RELATIONSHIPS BETWEEN SECOND LANGUAGE

LEARNING STRATEGIES AND LANGUAGE PROFICIENCY

IN THE CONTEXT OF LEARNER AUTONOMY AND SELF-
REGULATION

Rebecca L. Oxford
University of Alabama

ABSTRACT

The first section of the article provides a theoretical framework for

understanding the relationship of language learning strategies and lan-
guage proficiency. This framework is provided by two interrelated ideas:
the educational concept of learner autonomy and the psychological theory
of self-regulation. The second section presents research linking learning
strategy use with learning outcomes, both in general academic learning
and in language learning. The third section focuses on relationships be-
tween language learning strategy use and language proficiency in stud-
ies employing a well-known strategy use questionnaire, the Strategy In-
ventory for Language Learning. The final section raises issues for future
research related to language learning strategies and language proficiency
in the context of learner autonomy/self-regulation.

INTRODUCTION

Imagine a second or foreign language class with which you are somewhat or very
familiar. Now consider which students in this class are the most competent in learn-
ing the language. These particular students are likely to be using a wide variety of
language learning strategies and are probably on the way to becoming autonomous,
self-regulated learners, that is, learners who take significant responsibility for their
own learning.

Revista Canaria de Estudios Ingleses, No. 38, 1999, págs. 109-126

110 REBECCA L. OXFORD

Learning strategies are defined as “specific actions, behaviors, steps, or tech-

niques —such as seeking out conversation partners, or giving oneself encouragement
to tackle a difficult language task— used by students to enhance their own learning”
(Scarcella & Oxford, 1992, p. 63). The term strategy implies conscious movement
toward a goal. The penultimate goal of language learning strategies is to enable the
learner to accomplish individual learning tasks (Richards & Lockhart, 1996), and the
ultimate goal is to promote language proficiency (Tudor, 1996) so that the learner can
use the language outside the classroom.1
The purposes of this article are:

• To discuss learner autonomy and self-regulation as a theoretical framework for

understanding the role of language learning strategies;
• To synthesize findings from research on learning strategies and learning out-
comes, both in general and in the area of language learning;
• To highlight research results about the use of language learning strategies in
relation to proficiency in questionnaire-based studies employing the Strategy Inven-
tory for Language Learning; and
• To point out remaining issues to be explored by future research on language
learning strategies, language proficiency, and learner autonomy/self-regulation.

THEORETICAL FRAMEWORK: LEARNER AUTONOMY AND SELF-

REGULATION

The educational concept of learner autonomy is grounded in the social-cognitive

theory of self-regulation. Together, autonomy and self-regulation serve as an excel-
lent theoretical framework for understanding the nature and role of language learning
strategies.

LEARNER AUTONOMY AND LEARNING STRATEGIES

Learner autonomy is evident when the learner takes responsibility for his or her
learning. Many theorists in the area of second and foreign language learning have
discussed learner autonomy (see, for instance, Allwright, 1990; Holec, 1981, 1995;
Littlewood, 1996; Wenden, forthcoming). These theorists define learner autonomy in
slightly different ways.
Dickinson (1987) defines autonomy as a situation in which the learner takes over
his or her own language learning. Holec (1981) defines it as a learner’s ability to be
responsible for one’s learning. Littlewood (1996) says that autonomy is not just abil-
ity but also willingness to take responsibility. Allwright (1990) emphasizes that au-
tonomy involves not only ability and willingness but also action in the direction of
responsibility for learning. Little (1999) suggests that features of autonomy include
being able to perform a given task independently, with situational flexibility, in con-
texts beyond the immediate one, and (in formal learning environments) with con-
scious intention and reflection.
Building on these definitions, here is a comprehensive definition of learner au-
tonomy. Learner autonomy is the (a) ability and willingness to perform a language
 RELATIONSHIPS BETWEEN SECOND LANGUAGE LEARNING... 111

task without assistance, with adaptability related to the situational demands, with
transferability to other relevant contexts, and with reflection, accompanied by (b)
relevant action (the use, usually conscious and intentional, of appropriate learning
strategies) reflecting both ability and willingness. Learner autonomy leads to greater
achievement or proficiency. This situation can be summarized by “five A’s”:

Ability, attitude, + action = autonomy → achievement.

Learning strategies play a key role in autonomy. Numerous researchers in the

area of language learner autonomy identify learning strategies as relevant or even
crucial (Cotterall, 1995a, 1995b; Dickinson, 1992; Little, 2000; Littlewood, 1996;
Wenden, 1991). Strategies, when defined as above (specific actions or behaviors by
which the individual learner attempts to improve his or her language learning) reflect
the learner’s degree of autonomy and are mechanisms by which the learner develops
still greater autonomy.
Autonomy is a variable trait rather than a permanent state. In fact, Little says, the
development of higher cognition, i.e., cognition well above the level of sensory per-
ception, involves “a spiral process in which we move to new levels of autonomy only
by first moving through new phases of dependence” (Little, 1998). To put it another
way, higher-order cognitive development requires alternation between the social en-
gagement necessary to receive assistance and the separation necessary to allow indi-
vidual reflection. Learner autonomy does not refer to detachment of the individual
from other people, from institutions, nor from special materials. “Total detachment is
a principal determining feature not of autonomy but of autism.”.. (Little, 1991, 5).
Little, like Williams and Burden (1997), emphasizes the importance of social interac-
tion in cognitive development and second language learning.

SELF-REGULATION AND LEARNING STRATEGIES

That which is called autonomy in the foreign and second language field is often
known as self-regulation in the psychology field. Self-regulation is the heart of the social-
cognitive theory of Russian psychologist Lev Vygotsky (1978, 1986; for specific applica-
tions to language learning, see Little, 1999 and Scarcella & Oxford, 1992; and for still
other important aspects of self-regulation, see Iran-Nejad, 1990 and Prawat, 1998).
In Vygotsky’s theory, the goal of learning is to develop an independent, self-
regulated, problem-solving individual. This can occur only with the help of “more
capable others” (teachers, more competent peers, parents, or others), who offer as-
sistance to the learner. This assistance is metaphorically known as scaffolding, i.e.,
the external structure that supports and holds up a building that is under construction.
There comes a time when the edifice needs less and less external support or scaffold-
ing. When no longer essential, it is gradually removed. For instance, in higher-order
cognitive development, the “more capable other” removes the scaffolding bit by bit
from the individual learner as the learner becomes increasingly independent and self-
regulated.
Vygotsky’s theory also involves a different metaphor: the zone of proximal devel-
opment or ZPD. The ZPD is the difference between the learner’s actual performance
level and the potential performance level that could occur with the appropriate assist-
112 REBECCA L. OXFORD

ance from more capable others. The teacher, parent, or competent peer helps the learner
move through the ZPD until the learner has reached optimal performance. The ZPD
is more of a heuristic phenomenon than an operationally measurable thing; it is meant
to remind us of the great importance of the “social” part of “social cognition.”
Vygotsky’s psychological work on self-regulation involves learning strategies, al-
though he does not use the term strategies. Self-regulation, in Vygotsky’s view, is «the
process of planning, guiding, and monitoring one’s own attention and behavior» (Berk &
Winsler, 1995, 171). Planning, guiding, and monitoring, along with organizing and evalu-
ating, are among the essential learning behaviors that educators call metacognitive learn-
ing strategies —often corporately referred to as metacognition. (For more about
metacognitive learning strategies in the language field, see O’Malley & Chamot, 1990;
Oxford, 1990, 1996b; Wenden, 1991, forthcoming; Wenden & Rubin, 1987).
Metacognitive strategies are internalized via social interaction with more compe-
tent people in the environment, says Vygotsky’s theory. Likewise, with the right as-
sistance, the learner internalizes cognitive learning strategies, such as analyzing, syn-
thesizing, and evaluating, called by Vygotsky “higher-order cognitive functions”.
At the same time, social interaction requires what Oxford (1990, 1996b) terms
social learning strategies: asking questions, requesting assistance, and collaborating
with others via language, or social speech. Social speech (talking with others), ac-
cording to Vygotsky, encourages the learner to develop egocentric speech (talking to
oneself aloud), which in turn stimulates the development of inner speech (reflecting
metacognitive strategies that guide action) (Little, 1999).
In addition to the three general groups of learning strategies (metacognitive, cog-
nitive, and social) just described, Oxford (1990, 1996b) suggests three additional
kinds of strategies that might be part of language learner self-regulation. Affective
strategies (e.g., lowering anxiety through music, or rewarding oneself for good work)
help learners manage their emotions and motivation. Compensatory strategies (e.g.,
guessing from the context, making gestures to communicate unknown words) com-
pensate for or make up for missing knowledge. Memory strategies (e.g., remember-
ing through mental imagery or acronyms) are cognitive strategies that serve the spe-
cial function of embedding new information into long-term memory.
As we have seen, learning strategies are very important to learner autonomy and
self-regulation. These strategies also have a profound impact on the outcomes of learn-
ing, as demonstrated in the next section.

LEARNING STRATEGY USE, LEARNING OUTCOMES, AND STRATEGY

INSTRUCTION

Research shows many relationships between the use of learning strategies on the one
hand and learning outcomes (variously termed achievement, proficiency, or competence)
on the other. This section provides a brief summary of a number of investigations.

LEARNING STRATEGY USE AND POSITIVE LEARNING OUTCOMES

In subject areas not involving second or foreign languages, the use of learning
strategies is demonstrably related to achievement (Pressley & Associates, 1990;
 RELATIONSHIPS BETWEEN SECOND LANGUAGE LEARNING... 113

Weinstein, Goetz, & Alexander, 1988; see also the reviews included in Chamot,
Barnhardt, El-Dinary, & Robbins, 1996 and Chamot & O’Malley, 1996). Research has
repeatedly shown this relationship in content fields ranging from physics to reading and
from social studies to mathematics. In light of this remarkable association between
learning strategy use and positive learning outcomes, it is not surprising that students
who frequently employ learning strategies enjoy a high level of self-efficacy, that is, a
perception of being effective as learners (Bandura, 1997; Zimmerman & Pons, 1986).

LANGUAGE LEARNING STRATEGY USE AND LANGUAGE PROFICIENCY

Within the field of foreign and second language learning, authors (Cohen, 1990;
O’Malley & Chamot, 1990; Oxford, 1990, 1996b; Wenden & Rubin, 1987) have iden-
tified numerous links between the use of language learning strategies and proficiency
in the targeted language. In studies by Rubin (1975) and Naiman, Fröhlich, Stern,
and Todesco (1975), successful language learners consciously used certain types of
learning strategies, such as guessing from context, to become better learners. In other
studies, successful learners used a larger number and wider range of strategies, em-
ployed them more effectively, fitted them more closely to the task, and understood the
task better than did unsuccessful language learners (see the review in Chamot et al.,
1996). In an investigation by Nunan (1991), effective learners differed from ineffec-
tive learners in their greater ability to reflect on and articulate their own language
learning processes. Successful learners used strategies for active involvement more
frequently in Green and Oxford’s (1995) study.2
The frequency of use of language learning strategies is also associated with dif-
ferences between second and foreign language environments, which in turn appear to
be associated with the need for language proficiency. A second language is a lan-
guage learned in an environment where it is the primary language of daily communi-
cation for most people; some degree of competence or proficiency in the language is
needed for survival (example: learning English in the U.S. or Australia). A foreign
language is a language learned in a setting where it is not the usual medium of inter-
action for the majority of the population, and competence or proficiency is not needed
for survival purposes (example: learning Russian in France or Papua New Guinea).
Research shows that, in general, second language learners develop a wider range and
frequency of learning strategies than do foreign language learners, possibly because
second language learning environments simultaneously demand greater language
proficiency and offer greater environmental supports (Green & Oxford, 1995).

STRATEGY INSTRUCTION STUDIES

To increase language proficiency, some researchers and teachers have provided
instruction that helped students learn how to use more relevant and more powerful
language learning strategies (see, e.g., Chamot et al., 1996; Cohen et al., 1995; Cohen
& Weaver, 1998; Dadour & Robbins, 1996; Nunan, 1997). Positive effects of strategy
instruction have emerged for proficiency in speaking (Dadour & Robbins, 1996;
O’Malley, Chamot, Stewner-Manzanares, Kupper, & Russo, 1985) and reading (Park-
Oh, 1994), although results for listening have not been significant (O’Malley et al.,
1985). In other studies, strategy instruction has led to increased language learning mo-
tivation (Nunan, 1997) and greater strategy use and self-efficacy (Chamot et al., 1996).
114 REBECCA L. OXFORD

STRATEGY USE AND LANGUAGE PROFICIENCY IN STUDIES EMPLOYING

THE STRATEGY INVENTORY FOR LANGUAGE LEARNING

Research investigations involving a questionnaire or summative rating scale known

as the Strategy Inventory for Language Learning or SILL (Oxford, 1990) have produced
still more evidence demonstrating that strategy use relates to language proficiency.

ABOUT THE SILL

The SILL is highlighted here because it is the most widely used language learn-
ing strategy-assessment instrument in the world, although many other kinds of strat-
egy assessment tools are useful for a variety of purposes.3 The SILL also has well
documented reliability4 and validity.5
This questionnaire is currently used in at least 17 languages (Cantonese, English,
Finnish, French, Dutch, German, Japanese, Italian, Korean, Mandarin, Portuguese,
Russian, Serbo-Croatian, Spanish, Swedish, Thai, and Ukrainian). To date it has been
administered to between 9,000 and 10,000 students worldwide; has resulted in more
than 40 dissertations, theses, and other major studies; and is the basis of dozens of
research articles published in refereed journals. The SILL has two forms: a 50-item
questionnaire for people learning English as a second or foreign language and an 80-
item questionnaire for native English speakers learning other languages.
The instrument was originally designed for the U.S. Defense Language Institute,
which wanted to determine how the use of learning strategies influences the success
or failure of military personnel learning foreign languages (Oxford, 1986). However,
a far greater scope emerged for the SILL when it was adopted by various universities,
schools, agencies, and institutes around the globe.
The SILL is comprised of Likert-scaled items (scale 1-5, ranging from “never or
almost never” to “always or almost always”), with each item expressing a learning
strategy. It has the same general format as the Learning and Study Strategy Survey
(LASSI) (Weinstein et al., 1988). An example of a SILL item is: “I try to find patterns
in the language.”
Items in the SILL are grouped into the six categories of strategies mentioned
earlier: memory-related, cognitive, compensatory, metacognitive, affective, and so-
cial strategies. The first three categories are called “direct” strategies because they
directly involve processing or using the language that is being learned. The latter
three categories are collectively labeled “indirect” strategies because they do not in-
volve the language itself; instead they allow the learner to manage himself or herself
with regard to the following: planning, organizing, monitoring, evaluating, maintain-
ing motivation, lowering anxiety, and learning with others.
The SILL’s basic purpose is to provide a general picture of the individual learner’s
typical strategy use, rather than a specific portrayal of the strategies used by the learner
on a particular language task. Results of the SILL have been applied in different ways:

• as practical information for the teacher to improve language teaching and hence
learning (e.g., giving informal, on-the-spot strategy assistance to their students, ei-
ther individually or in groups; knowing which strategies should be taught via more
formal, systematic strategy instruction woven into regular language classes; deter-
 RELATIONSHIPS BETWEEN SECOND LANGUAGE LEARNING... 115

mining changes over time in the strategy use of their students; and identifying how
strategy use relates to language proficiency of individual students or classes),
• as practical information for “language counselors” in agencies and institutes
(e.g., using SILL results, along with data on language learning styles and motivation,
as a basis for offering formalized, private, one-to-one language counseling services
outside of the classroom for students experiencing serious difficulties in learning a
language), and
• as research data for building stronger theories of learning strategies and learner
autonomy (e.g., employing SILL results to obtain a group picture of learning strategy
use across many individuals; to study the internal factors of the SILL, to make com-
parisons of these factors across cultures; to determine how strategy use relates to
language proficiency and cultural beliefs about learner autonomy in various groups;
and to identify relationships between strategy use and a host of additional background
factors: language learning environment, gender, age, beliefs about learner and teacher
roles, intrinsic and extrinsic motivation, motivational goals, anxiety, and risk-taking).

EMPIRICAL RELATIONSHIPS BETWEEN SILL-MEASURED STRATEGY USE AND LANGUAGE

PROFICIENCY
Many SILL studies have shed light on relationships between strategy use and
language proficiency. These investigations involve several kinds of analyses:

• Prediction via Multiple Regression: With multiple regression, researchers have

tested the value of various types of strategies as predictors of language proficiency.
One of the key regression statistics is called R2, as R2 = .49. This figure, when ex-
pressed as a percentage, indicates that 49% of the variance in the thing that is pre-
dicted, such as language proficiency, is explained by the predictors (for instance, by
various types of language learning strategies and sometimes gender, motivation, or
other variables).6
• Correlation: Researchers have calculated correlations between the frequency of
strategy use and language proficiency. Correlations are expressed by the letter r and
show the strength (from .00, no correlation, to 1.00, complete correlation) and the
direction (positive or negative) of a relationship. For example, the correlation r = .70
between learning strategy use and language proficiency would mean that the rela-
tionship is positive and fairly strong, whereas the correlation r = -.15 would indicate
that the relationship is negative and very weak.7
• Studying Effects via Analyses of Variance (ANOVA) and Multiple Analyses of
Variance (MANOVA): Some investigators have employed ANOVA and/or MANOVA
to identify how the use of strategies in the SILL’s six categories (and other variables)
affect language proficiency. Other researchers have used these statistical techniques
to determine how language proficiency levels relate to the use of language learning
strategies of various types.

The following paragraphs give the statistically significant results of studies using
any of these analyses to determine relationships between strategy use and language
proficiency. Results are organized according to the type of learning environment:
second language learning versus foreign language learning. Remember that second
116 REBECCA L. OXFORD

language learning occurs in settings where the language is the main vehicle of every-
day communication (and survival) for most people, while foreign language learning
takes place in environments where this is not the case.

Second Language Learning Environments

Among 305 Afrikaans-speaking learners of English as a second language (ESL)
in South Africa, strategy use on the SILL strongly predicted language proficiency on
the Test of English as a Foreign Language (TOEFL), which is used in both second and
foreign language learning environments. In this study, 46% of the variance in TOEFL
scores was explainable by strategy use as measured by the SILL.8 Metacognitive strat-
egy use was the best predictor, followed by social and affective strategy use. In a
canonical correlation analysis conducted as part of the same investigation, strategy
category scores on the SILL and TOEFL section scores were strongly related, r = .73
(Dreyer & Oxford, 1996).
In a study of 141 adult ESL learners in the U.S., Phillips (1990, 1991) used ANOVA
and found a curvilinear relationship between frequency of strategy use on the SILL and
English proficiency levels on the TOEFL. Specifically, mid-proficient learners showed
higher overall strategy use than did high-proficient and low-proficient learners.
Green (1991) conducted a study of strategy use and language proficiency among
213 Puerto Rican university students who were studying English on the island. Puerto
Rico is included here as an ESL learning environment, although it has some EFL
features.9 Using the standardized English test known as the ESLAT,10 Green found an
average correlation of r = .35 between SILL scores and ESLAT scores. In that study,
using ANOVA, Green found a curvilinear relationship between strategy use and ESL
proficiency, just as Phillips (1990, 1991) had demonstrated.
In a later study with 374 Puerto Ricans learning English at the same university,
Green and Oxford (1995) found a linear relationship between English proficiency
levels on the ESLAT and four kinds of strategies: compensatory, cognitive,
metacognitive, and social. More proficient students used these strategies most fre-
quently. In this larger study, the curvilinear effect found in the earlier Puerto Rican
investigation was not present. Gender differences occurred in the 1995 study, favoring
women as more frequent strategy users.

Foreign Language Learning Environments

Multiple regression analysis showed that eight SILL items predicted 58% of the
variance in Comprehensive English Language Test (CELT) scores of 78 first-year
students of English as a foreign language (EFL) at a Japanese women’s college
(Takeuchi, 1993).11 Four of these items predicted proficiency positively (writing notes,
avoiding verbatim translation, analyzing words, paying attention), but four others
were negative predictors (questioning, flashcards, feelings in diary, looking for Eng-
lish-using opportunities). The researcher explained that these results largely stemmed
from cultural beliefs and values (Takeuchi, 1993).
In a different EFL study conducted in Japan, SILL scores and English language
proficiency self-ratings had a moderate correlation that averaged approximately
r = .30 (Watanabe, 1990). Participants in that study were university and college
students of EFL.
 RELATIONSHIPS BETWEEN SECOND LANGUAGE LEARNING... 117

SILL and TOEFL scores of 332 university-level EFL students in Korea produced
moderate correlations. These were: r = .33 (cognitive strategies), r = .30 (social strat-
egies), r = .28 (metacognitive strategies), r = .24 (memory strategies), r = .23 (affec-
tive strategies), and r = .21 (compensatory strategies) (Park, 1994). In this study,
ANOVA showed that high strategy users had higher TOEFL proficiency levels than
did medium and low strategy users (Park, 1994).
In Taiwan, Ku (1995) conducted a study to assess the value of SILL-gauged strat-
egy use for predicting EFL proficiency. Proficiency was measured via transformed,
standardized T-scores for a nationally normed EFL exam (for high school and univer-
sity students) and for EFL grades (for middle school students). The investigation
involved 904 Taiwanese students from these three educational levels. Results of mul-
tiple regression analyses demonstrated that strategy use scores in various SILL cat-
egories predicted EFL proficiency. Specifically, 21% of the variance in EFL profi-
ciency was explained by SILL-measured strategy use, with the best predictors being
cognitive and compensation strategies.12
Kato’s (1996) multiple regression study involved 40 English-speaking learners
of Japanese Kanji characters, which are known to be quite difficult. This research
took place at an Australian university. The SILL effectively predicted success on Kanji
test scores, with 51% of the variance in Kanji proficiency attributable to strategy use
as reported on the SILL.13 Best predictors were metacognitive and memory strategies.
SILL strategy-category scores and EFL course grades of 73 junior high students
in Turkey were related at a moderate level. Correlations were: r = .40 (cognitive strat-
egies), r = .36 (metacognitive strategies), and r = .24 (compensation strategies) (Ox-
ford, Judd, & Giesen, 1998).
For 110 university students majoring in English in Thailand, SILL compensatory
strategy frequencies and standardized EFL placement test scores were moderately
related, r = .38 (Mullins, 1991). A moderate, negative correlation emerged between
the frequency of affective strategy use and entrance exam scores, r = .32, thus raising
the question: Does anxiety lead to greater use of affective strategies and lower lan-
guage proficiency? Other research reported in Oxford (1996b) tends to support this
possibility.
In an investigation of a highly selective group of 262 adult, English-speaking
foreign language learners at the U.S. Foreign Service Institute (FSI), Oxford and
Ehrman (1995) reported that the SILL correlated with foreign language proficiency.
For metacognitive and cognitive strategies, the correlation was strong, r = .61. For
memory-related, compensatory, affective, and social strategies, the correlations were
more moderate.

SUMMARY AND DISCUSSION OF SILL RESULTS RELATED TO LANGUAGE PROFICIENCY

In short, language learning strategies do indeed make a significant difference in
language proficiency. With multiple regression analyses, SILL-gauged strategy use
explained 56% of the variance in EFL proficiency in a Japanese study, 51% of Kanji
proficiency in an Australian study, 46% of ESL proficiency in a South African study,
and 21% of EFL proficiency in a Taiwanese study. The average of these is 44%,
indicating that SILL-measured strategy use explains 44% of the variance in language
proficiency across these four studies. The multiple regression results indicate that (a)
118 REBECCA L. OXFORD

language proficiency can be predicted in both foreign language and second language
environments and (b) the SILL appears to be quite a useful predictor of language
proficiency.
It is instructive to compare these predictive strategy-and-language-proficiency
findings with results of a 1,650-student U.S. study involving a general learning strat-
egy questionnaire, the LASSI (mentioned earlier), and a general achievement test known
as the Preliminary Scholastic Aptitude Test or PSAT. The PSAT is a well-known aca-
demic achievement test designed for use with high school students who are preparing
for university admission. Overall, the SILL’s value in predicting language proficiency
(as shown by the ability to explain 44% of the variance in language proficiency across
four studies) is slightly greater than the combined value of the LASSI and the PSAT
for predicting high school grade point averages (as shown by the LASSI-PSAT combi-
nation’s ability to explain 38% of the variance in high school grades) (see Evertson,
Weinstein, Roska, Hanson & Laitusis, 1998).
Strong correlations emerged between language proficiency and SILL-assessed
strategy use among ESL learners in South Africa and among diplomats learning for-
eign languages in the U.S. (r = .73 and r = .61, respectively), although most of the
correlations between strategy use and language proficiency were somewhat lower for
EFL learners in five countries (r = .30 to .50). By way of comparison, the Evertson et
al. study (1998) in the U.S. showed correlations ranging only as high as r = .31 be-
tween sections of the LASSI and sections of the PSAT.
Some ANOVA and MANOVA studies showed linear relationships between strat-
egy use and language proficiency (i.e., greater strategy use frequency → greater
proficiency), while other such studies displayed curvilinear relationships. The mean-
ing of the curvilinear relationships is still open, but it might have something to do
with the automaticity with which some high-proficiency learners employ learning
strategies. Perhaps highly proficient learners in certain settings —especially second
language learning environments— might have less need for consciously deployed
strategies than less proficient learners (or learners in foreign language learning envi-
ronments, where resources and input in the language are not as prevalent). It might be
that highly proficient learners of a second language have, because of the necessarily
frequent use of the language and of language learning strategies, made their strategy
use so automatic that they no longer consciously employ these strategies.14
The above findings concerning SILL-assessed strategy use and language profi-
ciency are complemented by results of other studies involving the SILL. SILL factor
analytic results differ somewhat, although not completely, among various cultural
groups (e.g., People’s Republic of China, Taiwan, India, Japan, Egypt, U.S., Puerto
Rico) (Bedell, 1996; Hsiao, 1995; Oxford & Burry-Stock, 1995; Sheorey, 1999;
Takeuchi, 1997). Various researchers (e.g., Yang, 1992) have demonstrated that cul-
tural beliefs differ concerning the value of learner autonomy and learning strategies
in the development of language proficiency, so it is understandable that factor pat-
terns would differ somewhat across cultures.
Although there is insufficient space to explore other findings in detail here, many
SILL studies have shown additional results that might someday prove useful for un-
derstanding more about the relationships among learning strategy use, language pro-
ficiency, and learner autonomy/self-regulation. First, there are strong links between
 RELATIONSHIPS BETWEEN SECOND LANGUAGE LEARNING... 119

strategy use and motivation, and motivation is often related to language proficiency
and learner autonomy (see Oxford, 1996a and Schmidt, Boraie, & Kassabgy, 1996
for many sources that offer supporting evidence). Second, on SILL-derived instru-
ments, young children use some different strategies from older language learners, but
there is notable overlap (Gunning, 1997). Third, in at least a dozen studies in various
parts of the world, females have reported using language learning strategies, meas-
ured by the SILL and a variety of other instruments, significantly more frequently
than males (Zoubir-Shaw & Oxford, 1995), but these patterns are in certain Middle
Eastern cultures (Kaylani, 1996), suggesting that socialization might play a role. Fourth,
the SILL has been shown anecdotally to be a useful part of learner training/develop-
ment sessions, woven into regular classroom work (Ehrman & Oxford, 1989, 1990;
Oxford & Ehrman, 1995).
This section has presented information about the SILL and findings from an ar-
ray of SILL studies linking strategy use and language proficiency. We are now ready
to consider the many issues that remain to be explored.

WHAT REMAINS TO BE EXPLORED IN FUTURE RESEARCH

This section lists some key questions that need to be answered in future research
on language learning strategies, language proficiency, and learner autonomy/self-
regulation.

• Letting Teachers Know: How can we best help teachers understand the signifi-
cant influence that strategy use has on language proficiency —or, for that matter, on
academic performance in general? (See books by O’Malley & Chamot, 1990; Ox-
ford, 1990, 1996b; Wenden & Rubin, 1987; Pressley & Associates, 1990; Pressley
with McCormick, 1995; Weinstein et al., 1988; and others in language learning and/
or general educational psychology.)
• Assessing Language Learning Strategy Use: This article has shown the utility
of one form of language learning strategy assessment, the questionnaire. What about
other forms of strategy assessment? For what purposes and under what conditions
should they be employed? (For some initial ideas, see Cohen & Scott, 1996.)
• Helping Individuals Become Better Learners: What is the best way to teach
strategies and foster autonomy in a highly diverse, multilingual ESL classroom? In a
linguistically homogeneous foreign language classroom? In classrooms for other sub-
ject fields and with varied groups? How effective is strategy instruction? Does the
effectiveness of strategy instruction depend on learner motivation, institutional prac-
tices, cultural beliefs, familial beliefs, and other factors? Can or should an entire
language curriculum be based on learning strategies? (See Cohen, 1990; Cohen &
Weaver, 1998; Cohen, Weaver, & Yi, 1995; Dadour & Robbins, 1996; Nunan, 1997;
O’Malley & Chamot, 1990; Oxford, 1990, 1996b; Wenden & Rubin, 1987. Also con-
sider more general works by Pressley & Associates, 1990; Pressley with McCormick,
1995; Weinstein et al., 1988.)
• Making Sense of the Language Learning Environment: Why do students in sec-
ond and foreign language learning situations sometimes differ in their use of lan-
120 REBECCA L. OXFORD

guage learning strategies? How does the learning environment affect the frequency
of strategy use and choice of various kinds of strategies? Is motivation implicated in
these differences? (For some interesting thoughts on these matters, see Green & Ox-
ford, 1995; LoCastro, 1994.)
• Understanding the Role of Gender: Why are gender differences frequently re-
ported in language learning strategy research? What does this say, if anything, about
learner autonomy and self-regulation in males and females? What does it tell us, if
anything, about foreign and second language proficiency? (Gender role development
research would be very enlightening here; see Zoubir-Shaw & Oxford, 1995, for leads.)
• Comprehending the Influence of Age: What is the longitudinal, lifespan pro-
gression of strategy use and autonomy for individuals? How do strategy use and au-
tonomy relate to the general development of cognition and personality? (Erikson’s
psychosocial stage theory, described by Williams & Burden, 1997, and other lifespan
developmental theories might give hints. Long-term longitudinal studies of autonomy
and learning strategy use are sorely needed.)
• Revealing Cultural Effects: How much should we push learner autonomy and
strategy use if they are not a strong part of a culture’s belief system? How is the
individual’s zone of proximal development shaped by cultural influences? To what
extent can or should learners or teachers challenge anti-individualistic cultural be-
liefs? (Consider ideas about learner autonomy from Wenden, forthcoming, and Little,
1991, 1999; ZPD-related concepts from Vygotsky, 1978; and empirical research on
cultural belief systems from Hofstede, 1986.)

The challenge is here to help create more effective language learners who will be
able to use their new languages for actual communication. Three pieces of the puzzle
—learning strategy use, language proficiency, and learner autonomy or self-regula-
tion— must fit together closely and effectively in order for us to meet the challenge.
Therefore, we must continue to explore the remaining issues and questions raised
here concerning strategies, proficiency, and autonomy/self-regulation. The SILL, along
with other strategy assessment modes, can be of service in this quest.

Notes
1
Andrew Cohen (1997) distinguishes between strategies for language learning and strategies
for language use. However, Little (1999) and Oxford (1990) suggest that language learn-
ing and language use are not separable, because each occasion of language use is a stimu-
lus to further language learning, and each event of language learning is preparation for
language use.
2
In addition, language learning strategy use relates to choice of a non-technical academic
major and to gender (for details see Oxford, 1996b).
3
Other ways to identify the strategies people use to learn languages (and the frequency of use
of these strategies) include diaries written by the learner, dialogue journals that allow
written communication between learner and teacher, classroom discussions by groups of
learners, observations by the teacher, and (often videotaped) “think-alouds” in which a
learner mentions the strategies used while doing a particular language task. Each strategy
 RELATIONSHIPS BETWEEN SECOND LANGUAGE LEARNING... 121

assessment tool, including the SILL, has its own advantages and disadvantages and is
useful for specific research-related and instructional purposes (for details, see Cohen &
Scott, 1996; Nunan, 1995).
4
In a number of reliability studies (see Bedell, 1996; Ku, 19965; Oxford & Burry-Stock,
1995), the Cronbach alpha internal consistency index was .94-.98 for the 80-item SILL
when administered in English (or a translation) to language-homogeneous groups. Alpha
was .89-.90 for the 50-item version when administered in English to non-native English
speakers in groups containing many different mother tongues.
5
Various kinds of validity have been demonstrated for the SILL.
• Concurrent Validity: As expected, the SILL is significantly correlated with the Learning
and Study Strategy Inventory, the Modern Language Aptitude Test, the Learning Style
Profile, the Myers-Briggs Type Indicator, the Style Analysis Survey, the Affective Survey,
and other relevant tests (Dreyer & Oxford, 1996; Ehrman & Oxford, 1989, 1990; Oxford
& Ehrman, 1995).
• Content Validity: Two strategy experts independently matched SILL items with entries in
a comprehensive strategy taxonomy; resulting concordance of raters was 99% (Oxford,
1990).
• Social Desirability Response Bias: (a threat to validity): With the Marlowe-Crowne So-
cial Desirability Scale, the SILL showed no such bias in studies of EFL students (N=505,
Yang, 1992; N = 904, Ku, 1995).
6
Information ordinarily presented in multiple regression studies includes: the kind of regres-
sion used (forward selection, backward selection, stepwise); standard error; t for the null
hypothesis; significance of the overall regression model and for each predictor; multiple
R and R2 for the whole model; the percentage of variance in the dependent variable, say
language proficiency, explained by the predictors in the regression model; and the beta
weights of each predictor indicating its strength in the prediction.
7
The index r2 is the square of the correlation coefficient, r. Known as the coefficient of deter-
mination or the shared variance, r2 works in the same fashion for correlational analyses as
R2 does for multiple regression analyses. The coefficient of determination indicates how
much of the variance in one of the two correlates, say language proficiency, is explained
by the other correlate, such as the total SILL score. If r2 = .45, this means that 45% of the
variance is shared by the two correlates (which is the same as saying that the amount of
variance in language proficiency explained by learning strategy use, or vice versa, is 45%).
However, r2 is reported only infrequently in language learning studies.
8
Multiple R =.68 and R2 = .46. See the study itself for beta weights and other information.
9
Puerto Ricans normally speak Spanish during everyday communication and do not need to
know English to survive (hallmarks of an EFL setting); nevertheless, there is abundant
input in English, many opportunities to use English, and requirements for very long-term
studying of English (often viewed as signs of an ESL setting). This is definitely a “hybrid
situation”. Puerto Rico is classified here as an ESL learning environment for the sake of
convenience and because it does have some ESL attributes.
10
The ESLAT (ESL Achievement Test) is a TOEFL-like test created especially for Puerto Rico
by the TOEFL publisher.
11
Multiple R = .75 and R2 = .58. See the study itself for beta weights and other information.
12
Multiple R = .45 and R2 = .21. See the study itself for beta weights and other information.
13
Multiple R = .72, and R2 = .51. See the study itself for beta weights and other information.
122 REBECCA L. OXFORD

14
Some researchers state that if a strategy is used so often that it becomes automatic and
unconscious, then it is no longer a strategy but should be termed a process, procedure, or
procedural skill (O’Malley & Chamot, 1990).

Works Cited

Allwright, D., Autonomy in Language Pedagogy. CRILE Working Paper 6. Centre for
Research in Education, U of Lancaster, 1990.
Bandura, A. Self-efficacy: The Exercise of Control. New York: Freeman, 1997.
Bedell, D. “Cross-cultural Comparisons of Language Learning Strategies in the People’s
Republic of China and Other Countries.” Language Learning Strategies around the
World: Cross-cultural Perspectives. Ed. R. Oxford. Manoa: U of Hawaii P, 1996.
Berk, L.E. and A. Winsler, A., Scaffolding Children’s Learning: Vygotsky and Early
Childhood Education. Washington, DC: National Association for the Education
of Young Children, 1995.
Chamot, A.U., S. Barnhardt, P. El-Dinary and J. Robbins. “Methods for Teaching
Learning Strategies in the Foreign Language Classroom.” Language Learning
Strategies Around the World: Cross-cultural Perspectives. Ed. R. Oxford , Manoa:
U of Hawaii P, 1996.
— and J.M. O’Malley. “Implementing the Cognitive Academic Language Learning
Approach (CALLA).” Language Learning Strategies around the World: Cross-
cultural Perspectives. Ed. R. Oxford Manoa: U of Hawaii P, 1996.
Cohen, A.D., Language Learning: Insights for Learners, Teachers, and Researchers.
New York: Newbury/Harper Collins, 1990.
— “Language Learning Strategies and Language Use Strategies.” Lecture, Strate-
gies-Based Instruction Workshop, University of Minnesota, Minneapolis, USA.
1997.
— and K. Scott. “A Synthesis of Approaches to Assessing Language Learning Strat-
egies”. Language Learning Strategies Around the World: Cross-cultural Perspec-
tives. Ed. R. Oxford Manoa: U of Hawaii P, 1996.
— S.J. Weaver and T.Y. Yi. The Impact of Strategies-Based Instruction on Speaking a
Foreign Language. Research Report. Minneapolis, MN: National Foreign Lan-
guage Research Center, 1995.
— and S. J. Weaver. “Strategies-based Instruction for Second Language Learners”.
Learners and Language Learning. Anthology Series 39. Eds.W.A. Reyandya &
G.M. Jacobs Singapore: SEAMEO Regional Language Center, 1998. 1-25.
Cotterall, S. “Developing a Course Strategy for Learner Autonomy”. ELT Journal 49
(1995a): 219-227.
— “Readiness for Autonomy: Investigating Learner Beliefs”. System 23.2, (1995b):
195-205.
Dadour, E.S. and J. Robbins. “University-level Studies Using Strategy Instruction to Im-
prove Speaking Ability in Egypt and Japan.” Language Learning Strategies around
the World: Cross-cultural Perspectives. Ed. R. Oxford Manoa: U of Hawaii P, 1996.
Dickinson, L. Self-instruction in Language Learning. Cambridge, U.K.: Cambridge
UP, 1987.
 RELATIONSHIPS BETWEEN SECOND LANGUAGE LEARNING... 123

— Learner Autonomy 2: Learner Training for Language Learning. Dublin: Authentik,

1992.
Dreyer, C. and R. Oxford. “Learning Strategies and Other Predictors of ESL Profi-
ciency among Afrikaans-speakers in South Africa.” Language Learning Strate-
gies Around the World: Cross-cultural Perspectives. Ed. R. Oxford Manoa: U of
Hawaii P, 1996.
Ehrman, M. and R. Oxford. “Effects of Sex Differences, Career Choice, and Psy-
chological Type on Adults’ Language Learning Strategies.” Modern Language
Journal 73.1 (1989): 1-13.
— and R. Oxford. “Adult Language Learning Styles and Strategies in an Intensive
Training Setting.” Modern Language Journal 74 (1990): 311-26.
Everson, H.T., C.E. Weinstein, L. Roska, G.R. Hanson and V. Laitusis. “The Rela-
tionships among Measures of Students’ Academic Abilities and Strategic Learn-
ing Skills.” Paper presented at the symposium entitled “Issues in Strategic Learn-
ing in Higher Education,” the 24th International Congress of Applied Psychology,
San Francisco, California, 1998.
Green, J. “Language Learning Strategies of Puerto Rican University Students.” Paper
presented at the annual meeting of Puerto Rico Teachers of English to Speakers
of Other Languages, San Juan, Puerto Rico, 1991.
— and R. Oxford. “A Closer Look at Learning Strategies, L2 Proficiency, and Gen-
der.” TESOL Quarterly 29 (1995): 261-97.
Gunning, P. The Learning Strategies of Beginning ESL Learners at the Primary Level.
Unpublished Master’s Thesis, Concordia U, Montreal, 1997.
Hofstede, G, “Cultural Differences in Learning and Teaching.” International Journal
of Intercultural Relations 10 (1986): 301-20.
Holec, H. Autonomy and Foreign Language Learning. Oxford, U.K.: Pergamon. 1981.
— “On Autonomy: Some Elementary Concepts.” Discourse and Learning. Ed. P.
Riley New York: Longman, 1985.
Hsaio, T-Y. A Factor-analytic and Regressional Study of Language Learning Strate-
gies Used by University Students of Chinese, French, German, and Spanish.
Unpublished dissertation, U of Texas, Austin, 1995.
Iran-Nejad, A. “Active and Dynamic Self-regulation of Learning Processes.” Review
of Educational Research 60 (1990): 549-72.
Kato, F. “Results of an Australian Study of Strategy Use in Learning Japanese Kanji
Characters.” Unpublished manuscript, University of Sydney, 1996.
Kaylani, C. “The Influence of Gender and Motivation on EFL Learning Strategy Use
in Jordan.” Language Learning Strategies around the World: Cross-cultural Per-
spectives. Ed. R. Oxford Manoa: U of Hawaii P, 1996.
Ku, P.N. Strategies Associated with Proficiency and Predictors of Strategy Choice: A
Study of Language Learning Strategies of EFL Students at Three Educational
Levels in Taiwan. Unpublished Doctoral Dissertation, Indiana U, Bloomington,
1995.
Little, D. Learner Autonomy 1: Definitions, Issues, and Problems. Dublin: Authentik,
1991.
— “Learning as Dialogue: The Dependence of Learner Autonomy on Teacher Au-
tonomy.” System 23.2 (1995): 175-81.
124 REBECCA L. OXFORD

— “Learner Autonomy and Human Interdependence: Some Theoretical and Practical

Consequences of a Social-interactive View of Cognition, Learning and Language.”
Paper presented at the symposium “Learner Autonomy, Teacher Autonomy, Fu-
ture Directions,” held at the University of Nottingham 20-21 April 1998.
Littlewood, W. “ ‘Autonomy’: An Anatomy and a Framework.” System 24.4 (1996):
427-35.
LoCastro, V. “Learning Strategies and Learning Environments.” TESOL Quarterly,
28.2 (1994): 409-14.
Naiman, N., M. Fröhlich, H.H. Stern, and A. Todesco. The Good Language Learner.
Toronto: Ontario Institute for Studies in Education. 1975.
Nunan, D. Language Teaching Methodology. London: Prentice Hall, 1991.
— “Self-assessment as a Tool for Learning.” Bringing About Change in Language
Education. Eds. D. Nunan, R. Berry and V. Berry. Hong Kong: Department of
Curriculum and Instruction, U of Hong Kong, 1995.
— “Does Learner Strategy Training Make a Difference?” Lenguas Modernas 24
(1997): 123-42.
— Second Language Teaching and Learning. Boston: Heinle & Heinle, 1999.
O’Malley, J.M. and A.U. Chamot. Learning Strategies in Second Language Acquisi-
tion. Cambridge, UK: Cambridge UP, 1990.
— A.U. Chamot, G. Stewner-Manzanares, L. Kupper, and R.P. Russo. “Learning Strat-
egies Used by Beginning and Intermediate ESL Students.” Language Learning
25 (1985): 21-46.
Oxford, R.L. Development and Psychometric Testing of the Strategy Inventory for
Language Learning. ARI Technical Report No. 728. Alexandria, Virginia: U.S.
Army Research Institute for the Behavioral and Social Sciences, 1986.
— Language Learning Strategies: What Every Teacher Should Know. Boston: Heinle
& Heinle, 1990.
— ed. Language Learning Motivation: Pathways to the New Century. Manoa: U of
Hawaii P, 1996a.
— ed. Language Learning Strategies Around the World: Cross-cultural Perspectives.
Manoa: U of Hawaii P, 1996b.
— and J. Burry-Stock. “Assessing the Use of Language Learning Strategies World-
wide with the ESL/EFL Version of the Strategy Inventory for Language Learn-
ing.” System 23.1 (1995): 1-23.
— C. Judd, and J. Giesen, “Relationships among Learning Strategies, Learning
Styles, EFL Proficiency, and Academic Performance among Secondary School
Students in Turkey.” Unpublished manuscript, University of Alabama, Tusca-
loosa, 1998.
— Y.Y. Park-Oh, S. Ito and M. Sumrall. “Learning a Language by Satellite Televi-
sion: What Influences Student Achievement?” System 21 (1993): 31-48.
— and M. E. Ehrman “Adults’ Language Learning Strategies in an Intensive Foreign
Language Program in the United States.” System 23 (1995): 359-86.
— and Nyikos, M. “Variables Affecting Choice of Language Learning Strategies by
University Students”. Modern Language Journal 73 (1989): 291-300.
Park, G., “Language Learning Strategies: Why Do Adults Need Them?” Unpublished
manuscript, University of Texas, Austin, 1994.
 RELATIONSHIPS BETWEEN SECOND LANGUAGE LEARNING... 125

Park-Oh, Y. Self-regulated Strategy Training in Second Language Reading. Unpub-

lished doctoral dissertation, U of Alabama, Tuscaloosa, 1994.
Phillips, V. English as a Second Language Learner Strategies of Adult Asian Students
Using the Strategy Inventory for Language Learning. Unpublished Doctoral Dis-
sertation, U of San Francisco, San Francisco, 1990.
— “A Look at Learner Strategy Use and ESL Proficiency.” CATESOL Journal (1991):
56-67.
Prawat, R.S. “Current Self-regulation Views of Learning and Motivation Viewed
Through a Deweyan Lens: The Problems with Dualism.” American Educational
Research Journal 35.2 (1998): 199-224.
Pressley, M. and Associates. Cognitive Strategy Instruction That Really Improves Chil-
dren’s Academic Performance. Cambridge, MA: Brookline Books, 1990.
— C.B. McCormick. Advanced Educational Psychology for Educators, Researchers,
and Policymakers. New York: HarperCollins, 1995.
Richards, J.C. and C. Lockhart. Reflective Teaching in Second Language Classrooms.
Cambridge, UK: Cambridge UP, 1996.
Rubin, J. “What the ‘Good Language Learner’ Can Teach Us.” TESOL Quarterly 9
(1975): 41-51.
Scarcella, R.C. and R.L. Oxford. The Tapestry of Language Learning: The Individual
in the Communicative Classroom. Boston: Heinle & Heinle, 1992.
Schmidt, R., D. Boraie, and O. Kassabgy, “Foreign Language Motivation: Internal
Structure and External Connections.” Language Learning Motivation: Pathways
to the New Century. Ed. R. Oxford Manoa: U of Hawaii P, 1996.
Sheorey, R. “Learning Strategies of Indian College Students.” Research Compen-
dium: Language Learning Strategies in the Context of Autonomy. Ed. R. Oxford
Proceedings of the Invited Research Symposium on Language Learning Strate-
gies, Teachers College, Columbia University, New York, 1999.
Takeuchi, O. “Language Learning Strategies and Their Relationship to Achievement
in English as a Foreign Language.” Language Laboratory 30 (1993): 17-34.
— “Toward the Establishment of Country-by-country SILL (Strategy Inventory for
Language Learning) Norms.” Paper presented at the 37th Annual Conference of
the Language Laboratory Association of Japan, Kyoto, Japan, 1997.
Tudor, I. Learner-centredness as Language Education. Cambridge: Cambridge UP,
1996.
Vygotsky, L.S. Mind in society: The Development of Higher Psychological Proc-
esses. Cambridge, MA: MIT Press, 1978.
— Thought and Language. Cambridge, MA: MIT Press, 1986.
Watanabe, Y. External Variables Affecting Language Learning Strategies of Japanese
EFL Learners: Effects of Entrance Examinations, Years Spent at College/Univer-
sity, and Staying Overseas. Unpublished Master’s Thesis, Lancaster University,
Lancaster, 1990.
Wenden, A. “Promoting the Development of Autonomous Language Learners: Ex-
panding the Role of Foreign Language Teaching for the 21st Century.” Northeast
Conference Reports. Ed. M.A. Kassen. Middlebury, VT: Northeast Conference
on Foreign Language Teaching. Forthcoming.
— Learner Strategies for Learner Autonomy. Englewood Cliffs, NJ: Prentice Hall, 1991.
126 REBECCA L. OXFORD

— and J. Rubin, eds. Learner Strategies in Language Learning. Englewood Cliffs,

NJ: Prentice Hall, 1987
Weinstein, C., E.T. Goetz, and P.A. Alexander, eds., Learning and Study Strategies:
Issues in Assessment, Instruction, and Evaluation. San Diego: Academic Press,
1988.
Williams, M. and R.L. Burden. Psychology for Language Teachers. Cambridge, U.K.:
Cambridge UP, 1997.
Zimmerman, B.J. and M.M. Pons. “Development of a Structured Interview for As-
sessing Student Use of Self-regulated Learning Strategies.” American Educa-
tional Research Journal 23 (1986): 614-628.
Zoubir-Shaw, S. and R. Oxford. “Gender Differences in Language Learning Strategy
Use in University-level Introductory French Classes: A Pilot Study Employing a
Strategy Questionnaire.” Faces in the Crowd: Individual Learners in Multisection
Programs. Ed. C.A. Klee. Boston: Heinle & Heinle, 1995.
Yang, N-D. Second Language Learners’ Beliefs about Language Learning and Their
Use of Learning Strategies. Unpublished Dissertation, U of Texas, Austin, 1992.