You are on page 1of 21

Literate Actions, Reading Attitudes,

and Reading Achievement:


Interconnections Across Languages
for Adolescent Learners of English in
Korea
JUHEE LEE DIANE L. SCHALLERT
University of Texas University of Texas
Foreign Language Education Department of Educational Psychology
SZB 528–South D5700 SZB 504 D5800
Austin, Texas 78712 Austin, Texas 78712
Email: juheelee.carpediem@gmail.com Email: dschallert@austin.utexas.edu

Using the three components of attitude in Fishbein and Ajzen’s (1975) Reasoned Action Theory—
cognition (personal and evaluative beliefs), affect (feelings, emotions), and conation (behavioral
intentions)—the present study investigated relations among L1 and L2 reading attitudes, the relative
contributions of reading attitude and language proficiency to reading achievement, and various factors
that shape L2 reading attitude. Participants were 289 middle school students studying English in Korea.
Their language proficiency, reading achievement, and reading attitudes in L1 and L2 were measured.
Results indicated that measures of L1 reading attitude factors correlated with scores on the
corresponding L2 reading attitude factor. Despite the correlations between cross-linguistic reading
attitudes, reading attitudes in L1 and L2 exhibited different tendencies. Both cognitive attitude and
language proficiency predicted L1 reading comprehension, whereas language proficiency was the sole
contributor to L2 reading achievement. Reading frequency significantly contributed to reading attitude
and achievement scores in L1, but such patterns were not as clearly apparent in L2. Also, proficiency in
both L1 and L2 was associated with more positive reading attitudes and fewer negative feelings about
reading in the corresponding language. Finally, several factors predicted L2 reading attitude: access to
books, length of private instruction, teachers’ encouragement, and participant gender.
Keywords: reading attitudes; reading achievement; associations between L1 and L2 reading; reading
frequency; adolescent readers

SUCCESSFUL READING IN A SECOND OR factors, and metacognitive abilities on the L2


foreign language calls upon various abilities and reading process (e.g., Alptekin & Ercetin, 2010;
is influenced by several factors in the reader’s first Bernhardt & Kamil, 1995; Jimenez, Garcia, &
(L1) as well as second language (L2). In order to Pearson, 1996; Lee & Schallert, 1997; Tsai, Ernst,
explain L2 reading development, researchers & Talley, 2010). Although they acknowledge the
have investigated the effects of readers’ linguistic critical importance of cognitive factors and
knowledge of their two languages, cognitive linguistic knowledge in L2 reading development,
researchers have come to recognize the role that
affective variables play in engaging individuals in
The Modern Language Journal, 98, 2, (2014) L2 reading (Bernhardt, 2011; Grabe, 2009). Even
DOI: 10.1111/j.1540-4781.2014.12088.x fluent readers may develop negative attitudes
0026-7902/14/553–573 $1.50/0 toward reading and choose not to read when
© 2014 The Modern Language Journal
alternative tasks are available, a condition that
554 The Modern Language Journal 98 (2014)

McKenna, Kear, and Ellsworth (1995), in its Along with addressing definitional issues, many
extreme form, called aliteracy. Despite a growing researchers have offered different conceptualiza-
appreciation of the importance of reading atti- tions of what constitutes a reading attitude (Ajzen,
tudes, there has been surprisingly little empirical 1989; Ajzen & Fishbein, 1980; Mathewson, 1994;
research establishing the connection between McGuire, 1969; McKenna, 1994). Despite some
L2 reading attitude and L2 reading achievement. slight differences in approach, the general con-
In the few extant studies, critical components of sensus has converged on three constituents of
L2 reading attitude were not included in the reading attitude: cognitive (evaluative beliefs),
same design, making it difficult to examine the affective (feelings and emotions), and conative
full multidimensional nature of L2 reading (behavioral intentions) factors. The cognitive
attitudes. Also, most studies have involved adult component refers to the beliefs, opinions, or
L2 readers, who can be said to have fully evaluations that an individual has about reading,
developed attitudes toward reading, at least in and centers on the perceived instrumentality of
their first language. The purpose of this study was reading, such as linguistic, intellectual, or practi-
to investigate how reading attitudes of adolescent cal values associated with reading (e.g., “Reading
learners of English toward their native and target English is useful to get a good job in the future”).
languages are related to and affect each other, These instrumental values tend to develop from
and what other factors play a role in shaping L2 retrospective feelings or evaluations of prospec-
reading attitudes and influencing L2 reading tive outcomes from reading (McGuire, 1969). The
achievement. affective component is linked to feelings or emo-
tions that an individual experiences in connec-
tion with reading (McKenna et al., 1995). An
A THEORETICAL FRAMEWORK FOR example of a statement reflecting the affective
READING ATTITUDE CONSTRUCTS component is, “I feel anxious if I do not know all
the words in reading passages.” Such feelings or
A review of the literature reveals that defining emotions are likely to be influenced by surround-
reading attitude and grasping its multidimensional ing circumstances, such as family, friends, or the
nature have not been easy. Alexander and Filler learning environment (Paimin, Hadgraft, &
(1976) defined the term reading attitude as “a Alias, 2011). The last construct, conation, is
system of feelings related to reading which causes defined as the intentions one has to act toward
the learner to approach or avoid a reading a particular object (in this case to reading), “the
situation” (p. 1). Similarly, Smith (1990) wrote way in which a person (. . .) goes about acting” on
that reading attitude is “a state of mind, the goals they have adopted (Gerdes & Stromwall,
accompanied by feelings and emotions, that 2008, p. 236). For instance, an individual may
makes reading more or less probable” (p. 215). have a generally positive view of learning a second
Eagly and Chaiken (1993) stated that “attitude is a language but, lacking conative attitude, never
psychological tendency that is expressed by actually engage in any action that would make
evaluating a particular entity with some degree learning the language possible. In other words,
of favor or disfavor” (p. 1). These definitions conation entails showing through one’s actions
suggest that reading attitudes shape an individual’s how one feels about reading (e.g., “I go to a library
intentions, which are themselves connected to to borrow or read English books”).
actual reading behavior. An individual’s reading These three aspects of attitude were integral to
attitudes may be characterized at an overall level the tripartite approach represented by Fishbein
and simultaneously show differentiation toward and Ajzein’s (1975) Theory of Reasoned Action
particular genres or topics (McKenna et al., (TRA) framework, a psychological theory devel-
1995). Whether reading attitudes range from oped to explain the general connection among
general to specific, it is assumed that once they beliefs, attitudes, and actions. In the theory,
have become established, they show much cognitive beliefs are said to influence affective
constancy and less volatility, influencing reading attitudes that are then associated with behavioral
behavior either positively or negatively (McGuire, intentions (the conative component), themselves
1969). Because reading attitudes tend to develop expressed in actual behaviors. A reciprocal
from previous experiences at home and in school influence feedback loop exists such that actual
settings (Smith, 1990), these learned dispositions behaviors influence beliefs. For our study, the
become enduring and persistent regardless of TRA framework encouraged us to measure all
specific situations, and come to make up the three reading attitude constituents and to look for
individual’s cultural capital (Bourdieu, 1984). relationships among the three components.
Juhee Lee and Diane L. Schallert 555

L1 READING ATTITUDE AND ITS RELATION had more positive reading attitudes than male
TO READING PERFORMANCE students at all grades. Other studies have consis-
tently reported gender differences favoring girls
With our interest in the role of L2 reading over boys in their attitudes toward reading (Logan
attitudes in L2 reading performance, we turned & Johnston, 2009; Swalander & Taube, 2007;
first to the literature that has examined whether Worrell, Roth, & Gabelko, 2007).
and how attitudes toward reading in the native More recently, a more nuanced picture of
language are associated with reading perfor- gender differences in reading attitudes seems to
mance, as well as how reading attitudes are be emerging. For example, in a study examining
formed. reading attitudes of adolescents toward academic
In a review of the research on the connection and recreational reading in print and digital
between L1 reading attitude and reading achieve- formats, McKenna et al. (2012) reported that
ment, Roberts and Wilson (2006) reported on female adolescents had more positive attitudes
factors that give birth to L1 reading attitudes. Day toward academic reading but male students had
and Bamford (1998) claimed that attitudes are more positive attitudes toward recreational read-
formed from previous experiences related to ing, particularly in digital forms. Given that
reading, messages encountered in educational attitudes are formed in the crucible of cultural
settings, and cultural beliefs about reading. experiences, it is likely that gender differences in
Particularly, repeated success or failure with attitudes toward reading may differ depending on
reading significantly affects the development of the particular sociocultural context of the stu-
positive or negative attitudes toward reading dents being examined, even though the more
(Swanson, 1985). McKenna et al. (1995) found positive attitudes and higher performance of girls
that positive attitudes toward reading tend to and women over boys and men have been
change into negative ones as American children replicated surprisingly frequently across several
grow older, and this decline appears more countries (e.g., Logan & Johnston, 2009; Lummis
precipitous among struggling than successful & Stevenson, 1990; Sani & Zain, 2011). Our study,
readers. This phenomenon has been explained situated in a Korean middle school, afforded us an
as a component of what Stanovich (1986) called opportunity to take a crosscultural look at the
the Matthew effect, indicating that successful read- relation between L1 attitudes and L1 reading
ers at earlier stages of schooling tend to read more achievement, but also at a less frequently exam-
and continue to succeed in reading, whereas ined variable in the L2 reading literature: gender
those who struggle with reading in the early differences.
grades are not only likely to fall further behind in Important to our study was the literature on the
reading in later school years but are likely to avoid relationship between reading attitudes and read-
reading, choosing other activities instead (Cun- ing achievement, a literature that has produced
ningham & Stanovich, 1997). Therefore, through inconsistent results for L1 reading. Martinez,
past experiences with reading, children develop Aricak, and Jewell (2008) examined the associa-
not only emotions and beliefs about reading but tions among reading ability, reading attitude, and
also self-efficacy beliefs, and these aspects influ- reading achievement for 76 American elementary
ence the formation of L1 reading attitude. school students, through the Elementary Reading
In addition, from a socio–constructivist per- Attitude Survey (McKenna & Kear, 1990, 1999).
spective, individuals’ mental functioning cannot The results indicated that 4th grade reading
be separated from the social and cultural settings ability and reading attitudes were significant
in which the individual is situated, but is heavily predictors of 5th grade reading achievement,
affected by practices, norms, cultures, and per- together accounting for more than 50% of the
ceptions that the society values and expects variance, with ability the stronger predictor but
(Wertsch, 1991). In this regard, in explaining with attitude making up at least one third of the
why American girls and women are better at contributing explanation of reading achievement
reading and are more likely to report engaging in scores. Lazarus and Callahan (2000) used the
reading than boys and men, educational research- same survey instrument to measure the reading
ers have pointed to the social and cultural attitudes of 522 children diagnosed with learning
expectations individuals construct associated disabilities, who showed comparable or even
with their gendered selves and how these expect- more favorable reading attitudes than did non-
ations may have an impact on their attitudes diagnosed peers at low and average levels of
toward reading as an activity. McKenna and his reading achievement, suggesting that reading
colleagues (1995) reported that female students attitude may not depend on reading ability. Also,
556 The Modern Language Journal 98 (2014)

using longitudinal covariance structural model- assessment about their L2 reading ability and
ing, Kush, Watkins, and Brookhart (2005) re- their reported enjoyment of L2 reading were
ported that some poor readers with early positive significantly connected to reading comprehen-
reading attitudes held on to their positive sion, suggesting that nonlinguistic variables could
attitudes despite the challenges they experienced play a critical role in L2 reading performance.
with reading. Finally, in light of an early review by However, the study included only two nonlinguis-
Wigfield and Asher (1984) of studies that tic variables—self-assessment and enjoyment—
concluded that more able readers have more limiting the inferences that could be drawn about
positive reading attitudes than less able readers, it the role of L2 attitudes in reading performance.
is interesting to consider the broadly accepted Mills, Pajares, and Herron (2006) investigated the
finding that reading attitudes generally drop as contributions of students’ self-efficacy and anxiety
students advance in school (McKenna to reading achievement for 95 university-level
et al., 1995), implying that becoming a more English speakers enrolled in French classes for
practiced reader does not necessarily bring with it two semesters. Results indicated that reading self-
more positive attitudes toward reading. Thus, the efficacy positively predicted reading proficiency,
literature on L1 reading attitude and its connec- whereas reading anxiety was not significantly
tion to L1 reading achievement remains mixed. associated with reading performance. In a quali-
Regardless of the complexities that exist in the tative study of Latina bilingual college students,
L1 field per se, it would seem reasonable to expect Kamhi–Stein (2003) reported that the two women
a connection between attitudes toward reading in who held positive views of their first language and
a native language and attitudes toward reading in who viewed reading as meaning-construction
a second language. In fact, Day and Bamford allowed themselves to use their L1 when reading
(1998), in a list of possible factors influencing the in English whereas the two women who had more
acquisition and development of L2 reading negative views of the utility of Spanish to their
attitude, include L1 reading attitudes as a major academic success encountered more difficulties
source of influence. Readers with positive atti- as they read. Thus, although affective variables
tudes toward L1 reading might be expected to have been connected to L2 reading achievement,
hold favorable attitudes toward L2 reading, to the particular construct of L2 reading attitudes
which we turn next. was not the focus in these studies.
Another line of research has aimed to find
ROOTS OF L2 READING ATTITUDE AND contributing factors to L2 reading attitude. Based
RELATION TO READING PERFORMANCE on the analysis of first language reading attitudes
from Mathewson’s (1994) reading attitude model
Actual empirical studies focused on L2 reading and McKenna’s (1994) later development of that
attitudes are much more sparse. A few studies model, Day and Bamford (1998) identified four
examined affective variables associated with sources related to the development of L2 reading
reading and their connection to reading perfor- attitudes: a) L1 reading attitudes, b) positive or
mance; however, L2 reading attitude per se was negative experiences with learning to read other
not the focus in these studies. A second group of second languages (if any), c) general attitudes
studies has looked at contributors to L2 reading toward the target language, culture, and people,
attitude without measuring the full complement and d) the L2 classroom environment. Although
of factors making up attitudes. L2 proficiency may be necessary for fluent L2
In terms of the connection of affective variables reading, their model does not consider language
and L2 reading performance, a relevant study is proficiency as a contributor to the formation of L2
that of Saito, Horwitz, and Garza (1999) who reading attitudes.
reported that foreign language reading anxiety Without explicitly citing Day and Bamford’s
could be differentiated from general foreign (1998) contributors to L2 reading attitude,
language anxiety and was found to be negatively Camiciottoli (2001) included many factors in
correlated with reading performance. Their study her survey of 182 Italian adult learners of English
also demonstrated that students’ self-perceived that echo the Day and Bamford list: length of
difficulties in L2 reading were linked to levels of English (L2) study and private instruction, access
reading anxiety such that those who perceived to English books, L1 and L2 reading frequency,
reading in the second language as difficult tended experiences abroad, encouragement to read in
to have higher levels of reading anxiety than those English for pleasure, and self-perception of L2
who perceived it as less difficult. Similarly, reading ability. Results indicated that the amount
Brantmeier (2005) discovered that learners’ self- of L1 reading and experiences in the target
Juhee Lee and Diane L. Schallert 557

language culture were the most significant study adopted a tripartite approach to examine
predictors of L2 reading amount and attitude. both L1 and L2 reading attitudes, the impact of
Perhaps surprisingly, more years of L2 instruction L1 reading attitude on L2 reading attitude, and
were associated with a decrease in students’ self- ultimately, their connection to L2 reading
confidence about reading ability and to less achievement. We also included contributing
engagement in L2 reading, suggesting that Day factors to L2 reading that have been identified
and Bamford’s fourth factor, the quality of L2 in previous studies, such as the length of L2
instruction, is important in promoting L2 reading instruction, L2 proficiency, gender, and L2
attitudes. In sum, Camiciottoli found that L1 reading amount, to allow for a comprehensive
reading amount, experiences abroad, and the look at the construct of reading attitude. In
instructional setting predicted L2 reading addition, unlike earlier studies that have focused
attitudes. on adult L2 readers, our participants were
From what we could determine, Yamashita adolescent EFL students who might be described
(2004, 2007) is the only L2 researcher who has as still in the process of forming their L1 and L2
used Mathewson (1994) and McKenna’s (1994) reading attitudes. Thus, our study stood to
conceptualization of reading attitude to describe contribute to an understanding of attitude
the relationship between L1 and L2 reading development as well as to the relationship
attitudes. Leaving aside the conative component between attitude and achievement within and
because of difficulties in measuring it in the across languages. We organized our research
context of her study, she measured the subcom- questions to address three issues.
ponents of cognitive attitude (including intellec- With regard to predicting L2 reading attitudes
tual, practical, and linguistic values) and affective our research questions are:
attitude (including comfort and anxiety). Assum-
ing that affective variables developed in the native RQ.1.a. What are the relations between EFL
language may transfer to the L2 reading process, Korean adolescent students’ L1 and L2
Yamashita found that, although L1 reading reading attitudes?
attitudes differed from L2 reading attitudes RQ1.b. Besides L1 reading attitudes, what
among EFL adult learners, there was a significant additional factors predict L2 reading
attitudes?
connection between L1 and L2 reading attitudes,
a connection that remained strong for all learners
regardless of L2 proficiency. In these two studies, With regard to predicting reading achievement
the conative dimension of attitudes was left out, a our research questions are:
component that should be included in order to
provide a full picture of L2 reading attitude RQ.2.a. How are reading attitudes in L1 and L2
because it is related to what drives a student to associated with reading achievement in L1
engage in reading behavior. and L2, respectively?
Finally, gender is another factor that may be RQ2.b. What is the relative contribution of L1
associated with L2 reading attitude. Although and L2 reading attitudes and L2 language
proficiency to L1 and L2 reading
previous studies indicated that girls had more
achievement?
positive L2 reading attitudes, the reading attitude
surveys simply measured preference for reading
(Camiciottoli, 2001) or foreign language anxiety Finally, with regard to the role of reading
(Abu–Rabia, 2004) and failed to capture all frequency, our combined research questions are:
dimensions of L2 reading attitude. Thus, we
included gender in our study to test whether it RQ.3. What differences arise in L1 (L2) reading
would be a significant contributor to reading attitude and achievement among groups
attitude. of readers representing different levels of
L1 (L2) reading frequency, with L1 (L2)
language proficiency as a covariate?
CURRENT RESEARCH QUESTIONS

Earlier studies on L1 and L2 reading attitude


have identified three components of reading METHOD
attitude—cognition, affect, and conation—but these Participants
three dimensions have not been included in tests
of the connection between L2 reading attitudes Participants were 289 students (162 boys, 127
and L2 reading achievement. Thus, the present girls) learning English as a foreign language in
558 The Modern Language Journal 98 (2014)

Korea. They were 13 or 14 years old and attending tests, an English reading comprehension test, a
the second year of middle school (equivalent Korean language proficiency test, and a Korean
to U.S. 8th grade). Initially, 310 students were reading comprehension test. Table 1 displays
tested, but 21 data records were eliminated descriptive statistics of the tests, including mean
because they were incomplete. scores, standard deviations, and skewness and
Most elementary and middle schools in Korea kurtosis indices.
are required to follow a standardized curriculum Students’ English proficiency was measured by
including fixed lesson hours and textbooks. Our two tests. The first test, a 30-item standardized
participants had been receiving formal English English test composed of listening and reading
instruction for about two hours weekly from the comprehension multiple-choice questions, was a
3rd to 6th grade, with English lessons focused on component of the overall diagnostic assessment
basic communication skills through skits and undertaken by the educational district office at the
songs. Comprehensive English instruction, in- beginning of each academic year, to which we had
cluding reading, speaking, listening, writing, and access in the form of summative English scores
grammar, starts with the first middle school year from the school. The second test consisted of 70
(equivalent to U.S. 7th grade) with three or four vocabulary and grammar multiple-choice ques-
hours of lessons per week. Despite the relatively tions that we selected from ESL/EFL testing
similar English education across Korea, students’ websites specific to the proficiency level of the
English proficiency is likely to vary due to variable target participants. These 70 questions were
but pervasive access to private English instruction. reviewed by three English teachers, including
The majority of the participants (92.4%) had had one who was teaching the participants, for
some private instruction, and 53 students (18.3%) appropriateness. Scores on each test were con-
had received more than seven years of English verted to percent correct, and the resulting two
lessons when data were collected. Only 22 of 289 percentages averaged together to represent an
students had not experienced any extracurricular individual’s English proficiency score. In order to
English instruction. However, nearly all students address the reliability of the measures, we
(96.5%) reported either no experience of visiting calculated correlations between the overall mean
English speaking countries or visits that were for scores on the proficiency measures and the
less than a month, and only one student had lived students’ English achievement scores from the
in an English dominant culture for more than a previous semester that included measures of
year. This student and those who had been abroad reading, listening, speaking, writing, grammar,
for more than one month but less than a year and vocabulary. The high correlation (r ¼ .89)
(3.1%) were included in the data analysis because indicated that our measures were sufficiently
their data did not change results. Responses to the reliable for research purposes as measures of
demographic questionnaire, as well as the actual students’ English proficiency.
questions asked, are presented in tabular form in To measure students’ English reading compre-
the supplementary information available in the hension, we constructed a 30-item test by taking
online version of this article. multiple-choice questions from two types of
national level standardized English tests—high
Materials school entrance certificate tests for self-study
students, and high school entrance tests—admin-
Background Questionnaire. The questionnaire
istered for the past six to seven years by the Korea
asked participants about their background infor-
Institute for Curriculum and Evaluation, the
mation. Previous studies had identified several
office that develops and oversees national-level
factors that may be connected to reading attitudes
educational assessment in Korea. These standard-
and reading achievement (e.g., Camiciottoli, 2001;
ized English tests were chosen because they had
Day & Bamford, 1998), and we included nine items
been constructed and reviewed by assessment
in the questionnaire to measure gender, length of
experts and were assumed to be valid. In order to
formal and private English instruction, experience
ascertain whether the test items were appropriate
staying in English-speaking countries, experiences
for the target population in terms of test-taking
learning other foreign languages, access to English-
time and difficulty levels, this test was reviewed by
language books, teacher’s encouragement to read
two English teachers and then pilot tested with 27
books in English, and frequency of pleasure
middle school students who were at similar
reading in L1 and L2.
English proficiency levels as the participants.
Language Tests. The language measures in- Through a series of review procedures, 30
cluded five tests: two English language proficiency questions were finally selected, and scores were
Juhee Lee and Diane L. Schallert 559
TABLE 1
Descriptive Statistics for the Language Tests (N ¼ 289)

Skewness Kurtosis
Tests Items Min–Max Mean SD (SE) (SE)
Korean proficiency 30 12–100 63.1 20.5 .51 (.14) .58 (.29)
Korean reading 30 16–100 67.5 17.8 .68 (.14) .140 (.29)
English proficiency 1 30 12–96.5 66.5 23.9 .54 (.14) .85 (.29)
English proficiency 2 70 16–86 34.9 11.5 1.22 (.14) 2.09 (.29)
Combineda 1 & 2 100 15–91 50.7 16.3 .09 (.14) .81 (.29)
English reading 30 13–100 62.4 18.7 .38 (.14) .39 (.29)
Note. Min: Minimum score; Max: Maximum score; SD: Standard deviation; SE: Standard error;
a
the combined score of the English proficiency tests was used in analyses.

converted to percentage equivalents. The strong reading attitudes and 35 items focused on Korean
correlation (r ¼ .75) between scores on this test reading attitudes, with all items written in Korean.
and previous semester English achievement For each item, students rated the degree to which
demonstrated the reliability of the reading they endorsed the statement on a 5-point scale,
comprehension test. with 1 ¼ “Not at all true of me” and 5 ¼ “
Third, the students’ language proficiency in Completely true of me.” In order to make the
their first language and their knowledge of comparison of L1 to L2 reading attitude possible,
Korean grammar, vocabulary, and reading com- we used the same items on each survey modified
prehension was assessed using a standardized only to indicate whether English or Korean
measure provided by the district educational reading was being assessed. Thus, students
office. This test was also a part of overall diagnostic responded to items like “I feel tired when reading
evaluation administered by the educational office English books” or “I feel tired when reading
in the region. Finally, Korean reading compre- Korean books.” Adopted or adjusted from previ-
hension was measured through 30 multiple- ous L1 and L2 reading attitude questionnaires,
choice items from the Test of Proficiency in items addressed Mathewson’s (1994) three com-
Korean made by the National Institute for ponents of reading attitude: feelings (affective),
International Education in Korea. This test is a evaluative beliefs (cognitive), and action readiness
national and international level test used to assess (conative). Some affective items were adapted
Korean language proficiency of learners of from Yamashita (2004, 2007), with several addi-
Korean as a foreign language. We chose this test tional items added such as comfort, anxiety,
because the types of reading comprehension enjoyment, and avoidance, to provide a comprehen-
questions were parallel to the reading compre- sive measure of the affective component of attitude
hension items in English, requiring less content (Stokmans, 1999; Yamashita, 2004, 2007). In terms
knowledge of the passages but measuring various of the cognitive domain, we again borrowed items
aspects of reading comprehension ability. The from Yamashita (2007) measuring intellectual,
selected items were reviewed by an experienced practical, and linguistic attitudes, as well as adapt-
Korean language teacher for suitability for our ing some self-assessment items from Brantmeier
study participants. Again, to verify reliability of (2005) to measure Mathewson’s evaluative beliefs
these two measures, we calculated correlations dimension. Finally, we created conative items
between the scores of each test and the partic- guided by studies of L1 conative attitude for
ipants’ Korean language achievement scores reading (Tait–McCutcheon, 2008; van Schooten,
assessed in the previous semester at school. The de Glopper, & Stoel, 2004). Items in these two
strong correlations suggested that the Korean sections of the questionnaire were factor analyzed,
proficiency test (r ¼ .81) and the Korean reading and we report the factor analysis results as well as
comprehension test (r ¼ .74) were consistently reliabilities of subscales in the next section.
and dependably measuring students’ Korean
language ability.
Procedure
Measuring L1 and L2 Reading Attitudes. Our
next set of measures assessed different dimen- At the beginning of the academic school year,
sions of L1 and L2 attitudes, using two sets of the five language tests were administered in paper
parallel items: 35 items focused on English form either as part of the all-day district-wide test
560 The Modern Language Journal 98 (2014)

administration or during relevant regular classes Finally, items associated with students’ percep-
(either Korean or English), distributed across tions about their own reading competence loaded
three weeks so as not to take more than 45 on Factor 5, self-assessment.
minutes of instructional time per week. Students Also, we calculated the internal consistency
were told that the test scores would not affect their estimate of reliability for each factor (see Tables 2
school grades but were encouraged to do their and 3). The L1 and L2 reading attitude factors
best. In the fourth week, the attitude question- demonstrated good internal consistency, ranging
naires were administered, again in paper form, from .58 to .89, with the anxiety factor showing
during a session of a social studies class. relatively lower reliability in both L1 and L2,
Instructions for each measure and for how to a ¼ .62 and .58, respectively.
answer the 5-point rating scales were given in
Korean, and the students were encouraged to ask
questions if any survey item puzzled them. Research Questions 1a and 1b: Predicting L2
Reading Attitudes from L1 Reading Attitudes as Well
RESULTS as from Additional Reading Factors

Factor Analysis to Identify L1 and L2 Reading We first examined the connection between L1
Attitude Factors and L2 reading attitudes from two perspectives: by
calculating the contribution of L1 attitudes to L2
In order to identify components of reading attitudes in multiple regressions and sequential
attitude, we began with two separate factor regressions and by testing for differences between
analyses for L1 and L2 reading. Following L1 and L2 attitudes. We next tested several factors
Tabachnick and Fidell (2001), we used the that had been identified in the previous literature
following criteria in extracting factors: Items as predictors of L2 reading attitude.
with a minimum eigenvalue of 1.0 and that The results of simple bivariate correlations
loaded on a factor at a level of .45 and above were between L1 and L2 reading attitude factors
kept; items that loaded on multiple factors at .32 confirmed no violation of the multicolinearity
or higher or with little theoretical connections to assumption that would preclude the use of
the factor were deleted. Results indicated the multiple regression. We also checked whether
same five-factor solution (with varimax rotation) our data satisfied other regression assumptions,
for both surveys (29 items for L1; 30 items for L2). including normal distribution, homogeneity of
Tables 2 and 3 present the factor loadings of variance, homoscedasticity, and linearity, and
items, and the mean scores and standard devia- verified that all assumptions were met (Tabach-
tions of each reading attitude factor. The five nick & Fidell, 2001). The correlation between
factors of L1 and L2 reading attitudes accounted attitude factors within and across languages were
for 52.5% and 54.7% of the total variance, of low to moderate size (see Table 4). The
respectively. Because the two questionnaires strongest associations appeared between corre-
were almost identical, it is not surprising that sponding L1 and L2 reading attitude factors,
the same five factors were extracted. Factor 1, ranging from the cognitive attitude factor
named cognitive attitudes, was associated with (r ¼ .71) to the negative affect factor (r ¼ .44).
intellectual, practical, and linguistic values that Without claiming a causal connection between L1
students expected to achieve through reading. reading attitudes and L2 reading attitude, we saw
Factor 2, conative attitudes, obtained high loadings these associations as suggestive of an underlying
from items that measured students’ actual read- attitude tendency that connects particular atti-
ing behaviors, their intentions to read in L1 or L2. tude factors in a student’s L1 with the same factor
Factors 3 and 4 were related to negative feelings in the student’s L2.
associated with reading. Initially, we had expected Results of simultaneous multiple regressions
enjoyment, avoidance, comfort, and anxiety either to revealed that the five L1 reading attitude factors,
load on one overarching factor or on several that is, the factors of L1 cognitive, conative,
independent factors. However, items measuring negative affect, anxiety, and self-assessment,
enjoyment were dropped because they did not entered as independent predictors, explained
meet the factor analysis criteria. Also, the items approximately 22% to 51% of the total variance
connected to avoidance and discomfort loaded for each of the five L2 reading attitude factors.
together on Factor 3, which we labeled negative These results suggest that variations in L2 reading
affect, and separate from anxiety items, which attitudes could be explained by L1 reading
loaded on Factor 4 and which we labeled anxiety. attitude factors (see Table 5). Interestingly, for
Juhee Lee and Diane L. Schallert 561
TABLE 2
Five-Factor Solution for Reading Attitudes in Korean

Loadings M SD
Factor 1: Cognitive attitudes about L1 reading 3.54 .73
K12. I can acquire broad knowledge if I read books. .785 3.67 1.10
K22. I can improve my sensitivity to the Korean language if I read books. .730 3.54 1.03
K35. I can develop reading ability if I read books. .705 3.61 .97
K8. Reading books is useful to get a good job in the future. .694 3.71 1.05
K18. Reading books is useful to get a good grade in class. .675 3.42 1.02
K27. I can acquire vocabulary if I read books. .651 3.52 1.00
K26. I can become more knowledgeable if I read books. .649 3.43 .97
K6. I can develop writing ability if I read books. .597 3.64 1.09
K16. I get to know different values if I read books. .547 3.34 1.06
a ¼ .87
Factor 2: Conative attitudes about L1 reading 2.90 .81
K14. I go to a library to borrow or read books. .733 2.50 1.17
K13. I like to read books in my spare time. .679 2.55 1.10
K24. I try to find time for reading. .603 2.80 1.06
K21. I want to read many books in the future. .540 3.33 1.13
K33. I sometimes use the Internet to read stories that interest me. .457 2.98 1.24
K15. If someone tells me that he or she likes a book very much, .425 3.25 1.25
I am going to read it too.
a ¼ .87
Factor 3: Negative affect about L1 reading 2.57 .74
K19. Reading books is dull. .660 2.70 1.21
K9. I want to avoid reading as much as possible. .656 2.45 1.07
K17. I feel tired if I read books. .635 2.68 1.13
K11. I do not want to read even if the content is interesting. .607 2.17 1.08
K5. I don’t mind even if I cannot understand the book content .600 2.69 1.07
entirely.
K30. If I do not understand content in reading, I skip the part. .574 2.75 1.13
K25. When I read books, I find it difficult to concentrate. .560 2.51 1.07
a ¼ .79
Factor 4: Anxiety about L1 reading 2.38 .76
K20. I sometimes feel anxious that I may not understand even if I read. .794 2.37 1.05
K3. I feel anxious if I don’t know all the words in reading passages. .694 2.27 1.06
K34. I feel anxious when I’m not sure whether I understood the book .574 2.51 1.00
content.
a ¼ .58
Factor 5: Self-assessment about L1 reading 3.02 .71
K28. I am good at understanding what I read. .583 3.25 1.00
K4. My grades for Korean classes at middle school are very good. .536 2.84 .94
K23. I feel confident when I am reading books. .529 3.02 .99
K32. I read fast and understand quickly. .491 2.98 1.08
a ¼ .66
Note. M: Mean; SD: Standard deviation; a: Cronbach’s alpha; Percentage variance explained: total variance,
52.50; cognitive, 17.04; conative, 11.40; negative affect, 10.80; anxiety, 6.24; self-assessment, 7.03.

three of the five L2 reading attitude factors (L2 L2 negative affect was predicted by L1 self-
cognitive, conative, and anxiety), the only signifi- assessment (negatively) as well as by L1 negative
cant predictor was the corresponding L1 attitude affect; and (b) L2 self-assessment was predicted by
factor. For instance, students with positive cogni- L1 cognitive factor as well as by L1 self-assessment.
tive attitudes toward L1 reading were likely to This result suggests that reading attitudes estab-
score high on the L2 cognitive attitude factor lished in one’s native language may transfer to, or
(B ¼ .67). For two of the L2 attitude factors, a at least correlate with, L2 reading attitudes. Next,
second predictor, in addition to the correspond- to calculate estimates of the unique proportion of
ing L1 factor, made a significant contribution: (a) each significant predictor, we used sequential
562 The Modern Language Journal 98 (2014)
TABLE 3
Five-Factor Solution for Reading Attitudes in English

Loading M SD
Factor 1: Cognitive attitudes about L2 reading 3.27 75
E22. I can develop my English reading ability if I read English. .773 3.28 1.04
E10. I can develop my English writing ability if I read English. .756 3.25 1.10
E23. I can improve my sensitivity to the English language if I read English. .730 3.21 1.01
E32. I can acquire broad knowledge if I read English. .724 3.29 1.03
E15. Reading English is useful to get a good grade in class. .691 3.25 1.06
E8. I can become more knowledgeable if I read English. .672 3.27 1.09
E12. I can acquire English vocabulary if I read English. .671 3.53 1.04
E16. Reading English is useful to get a good job in the future. .669 3.47 1.02
E14. I get to know different values if I read English. .577 2.88 .96
a ¼ .89
Factor 2: Conative attitudes about L2 reading 2.32 .73
E29. I go to a library to borrow or read English books. .756 1.91 .96
E9. I like to read English books in my spare time. .710 1.90 .91
E21. I sometimes visit English websites and read them on the Internet. .654 2.09 1.09
E2. I try to find time for reading in English. .612 2.27 .95
E35. During my vacation I want to read at least one English book. .545 2.72 1.23
E26. If someone tells me that he or she likes an English book very much, .507 2.74 1.11
I am going to read it too.
E18. I want to read many English books in the future. .499 2.58 1.11
a ¼ .82
Factor 3: Negative affect about L2 reading 2.99 .71
E13. I want to avoid reading in English as much as possible. .702 3.07 1.08
E11. Reading English is dull. .700 3.05 1.05
E34. If I do not understand content in reading, I skip the part. .687 3.21 1.09
E1. I feel tired if I read English. .676 2.85 1.07
E31. When I read in English, I find it difficult to concentrate. .615 2.80 1.07
E3. I do not want to read in English even if the content is interesting. .512 2.89 1.10
E17. Reading English is troublesome. .501 3.20 1.10
E5. I don’t mind even if I cannot understand the book content entirely. .467 2.88 1.07
a ¼ .81
Factor 4: Anxiety about L2 reading 2.73 .85
E20. I feel anxious if I don’t know all the words in reading passages. .796 2.82 1.14
E7. I sometimes feel anxious that I may not understand what I read. .785 2.71 1.12
E4. I feel overwhelmed whenever I see a whole page of English in .520 2.67 1.10
front of me.
a ¼ .62
Factor 5: Self-assessment about L2 reading 2.67 .89
E27. My grades for English reading tests at middle school are very good. .682 2.57 1.06
E28. I feel confident when I am reading in English. .654 2.63 1.03
E25. I am good at reading in English. .616 2.81 1.16
a ¼ .75
Note. M: Mean; SD: Standard deviation; a: Cronbach’s alpha; Percentage variance explained: total variance,
54.66; cognitive, 17.29; conative, 11.33; negative affect, 12.81; anxiety, 6.39; self-assessment, 6.84.

regression analysis. Each predictor was entered Bonferroni correction that established the signif-
into the model as a second step controlling for all icance level at .01 to correct for an inflated Type 1
other predictors entered in together in a first error, we calculated matched t-tests and found
block (Warner, 2013). The effect sizes for that students were reporting different reading
significant individual predictor variables ranged attitudes for reading in English and Korean (see
from 2% to 28% of the total variance. Table 6). L1 reading attitudes were higher than
Another approach to addressing the first L2 reading attitudes for the cognitive, conative,
research question was to examine differences and self-assessment factors, with differences all
between L1 and L2 reading attitudes. Using a statistically significant. Thus, these EFL middle
Juhee Lee and Diane L. Schallert 563
TABLE 4
Intercorrelations of the Reading Attitude Factors for L1 (Korean) and L2 (English) Reading

Scale Factors K1 K2 K3 K4 K5 E1 E2 E3 E4
K Cognitive (K1) –
K Conative (K2) .50 –
K Negative affect (K3) .30 .56 –
K Anxiety (K4) .06 .19 .01 –
K Self-assessment (K5) .56 .59 .43 .05 –
E Cognitive (E1) .71 .36 .21 .09 .40 –
E Conative (E2) .41 .58 .37 .14 .32 .47 –
E Negative affect (E3) .24 .29 .44 .00 .17 .36 .55 –
E Anxiety (E4) .12 .19 .09 .46 .07 .12 .05 .19 –
E Self-assessment (E5) .45 .42 .35 .02 .47 .47 .53 .50 .09

p < .05.
Boldface coefficients indicate corresponding correlations across L1 and L2.

school students seemed to expect higher levels of previous research, we tested variables that had
cognitive outcomes (linguistic, intellectual, prac- been found to contribute to the L2 reading
tical) from L1 reading than L2 reading, were attitude: length of L2 formal and private instruc-
more willing to read in their native than in their tion, experiences of learning foreign languages
second language, and had higher self-confidence other than English, L1 and L2 reading frequency
about L1 reading compared to L2 reading. As for as separate factors, teachers’ encouragement to
affective factors, students had more negative read L2 texts, gender, and access to reading
affect toward English than Korean reading, and resources. These demographic variables, along
higher anxiety in L2 than L1 reading. As shown in with the measure of L2 proficiency as another
Table 6, calculations of Cohen’s d, a common well-established contributor to L2 reading atti-
measure of effect size and indicator of the tude, were entered into simultaneous multiple
practical import of a significant difference, regressions as predictors of each of the five L2
indicated that effect sizes ranged from .36, a reading attitude factors (see Table 7).
low–moderate effect size associated with less than Of the nine variables, L2 proficiency, L2 reading
1/3 of a point difference between means, to .75, a frequency, and access to English books significantly
high–moderate effect size associated with 3/5 of a contributed to four of the L2 reading attitude
point on the 5-point scale. Note that although factors (excluding the anxiety factor), whereas L1
these means differed significantly, they were reading frequency was not a statistically significant
neither very high nor very low in absolute terms predictor of any of the factors. Furthermore,
on the 5-point scale, ranging from 3.54 for the L1 length of L2 instruction had almost no impact on L2
cognitive factor to 2.32 for the L2 conative factor. reading attitude, although length of L2 private
Finally, we tested the contributions of various instruction was negatively linked to conative
other factors, in addition to L1 reading attitude, attitude, suggesting that the longer students
in predicting L2 reading attitude. Guided by received private L2 instruction, the less they

TABLE 5
Contribution of L1 Reading Attitude to L2 Reading Attitude

Dependent variable Significant predictors B (beta) t R2inc Total R2 Total F


E1 Cognitive K1 Cognitive .67 (.65) 12.60 .28 .51 59.06
E2 Conative K2 Conative .44 (.49) 7.14 .17 .36 31.19
E3 Negative affect K3 Negative affect .32 (.33) 5.17 .07 .24 17.42
K5 Self-assessment .17 (.17) 2.36 .02
E4 Anxiety K4 Anxiety .50 (.45) 8.32 .19 .22 16.36
E5 Self-assessment K5 Self-assessment .46 (.37) 5.29 .07 .25 19.23
K1 Cognitive .24 (.20) 3.11 .03

p < .05.
Note. K: Korean (L1); E: English (L2); standardized betas are in parentheses; R2inc: R squared incremental refers to
R square change, and is an effect size index for each individual predictor variable.
564 The Modern Language Journal 98 (2014)
TABLE 6
Pairwise Comparisons Between L1 and L2 Reading Attitude

Korean (L1) M (SD) English (L2) M (SD) t Cohen’s d



K1 Cognitive 3.54 (.73) E1 Cognitive 3.27 (.75) 8.17 0.36
K2 Conative 2.90 (.81) E2 Conative 2.32 (.73) 13.97 0.75
K3 Neg. affect 2.57 (.74) E3 Neg. affect 2.99 (.71) 9.52 0.58
K4 Anxiety 2.38 (.76) E4 Anxiety 2.73 (.85) 7.15 0.43
K5 Self-assess 3.02 (.71) E5 Self-assess 2.67 (.89) 7.21 0.43

p < .05 (Bonferroni adjustments)
Note. K: Korean (L1); E: English (L2); M: Mean; SD: Standard deviation.

were self-directed in choosing to read English attitudes. Because these middle school students’
books. Pressures associated with private English foreign language curriculum solely focused on
instruction from an early age, including private English, only 23% of the sample (N ¼ 66) had
tutoring or special cram schools, may have led to experienced learning other foreign languages.
negative attitudes toward L2 reading. In contrast, The last of the variables that showed some
access to English books contributed to developing degree of association with any of the attitude
positive reading attitudes (cognitive, B ¼ .10; factors was gender. Girls reported more negative
conative, B ¼ .15; negative affect, B ¼ .09; self affect (B ¼ .18) and lower self-assessment (B ¼
assessment, B ¼ .13), and teachers’ encourage- .27) toward L2 reading compared to boys, when
ment for L2 reading was associated with positive other factors (e.g., L2 proficiency, L2 reading
cognitive and conative attitude factors (B ¼ .13 frequency, and access to English books) were
and .09, respectively). Our participants were taken into account.
adolescent L2 learners who perhaps were suscep- When we used sequential regression to measure
tible to the influence of those around them, the contribution of individual predictors to each
including their teachers. reading attitude factor, we found that although
Contradicting Day and Bamford (1998), experi- each L2 reading attitude factor (except for
ences of learning to read other second languages was not anxiety) was accounted for by the composite of
a significant predictor of these students’ reading the nine variables from 22% to 37% of the total

TABLE 7
Factors Predicting L2 Reading Attitude

Dependent variable Significant predictors B (beta) t R2inc Total R2 Total F



E1 Cognitive L2 proficiency .01 (.28) 4.33 .05 .23 9.04
Teachers’ encouragement .13 (.18) 3.03 .03
Access to English books .10 (.14) 2.34 .02
E2 Conative Access to English books .15 (.22) 3.84 .04 .30 13.29
L2 reading frequency .14 (.22) 3.84 .04
L2 proficiency .01 (.19) 3.09 .02
Private L2 instruction .09 (.16) 2.50 .02
Teachers’ encouragement .09 (.13) 2.39 .01
E3 Negative affect L2 proficiency .01 (.32) 4.83 .07 .22 8.69
L2 reading frequency .15 (.22) 3.77 .04
Gendera .18 (.12) 2.21 .01
Access to English books .09 (.14) 2.25 .01
E4 Anxiety .02 .48
E5 Self-assess L2 proficiency .02 (.41) 6.98 .11 .37 17.95
L2 reading frequency .15 (.19) 3.46 .03
Access to English books .13 (.16) 2.85 .02
Gendera .27 (.15) 3.00 .02

p < .05.
Note. K: Korean (L1); E: English (L2); standardized betas are in parentheses; R2inc: R squared incremental refers to
R square change, and is an effect size index for each individual predictor variable.
a
Gender was coded with boys ¼ 1 and girls ¼ 2.
Juhee Lee and Diane L. Schallert 565

variance, the proportion of the total variance scores (B ¼ 4.20) as well as L2 reading scores
explained by individual predictors was relatively (B ¼ 6.62). Another way of interpreting the results
low, ranging from 1% to 11% (see Table 7). of the two regressions is to consider the combina-
In sum, although there were differences tion of contributors that predicted each of the
between L1 and L2 reading attitudes, significant achievement scores. L2 reading achievement was
correlations existed between the corresponding predicted by the combination of L2 self-assess-
L1 and L2 attitude constructs and a sizable ment and L1 cognitive attitude factors whereas L1
proportion of L2 reading attitudes was predicted reading achievement was predicted by L1 cogni-
by L1 reading attitude. In addition, several other tive, L1 negative affect, and L2 self-assessment
factors were associated with L2 reading attitude, factors. Again, although the proportion of
including access to English reading materials, variance explained by the composite of L1 and
teacher encouragement, gender, and length of L2 L2 reading attitudes was approximately one
private instruction. Note that two other predictors quarter of the total variance, the contribution of
of L2 reading attitude, L2 proficiency and L2 each individual reading attitude factor to reading
reading frequency, have as yet not been fully achievement was rather small in both L1 and L2,
discussed because these two variables are ad- explaining from 2% to 5% of the variance.
dressed in more depth in the next two sections. In order to examine the relative contributions
of reading attitude and language proficiency to
reading achievement in L1 and L2, multiple
Research Questions 2a and 2b: Associations Between regressions were conducted. When L2 proficiency
Reading Attitude and Reading Achievement, Taking and L2 reading attitude were entered into the
into Account the Contributions of Proficiency multiple regression simultaneously as predictor
variables of L2 reading achievement, only L2
Addressing this research question meant relat- proficiency remained as a significant factor
ing each attitude factor to reading achievement (B ¼ .83) (see Table 10). Similarly, when L1
scores in both languages, again using simple proficiency and L1 reading attitude factors were
bivariate correlations followed by multiple re- entered into the multiple regression simulta-
gressions. Results indicated significant correla- neously as predictor variables of L1 reading
tions between most attitude factors and reading achievement, L1 proficiency was a critical con-
scores in L1 and L2 (see Table 8). The only factor tributor (B ¼ .58), but so was L1 cognitive attitude
that showed no relation with either reading score (K1) (B ¼ 5.00). The findings indicated that L2
was the anxiety factor in both L1 and L2. The two proficiency was the sole contributor to L2 reading
strongest relationships were represented by L1 scores, explaining approximately 58% of the
cognitive attitude (K1) linked to Korean reading variance, whereas both linguistic and nonlinguis-
scores (r ¼ .47) and L2 self-assessment (E5) and tic (attitude) variables predicted approximately
English reading scores (r ¼ .42). 60% of the total variance of L1 reading
Similar results were obtained when L1 and L2 performance.
reading attitude factors were entered in multiple Next, to assess the effect size of individual
regression analyses simultaneously as predictor predictor variables on reading achievement
variables (see Table 9). In both regressions, scores, we conducted follow-up analyses using
cognitive attitude toward L1 reading was a hierarchical regression. After accounting for the
significant predictor, not only of L1 reading rest of the predictors (e.g., attitude factors) in a
achievement (B ¼ 7.97) but also of L2 reading first block, we entered each significant predictor
scores (B ¼ 5.40), whereas the L2 self-assessment (e.g., proficiency) in a second step. L1 proficiency
attitude factor significantly predicted L1 reading and L2 proficiency were the most explanatory

TABLE 8
Correlations Between Reading Attitude Factors and Reading Achievement in L1 and L2

K3 E3
Reading K1 K2 Neg. K4 K5 E1 E2 Neg. E4 E5
score Cog. Cona. affect Anxiety Self-assess Cog. Cona. affect Anxiety Self-assess
Korean .47 .32 .28 .02 .33 .35 .17 .12 .03 .29
English .33 .20 .20 .09 .29 .31 .26 .30 .02 .42

p < .01.
Note. K: Korean (L1); E: English (L2).
566 The Modern Language Journal 98 (2014)
TABLE 9
Contribution of L1 and L2 Reading Attitude to L1 and L2 Reading Achievement

Dependent variable Predictor variable B (beta) T R2inc Total R2 Total F



Korean reading K1 Cognitive 7.97 (.33) 3.98 .04 .28 10.57
K3 Negative affect 4.31 (.18) 2.59 .02
E5 Self-assess 4.20 (.21) 3.00 .02
English reading E5 Self-assess 6.62 (.31) 4.36 .05 .23 8.48
K1 Cognitive 5.40 (.21) 2.49 .02

p < .05.
Note. K: Korean (L1); E: English (L2); Standardized betas are in parentheses.

factor, accounting for 35% and 37% of the total levels had different levels of reading attitudes.
variance in L1 and L2 reading scores, respectively. Specifically, in terms of awareness of cognitive
Although cognitive attitude factor (K1) was the values related to reading, intention to read books
only significant factor among the reading attitude (the conative factor), lower negative affect
factors, its individual effect on the Korean reading associated with reading, and higher self-assess-
score was limited to 3% of the total variance in ment of one’s reading competence, students at
reading achievement. However, it should be higher proficiency levels had more positive
noted that there remained variance that was not reading attitudes, in both L1 and L2, compared
explained by proficiency, which accounted for to those at lower proficiency levels.
25% and 21% respectively of the variance in each In sum, results indicated that language profi-
L1 and L2 reading score. This proportion of the ciency had a strong association with reading
variance was assumed to be due to the combina- attitude in both L1 and L2. Students who scored
tion of reading attitude factors even though high on measures of language proficiency tended
individual attitude factors were not separately to report positive reading attitudes and less
significant contributors. negative feelings about reading; this was true in
Given that reading achievement was predicted both languages.
by reading attitude and language proficiency, we
added a follow-up analysis to determine whether
reading attitude differed as students’ levels of Research Question 3: Associations Among Reading
language proficiency increased. For this analysis, Frequency, Reading Attitude, and Reading
we computed Pearson product–moment correla- Achievement
tion coefficients and drew a series of scatterplots
to illustrate the relationship between proficiency Although the importance of reading frequently
scores and each of the five reading attitude factors has often been endorsed as a route to improving
in L1 and L2, respectively (see Figure 1). Except reading attitudes as well as reading scores, there
for the anxiety factor, all the correlations between has been a paucity of research on the effects of
proficiency and attitude factors were statistically reading frequency on achievement and attitude
significant at p < .01: for L1: cognitive r ¼ .40, in L1 and L2. In our study, we used students’
conative r ¼ .27, negative affect r ¼ .26, and self- survey responses about how often they read books
assessment r ¼ .36; for L2: cognitive r ¼ .37, in Korean and in English for pleasure on a five-
conative r ¼ .28, negative affect r ¼ .34, and point scale, with 1 indicating that the person
self-assessment r ¼ .50. Such correlations indicat- “almost never” read for pleasure; 2 indicating that
ed that students of different language proficiency he/she read “once a month”; 3 indicating the

TABLE 10
Contributions of Language Proficiency and Reading Attitude to Reading Achievement

Dependent variable Predictor variable B (beta) T R2inc Total R2 Total F



Korean reading score L1 proficiency .58 (.66) 15.74 .35 .60 69.99
Cognitive attitude (K1) 5.00 (.20) 4.20 .03
English reading score L2 proficiency .83 (.72) 15.75 .37 .58 64.19

p < .001.
Note. K: Korean (L1); E: English (L2); standardized betas are in parentheses; R2inc: R squared incremental refers to
R square change, and is an effect size index for each individual predictor variable.
Juhee Lee and Diane L. Schallert 567
FIGURE 1
Scatterplots of L1 and L2 Reading Attitude Scores Regressed on L1 and L2 Proficiency

person read “once a week”; 4 indicating he/she separate MANCOVAs with language proficiency
read “3–4 times a week”; and 5 indicating that the as covariates. For reading scores as dependent
person read for pleasure “almost every day.” variables, we used two ANOVAs to determine
Responses to these items were used to create five whether the reading frequency groups differed.
groups, separately for L1 and L2, representing ANOVA, and not MANCOVA, was appropriate
reading frequency levels. For the reading attitude here because the high correlation between
factors as dependent variables, we used two proficiency and reading achievement (.75 and
568 The Modern Language Journal 98 (2014)
FIGURE 1
(Continued)

.76 for L1 and L2) meant that little variance would cant differences between the two lower groups
be left unaccounted for once proficiency was and the three higher group, and generally no
partialled out (Warner, 2013). difference between the first and second groups
In L1 reading, the results indicated that there and between the fourth and fifth group. There
was a large difference in reading scores between was a dramatic increase in conative attitude from
those who almost never read (Group 1) and all those who never read to those who read every day.
other groups, who reported reading once a Because conative attitude should be highly
month or more often (see Table 11). Almost no predictive of students taking action for reading,
differences were found in reading achievement it is perhaps not surprising that students report-
among Groups 2, 3, 4, and 5, indicating that ing different levels of reading frequency should
occasional readers did not achieve significantly report such strikingly different conative attitudes.
different scores from those who read every day. As In L2 reading, because there were only 8
for reading attitudes, significant main effects for students who read in English every day for fun, we
frequency of reading for pleasure were found for combined groups 4 and 5 in order to make the
four of the five attitude factors, cognitive, groups of students more comparable (see Ta-
conative, negative affect, and self-assessment, ble 11). Despite the similarity to results for L1
with only anxiety showing no difference. These reading, reading scores and attitude were not
results suggested that as students reported linearly associated with frequency. For instance,
reading more often, they recognized the cognitive although those who read once a month or more
value of reading better, had less negative affect often achieved statistically significantly higher
toward reading, and evaluated their reading reading scores than those who never read for
competence as higher than those who read less. pleasure, more frequent reading groups did not
Pairwise comparisons generally indicated signifi- always have higher reading achievement than less
Juhee Lee and Diane L. Schallert 569
TABLE 11
Reading Attitude and Achievement in L1 and L2 by Frequency Grouping of Students

L1 (Korean) Groups (N)


G1 (74) G2 (68) G3 (69) G4 (51) G5 (27) Total F
Reading scores 61.42 69.57 a
68.59 a
68.88 a
73.74 a
3.47
Reading Attituded
Cognitive 3.27a 3.50 ab
3.68bc 3.64 bc 3.86 c
3.82
Conative 2.31 2.90 a
3.01 a 3.31 b 3.45 b
19.65
Negative affect 2.96 2.55 a
2.47 ab 2.35abc 2.17 bc
8.40
Anxiety 2.40 a 2.32 a
2.44 a 2.42 a 2.27 a
.40
Self-assessment 2.77 a 2.93 a
3.04 a 3.29 b 3.41 b
6.11

L2 (English) Groups (N)


G1 (176) G2 (54) G3 (29) G4 and G5 (30) Total F
Reading scores 59.39 66.74 a
68.41 a
66.83 a
3.22
Reading Attituded
Cognitive 3.13 a
3.57 b
3.30 abc 3.49 abc
4.26
Conative 2.08 2.65 a
2.62ab 2.78 ab
16.53
Negative affect 3.16 2.78 a
2.83 ab 2.55 ab
7.96
a a
Anxiety 2.73 2.85 2.68 a 2.61 a
.56
Self-assessment 2.44 2.95 a
2.93 ab 3.23 ab
9.13

p < .05. Overall N ¼ 289.
Note. dCovariates: Korean (and English) language proficiency test scores. Means that share the same letter are not
statistically different from each other.

frequent reading groups. In addition, the in- found to explain the most variance in reading
crease in reading attitude scales was not as attitudes were language proficiency (both L1 and
dramatic or as simple as we had found for L1 L2) and frequency of reading (in L1 and L2).
reading despite the fact that L2 reading frequency Thus, a final step was to examine whether L1 and
was a significant main effect for all areas of L2 L2 reading attitude predicted reading perfor-
reading attitude except anxiety. mance in each language, taking language profi-
ciency into account, and to establish what
DISCUSSION associations existed among reading frequency,
reading attitudes, and reading achievement in L1
Whereas previous studies have consistently and L2.
found evidence of cross-linguistic transfer from
L1 to L2 by demonstrating that L1 literacy and
oral skills are associated with L2 proficiency, L2 L1 and L2 Reading Attitudes
achievement, and L2 aptitude (Ganschow &
Sparks, 2001; Sparks et al., 2009, 2012), few From two factor analyses, five components of
studies have investigated affective factors that may reading attitudes were specified for both English
similarly transfer from L1 to L2 and the contri- and Korean reading: cognitive, conative, negative
butions of these affective factors to predicting affect, anxiety, and self-assessment attitude fac-
linguistic achievement. Designed to provide a tors. Through multiple regression analysis, the
more nuanced conceptualization of the con- five constructs making up students’ attitudes
structs of L1 and L2 reading attitudes based on toward reading Korean were found to contribute
Fishbein and Ajzen’s (1975) tripartite approach, significantly to the corresponding components of
our study examined several aspects of the English reading attitudes, suggesting the possibil-
connection between attitude and reading for ity of cross-linguistic transfer in the affective
middle school students in Korea learning English. domain and corroborating the reports of Day and
This examination first took up what relationships Bamford (1998) and Yamashita (2004, 2007) that
obtained among different components of reading reading attitudes developed in the native lan-
attitudes, both within and across languages, and guage may shape L2 attitudes.
what factors might be associated with L2 reading Despite the close associations between L1 and
attitudes. From among these factors, the two we L2 reading attitudes, the degree of favorability for
570 The Modern Language Journal 98 (2014)

reading differed in each language. These EFL that has forced young Korean students to study
adolescent students believed that they would gain English until late at night in cram schools, might
more cognitive benefits from reading in Korean, have an unintended impact on students’ attitudes
more willingly engaged in L1 reading, had lower towards English reading. Our results indicated
negative affect and anxiety in L1 reading, and that years of English study in formal schooling was
evaluated their own reading ability more highly not a significant predictor, but length of private
for reading in Korean than in English. Such English instruction was a negative contributor to
results may not be surprising. Although the conative attitudes for English reading (perhaps
students expected positive cognitive outcomes simply by way of displacing any time a student
from reading in English in terms of linguistic, could spend in pleasure reading). This finding
practical, and intellectual aspects, their insuffi- confirmed Day and Bamford’s (1998) argument
cient L2 proficiency seemed connected to higher that enjoyable experiences in the second lan-
negative affect and anxiety, lower self-assessment, guage classroom environment, including teach-
and a reluctance to read in English compared to ers, peers, learning materials, and activities, can
L1 reading. The results are also consistent with shape positive reading attitudes in L2, whereas
previous studies that have found that EFL learners unpleasant impressions or ongoing experiences
felt more comfortable and less anxious when may develop into negative attitudes.
reading in their first language and perceived Results of our study related to gender effects
more intellectual value from L1 than L2 reading were complex. Inconsistent with previous studies
(Yamashita, 2004, 2007). in the first language (e.g., Logan & Johnston,
In addition, we examined specific factors that 2009; McKenna et al., 1995; Swalander & Taube,
may have an effect on attitudes toward L2 reading 2007), girls had more negative affect (discomfort,
through multiple regressions. The significant avoidance) toward English reading and lower self-
predictors of L2 reading attitudes included L2 assessment, even though they significantly out-
proficiency, L2 reading frequency, access to performed boys in terms of English proficiency
English books, the length of private L2 instruc- (girls: 54.6, boys: 47.7, t(287) ¼ 3.57, p < .001)
tion, teachers’ encouragement, and gender. and English reading comprehension (girls: 65.7,
Because L2 proficiency and L2 reading frequency boys: 59.9, t(287) ¼ 2.64, p < .01). Our results
explained a large proportion of the variance of L2 were similar to a crosscultural study of gender
reading attitudes, we conducted several analyses effects in beliefs about school performance by
with these factors, and these are discussed after we Stetsenko et al. (2000) who reported that girls
discuss the other predictors. between the ages of 8 and 13 from six countries
First, we should note that access to English (Czech Republic, Germany, Japan, Russia,
books significantly contributed to cognitive atti- Switzerland, and the United States) and from
tudes, negative affect (negatively), and self- several contexts were similar in that they had a
assessment in L2 reading, whereas the length of tendency to undervalue their talent for good
private L2 instruction negatively predicted cona- performance even though their actual achieve-
tive attitudes toward L2 reading. That is, when ment tended to be better than boys. However, in
English reading resources are available to stu- some contexts (e.g., Prague, and Berlin at some
dents, they are likely to build positive attitudes data gathering times), the connection between
toward L2 reading, implying that self-initiated beliefs and performance tended to be lower than
reading behavior may be more naturally shaped in others (e.g., Moscow, Los Angeles, West
by being exposed to sufficient reading resources Berlin). Also, in some contexts (e.g., Bern), there
rather than by being sent to private L2 instruction were no differences between boys and girls in
at an early age. Unlike Camiciottoli (2001) who either beliefs or performance, pointing to the
found a negative effect of years of previous complexity of culturally-situated gender differ-
English study on reading attitudes among Italian ences in school performance.
EFL university students, we included both the As a final factor predicting L2 reading attitudes,
years of school English education and the length we were somewhat surprised to find that L1
of private English lessons in the multiple regres- reading frequency was not a significant contribu-
sion model because the majority of the Korean tor to any L2 reading attitudes. We had assumed
students had received private English instruction that L1 reading frequency might have an indirect
from an early age in addition to school education. impact on the development of L2 reading
In particular, we assumed that the Korea-specific attitudes because L1 reading frequency seems to
educational context, recently labeled as display- be associated with L1 reading attitudes, and L1
ing collective neurosis of English fever (Kim, 2002) reading attitudes contribute to the formation of
Juhee Lee and Diane L. Schallert 571

L2 reading attitudes, according to Day and frequency in L1 and L2 was associated with
Bamford (1998). In addition, Sparks et al. reading achievement and reading attitudes within
(2012) reported that L1 print exposure contrib- each language. With regard to the associations
uted to 3 to 10% unique variance to the between reading frequency and reading achieve-
prediction of L2 proficiency, including reading ment, those who read almost never scored
comprehension, word decoding, and oral and significantly lower in reading comprehension
writing skills. Thus, we had speculated that those than occasional or everyday readers in both L1
students who engage in L1 reading to a higher and L2. Previous studies (Stanovich, 1986, 2000)
extent are the same students with higher L2 have also indicated a causal relationship between
proficiency, who in turn have more favorable L2 reading amount and reading achievement by
reading attitudes. However, although L1 reading demonstrating that those who read more
frequency may contribute to some of the variance achieved a higher level of language-related skills,
in either L1 reading attitudes or L2 proficiency, its including reading comprehension. However, we
power and impact were too insufficient and need to consider why occasional, frequent, and
indirect to account for L2 reading attitudes. everyday readers obtained similar reading com-
prehension test scores in both L1 and L2. A
Reading Attitudes, Language Proficiency, and possible reason for this phenomenon may be that
Reading Achievements the total amounts of reading may be similar
between those who read once a month and those
Both L1 and L2 reading attitudes were signifi- who read every day.
cant predictors of reading achievement in each As for the relation between reading frequency
language although the components of attitudes and reading attitudes, an obvious pattern ob-
that were significant predictors were different. tained in both L1 and L2: Those who reported
Specifically, for L1 reading attitudes, it was the more frequent reading perceived more cognitive
cognitive component, that is, students’ percep- value in reading, more voluntarily engaged in
tions of the linguistic, intellectual, and practical reading, and had lower negative affect and higher
benefits of reading, that predicted both L1 and L2 self-assessment about their reading ability than
reading achievement, whereas for L2 reading those who read less. Such patterns were more
attitudes, negative affect and self-assessment were noticeable in L1 reading attitudes than L2,
the predictors of L1 and L2 reading achievement. possibly because insufficient linguistic knowledge
However, when language proficiency was en- and fewer automatic decoding skills in L2 reading
tered into the multiple regression with L1 and L2 slow down the reading speed that connects full
reading attitudes as independent variables, a large involvement in reading with the development of
proportion of the variance in reading achieve- positive affect and willingness to read more.
ment was explained by language proficiency,
particularly for L2 reading achievement. We IMPLICATIONS
also found a significant association between
language proficiency and reading attitudes in Second/foreign language reading involves
both L1 and L2, which is inconsistent with complex abilities and constructs in two language
Yamashita’s (2007) findings that L2 proficiency systems. Many attempts have been made to
was not a significant predictor of L2 reading examine the multidimensional nature of L2
attitudes except for L2 comfort. In our study, reading, particularly focusing on L1 literacy and
students’ reading attitudes, except for anxiety, L2 knowledge. However, as Bernhardt (2005)
were associated significantly with language profi- pointed out, only half of L2 reading is explained
ciency in both L1 and L2. The patterns across the by linguistic knowledge, with the other half
two languages were remarkably similar, with remaining unexplained. Nonlinguistic variables,
higher levels of proficiency associated with higher such as L2 reading attitudes, may begin to
scores on the positive attitude factors (cognitive, contribute to explaining the other half. This
conative, and self-assessment) and with lower study was intended to identify the connections
scores on the negative affect factor. between L1 and L2 reading attitudes and their
relation to L2 reading achievement.
Reading Attitudes, Reading Frequency, and Reading The results of this study may have implications
Achievements for L2 teachers as well as L2 learners. With its
demonstration of the connection between L2
Although L1 reading frequency was not a signi- learners’ attitudes for L1 and L2 reading,
ficant predictor of L2 reading attitudes, reading and between reading attitudes and reading
572 The Modern Language Journal 98 (2014)

achievement in each language, the findings of this Cunningham, A. E., & Stanovich, K. E. (1997). Early
study may help L1 and L2 readers recognize the reading acquisition and its relation to reading
importance of nonlinguistic variables such as experience and ability 10 years later. Developmental
attitudes in L2 reading. Moreover, the multidi- Psychology, 33, 934–945.
Day, R., & Bamford, J. (1998). Extensive reading in the
mensionality of reading attitude that the results
second language classroom. Cambridge: Cambridge
demonstrated may give teachers a basis for University Press.
understanding students’ positive or negative Eagly, A. H., & Chaiken, S. (1993). The psychology
attitudes. Thus, teachers who recognize what of attitudes. Fort Worth, TX: Harcourt Brace
aspects of reading attitudes are positively or Jovanovich.
negatively related to students’ reading behaviors Fishbein, M., & Ajzen, I. (1975). Belief, attitude, intention,
and ultimately, their reading comprehension may and behavior: An introduction to theory and research.
take different approaches in encouraging their Reading, MA: Addison–Wesley.
students to read. Finally, this study may pave the Ganschow, L., & Sparks, R. (2001). Learning difficulties
way for future research in an area that has so far and foreign language learning: A review of
research and instruction. Language Teaching, 34,
been underexplored, that of L2 reading attitudes
79–98.
including the three domains of cognition, affect, Gerdes, K. E., & Stromwall, L. K. (2008). Conation: A
and conation for adolescent L2 readers. missing link in the strengths perspective. Social
Work, 53, 233–242.
Grabe, W. (2009). Reading in a second language: Moving
REFERENCES
from theory to practice. New York: Cambridge
University Press.
Abu–Rabia, S. (2004). Teachers’ role, learners’ gender Jimenez, R. T., Garcia, G. E., & Pearson, P. D. (1996).
differences, and FL anxiety among seventh-grade The reading strategies of bilingual Latina/o
students studying English as a FL. Educational students who are successful English readers:
Psychology, 24, 711–721. Opportunities and obstacles. Reading Research
Ajzen, I. (1989). Attitude structure and behavior. In Quarterly, 31, 90–112.
A. R. Pratkanis, S. J. Brecker, & A. G. Greenwald Kamhi–Stein, L. D. (2003). Reading in two languages:
(Eds.), Attitude structure and function (pp. 241– How attitudes toward home language and beliefs
274). Hillsdale, NJ: Lawrence Erlbaum. about reading affect the behaviors of “underpre-
Ajzen, I., & Fishbein, M. (1980). Understanding attitudes pared” L2 college readers. TESOL Quarterly, 37,
and predicting social behavior. New York: Prentice 35–71.
Hall. Kim, Y. M. (2002). Collective neurosis of English fever.
Alexander, J. E., & Filler, R. C. (1976). Attitudes and Education Review, 9, 56–64.
reading. Newark, DE: International Reading Kush, J. C., Watkins, M. W., & Brookhart, S. M. (2005).
Association. The temporal–interactive influence of reading
Alptekin, C., & Ercetin, G. (2010). The role of L1 and L2 achievement and reading attitude. Educational
working memory in literal and inferential com- Research and Evaluation, 11, 29–44.
prehension in L2 reading. Journal of Research in Lazarus, B. D., & Callahan, T. (2000). Attitudes toward
Reading, 33, 206–219. reading expressed by elementary school students
Bernhardt, E. B. (2005). Progress and procrastination in diagnosed with learning disabilities. Reading
second language reading. Annual Review of Applied Psychology, 21, 271–282.
Linguistics, 25, 133–150. Lee, J. W., & Schallert, D. L. (1997). The relative
Bernhardt, E. B. (2011). Understanding advanced second- contribution of L2 language proficiency and L1
language reading. New York: Routledge. reading ability to L2 reading performance: A test
Bernhardt, E. B., & Kamil, M. L. (1995). Interpreting of the threshold hypothesis in an EFL context.
relationships between L1 and L2 reading: Consol- TESOL Quarterly, 31, 713–739.
idating the linguistic threshold and the linguistic Logan, S., & Johnston, R. (2009). Gender differences in
interdependence hypotheses. Applied Linguistics, reading ability and attitudes: Examining where
16, 15–34. these differences lie. Journal of Research in Reading,
Bourdieu, P. (1984). Distinction: A social critique of the 32, 199–214.
judgement of taste. London: Routledge. Lummis, M., & Stevenson, H. W. (1990). Gender
Brantmeier, C. (2005). Nonlinguistic variables in differences in beliefs and achievement: A cross-
advanced second language reading: Learners’ cultural study. Developmental Psychology, 26, 254–
self-assessment and enjoyment. Foreign Language 263.
Annals, 38, 494–504. Martinez, R. S., Aricak, O. T., & Jewell, J. (2008).
Camiciottoli, B. C. (2001). Extensive reading in English: Influence of reading attitude on reading
Habits and attitudes of a group of Italian university achievement: A test of the temporal–interaction
EFL students. Journal of Research in Reading, 24, model. Psychology in the Schools, 45, 1010–
135–153. 1022.
Juhee Lee and Diane L. Schallert 573
Mathewson, G. C. (1994). Model of attitude influence of L2 proficiency? Language Learning, 62, 473–
upon reading and learning to read. In R. B. 505.
Ruddel, M. R. Ruddel, & H. Singer (Eds.), Stanovich, K. E. (1986). Matthew effects in reading:
Theoretical models and processes of reading (4th ed., Some consequences of individual differences in
pp. 1131–1161). Newark, DE: International Read- the acquisition of literacy. Reading Research Quar-
ing Association. terly, 21, 360–407.
McGuire, W. J. (1969). The nature of attitudes and Stanovich, K. E. (2000). Progress in understanding reading:
attitude change. In G. Lindzey & E. Aronson Scientific foundations and new frontiers. New York:
(Eds.), The handbook of social psychology (pp. 108– Guilford.
141). New York: Rinehart and Winston. Stetsenko, A., Little, T. D., Gordeeva, T., Grasshof, M., &
McKenna, M. C. (1994). Toward a model of reading Oettingen, G. (2000). Gender effects in children’s
attitude acquisition. In E. H. Cramer & M. Castle beliefs about school performance: A cross-cultural
(Eds.), Fostering the life-long love of reading: The study. Child Development, 71, 517–527.
affective domain in reading education (pp. 18–40). Stokmans, M. J. W. (1999). Reading attitude and its
Newark, DE: International Reading Association. effect on leisure time reading. Poetics, 26, 245–261.
McKenna, M. C., Conradi, K., Lawrence, C., Jang, B. G., Swalander, L., & Taube, K. (2007). Influences of family
& Meyer, J. P. (2012). Reading attitudes of middle based prerequistes, reading attitude, and self-
school students: Results of a U.S. survey. Reading regulation on reading ability. Contemporary Educa-
Research Quarterly, 47, 283–306. tional Psychology, 32, 206–230.
McKenna, M. C., & Kear, D. J. (1990). Measuring Swanson, B. B. (1985). Teacher judgments of first-
attitudes toward reading: A new tool for teachers. graders’ reading enthusiasm. Reading Research and
The Reading Teacher, 46, 626–639. Instruction, 25, 41–46.
McKenna, M. C., & Kear, D. J. (1999). Garfield revisited: Tabachnick, B. G., & Fidell, L. S. (2001). Using
Unlimited extension of permission to copy the multivariate statistics. Boston: Allyn and Bacon.
ERAS. Reading Teacher, 53, 244. Tait–McCutcheon, S. L. (2008). Self-efficacy in mathe-
McKenna, M. C., Kear, D. J., & Ellsworth, R. A. (1995). matics: Affective, cognitive, and conative domains
Children’s attitudes toward reading: A national of functioning. In M. Goos, R. Brown, & K. Makar
survey. Reading Research Quarterly, 30, 934–956. (Eds.), Navigating currents and charting directions:
Mills, N., Pajares, F., & Herron, C. (2006). A reevaluation Proceedings of the 31st Annual Conference of the
of the role of anxiety: Self-efficacy, anxiety, and Mathematics Education Research Group of Australasia,
their relation to reading and listening proficiency. (Vol. 2, pp. 507–513). Brisbane, Australia: MERGA.
Foreign Language Annals, 39, 276–295. Tsai, Y. R., Ernst, C., & Talley, P. C. (2010). L1 and L2
Paimin, A. N., Hadgraft, R., & Alias, M. (2011, October). strategy use in reading comprehension of Chinese
An examination of learning strategy, interest, intention EFL readers. Reading Psychology, 31, 1–29.
and academic performance: Case studies of Australia van Schooten, E., de Glopper, K., & Stoel, R. D. (2004).
and Malaysia. Paper presented at the Research Development of attitude toward reading adoles-
in Engineering Education Symposium, Madrid, cent literature and literary reading behavior.
Spain. Poetics, 32, 343–386.
Roberts, M. S., & Wilson, J. D. (2006). Reading attitudes Warner, R. M. (2013). Applied statistics: From bivariate
and instructional methodology: How might through multivariate techniques (2nd ed). Thousand
achievement become affected? Reading Improve- Oaks, CA: SAGE.
ment, 43, 64–69. Wertsch, J. V. (1991). A sociocultural approach to
Saito, Y., Horwitz, E. K., & Garza, T. J. (1999). Foreign socially shared cognition. In L. B. Resnick, J. M.
language reading anxiety. Modern Language Levine, & S. D. Teasley (Eds.), Perspectives on socially
Journal, 83, 202–218. shared cognition (pp. 85–100). Washington, DC:
Sani, A. M., & Zain, Z. (2011). Relating adolescents’ American Psychological Association.
second language reading attitudes, self-efficacy for Wigfield, A., & Asher, S. R. (1984). Social and motivational
reading, and reading ability in a non-supportive influences on reading. New York: Longman.
ESL setting. Reading Matrix, 11, 243–254. Worrell, F. C., Roth, D. A., & Gabelko, N. H. (2007).
Smith, M. C. (1990). A longitudinal investigation of Elementary reading attitude survey (ERAS) scores
reading attitude development from childhood in academically talented students. Roeper Review,
to adulthood. Journal of Educational Research, 83, 29, 119–124.
215–219. Yamashita, J. (2004). Reading attitudes in L1 and L2,
Sparks, R., Patton, J., Ganschow, L., & Humbach, N. and their influence on L2 extensive reading.
(2009). Long-term crosslinguistic transfer of skills Reading in a Foreign Language, 16, 1–19.
from L1 to L2. Language Learning, 59, 203–243. Yamashita, J. (2007). The relationship of reading
Sparks, R., Patton, J., Ganschow, L., & Humbach, N. attitudes between L1 and L2: An investigation of
(2012). Do L1 reading achievement and L1 adult EFL learners in Japan. TESOL Quarterly, 41,
print exposure contribute to the prediction 81–105.

You might also like