’s some water. The frogs
must have been happy about that, for they all sent him off with a smile on their
faces. The end.
Akakura 33

Elicited imitation task (Version A): Material provided to test-takers

1. PRACTICE

Picture of a boy getting dressed, while his Picture of one of the pets, a frog, jumping
pets watch him. into the boy’s coat pocket without the
boy noticing it.

1. dog, turtle, frog   2. … jumped into …

Picture of the boy waving good-bye as he leaves the house with his family. The frog
is in his coat pocket, also waving good-bye to the dog and turtle.

3. good-bye to, pet

End of practice
2. Task (Version A)
 Actual test had pictures above the words listed below

1. Once in …, boy, frog …

2. During the night when …, unexpected thing …
3. I think … dog, extinct, fifty years.
4. wondering if the frog …, accident.
5. You see, dog, suspicious.
6. badly bitten, angry animal
7. One of the insects, attack, bee
8. What flew out, old …, oak tree.
9. It could be, owl, die out if they’re …
10. … bird that rules the sky.
11. deer, big strong animal, little boy
12. deer, stopped in front of …
13. dog, usually, good swimmer.
14. frog likes …

Oral production Task (Version A)

Material was the same series of pictures as the Elicited Imitation Test.
34 Language Teaching Research 16(1)

Appendix 2
Measures of explicit knowledge
Grammaticality judgement test and Metalinguistic knowledge test (Writing task Version A)
PART 1. Instruction
You will see sentences with one underlined portion. You are asked to do 2 things.
(1) Please decide if the underlined words are correct in English or not.
Tick  one box that describes your answer.

 Practice 1. We hope you will enjoying visiting Paris.

Not correct Probably not correct Probably correct Correct

   

(2) Then indicate how you made your judgement.

 How did you decide?    By Feel    By Rule

There is no ‘right’ vs. ‘wrong’ or ‘better’ way to make your judgement.

Tick  ‘By Feel’ … If you knew your answer immediately, just by reading the sentence.
or
If you did not decide immediately, but you did not rely on any grammatical rule to
make
your judgement.
Tick  ‘By Rule’ … If you did not decide immediately, and you relied on a grammatical
rule to make your
judgement.
or
If you changed your answer from what you first ‘felt’, after thinking of a grammatical
rule.

An example of a grammatical rule is: Verbs in the past tense often end in ‘-ed.’

 Practice 2. I went to a flower shop. There were many flower in the shop window.

Not correct Probably not correct Probably correct Correct

   

How did you decide?    By Feel    By Rule

Do you have any questions? If you do, please ask now.

Once you complete an item, you MUST NOT go back and change your answers.
Akakura 35

You can take as much time as you need for each item.
There are 20 questions in part 1.
Please turn over the page when you are ready to begin.

Test sentences

1. A religious ceremony may be perceived as a strange custom if one is unfamiliar with

that religion.
2. “What are you drawing?” asked mum.
“It’s supposed to be a elephant,” said Hannah crossly.
3. World Expo 2005 deals with how we humans can live on the planet without
destroying
the environment.
4. Dolphin feeds on their mother’s milk.
5. Among New Zealand’s rare birds is black robin.
6. Astrid Klein was trained at art school but began as writer.
7. An increase in the number of children born will create an additional demand for
education.
8. It had rained, and so ground was wet.
9. A penguin uses its wings when swimming under water.
10. One of the famous flightless birds of New Zealand is the kiwi.
11. Hiding the mouth with one’s hands when laughing is a gesture seen among many
Japanese.
12. “What did you have for breakfast?” asked Mary. “I had a egg,” said James.
13. The Chinese celebrate the Mid-Autumn Festival when the moon is said to be at its
brightest.
14. Whale feeds on its mother’s milk.
15. Among New Zealand’s rare birds is brown kiwi.
16. Pablo Picasso is painter and artist who changed the history of 20th Century art.
17. An increase in the number of the elderly will create an additional need for medical
facilities.
18. It had rained for a whole week, but today sun came out at last.
19. A cat has good vision in the dark.
20. A famous New Zealand bird that no longer exists is the moa.

PART 2. Instruction
You will see sentences containing errors. The part of the sentence that contains the error
is underlined. You are asked to do 2 things.

(1) Please correct the underlined error. Do not change any other parts of the sentence.
(2) Then indicate how you made your decision.

Here is an example.

We hope you will enjoying visiting Paris.

36 Language Teaching Research 16(1)

(1) Correct the error. [ enjoy ]

(2) How did you decide?    By Feel    By Rule

Practice 1.
  I went to a flower shop.
  There were many flower in the shop window.

(1) Correct the error. [ ]

(2) How did you decide?    By Feel    By Rule

Practice 2.
  Excuse me.
  I cannot remember if I have been met you before.
(A conversation at a cocktail party)

(1) Correct the error. [ ]

(2) How did you decide?    By Feel    By Rule

Do you have any questions? If you do, please ask now.

Once you complete an item, you MUST NOT go back and change your answers.
You can take as much time as you need for each item.
There are 10 questions in Part 2.
Please turn over the page when you are ready to begin.

Sentences in the test:

1. In the first modern Olympics held in Athens in 1896, New Zealand won the bronze
medal.
2. “What is that noise?” asked Dora.
“I think it is a airplane,” said Peter, looking up at the sky.
3. Do people speak English in a country where you were born?
4. Cat is said to have nine lives.
5. Among the animals protected by the wild life protection agency is panda.
6. Katherine Mansfield is the New Zealand writer.
7. “What did you buy for her birthday?” asked June.
  “I bought her a expensive watch,” replied John.
8. Which city is a capital of New Zealand, Auckland or Wellington?
9. Cow produces milk.
10. Among the imaginary animals, dragon is particularly interesting.

Please state why the underlined words are incorrect. Do not change any other parts of the
sentence. Do not give the same reason for more than one answer.
Akakura 37

example: There were many flower in the shop window.

You need to … use the plural form of the noun ‘flowers’
because … ‘many’ indicates that there’s more than one flower.
_________________________________________________________________

1. In the 1996 Olympics held in Atlanta, New Zealand won the gold medal.
You need to …
because …

2. The thief stole a expensive vase from the living room.

You need to …
because …

3. Where was last country you visited before you came back?
You need to …
because …

4. Panda likes to eat bamboo shoots.

You need to …
because …

5. A kiwi bird is in danger of becoming extinct.

You need to …
because …