International Journal of

Bilingualism
http://ijb.sagepub.com/

Psychological and sociodemographic correlates of communicative

anxiety in L2 and L3 production
Jean-Marc Dewaele
International Journal of Bilingualism 2002 6: 23
DOI: 10.1177/13670069020060010201

The online version of this article can be found at:

http://ijb.sagepub.com/content/6/1/23

Published by:

http://www.sagepublications.com

Additional services and information for International Journal of Bilingualism can be found at:

Email Alerts: http://ijb.sagepub.com/cgi/alerts

Subscriptions: http://ijb.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://ijb.sagepub.com/content/6/1/23.refs.html

>> Version of Record - Mar 1, 2002

What is This?

Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
 IJB Volume 6anxiety
Communicative • Number 1 •and
in L2 March 2002, 23 – 38 23
L3 production

Psychological and sociodemographic

correlates of communicative anxiety in
L2 and L3 production*
Jean-Marc Dewaele
Birkbeck College, University of London

Acknowledgments*
We would like to thank the director of the Koninklijk Atheneum I Centrum Brugge, Mrs. Monique Devisscher,
and her staff for their cooperation and help in the gathering of the data for the present study.

Abstract
This paper analyses foreign language anxiety in the French L2 and English Key words
L3 speech production of 100 Flemish students. The findings suggest that
foreign language anxiety is not a stable personality trait among experienced foreign language
language learners. The societal as well as the individual contexts were found anxiety
to determine levels of communicative anxiety. The perception of French as
the former prestige language in Flanders and its function as a social marker L2
was found to be linked to the participants’ social class, which was, in turn,
linked to levels of anxiety in French — but not in English. This social effect L3
appeared to be a stronger predictor of communicative anxiety in French than
three personality variables (extraversion, neuroticism, and psychoticism). personality
Psychoticism, extraversion, and, to a lesser extent, neuroticism did however
significantly predict levels of communicative anxiety in English L3 produc-
tion. Students who scored high on the extraversion and psychoticism scales reported signif icant lower levels
of communicative anxiety in English. Those who scored low on the neuroticism scale also tended to report
lower levels of communicative anxiety in English. The same pattern emerged for communicative anxiety in
French without reaching statistical significance.

1 Introduction
Foreign language anxiety is one of the most important predictors of foreign language achieve-
ment (Onwuegbuzie, Bailey, & Daley, 1999). It can be defined as “the feeling of tension
and apprehension specifically associated with second language contexts, including speaking,
listening, and learning” (MacIntyre & Gardner, 1994, p. 284). More recently, Onwuegbuzie,
Bailey, and Daley (2000) described foreign language anxiety as “a form of situation-specific
anxiety” (p. 88). Although much research has been done on foreign language anxiety, its
complex and multidimensional nature (Young, 1991) could explain why this field still remains
underdeveloped (Aida, 1994). Such research is vital because we need to understand why a
language learner feels anxious in order to control such anxiety and alleviate its effects
(Horwitz, 1996; MacIntyre & Gardner, 1991a, b; Onwuegbuzie et al., 2000; Scovel, 1991).

Address for correspondence

Department of French, Birkbeck College, University of London, 43 Gordon Square, London WC1H 0PD;
e-mail: <[email protected]>
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
 24 J.-M. Dewaele

Philipps (1991) warns that language anxiety in the classroom causes negative affective reac-
tions and can induce negative attitudes and discourage students from continuing their language
study. Teachers need to be aware that language anxiety plays a role in overall strategy use
and the use of certain types of strategies in the language classroom (MacIntyre & Noels, 1996).
Research on foreign language anxiety will thus have important implications for the teaching
as well as the learning of foreign languages.

2 Previous research
In their exploratory study of the relations between language anxiety and other anxieties in
speaking English L1 and French L2, MacIntyre and Gardner (1989) found two orthogonal
factors which they labeled “General Anxiety” and “Communicative Anxiety.” The former
factor included scales of Trait, State, and Test Anxiety; the latter was defined by French
Class, French Use, English Class and Audience anxieties. No relationship was found between
General Anxiety and French vocabulary learning and production but significant negative
effects emerged on the learning of French L2 vocabulary. A similar effect appeared in tests
of the written and oral production of similar items from long-term memory. A close analysis
of the correlations between the individual anxiety scales and the language production measures
revealed that the French anxiety scales were responsible for the effects observed for the
Communicative Anxiety factor. The results confirmed earlier findings by Horwitz (1986) and
Gardner, Moorcroft, and MacIntyre (1987) on the orthogonal nature of Language Anxiety and
Trait Anxiety. Similar results emerged from MacIntyre and Gardner’s (1991b) study into the
factor structure underlying 23 scales assessing both language anxiety and other forms of
anxiety. French L2 tasks were judged to be more anxiety-provoking than their English L1 equiv-
alents by 95 students who had had an average of eight years of courses in French as a L2 (p. 518).
Subjects with higher levels of language anxiety in the L2 (but not the L1) obtained signifi-
cantly lower scores on a Digit Span test (a measure of short-term memory) and on a Thing
Category test (vocabulary production). The authors suggest that impaired performance among
more anxious students could be related to short-term memory loss and problems in the long-
term memory retrieval, both attributable to anxiety (p. 530). MacIntyre and Gardner (1994)
proposed a 3-stage model of foreign language anxiety. Foreign language students can expe-
rience fear at the input stage when they are presented with new information in the foreign
language. This anxiety can have detrimental effects on their concentration and on their ability
to encode the linguistic stimuli. Secondly, anxiety at the processing stage can debilitate
cognitive operations performed on external stimuli and memory processes. The students
may experience a reduced ability to understand messages and learn new vocabulary. Finally,
anxiety at the output stage can interfere with the retrieval of previously learned material and
might hinder the students’ ability to produce the foreign language (MacIntyre & Gardner,
1994). The authors also found that language anxiety “tends to correlate with measures of perfor-
mance in the second language but not in the native language” (p. 301). The authors conclude
that the “potential effects of language anxiety on cognitive processing in the second language
may be pervasive and may be quite subtle” (p. 301). Further research generally supported
the three-stage model of foreign language anxiety (Onwuegbuzie, et al., 2000).
MacIntyre (1999) reviewed the literature on foreign language anxiety and concluded
that a moderate negative relationship exists between language anxiety and various measures
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
 Communicative anxiety in L2 and L3 production 25

of language achievement. Interestingly enough, language anxiety does not necessarily diminish
as the learner progresses. Saito and Samimy (1996) found that anxiety appeared highest
among advanced subjects and lowest among intermediate subjects learning Japanese as a L2
in the U.S.A. The authors believe that this shift may be due in part to the increased degree
of difficulty and the decreased instructional time encountered in advanced courses. Findings
confirm the negative impact of anxiety on performance.

2.1
Factors associated with foreign language anxiety
In their study of 210 university students enrolled in French, Spanish, German, and Japanese
courses, Onwuegbuzie, et al. (1999) examined the effect of 26 factors on foreign language
anxiety. The authors identified seven variables that significantly predict foreign language anxiety
(i.e., age, academic achievement, prior history of visiting foreign countries, prior high school,
experience with foreign languages, expected overall average for current language course,
perceived scholastic competence, and perceived self-worth). These variables account for 40%
variance in foreign language anxiety (p. 232).
Several researchers suggest that skill in one’s native language (e.g., reading, vocabu-
lary, group achievement) may affect anxiety levels in the foreign language (Ganschow & Sparks,
1996; Sparks, Artzer, Ganschow, Siebenhar, Plageman, & Patton, 1998): “Students who have
overt or subtle native-language difficulties in reading, writing, listening, and speaking are likely
to experience similar difficulties in learning a foreign language” (Sparks et al., 1998, p. 209).
Foreign language anxiety has also been found to be linked to interpersonal relation-
ships. In a recent study, Dörnyei and Kormos (2000) failed to find an expected negative
correlation between language anxiety and English L2 use in dyadic conversations between 46
Hungarian students. The authors argue that one interlocutor’s degree of language anxiety
might in fact influence the other interlocutor: if both are anxious this variable will signifi-
cantly affect the L2 output, whereas, “if the interlocutor is sufficiently confident, h/she might
‘ pull along’ the more anxious speaker and therefore the impact of anxiety may not reach
statistical significance” (Dörnyei & Kormos, 2000, p. 296).
In the present study, we will focus on the effect of social and individual contexts
(MacIntyre, Dörnyei, Clément, & Noels, 1998) on foreign language anxiety, and more specif-
ically its effects on speaking a foreign language, as this is generally considered to be the
most anxiety-provoking of second language activities (MacIntyre & Gardner, 1991a; Phillips,
1992; Young, 1990). It is thus a study on the links between output anxiety (MacIntyre &
Gardner, 1994) and enduring influences (e.g., intergroup relations, learner personality) which
represent “stable, long-term properties of the environment or person that would apply to
almost any situation” (MacIntyre et al., 1998, p. 546).

2.2
The societal context
The societal context refers to the intergroup climate in which interlocutors evolve (MacIntyre
et al., 1998). The authors distinguish “two complementary dimensions concerned with the
structural characteristics of the community and their perceptual and affective correlates”
(p. 555).
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
26 J.-M. Dewaele

The structural characteristics concern ethnolinguistic vitality and communication

networks that either favor or do not favor the use of the L2. Ethnolinguistic vitality reflects
the socioeconomic power of a community and the extent to which its members are represented
in sociopolitical institutions (p. 555). Communication networks refer to the groups with
which there is regular communication; these can consist of L1 and L2 subnetworks. Perceptual
and affective correlates determine “attitudes and values regarding the L2 community and the
motivation to adapt and reduce social distance between ethnic groups” (p. 556). Intergroup
relations are often tense and this will affect L2 learning and communication. Prejudiced atti-
tudes might demotivate L2 learners. Gardner (1985) found that parents’ attitudes are highly
influential on their children’s language-learning motivation. Parents who encourage their
children to learn an L2 might nullify this by showing negative attitudes toward the L2 group
in question.
Likewise, the socioeconomic status of the parents is important in predicting attitudes
towards learning and academic performance (Furnham & Heaven, 1998). Parents with high
socioeconomic status have certain values and beliefs that increase the likelihood of academic
success (Argyle, 1994). One could expect, by extrapolation, that language learning success
is also determined by the socioeconomic status of the parents.

2.3
The individual context
The individual context refers to stable personality characteristics found to be particularly rele-
vant to communication (MacIntyre et al. , 1998, p. 555). The causes of personality traits
“have always been acknowledged to be both biological and social” (Furnham & Heaven,
1998, p. 32). According to these authors, extraversion-introversion and levels of anxiety seem
to be the best predictors of performance, depending “on the type of task being completed,
the presence of distracting factors, and the complexity of the task” (p. 82).
Extraversion– Introversion. Extraverts tend to be sociable, outgoing, gregarious, talkative,
underaroused individuals (Furnham & Heaven, 1998, p. 325), while introverts tend to be
reserved, quiet, and unassertive. Studies on language and extraversion are relatively few in
number (Dewaele & Furnham, 1999, 2000, 2001; Furnham, 1990; MacIntyre & Charos,
1996). Negative publicity for trait extraversion within the field of applied linguistics resulted
from one seriously flawed — but unchallenged — study by Naiman, Frohlich, Stern, and
Todesco (1978) on personality variables and language learning, where extraversion scores
determined with the Eysenck Personality Inventory (EPI) were found not to correlate with
written language test results (Dewaele & Furnham, 1999). Dewaele and Furnham (1999)
argued that if Naiman et al. (1978) had used a wider variety of more sophisticated linguistic
variables, covering not only written language but also natural communicative oral language,
they might have found that the construct validity of the EPI was not to blame for the lack of
expected correlations. This is what Dewaele and Furnham (2000) did using a corpus of
French interlanguage from Dutch L1 students. Correlational analyses between extraversion
scores and six linguistic variables reflecting fluency and accuracy revealed that extravert
multilinguals are more fluent than introvert multilinguals, especially in stressful interpersonal
situations. It was argued that some cognitive and physiological characteristics associated
with extraversion, such as superior short-term memory and better resistance to stress, can
explain the superior fluency of the extraverts’ speech production (Dewaele & Furnham,
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
 Communicative anxiety in L2 and L3 production 27

2000). The extraverts’ superior fluency in stressful situations could also be linked to their
lower communicative anxiety in the L2 and their inclination to take risks (Ely, 1986;
Horwitz, 1986). The finding that introverts use significantly fewer colloquial words than
extraverts in conversations in French interlanguage was also interpreted as an illustration of
the introverts’ higher levels of language anxiety (Dewaele & Furnham, 2001). Eysenck and
Eysenck (1985) and Matthews and Dorn (1995) suggest that extraverts are underaroused,
introverts overaroused. These arousal levels have been linked to the proportion of the neuro-
transmitters dopamine and norepinephrine in certain brain areas (Depue & Collins, 1999;
Lieberman, 2000). The introverts’ higher level of cortical arousal and greater fear of
punishment may make them behave in a cautious manner. Anxiety can explain performance
differences between extraverts and introverts, according to Furnham and Heaven (1998).
Anxiety “has an impact on one’s ability to perform certain tasks, and this relationship is
mediated by arousal” (p. 82). Earlier research had already shown that introverts tend to be
more socially anxious (Cheek & Buss, 1981) and that high anxiety leads to increased atten-
tional selectivity and reduced attentional capacity (Eysenck, 1979, 1981; Fremont, Means,
& Means, 1976). Highly anxious subjects have a reduced elaboration of encoding under high
anxiety (Eysenck, 1981, p. 183). Eysenck (1979) reconceptualized anxiety in terms of cogni-
tive interference. He suggested that the anxious person divides his/ her attention between
task-related cognition and self-related cognition, making cognitive performance less efficient.
Eysenck suggested that the anxious person tries to compensate for reduced efficiency by
increased effort. Markham and Darke (1991) found that high anxiety inhibited verbal — but
not spatial — reasoning tasks when the processing demand was high, and concluded that
cognitive self-concern has different effects on the verbal and visual domains of the working
memory system. We have seen earlier that MacIntyre and Gardner (1991b) develop this
explanation to explain the deficit in more anxious students, without referring explicitly to
extraversion. In a more recent article however, MacIntyre and Gardner (1994) point out that
Eysenck’s theory “is able to explain the negative effects observed for language anxiety”
(p. 285). Pursuing this line of research MacIntyre and Charos (1996) used path analysis to
investigate the role of global personality traits on self-reported frequency of communication
in a second language. They found significant negative correlations between extraversion and
language anxiety (p. 19). A negative correlation emerged between language anxiety and will-
ingness to communicate, suggesting that introverts communicated less in their French L2 than
extraverts (p. 18).
Furnham and Medhurst (1995) analyzed the link between personality types and seminar-
type (or classroom) activities. They found quite predictably that extraverts are much better
suited to seminar-type activities than other personality types. They participate more and are
less averse to making oral presentations. The overall best predictor, however, was the partici-
pant’s outcome on Eysenck’s psychoticism scale.
Neuroticism— Emotional stability. Neuroticism (N) is the second major personality
domain in Eysenck’s model of personality. It could be described as a minor nervous disorder.
Persons scoring high on this scale tend to suffer from “anxiety, phobia, depression, and
hypochondriasis” (Furnham & Heaven, 1998, p. 326). Those with low scores on N can be
described as calm, contented, and unemotional (Goldberg, 1992). MacIntyre and Charos
(1996) did not find a path from emotional stability to language anxiety. They interpret this
as proof that:
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
 28 J.-M. Dewaele

general trait anxiety, as would be reflected in the emotional stability factor, is not typically
associated with language anxiety. This supports the assertion that it is the social and commu-
nicative demands of L2 interaction, and not a predisposition to nervousness, that drive
language anxiety. (p. 19)

In their study of personality differences in Belgian university students according to their

major area of study, De Fruyt and Mervielde (1996) found that language majors scored
highest on the neuroticism scale and lowest on extraversion.
Psychoticism. Psychoticism (P) is the third major personality domain in Eysenck’s model
and measures tough-mindedness (Furnham & Heaven, 1998, p. 230). Persons scoring high
on psychoticism “tend to be hostile, cold, aggressive, and have poor interpersonal relations”
(p. 327). Furnham and Medhurst (1995) found that individuals with low scores on the P scale
were more likely to have good oral and written expression, were more motivated, and partic-
ipated more actively in seminars.

3 Rationale for the present study

Almost all studies on language anxiety have examined it as “a stable personality trait, among
experienced language learners” (MacIntyre & Gardner, 1991, p. 297). Another possibility,
of course, is that the stability of the trait could be related to the fact that the studies in ques-
tion considered only one L2 (Cheng, Horwitz, & Schallert, 1999; Horwitz, 1986; MacIntyre,
1999; MacIntyre & Charos, 1996; MacIntyre & Gardner, 1989, 1991a, b, 1994; Onwuegbuzie
et al., 2000).
The present study will investigate whether foreign language anxiety can be considered
as a stable personality trait by extending the analysis to anxiety in speaking an L2 and an
L3.
The purpose of the current study is also to investigate psychological (extraversion,
psychoticism, and neuroticism) and sociodemographic correlates (gender, social class) of
communicative anxiety in the foreign languages of advanced learners. We also decided to
investigate the extent to which self-perceived competence in the foreign language, types and
frequency of contact with the foreign language, and parental attitudes are related to language
anxiety.

4 Method
4.1
Participants
The participants included 100 pupils in their last year of secondary education at the Koninklijk
Atheneum I in Bruges, Belgium. The sample consisted of 49 males and 51 females. The
ages of the participants ranged from 17 to 21 (M = 17.8, SD = .09). Ninety-three participants
had Dutch as an L1, four had both Dutch and French as an L1, one had Chinese and Dutch
as an L1, one had Tunisian Arabic as an L1 and one was a native speaker of Thai. All the
pupils had had formal instruction in Dutch, in French (between 3 and 5 hours a week, starting
at age 10), and in English (between 2 and 4 hours a week, starting at age 12 or 14). Forty-
four pupils had also studied German and 15 pupils had learned Spanish.
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
 Communicative anxiety in L2 and L3 production 29

4.2
Materials
The materials included the short version of the Eysenck Personality Questionnaire (EPQr)
(Eysenck, Eysenck, & Barrett, 1985) which yielded the following results (N = 100): psycho-
ticism (P): M = 3.96, SD = 2.07, extraversion (E): M = 8.39, SD = 3.10, neuroticism (N):
M = 6.18, SD = 3.39.
Participants also completed a sociodemographic questionnaire based on Baker (1992)
relating to attitudes, motivation1, communicative anxiety, frequency and types of contact
with French and English, and parental attitudes towards these two languages.
Self-perceived competence and communicative anxiety in French and English were
measured with a three-point Likert response format. The communicative anxiety scale
measures apprehension at speaking French and English in public. The questionnaire also
contained items about frequency of communication, types of contact in French and English,
and parental attitudes using a True/False format.
Other sociodemographic data comprised gender (51 females and 49 males) and the
family’s social class which was determined through the highest level of education attained
by one of the parents (Preston, 1989). Thirty-nine participants thus fell into category 1
(degree of secondary education or less), 42 into category 2 (degree of further education,
maximum length being 3 years) and 19 into category 3 (university degree, obtained after a
minimum of 4 years of study).
The pupils’ grades for French and English were also collected. These grades were based
on written work including grammar exercises and essays. The grades for French ranged
between 36 and 88 out of 100 (M = 60.5, SD = 9.9). The grades for English ranged between 38
and 91 out of 100 (M = 65.0, SD = 13.0).

5 Results
The results will be examined in three sections. First, mean scores for levels of communica-
tive anxiety, frequency of communication, self-perceived competence, grades, type, and
frequency of contacts in French and English will be analyzed and compared through a series
of t-tests. Next, the relationship between gender, social class, the scores on the psychoticism,
extraversion, and neuroticism scales and the mean scores for communicative anxiety in
French and English will be examined through t-tests and Pearson product-moment correla-
tion coefficients. As a correlation analysis assesses the contribution of an independent
variable, in isolation, to a dependent variable, it is unable to illuminate phenomena involving
multiple effects (Bailey, Onwuegbuzie, & Daley, 2000).
Thus, a multiple regression analysis will be used in order the estimate the predictive
power of the three personality dimensions and the social dimension (i.e., independent vari-
ables) on the scores for communicative anxiety in French and English (i.e., dependent
variables).

1
The data on motivation and attitudes have not been used in the present study as they do not reflect “basic” dimensions
(cf. MacIntyre & Charos, 1996).

The International Journal of Bilingualism

Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
30 J.-M. Dewaele

5.1
Mean scores of variables
The analysis of the mean scores for communicative anxiety in French and English yielded
an unexpected result. Participants expressed a significantly greater communicative anxiety
in French than in English (see Table 1). The result is surprising because formal instruction
in French starts earlier and is more intensive than formal instruction in English. Similarly,
participants judged their competence in English to be significantly higher than in French.
These judgments about competence are reflected in the pupils’ grades which turned out to
be significantly higher for English than for French.

Table 1
Levels of com m unicative anxiety, self-perceived com petence, and grades in French and English (M eans (M) and
standard deviation (SD))

Variable N French English Value of t

M SD M SD
Level of communicative anxiety 99 1.92 .59 1.63 .56 3.98***
Self-perceived competence 99 2.16 .54 2.46 .52 –4.45***
Grades 96 60.5 9.9 65.0 13.0 –3.55***
*p < .05 **p < .01 ***p < .001

Less than half of the participants stated that they used French regularly outside school,
a slightly higher number using English regularly outside school. This difference is not signif-
icant (Table 2). Flanders being on the linguistic crossroads in Europe 2, we asked the participants
if they watched English and French television channels regularly. The participants reported
watching English channels significantly more frequently than French ones (Table 2).
Significantly more participants listen regularly to English songs than French ones; and a
significantly higher number read books in English rather than in French (Table 2). Parental
attitudes to French and to English were the same, with 99 out of 100 participants reporting
positive attitudes towards both languages (Table 2).
To sum up, English seems to be the dominant foreign language among our partici-
pants, although parental attitudes are equally positive towards both languages.

5.2
Relationships between independent and dependent variables
A Pearson correlation analysis revealed a significant relationship between social class and
communicative anxiety in French but not in English (Table 3). Participants from lower social
classes reported higher levels of communicative anxiety in French. Gender did not have any
effect on communicative anxiety in French or English (Table 3). Self-perceived competence
in French and English was found to be strongly linked to communicative anxiety in these
languages, which confirms earlier research on this subject (Table 3). A Pearson correlation
analysis between the scores on the psychoticism, extraversion, and neuroticism scales and

2
Flemish families have been able to watch foreign television stations for at least 15 years through regular cable connec-
tion. The current choice comprises Dutch, French, British and American English, German, Spanish, and Italian channels.

The International Journal of Bilingualism

Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
 Communicative anxiety in L2 and L3 production 31

Table 2
D ifferences in attitudes, type and frequency of co ntact with French and English for 100 Flem ish high-school
students (t- test)

Variable French English Value of t

n= n= (D F = 9 9 )
Use the target-language (TL) regularly outside school 43 49 1.14
Watch TL television channels regularly 28 84 10.80***
Listen to TL songs regularly 31 100 6.66***
Read TL books regularly 24 54 5.05***
Positive parental attitudes towards TL 99 99 .57
*p < .05 **p < .01 ***p < .001

the mean scores for communicative anxiety in French and English also yielded intriguing
results. The links between personality scores and communicative anxiety in French were
nonsignificant (Table 3). However, highly significant correlations emerged between scores
on the three personality scales and communicative anxiety in English (Table 3). This suggests
that extraverts (high E), less anxious (low N) and colder, more hostile participants (high P)
experience significantly less communicative anxiety in English. The pattern is similar but
nonsignificant for communicative anxiety in French.

Table 3
Relations between independent variables and com m unicative anxiety in French and English ( t-test and Pearson
correlation analysis)

Independent variable Communicative anxiety in French Communicative anxiety in English

Social class r (99) = –.20* r (99) = .09

Gender t (98) = –.69 t (98 )= –1.02
Self-perceived competence r (99) = –.49*** r (99) = –.55***
Psychoticism r (99) = –.11 r (99) = –.30**
Extraversion r (99) = –.10 r (99) = –.23*
Neuroticism r (99) = .17 r (99) = .22*
*p < .05 **p < .01 ***p < .001

These results suggest that the link between independent and dependent variables
varies according to the particular foreign language. The next section will identify the
predictors of communicative anxiety in French and English. We excluded the variables
“gender” which does not appear to affect communicative anxiety and “self-perceived
competence” which might not be a truly independent variable but rather the negative
image of communicative anxiety.

5.3
Prediction of communication anxiety in French
The regression of psychoticism, extraversion, neuroticism, and social class is close to statis-
tical significance: (N = 100) (R2 = .09, F = 2.26, df (4, 95), p < .069). No individual personality
variable is a significant predictor: psychoticism: (beta= – .09, t(95)= – .92, p =ns); extraversion:
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
 32 J.-M. Dewaele

(beta = – .08, t (95) = – .81, p = ns) and neuroticism: (beta = .17, t (95) = 1.7, p < .089). Social
class, however, is a significant predictor: (beta = – .20, t (95) = – 2.01, p < .047).
These four variables combined to explain 9% of the total variation. Using Cohen’s
(1992) criteria for assessing the predictive power of a set of independent variables in a
multiple regression model, the proportion of variance indicates a small effect size3.

5.4
Prediction of communication anxiety in English
The regression of psychoticism, extraversion, and neuroticism, is highly significant (N = 100)
(R2 = .20, F = 6.10, df (4, 95), p < .001). Both psychoticism and extraversion are significant
predictors (beta = – .35, t (95) = – 3.72, p < .001) and (beta = – .23, t (95) = – 2.43, p < .017).
Neuroticism is a weaker predictor: (beta = .17, t (95)= 1.9, p < .07). Social class is not a signifi-
cant predictor: (beta = .11, t (95) = .08, p = ns). According to Cohen’s (1992) criteria, the
proportion of variance explained (20%) indicates a medium effect size.

6 Discussion
Studies on communicative anxiety in foreign language production, typically report higher
levels of anxiety in speaking the second language compared to the first language. It is assumed
that both competence and self-perceived competence in a language affect language anxiety.
As English is being taught as a third language in Flemish schools, starting later in the school
curriculum and with fewer contact hours, one would expect the pupils to have become more
proficient in French and, accordingly to rate their competence in that language higher than
in English. This should logically lead to lower levels of communicative anxiety in French,
especially given the fact that they do not report a significantly higher usage of English outside
the school gates. If pupils have equal opportunities for interaction in French and English, they
could be expected to have similar levels of perceived competence (MacIntyre & Charos,
1996). The omnipresence of English music, film, and multimedia in Flanders might explain
these apparently contradictory findings in our data. Daily contact with highly valued Anglo-
American culture might create a sense of linguistic security and ease when speaking English.
Another factor easing levels of anxiety in English could be the fact that the interlocutors might
often be non-native speakers themselves. Conversations in French are probably more often
conducted with native speakers, hence the inequality in the status of the interlocutors (Py,
1995). Native speakers of French might also react differently to errors in French than native
speakers of English who are notoriously more relaxed about their language (Dewaele, 1999;
Wise, 1997). A few unpleasant experiences with native speakers of French might significantly
raise the students’ levels of anxiety in that language.
A second important finding is the variable effect of personality on communicative
anxiety, especially as the latter was generally assumed to be stable. The fact that extraversion,
neuroticism, and psychoticism are significant predictors for communicative anxiety in
English but not in French is puzzling. It is also surprising that social class is a predictor for
communicative anxiety in French but not in English. Could these two findings be related?

3
According to Cohen (1992) squared partial correlations values between 2 and 12.99% suggest small effect sizes,
values between 13 and 25.99% indicate medium effect sizes, and values of 26% and greater suggest large effect sizes.

The International Journal of Bilingualism

Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
 Communicative anxiety in L2 and L3 production 33

For this we have to delve into the turbulent history of French in Flanders. French has long
been spoken by members of higher social classes in Flanders and is still the first language
(together with Dutch) for a minority of Flemings (Baetens Beardsmore, 1993; Dewaele,
2000; Willems, 1997). The hostility of many Flemings towards French is not so much the result
of the present— sometimes strained — relationship with the Walloons and the French speakers
of Brussels within the federal kingdom of Belgium, but rather the perception that French was/is
socially superior and that speaking French in Flanders is interpreted as a sign of ostentation
and of disregard for Dutch. Flemings remember the time well when not being able to speak
French fluently was a cause of ridicule and a bar to social and professional promotion. One
can assume that until a few decades ago, Flemings did in fact experience higher levels of anxiety
when speaking French if they were not fluent in it. Since Flemings from higher social classes
were more likely to know that language well, they would have had much lower levels of
anxiety. Less privileged Flemings, on the other hand, would have experienced high levels of
language anxiety because of their lower fluency. This would have resulted in a wider social
gap with their French-speaking interlocutor. Could this old, socially motivated language
anxiety have trickled down the generations? If that were the case, it would have to be at an
unconscious level because the participants reported overwhelmingly positive parental atti-
tudes to both English and French (98 out of 100 for both languages)4.
The small and medium effect sizes of our independent variables on communicative
anxiety in French and English suggest that communicative anxiety is a highly complex, multi-
dimensional concept. Causes for anxiety in one language do not necessarily have the same weight
in another language. The effects of extraversion, psychoticism and neuroticism on language
anxiety might not be significant in French, but the pattern is similar to the one in English.
The anxious foreign language speaker is nearer the introvert end of the continuum,
which follows logically from the observation that introverts tend to be reserved, quiet, and
unassertive in contrast to the more outgoing and talkative extraverts (Furnham & Heaven, 1998).
The extraverts’ more optimistic side might limit their fear of speaking a foreign language. The
degree of extraversion is moreover linked to self-perceived competence in English (r = .19,
p < .06), which is linked negatively to communicative anxiety in English (r = – .55, p < .00).
This does indeed confirm earlier findings, that is, that anxious foreign language students tend
to underestimate their competence relative to less anxious students, who tend to overestimate
their competence (MacIntyre, Noels, & Clément, 1997).
While we do not doubt the independent nature of general trait anxiety and language
anxiety, our results show a moderate positive relationship between both5. Those speakers
who scored higher on the N scale (showing signs of anxiety, phobia, depression, and hypo-
chondriasis) also tended to report higher levels of language anxiety in English. This result

4
This highly surprising finding will be investigated in more detail in a separate study on the students’ attitudes and
motivations.
5
A similar relationship exists between anxiety and introversion (Eysenck, 1981) although both dimensions are independent:
the anxiety dimension, as measured by tests such as the Manifest Anxiety Scale ( … ), lies within the ( … )
space defined by introversion-extraversion ( … ), correlating approximately + 0.3 to + 0.4 with the intro-
version end of introversion-extraversion dimension. (p. 167)
Dewaele (2002) argues that anxiety and introversion both affect catecholamine activity (i.e., levels of norepinephrine
and dopamine) which is linked to capacity and/or efficiency of the short-term memory.

The International Journal of Bilingualism

Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
34 J.-M. Dewaele

contradicts earlier findings of the absence of a link between emotional stability and language
anxiety (MacIntyre & Charos, 1996). It thus seems that “the social and communicative
demands of L2 interaction drive language anxiety” (1996, p. 19), but that a predisposition to
nervousness can reinforce the level of language anxiety. Higher levels of general trait anxiety
(and language anxiety) do not seem to deter students from choosing language majors (De
Fruyt & Mervielde, 1996).
However, Furnham and Medhurst’s (1995) finding that individuals with low scores on
the P scale were more likely to have good verbal skills does not mean that they are free of
language anxiety. Low scores on the P scale are indicative of good interpersonal relations
and a warm and friendly disposition; it also means being attentive to the needs of inter-
locutors and hence being worried that one’s own delivery in a foreign language might not be
good enough. People with higher scores on the P scale are colder and more hostile, so it
seems logical that they should experience less communicative anxiety in a foreign language
because they care less whether or not they have been properly understood.
Our findings concerning the variable effects of sociodemographic and psychological
correlates of communicative anxiety in French and English have important implications for
the teaching of these languages in Flanders. While all formal language instruction should take
place in a low-threat environment, the teaching of French would benefit most from this more
relaxed approach. Speaking French is perceived as being a more anxiety-provoking activity,
especially for students from lower social classes, than speaking English. Teachers of French
need to convince their Flemish students that French lost its function as social marker decades
ago and that a less than native proficiency in French has no social importance. They need to
show that French is “as easy,” as interesting and as useful as English. To do this, French
teachers need to overcome a number of major obstacles. Unlike English, which is inextri-
cably linked to youth culture on radio (pop music), television (MTV, popular British and
American soap series like the Simpsons), in cinemas (English-speaking films in original
version with subtitles) and on the Internet, contact with French and French culture is mostly
restricted to French classes at school. The main source of information about anything French
is the student’s language-teaching manual. The French textbooks used in Flemish schools follow
the communicative competence models as set out in the official curriculum (Leerplan, 1997a;
1997b) but offer little more than stereotypical speech acts on how to buy bread and find
your way to the station, reinforcing cultural stereotypes along the way. These books are
incredibly boring, uninspiring, bland, totally lacking in authenticity, and they probably rein-
force the students’ resentment and language anxiety in French (Dewaele & Dewaele, 2000).
We would argue that this is the result of stifling rules and directives in the Flemish curriculum
that severely limit the freedom of authors and discourage publishers from accepting projects
that do not strictly conform to the rules (in, so far, a very small market). Similar concerns
have been voiced for the teaching of German in Flanders (Sercu, 2000). Dewaele and Dewaele
(2000) plead for a radical change in the Flemish curriculum that would allow publication of
French language manuals with a wide variety of authentic material (French rap, popular
film, comics, rhymes, as well as poems, extracts from plays etc.) from a wider variety of
countries (Québec, French-speaking African countries, Switzerland, Belgium… ). The societal
causes for the higher levels of communicative anxiety reported in French can only be reme-
died by changes in educational policy. These changes will be deemed to have been successful
The International Journal of Bilingualism
Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
 Communicative anxiety in L2 and L3 production 35

when communicative anxiety in French is predicted only by personality variables, and no longer
by social class.

7 Conclusion
We have established that foreign language anxiety might in fact not be a stable personality
trait among experienced multilingual language learners. Indeed, anxiety levels were found
to be quite different for L2 and L3 speech production. Both the societal and the individual
contexts were found to determine levels of communicative anxiety. The perception of French
as the former prestige language in Flanders and its function as a social marker were found
to be linked to the participants’ social class. The higher level of anxiety in French — but not
English— experienced by participants from lower social classes could be seen, in the absence
of a better explanation, as the left-overs of a bygone era when the quality of one’s French reflected
one’s social and professional standing. This social effect appeared to be a stronger predictor
of communicative anxiety in French than the three personality variables (extraversion, neuroti-
cism, and psychoticism). Psychoticism, extraversion, and, to a lesser extent, neuroticism did
predict communicative anxiety in English.
These findings raise a number of interesting questions and provide useful directions
for future studies of foreign language anxiety. To begin with, future research needs to compare
levels of communicative anxiety in speakers who are at least trilingual. Future studies also
need to look at the variable effects of societal and individual factors on foreign language anxiety.
Also, the issue of the “hidden” perception of French as a socially superior language by
members of lower social classes in Flanders merits further investigation. We hope that we
have provided a useful first step in this research by showing that foreign language anxiety
is an extremely complex phenomenon, resulting from an intricate mix of societal and indi-
vidual factors. Finally, this study also has important educational implications. We have argued
that changes need to be implemented in the official Flemish school curriculum in order to
make language courses more attractive and more interesting to students, as this seems a
necessary prerequisite for better teaching and learning, which would, in turn, boost self-
confidence and reduce levels of foreign language anxiety.

Received: July, 2000; revised: January, 2001; accepted: February, 2001

The International Journal of Bilingualism

Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
36 J.-M. Dewaele

References
AIDA, Y. (1994). Examination of Horwitz, Horwitz and Cope’s construct of foreign language
anxiety: The case of students of Japanese. The Modern Language Journal, 78, 155 – 168.
ARGYLE, M. (1994). The psychology of social class. London: Routledge.
BAETENS BEARDSMORE, H. (1993). Belgium. In U. Ammon, K. J. Mattheier & P. Nelde (Eds.),
Conceptions plurilingues dans l’enseignement européen, Sociolinguistica, 7, 12 – 21.
BAILEY, P., ONWUEGBUZIE, A., & DALEY, C. E. (2000). Using learning style to predict foreign
language achievement at the college level. System, 28, 125 – 133.
BAKER, C. (1992). Attitudes and language. Clevedon: Multilingual Matters.
CHEEK, J. M., & BUSS, A. H. (1981). Shyness and sociability. Journal of Personality and Social
Psychology, 41, 330 – 339.
CHENG, Y., HORWITZ, E. K., & SCHALLERT, D. L. (1999). Language anxiety: Differentiating
writing and speaking components. Language Learning, 49, 417 – 446.
COHEN, J. (1992). Statistical power analysis for the behavioral sciences. New York: John Wiley.
De FRUYT, F., & MERVIELDE, I. (1996). Personality and interest as predictors of educational
streaming and achievement. European Journal of Personality, 10, 405 – 425.
DEPUE, R. A., & COLLINS, P. F. (1999). Neurobiology of the structure of personality: Dopamine,
facilitation of incentive motivation, and extraversion. Brain and Behavioral Sciences,
22, 491 – 569.
DEWAELE, J.-M., (1999). Is it the corruption of French thought processes that purists fear? A
response to Henriette Walter. Current Issues in Language and Society, 6, 231 – 234.
DEWAELE, J.-M., (2000). Néerlandophones— Francophones en Belgique: Liberté, égalité, fraternité?
In E. Labeau (Ed.), France-Belgique: Des ennemis de la langue de chez nous? (pp.39 – 47).
Québec: International Center for Research on Language Planning, Université Laval.
DEWAELE, J.-M., (2002). Individual differences in L2 fluency. The effects of neurobiological
correlates. In V. Cook (Ed.), The nature of the L2 user (pp.219– 250). Clevedon: Multilingual
Matters.
DEWAELE, J., & DEWAELE J.-M. (2000). L’invasion de la banalité dans les manuels de français en
Flandre. In J. Eichperger (Ed.), Roeland: Dynamiek en talen voor de jeugd, dynamics and
languages for youth (pp. 209 – 211). Gent: Roeland v. z.w.
DEWAELE, J.-M., & FURNHAM, A. (1999). Extraversion: The unloved variable in applied
linguistic research. Language Learning, 49, 509 – 544.
DEWAELE, J.-M., & FURNHAM, A. (2000). Personality and speech production: A pilot study of
second language learners. Personality and Individual Differences, 28, 355 – 365.
DEWAELE, J.-M., & FURNHAM, A. (2001). Personality and the use of colloquial words among non-
native French speakers. Unpublished paper, University of London.
DÖRNYEI, Z., & KORMOS, J. (2000). The role of individual and social variables in oral task
performance. Language Teaching Research, 4, 275 – 300.
ELY, C. M. (1986). An analysis of discomfort, risktaking, sociability, and motivation in the L2
classroom. Language Learning, 36, 1 – 25.
EYSENCK, H. J. (1979). Anxiety, learning and memory: A reconceptualization. Journal of Research
in Personality, 13, 363 – 385.
EYSENCK, M. W. (1981). Learning, memory and personality. In H. J. Eysenck (Ed.), A model for
personality (pp. 169 – 209). Berlin: Springer Verlag.

The International Journal of Bilingualism

Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
 Communicative anxiety in L2 and L3 production 37

EYSENCK, H. J., & EYSENCK, M. W. (1985). Personality and individual differences. A natural
science approach. New York and London: Plenum Press.
EYSENCK, S., EYSENCK, H. J., & BARRETT, P. (1985). A revised version of the Psychoticism
scale. Personality and individual differences, 6, 21 – 29.
FREMONT, T., MEANS, G. H., & MEANS, R. S. (1976). Anxiety as a function of task performance
feedback and extraversion-introversion. Psychological Reports, 27, 455 – 458.
FURNHAM, A. (1990). Language and personality. In H. Giles & W. P. Robinson (Eds.), Handbook
of language and social psychology (pp. 73 – 95). Chichester: John Wiley and Sons.
FURNHAM, A., & HEAVEN, P. (1998). Personality and social behavior. London: Arnold.
FURNHAM, A., & MEDHURST, S. (1995). Personality correlates of academic seminar behavior: A
study of four instruments. Personality and Individual Differences, 19, 197 – 208.
GANSCHOW, L., & SPARKS, R. (1996). Anxiety about foreign language learning among high
school women. Modern Language Journal, 80, 199 – 212.
GARDNER, R. C. (1985). Social psychology and second language learning: The roles of attitudes and
motivation. London: Edward Arnold.
GARDNER, R. C., MOORCROFT, R., & MacINTYRE, P. D. (1987). The role of anxiety in second
language performance of language drop-outs (Research Bulletin No 657). London, Ontario:
The University of Western Ontario, Department of Psychology.
HORWITZ, E. K. (1986). Preliminary evidence for the reliability and validity of a foreign language
anxiety scale. TESOL Quarterly, 20, 559 – 562.
HORWITZ, E. K. (1996). Even teachers get the blues: Recognizing and alleviating language
teachers’feelings of foreign language anxiety. Foreign Language Annals, 29, 365 – 372.
Leerplan Secundair Onderwijs: Frans (1997a). Onderwijssecretariaat van de Steden en Gemeenten
van de Vlaamse Gemeenschap v. z.w., Brussel: OVSG.
Leerplan Secundair Onderwijs: Frans Eerste Graad (1997b). Vlaams Verbond van het Katholiek
Secundair Onderwijs, Brussel: Licap.
LIEBERMAN, M. D. (2000). Introversion and working memory: Central executive differences.
Personality and Individual Differences, 28, 479 – 486.
MacINTYRE, P. D. (1999). Language anxiety: A review of the research for language teachers. In D.
J. Young (Ed.), Affect in foreign language and second language teaching: A practical guide to
creating a low-anxiety classroom atmosphere. Boston: McGraw-Hill.
MacINTYRE, P. D., & CHAROS, C. (1996). Personality, attitudes, and affect as predictors of second
language communication. Journal of Language and Social Psychology, 15, 3 – 26.
MacINTYRE, P. D., & GARDNER, R. C. (1989). Anxiety and second language learning: Towards a
theoretical clarification. Language Learning, 39, 251 – 275.
MacINTYRE, P. D., & GARDNER, R. C. (1991a). Methods and results in the study of foreign
language anxiety: A review of the literature. Language Learning, 41, 25 – 57.
MacINTYRE, P. D., & GARDNER, R. C. (1991b). Language anxiety: Its relationship to other
anxieties and to processing in native and second language. Language Learning, 41, 513 – 534.
MacINTYRE, P. D., & GARDNER, R. C. (1994). The subtle effects of language anxiety on cognitive
processing in the second language. Language Learning, 44, 283 – 305.
MacINTYRE, P. D., & NOELS, K. A. (1996). Using social-psychological variables to predict the use
of language learning strategies. Foreign Language Annals, 29, 373 – 386.
MacINTYRE, P. D., DÖRNYEI, Z., CLÉMENT, R., & NOELS, K. A. (1998). Conceptualizing
willingness to communicate in a L2: A situational model of L2 confidence and affiliation.
Modern Language Journal, 82, 545 – 562.

The International Journal of Bilingualism

Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014
38 J.-M. Dewaele

MacINTYRE, P. D., NOELS, K. A., & CLÉMENT, R. (1997). Biases in self-ratings of second
language proficiency: The role of language anxiety. Language Learning, 47, 265 – 287.
MARKHAM, R., & DARKE, S. (1991). The effects of anxiety on verbal and spatial task
performance. Australian Journal of Psychology, 43, 107 – 111.
MATTHEWS, G., & DORN, L. (1995). Cognitive and attentional proceedings in personality and
intelligence. In D. H. Saklofske & M. Zeidner (Eds.), International handbook of personality and
intelligence (pp. 367 – 396). New York: Plenum Press.
NAIMAN, N., FROHLICH, M., STERN, H. H., & TODESCO, A. (1978). The good language learner.
Toronto: Ontario Institute for Studies in Education.
ONWUEGBUZIE, A., BAILEY, P., & DALEY, C. E. (1999). Factors associated with foreign
language anxiety. Applied Psycholinguistics, 20, 217– 239.
ONWUEGBUZIE, A., BAILEY, P., & DALEY, C. E. (2000). The validation of three scales measuring
anxiety at different stages of the foreign language learning process: The input anxiety scale, the
processing anxiety scale, and the output anxiety scale. Language Learning, 50, 87 – 117.
PHILLIPS, E. M. (1991). Anxiety and oral competence: Classroom dilemma. French Review,
65, 1 – 14.
PHILLIPS, E. M. (1992). The effects of language anxiety on student oral test performance and
attitudes. The Modern Language Journal, 76, 14 – 26.
PRESTON, D. (1989). Sociolinguistics and second language acquisition. Oxford: Basil Blackwell.
PY, B. (1995). Quelques remarques sur les notions d’exolinguisme et de bilinguisme. Cahiers de
praxématique, 25, 79 – 95.
SAITO, Y., & SAMIMY, K. K. (1996). Foreign language anxiety and language performance: A study
of learner anxiety in beginning, intermediate, and advanced-level college students of Japanese.
Foreign Language Annals, 29, 239 – 251.
SCOVEL, T. (1991). The effect of affect on foreign language learning: A review of the anxiety
research. In E. K. Horwitz & D. J. Young (Eds.), Language anxiety from theory and research to
classroom implications (pp. 15 – 23), Englewood Cliffs, NJ: Prentice Hall.
SERCU, L. (2000). Acquiring intercultural competence from textbooks. The case of Flemish students
acquiring German. Leuven: Acco.
SPARKS, R., ARTZER, M., GANSCHOW, L., SIEBENHAR, D., PLAGEMAN, M., & PATTON, J.
(1998). Differences in native language skills, foreign-language aptitude, and foreign-language
grades among high-, average-, and low-proficiency foreign language learners: Two studies.
Language Testing, 15, 181 – 216.
WILLEMS, D. (1997). Le français en Flandre. In D. Blampain et al. (Eds.), Le français en
Belgique: Une langue, une communauté (pp. 261 – 272). Louvain-la-Neuve: Duculot.
WISE, H. (1997). The vocabulary of modern French. Origins, structure and function. London:
Routledge.
YOUNG, D. J. (1990). An investigation of students’ perspectives on anxiety and speaking. Foreign
Language Annals, 23, 539 – 553.
YOUNG, D. J. (1991). Creating a low-anxiety classroom environment: What does language anxiety
research suggest? The Modern Language Journal, 75, 426 – 439.

The International Journal of Bilingualism

Downloaded from ijb.sagepub.com at Nat. Taichung Univ. of Sci. & Tech. on April 22, 2014