INTROVERSION-EXTRAVERSION A N D THE EFL

PROFICIENCY OF JAPANESE STUDENTS

Deborah Busch
Cornell University'

This study explores the relationship between the introversion-extraversion tendencies

of Japanese studentsand their proficiency in English a s a foreign language(EFL). I t was
hypothesized that in an E F L situation. extraverted students would attain a higher
proficiency in English because they may takeadvantage oft he few available opportunities
t o receive input in Englishand practice the language with nativespeakers. Inordertotest
the hypothesis that extraverts are more proficient in English. 80 junior college English
students and 105 adult school English students took a standardixd English test.
completed a personality questionnaire. and provided information o n a biodata form. In
addition. 45 of the junior collegc students participated in English oral interviews which
were then rated for proficiency by two evaluators.
The hypothesis that extraverts are more proficient in English was not supported. In
fact. statistical analysis revealed that extraversion had a significant negativecorrelation
with pronunciation. a subcomponent of the oral interview test. In addition, introverts
tended to have higher scores on the readingand grammar component s o f t hestandardized
English test.
Even though introverts tended to score better on most of the English proficiency
measures. it was found that junior college males who had tendencies towards extraversion
had higher oral interview scores. Extraversion also correlated positively with length of
time spent studying English at theadult school. These findings are discussed with respect
to cultural factors predominant in Japanese society and psychological theory.

The assumption that there is a relationship between extraversion and

proficiency in a foreign language is widely held by teachers, researchers,
and students of second languages. For example, 31% of the students who
were considered to be good language learners in the study by Naiman,
Frohlich, and Stern (1975) stated that extraversion was helpful in acquiring
oral skills. Several researchers have propounded hypotheses as to why
extraverts are probably better language learners (especially in acquiring
oral skills). Rubin (1975) describes the good language learner as having a
strong desire to communicate which, according t o psychologists, is also
considered t o be a characteristic of extraverted people. She further suggests
that if good language learners are successful in communicating, their
motivation t o acquire the necessary tools t o continue communicating will
increase. This implies that motivation is triggered by success in the second

' A preliminary version of this paper was completed while the author was a master's degree
candidate at the University of California. Los Angeles.

109
 I10 Language Learning Vol. 32, No. 1

language, not success resulting from motivation. Rossier ( 1975) supports

this position in his contention that although students may be motivated to
learn a language and integrate with a certain language group, they may not
be able to implement this desire efficiently if they d o not have an intense
desire to communicate and interact in general. This, he says, may explain in
part why certain students who appear to be highly motivated fail to attain
proficiency in a foreign language. In such cases, personality factors may
actually inhibit individuals from fulfilling their desires. For example,
extraverted people may have an obvious advantage over their introverted
counterparts in learning a second language because they create more
opportunities to practice the language, obtain input, and experience
success in communicating (Rossier 1975, Rubin 1975, Seliger 1977). The
introverted person may rarely create such situations, and thus opportuni-
ties to practice, obtain input, and experience success in the language would
be limited.
Other researchers have questioned the assumption that extraverts are
better foreign language learners. Brown (1973:236) states that “it is a
common belief among teachers in general, particularly in Western society,
that introversion is an undesirable behavior,” and that it is better to be an
outgoing, talkative person. In language classrooms, quiet, reserved
personalities are often treated as problems and language teachers “seek
ways of encouraging extraversion.” However, he questions whether
introverts are qualitatively less proficient than extraverts. Although many
people intuitively believe that extraverts are probably better at learning
languages, it could be conjectured that introverts might be better at
learning to read and write in a second language.
Psychologists have written volumes on the subject of introversion-
extraversion, and it is necessary to define the concept as it is used in this
paper. Although there are minor differences between the various opera-
tional and theoretical definitions, the basic construct remains the same.
The most highly developed theory comes from Hans Eysenck (Morris
1979), who contends that the basic difference between extraverts and
introverts is biological, rooted in the reticular activating system of the
brain. This system, which monitors incoming neural impulses resulting
from environmental stimulation, controls the arousal level of the cortex of
the brain. Introverts are held to have higher levels of cortical arousal
compared with extraverts. These different arousal levels cause introverts
and extraverts to have different behavioral and attitudinal preferences and
tendencies. Assuming that both groups function best at a moderate level of
 arousal, extraverts tend to seek stimulation from the environment to
increase arousal level while introverts attempt to seek a reduction of
stimulation. The behavioral differences are such that extraverts seek out
the presence of other persons, enjoy social activities and talking, tend to act
aggressively and impulsively, and crave excitement. O n the other hand,
introverts learn social inhibitions since social situations are most likely to
be overstimulating for them. They tend to be introspective, reserved,
unimpulsive, unaggressive, and prefer reading to talking with people
(Eysenck and Eysenck 1968). To measure these characteristics, Eysenck
and Eysenck have developed several personality measures which are
discussed in the studies below. (Incidentally, the spelling of extraversion
used in this paper is adopted from Eysenck.)
Although hypotheses have been proposed about the relationship of
introversion-extraversion to foreign language learning, very few studies
have been undertaken to investigate this intriguing topic. Of the few, the
studies done by Naiman, Frohlich, and Stern (1975) and Naiman,
Frohlich, Stern, and Todesco (1978) respresent the most ambitious
attempts to identify the characteristics of good second language learners.
The rationale for investigating the role of personality in second language
learning (in this case, French) was to learn whether certain personality
characteristics might affect learning strategies and outcomes, i.e., second
language proficiency. Using the Eysenck Personality Inventory (1968) to
measure the introversion-extraversion tendencies of junior and senior high
school students, they found no significant correlations between extra-
version and second language proficiency as measured by a battery of
standardized tests. They did, however, find that certain types of extra-
verted behavior, such as calling out answers and handraising, correlated
positively with second language proficiency. They also found that overall
classroom personality, e.g., tendencies for a student to call out answers or
remain silent, did not correlate significantly with any of the personality
tests. The explanation for this was that the personality tests might not be
measuring what they purport to measure, or that classroom personality
might be different from overall personality.
Another study which utilized one of Eysenck’s personality scales was
undertaken by Rossier ( 1975) to determine whether introversion-extra-
version was a significant variable in the learning of English as a second
language by Spanish speaking high school students in the United States.
Rossier (197.55) hypothesized that “certain personality components
influence the ESL (English as a Second Language) learner’s ability to be
I I2 Language Lmrning Vol. 32, N o . I

exposed to authentic English language behavior and therefore influence his

progress in learning the language.” His study examined the relationship
between introversion-extraversion, contact with English speakers, and
English proficiency as measured by a battery of standardized tests, for
example, the Iowa Test. A positive correlation was found between
extraversion as measured by the Eysenck Personality Questionnaire and
oral English fluency as judged by three raters when variables representing
the written aspects of English and the length of time spent in the United
States were controlled. However, he never stated whether oral fluency
refers to grammatical correctness, ability to communicate, o r to speed and
lack of hesitation in speaking. I t could be hypothesized that extraverts,
because of their impulsiveness, speak more but also make more mistakes.
Using a different instrument-the Marlowe-Crowne Reserved vs.
Outgoing Personality Scale-Chastain ( 1975) found that a n outgoing
personality was one of the factors most significantly related t o the final
course grade of American students studying French, German, and Spanish
in high school. However, whether the grade in the course is synonymous
with proficiency is open t o question: The teacher may have been rewarding
the more extraverted students.
Besides the studies of Naiman et al. (1975, 1978), the only other
published study that has relied on classroom observation rather than on
personality scales t o determine levels of introversion-extraversionis that of
Seliger (1977). In order t o test the hypothesis that students who initiate
language interactions (High lnput Generators, or HIGs) are higher
achievers in second language learning, Seliger devised a n experiment in
which six students were observed in a classroom situation. He found that
the HlGs scored significantly higher than the Low lnput Generators
(LIGs-students who are passive in language interaction situations) on the
final exam, but not on the pretest. In addition, he found that HlGs had
more contact with the second language outside of the classroom. Thus, he
concluded, active learners who utilize all opportunities-both formal and
informal-to learn the second language develop their second language at a
faster rate.
The results of the studies using introversion-extraversion scales are
inconclusive, perhaps due t o the fact that so few studies have been
undertaken up until now. Several of the studies involving the scales
(Chastain 1975, Naiman et al. 1975, 1978, Rossier 1975) have dealt with
speakers of European languages learning other European languages as a
second language. When it is considered that national levels of extraversion
 Busch 113

are higher for Western countries than for a country such as Japan
(Iwawaki, Eysenck, and Eysenck 1977, Lynn and Hampson 1975), it is
conceivable that extraversion might not be as significant a variable for
Westerners as it might be for the Japanese. The fact that Japanese students
have the lowest average scores when compared t o other groups on
standardized tests of English proficiency (Chastain 1975, Farhady 1978,
lrnamura 1978) seems t o support this hypothesis. In countries where
introversion levels are high, the “cultural and social barriers” preventing a
person from going out and getting input in the second language may be far
greater. These barriers may becompounded in a situation where the second
language is not spoken outside of class, as in a n English as a foreign
language (EFL) situation. In a n English a s a second language (ESL)
situation, even introverted students may often find themselves in situa-
tions where they must communicate in English. R o s i e r ( 1975) reports that
such situations are important for the learning of a second language.
However, in a n E F L situation, introverts may get the opportunity t o
practice only in the classroom. Since there is rarely an instance where
introverted students feel a n overwhelming desire to communicate in a n
ESL (or EFL)situation(Rubin 1975),it would seem that they are usuallyat
a disadvantage when it comes to creating opportunities to practice the
second language unless they can take advantage of the input generated by
the more extraverted students (Larsen-Freeman 1980, personal com-
munication). Extraverted students, on the other hand, will communicate
n o matter where they are (Rubin 1975).
The rationale for undertaking this study in a n E F L environment is that
under the conditions described above, overall introversion-extraversion
tendencies as measured by a personality inventory may produce significant
correlations with proficiency in English as a foreign language. In a n EFL
situation such as exists in Japan, it would seem that extraverted students
would attain a higher proficiency in English, especially oral English,
because they take advantage of the few opportunities they get t o receive
input in English and practice the language both inside and outside the
classroom (e.g., initiating contacts with foreigners). The purpose of this
study, then, was t o learn whether Japanese who have general tendencies
towards extraversion, defined as high sociability and impulsivity, tend to
be more proficient in English as measured by a standardized proficiency
test and a n oral interview. Based upon this, the research questions were:
I . Is there a relationship between the introversion-extraversion ten-
dencies of Japanese learners of English and their performance on a
114 Language Learning Vol. 32, No. 1

standardized English proficiency test consisting of four subtests?

2. Is there a relationship between the introversion-extraversion ten-
dencies of Japanese learners of English and their performance o n a n
oral interview test?

METHOD

Two groups of Japanese students were randomly selected t o participate

in this study. One group consisted of students from the Kobe YMCA
Junior College; the other group consisted of students from the Kobe
YMCA Night School. The two groups differed in terms of age, motivation,
occupation, and educational attainment. All 185 subjects took a profi-
ciency exam that was developed for the YMCAs in Japan, and answered
the questions on the Eysenck Personality Inventory (1968). In addition, 45
students from the junior college group were randomly selected to
participate in oral interviews conducted by the author of this paper. The
interviews were taped and then rated for English proficiency by two
different judges.

Subjects

The first group of subjects consisted of 80 junior coliege students, 17

males and 63 females, ranging in age from 18 to 24. All had had six years of
English taught by the grammar translation method in thejunior and senior
high school. At the YMCA the students receive two to three hours of
English instruction daily in addition to business courses such as typing and
trade regulations. The function of the junior college is to prepare students
for occupations in Japanese businesses; however, the students’ reasons for
studying English are varied. Many of them attend the YMCA as a n
alternative to four-year colleges because gaining admission into universi-
ties is very difficult. Many students hope that studying English will help
them find jobs as secretaries in business and trading companies, most of
which require some knowledge of English. Some plan to study abroad and
are learning English in order to converse with English speaking people.
Still others could not think of a reason for studying English.
In addition to the junior college group, 105 subjects were selected f r o m a
group of 300 YMCA Night School students ranging in age from 17 to 69.
Like the junior college students, the night school group had had six years of
 Busch 115

English in junior a n d senior high school. In addition, some had studied

conversational English a t other schools or a t the YMCA. The students,
who a r e businessmen, housewives, four-year college students, and teach-
ers, attend English classes a t night for six hours per week during fifteen-
week terms. T h e night school emphasis is o n speaking and listening skills
taught primarily by the audio-lingual method. Seven levels of instruction
are offered, ranging from elementary t o advanced. Students entering a t the
elementary level can expect t o progress through all leyels in about three
years after which they are awarded a certificate of graduation. The classes
are taught by both Japanese a n d native English speaking teachers. A cross-
section of the seven levels was obtained for this study.

Instruments and procedures

The instruments used in this study were the Eysenck Personality

Inventory, the YMCA Test of English Proficiency, a n oral interview, a n d a
biodata form. Further description and the rationale for the use of each
instrument are given below.
The Eysenck Personality Inventory (EPI) measures two major per-
sonality dimensions, introversion-extraversion and neuroticism-stability,
which, according t o Eysenck, account for most of the variance in the
personality domain. In this study, only the introversion-extraversion scale
was used, although the neuroticism scale was administered as well. The
extraversion scale measures individual levels of impulsivity and sociability,
as these two factors are believed to be the main subcomponents of
extraversion.
The EPI was chosen because it is easy to administer (it requires only ten
minutes) and is easy to score. In addition, although it was developed for use
with British and American subjects, it has been validated in other
countries. The Eysenck Personality Questionnaire was used in J a p a n
(Iwawaki et ak. 1977), but since it is not a s well validated as the EPI, it was
not chosen for this study. Also, there are only minor differences in the
extraversion scales of the two instruments. In addition, the EPI has been
used in other studies that have investigated the relationship between
English as a second language proficiency and introversion-extraversion
(Naiman et al. 1975, 1978, Rossier 1975) and can provide a good basis for
comparison.
In order to avoid language problems, the EPI was translated-back-
translated three times by three different Japanese-English bilinguals using
116 Language Learning Vol. 32, No. 1

one of the procedures described by Brislin (1970). After two back-

translated versions of the EPI were obtained, both versions were compared
by a fourth bilingual. When there was a discrepancy in meaning between
the two versions, the best translation was determined by comparing both
versions with the original English. Although the EPI was developed for a
British/ American population, the Japanese translators felt that Japanese
students would have no difficulty relating their experiences to the questions
asked on the EPI. The final form, which included biodata information, was
then administered to the students during class time by regular teachers.
The reliability coefficient of the extraversion scale for both the night
school and the junior college was alpha .7 1, slightly lower than the split half
coefficient of .75 reported for American college students (Eysenck and
Eysenck 1968). Since the EPI was not developed for a Japanese popula-
tion, the lower reliability coefficient is not surprising.
The biodata information that was collected included age, sex, occupa-
tion, years of English study, months of study at the YMCA, level at the
YMCA, and amount of time spent in English speaking countries. A few of
the subjects did not provide all of the information requested.
The YMCA Test of English Proficiency (Japanese YMCA Language
Education Standardization Committee 1979) was developed as a place-
ment and proficiency test for the YMCA population in Japan and has been
validated on approximately 6000 students in various YMCAs throughout
Japan. There are four sections of the test. Part 1 consists of 60 discrete-
point multiple-choice items that test vocabulary and grammar. Part 2 is a
50-item multiple-choice cloze passage. The third part is an aural compre-
hension test having 180 points, and the fourth subtest is a dictation having
110 points. Since the YMCA English programs emphasize speaking and
listening skills, the YTEP has less emphasis on writing, reading, and
grammar. Descriptive statistics of the YTEP for all Japanese YMCAs as
compared with the Kobe YMCA population are given in Table I . The
average scores for the Kobe YMCA are nearly the same as the average for
the entire YMCA population.
Oral interviews were administered to 45 randomly selectedjunior college
students. The first six interviews were treated as practice interviews to
allow the interviewer to develop practice in the interviewing technique.
These six interviews were eliminated as part of the data for this study; thus,
the sample size for the oral interviews was 39.
The interview itself consisted of three sections. In the first part, which
lasted about five minutes, the interviewer used a standard set of questions
 Busch 117

Table I
YTEP [Link].3

Japanese YMCAs Kobe YMCA

- ( N = 5939) - ( N = 398)
Max. X Range X Range
~

Gram/ Vocab 60 30.5 0-60 31.1 8-58

Reading 50 28.0 0-50 28.5 8-50
Aural I80 103.9 0-177 101.5 34- I 70
Dictation I10 59.3 0-1 10 58.7 0- 106
Total score 400 221.9 0-396 219.9 80-375
Now i%u \landard deviation information was Included: ihrrefore. n o ~iati\iicalanalysl\ could be done 10 compare the
groups.

that became progressively more difficult. All interviewees were asked the
same questions, although the discourse expanded on some topics depend-
ing upon the interviewee’s answers. Then, the interviewer asked the
interviewee t o describe three preselected pictures. In the final part of the
interview, the interviewees were asked if they had any questions. Some
subjects responded: others did not. Each interview took approximately
fifteen minutes and all interviews were recorded for future rating. In order
to reduce the students’anxiety, the tape recorder was hidden under adesk,
although the student was able t o see the microphone. The interviewer
avoided pressing any of the buttons on the tape recorder while the
interviewee was in the room.
Approximately two months after the interviews were recorded, two
raters rated all of the tapes for comprehension, pronunciation, fluency, and
grammar using a n adapted version of the rating instrument of Mullen
(1978). Before traveling to Japan, the researcher tested the instrument on
Japanese junior college students in the United States and found that it
discriminated well among students at the lower end of English proficiency.
For this reason, it was chosen for use in Japan. Using the rating instrument,
the raters judged the subjects on their English ability in the four
subcategories on a nine-point scale. The scores for each category were then
averaged to obtain each subject’s true rating. The overall oral proficiency
score was calculated by adding the averaged subtest scores. These
combined ratings were then used for all statistical analyses that followed.
The raters both had had three years of experience in teaching English as a
second language, mostly in Japan. (Rater I is the author of this paper.)
Both are native English speakers who also speak Japanese. A training
session was held in which both raters rated the six practice interviews
 118 Language Learning Vol. 32, No. I

together in order to set standards for the evaluation of the remaining

interview tatJes.
Interrater reliability for the ratings of the oral interviews is reported in
Table 2. While the correlation coefficients are not exceedingly high, they
are all significant at the .001 level and therefore the averaged scores of both
raters can be considered reliable.

Table 2
Inrerrarer reliabiliry for the oral interviews (p < ,001, N = 39)

Component Coefficient
Overall proficiency .7 I

Comprehension .63

Pronunciation .54

Fluency .66

Grammar .61
~

Uoie Peanon Product-Moment Coefficients

Rater 2 tended to give lower scores overall than did Rater 1, and the
differences are reported in Table 3. Also, Rater 2's scores did not correlate
as highly with the components of the YTEP as did the scores of Rater I
(Table 4). However, all of the combined oral interview scores had
significant positive correlations with the scores on the components of the
YTEP. As might be expected, the correlations between the reading and
grammar scores of the YTEP and the oral interview scores are substantially
lower than the correlations between the listening sections of the YTEP and
the oral interview. The fact that there is a high correlation between the
YTEP aural section and the oral interview provides further proof that the
oral interview rating instrument is providing accurate estimates of the
students' oral proficiency.
All of the data were collected duringclass timeat the Y MCA. The YTEP
and the EPI were administered by the usual classroom teachers. For the
oral interview, students were sent one by one to the interviewing room,
usually a quiet language laboratory. Before participating in the research
project, all of the students were told that a former YMCA teacher would be
conducting a research project to help Japanese students learn English. All
students had the option of refusing to participate.
 Busch 119

Table 3
Average oral interview ratingsfor both raters

Rater
- I Rater
- 2
Max. X s X S

Comprehension 9 5.8 I .8 4.0 1.1

Pronunciation 9 5.6 1.3 3.6 I .o
Fluency 9 4.6 1.5 3.2 1.1
Grammar 9 4.0 I .6 3.2 1.2
Overall 36 20.0 5.7 14.0 4. I

Table 4
Resulis of Pearson correlation between raters' scores and YTEP components (p < .001
except where indicaied)

Rater I Rater 2 Combined

Gram/ Voca b .5 I .4 I * .50
Reading .48 .44* SO
Aural .73 .57 .7 I
Dictation .69 .42* .62
Total score .78 .58 .75
'p < .01

RESULTS

Junior college findings

For both the junior college and night school data, descriptive statistics
were calculated using programs of the Statistical Package for the Social
Sciences (Nie et al. 1975). Table 5 presents the junior college descriptive
statistics for each instrument. A preliminary review of the data revealed
that there was a large difference between male and female mean extraver-
sion scores. A comparison of the male and female scores produced a t-value
of 13.2 which was significant beyond the .01 level. It should be noted that in
each year (Table 6), the males had lower average scores on the YTEP than
did females. The average YTEP scores of first year males fell below both
the Japanese national average and the Kobe YMCA average (see Table 1);
however, all other means were higher than the population averages.
Although males tended to have higher extraversion scores, they tended to
have lower YTEP scores while the females as a group were more
introverted and had higher YTEP scores.
 Table 5
Junior college descriptive statistics for all instruments

- 1st year - 2nd year - Combined

Maximum X S X S X S Range
YTEP totala 400 227.3 6.4 277.5 6.5 250.5 47.4 I4 1-348
Gram/ Vocab 60 31.4 I .4 35.7 1.8 33.4 10.1 10-56
Reading 50 29. I 1.1 31.9 I .2 30.4 7.6 12-46
Aural I80 102.4 3.0 125.5 3.3 113.1 23.2 63- I58
Dictation I10 64.3 2. I 84.5 1.7 73.7 15.9 30-106
Oral proficiency’
Total 36 14.6 .9 19.3 .9 17.0 4.5 7.5-27.5 6
Comprehension 9 4. I .3 5.7 .2 4.9 I .3 2.0-8.0 $
Pronunciation 9 4.0 .2 5. I .2 4.6 I .o 2.5-7.0 6
Fluency 9 3.4 .3 4.4 .i 3.9 I .2 1.5-7.0 %
Grammar 9 3. I .2 4. I .3 3.6 1.2 1.5-6.5
Extraversion 24 12.5 .5 13.7 .7 13.1 3.9 4-2 I L-
3
a 9’
n = 43 for firat year: n = 37 for second year 09
b
n = 19 for first year. n= 20 for second year
 Busch 121

'?
r.
r-
N

L
*o r-.
A M

-0 I1
25
122 Language Learning Vol. 32, No. I

In order to determine whether extraversion and English proficiency are

significantly related, a Pearson Correlation Matrix was constructed using
the following variables:
I . Comprehension (from oral interview)
2. Pronunciation (from oral interview)
3. Fluency (from oral interview)
4. Grammar (from oral interview)
5 . Overall Oral Proficiency (from oral interview)
6 . YTEP Grammarl Vocabulary
7. YTEP Reading
8. YTEP Aural
9. YTEP Dictation
10. YTEP Total
11. Extraversion (EPI)
Table 7 gives the results of the Pearson Correlation for extraversion and
English proficiency scores. In many cases there is an inverse relationship
between extraversion and YTEP component scores. None of these
relationships are significant at the .05 level; however, the negative
relationship between extraversion and grammar/ vocabulary, and between
extraversion and reading approaches significance (p < .07).

Table 7
Pearson correlations between EPI scores and English proficiency scores

Variables Coefficient Significance

YTEP totala .oo ,499
Gram/Vocab -.I8 ,051
Reading -.I6 ,069
Aural .09 ,213
Dictation
b
.oo ,297
Oral proficiency total -.26 ,056
Comprehension -.I7 ,140
Pronunciation -.38 ,009
Fluency -.22 ,087
Grammar -.26 ,091
a" = 80

b" = 39

For the subcomponents of the oral proficiency test, the only significant
correlation was between extraversion and pronunciation. Students with
tendencies towards introversion had better English pronunciation. All
other relationships between extraversion and oral proficiency were
 Busch 123

negative but not significant. However, Oral Proficiency Total, Pronuncia-

tion, Fluency, and Grammar nearly reached the .05 level of significance.
Thus, the findings presented here reveal a tendency for introverts to have
higher English proficiency scores.
In addition to the Pearson Product-Moment Test for significance
between extraversion and English proficiency, a chi-square test using the
Crosstabs program was run. It was thought that significant differences
between introverts and extraverts on the YTEP and oral proficiency
interview might appear if the scores on all three instruments were divided
into three categories: high, middle, and low. Thus, scores in the top 27% for
all instruments were classified as high (i.e., high extraversion, high YTEP,
and high oral proficiency). Likewise, the scores in the lowest 27% were
classified as low and the remaining 46% were designated as middle. The
results are presented in Tables 8 and 9. N o significant differences in English
proficiency as measured by the YTEP and oral interview were revealed.
Table 8
Chi-square of YTEP scores for levels of extraversion

YTEP scores
High Middle Low . Totals
Extraversion
High 4 9 6 19
Middle I1 20 8 39
Low 6 8 8 22
Totals 21 31 22 80
x' = 2.38. p = .66

Table 9
Chi-square of oral proficiency: scores f o r level3 of extraversion

YTEP scores
High Middle Low Totals
Extraversion
High 3 5 3 I1
Middle 3 II 4 18
Low 4 3 3 10
Totals 10 19 10 39
X' = 2.86. p = .58

Since there were large differences in average extraversion scores for

males and females, it was conjectured that extraversion might be a more
important variable in attaining English proficiency depending upon sex.
Therefore, Crosstabs for extraversion and oral proficiency, and extraver-
 124 Language Learning Vol. 32, No. I

sion and YTEP scores were run for males and females. The results show no
significant differences for females at different levels of extraversion (Table
10).

Table 10
Chi-square of oral proficiency1 YTEP and extraversion f o r females
~

Oral proficiency scores

High Middle Low Totals
~

Extraversion
High I 5 3 9
Middle I 4 4 9
Low 1 2 I 4
Totals 3 11 8 22
X' = .90. p = 92
YTEP scores
High 2 9 6 17
Middle 7 17 6 30
Low 4 6 6 16
Totals 13 32 18 63
X ' = 3 16. p = 53

For males, however, the Crosstabs program revealed a significant

difference in oral proficiency scores between introverts and extraverts
(Table 11). Males having higher levels of extraversion tended to have
higher oral proficiency scores. No significant differences were found
between male introverts and extraverts and performance on the YTEP.
Table I !
Chi-square of oral proficiency1 Y TEP and extraversion for males
~

Oral proficiency scores

High Middle Low Totals
Extraversion
High 2 0 0 2
Middle 2 7 0 9
Low 3 I 2 6
Totals 7 8 2 17
x?= 10.17. p = .04
YTEP scores
High 2 0 0 2
Middle 4 3 2 9
Low 2 2 2 6
Totals 8 5 4 17
X? = 2.83, p = .59
 Table 12
Night school descriptive statistics for all instruments

Level Combined
- I - 2 - 3 4 - 5 - 6 - 7
X S X S X S 2 S X S X S X S R S

I. 26.8 3.1 23.8 1.2 24.7 2.7 23.6 1.4 24.1 1.2 22.2 .7 26.7 2.6 24.6 8.0
2. 7.5 .4 8.1 .5 8.1 .4 9.3 .7 9.3 .3 9.8 .4 9.4 .3 8.7 1.8
3. 3.8 .3 6.1 .9 6.8 1.2 16.5 2.1 11.9 2.0 17.8 2.9 27.0 3.4 12.1 10.4
4. 22.4 1.3 23.6 2.1 26.9 2.5 30.9 3.2 37.7 2.3 43.7 3.8 46.2 1.3 32.2 12.3
5. 20.0 1.2 22.1 2.2 28.5 1.6 31.4 2.0 33.1 2.0 35.8 2.3 36.8 2.0 29.0 9.3
6. 66.5 3.5 89.9 3.9 97.0 5.0 97.8 5.8 120.5 5.2 129.0 5.1 127.7 6.0 101.9 28.3
7. 36.8 2.9 49.1 4.2 52.8 2.8 57.7 3.9 69.6 3.8 72.6 2.7 81.9 8.6 58.6 21.7
8. 145.7 7.0 184.8 10.2 205.2 8.1 217.8 11.8 260.9 10.8 281.2 8.1 292.5 11.7 221.6 62.0
9. 12.9 1.1 12.4 .9 14.2 1.0 13.3 1.0 12.9 1.0 15.3 1.4 14.6 1.0 13.6
N 18 16 17 13 15 13 13 I I5
I
?2-
I Age 6. YTEP Aural. maximum = 180
2. Year, of English study 7. YTEP Dictation. maximum = 110
3. Number of monthr at the Y M C A H. Y'I EP total score. maximum = 400
4 YTEP Vocabulary/Grammar: maximum = 60 9. EPI extraversion: maximum = 24
5 YTEP Reading; maximum = 50

4
h,
bl
126 Language Learning Vol. 32, No. 1

Night school findings

Table 12 gives the night school descriptive statistics for all of the
instruments. The division of the data by levels shows that the average
scores on the YTEP increase by level, indicating that the YTEP distin-
guishes students according to proficiency by level. The table also reveals
differences in extraversion scores according to level.
A Pearson Correlation Matrix was constructed for the night school
and included the following variables:
I . Level at the YMCA ( 1 to 7)
2. Age
3. Years of English Study
4. Months of Study at the YMCA
5 . YTEP Grammar/Vocabulary
6. YTEP Reading
7. YTEP Aural
8. YTEP Dictation
9. YTEP Total
10. Extraversion (EPI)
Table 13 reports the correlations between extraversion and all other
variables. For all of the measures of English proficiency, there are no
significant correlations with extraversion. It should also be noted that all of
the relationships are positive. This contrasts with the negative relationships
between extraversion and YTEP sbores found in the junior college data.
Age, level, and number of months spent studying at the YMCA all had
significant but weak correlations with extraversion. Older people tended to
be more introverted. Students who had been studying at the YMCA for
longer periods of time tended to have higher tendencies towards extraver-
sion as did those who were in higher levels.
Table 13
Pearson correlation hetu-een extraversion and other variables ( N = 105)

Variable Coefficient Significance

Age -.20 .02

Level .I6 .05
Years of English -.03 .40
Months at YMCA .21 .02
Gram/Vocab .I0 .I6
Reading .I0 .on
Aural .06 .26
Dictation .03 .36
Total .on .21
 Busch 127

Since the correlation analysis produced a significant correlation between

level and extraversion, a one-way analysis of variance was programmed in
order t o determine where the differences between levels were occurring.
However, the analysis of variance showed no difference in extraversion
scores between levels (Table 14).
Table 14
Analysis of variance qj' extraversion b.s level (N = 105)

df ss MS F P
Between 6 95.3 15.9 .96 .45
Within 98 161 1.5 16.4
Total I04 1706.8

An analysis of variance was also run t o determine whether there were

significant differences in extraversion for scores for males and females in
the night school sample. The mean extraversion scores for males and
females were almost identical and no significant differences were found. I n
addition, male and female mean scores on the YTEP did not differ
significantly, although the mean for males was slightly lower. (See Table
15.)

Table I S
Night school [Link] statisriu for males and females

Extraversion YTEP
N X S X S

Males 54 13.5 4.2 215.3 52.8

Females 51 13.7 3.9 228.3 70.4
Total I05 13.6 4. I 221.6 62.0

Other significant correlations that were unrelated t o extraversion are

reported in Table 16. Level, years of English study, and months of
attendance at the YMCA all correlated significantly with scores on the
YTEP.

Table 16
Pearson correlation betHseen YTEP total scores and other variables (p < .001)

Variable Coefficient
Level .80
Years of English .42
Months at YMCA .34
 128 Language Learning Vol. 32, No. I

DISCUSSION

It is now possible to discuss the research questions posed earlier. The first
question was, Is there a relationship between the introversion-extraversion
tendencies of Japanese learners of English and their performance on a
standardized English proficiency test consisting of four subtests? The
results of this study reveal no significant relationship between introversion-
extraversion and performance on the YTEP. In addition, although not
significant, many of the relationships between extraversion and YTEP
scores were found to be negative for the junior college group. This negative
trend almost reached significance for reading and grammar.
The best predictors of YTEP test scores in this study appear to be months
of study at the YMCA, years of English study, and level at the YMCA. In
addition, length of study at the YMCA had a weak but significant
correlation with extraversion. While extraversion does not relate directly
to YTEP scores, it does appear that extraverts are more likely to continue
attending classes at the YMCA. Extraverts have a general sociability drive
which may explain why they have less tendency to discontinue classes.
Since the classes at the YMCA emphasize speaking skills, perhaps
introverted students are more inclined to drop out.
The second research question was, Is there a relationship between the
introversion-extraversion tendencies of Japanese learners of English and
their performance on an oral interview test? The Pearson Product-Moment
matrix produced only one significant correlation between extraversion and
oral proficiency. Introverts were found to be significantly better in English
pronunciation. This implies that introverts take more time to pronounce
correctly, probably because they are less impulsive. Since one of the traits
being measured by the EPI is impulsivity, it is quite possible that lack of
impulsivity covaries with good pronunciation. Another explanation for the
significant relationship between extraversion and pronunciation, but not
with other oral interview components, is that the interrater reliability was
lower for pronunciation than for all of the other components. However, an
examination of the correlations between each rater’s pronunciation ratings
and extraversion scores reveals significant relationships (for Rater 1, p <
.01; for Rater 2, p < .05). Thus, it seems unlikely that low interrater
reliability can account for the significant correlations between extraversion
and pronunciation.
One other significant finding was that extraverted junior college males
had higher oral proficiency scores. The researcher hesitates to draw
 Busch 129

conclusions from this finding because of the fact that the sample size of
males was very small, only 17 subjects, compared to 63 females. However,
this finding does seem to suggest that introversion-extraversion tendencies
may have importance for the learning of spoken English for males but not
for females. An explanation for this is discussed later in this paper.
Although this study has produced some significant relationships be-
tween extraversion and EFL proficiency, these findings and lack of
correlations for other variables require further explanation. One factor
that may account for the negative relationship between extraversion and
English proficiency is that personality factors that contribute to overall
academic achievement probably apply to foreign language learning also.
Researchers have found that introverts tend to do better in high school and
college (Cattell and Cattell 1969, Entwistle 1972, Kelly 1973, Power 1977).
The results of the junior college study tend to support the view that
introverts do better academically (e.g., junior college introverts tended to
d o better in English grammar and reading).
Another explanation for the overall lack of significant correlations
between extraversion and English proficiency, in particular oral skills, is
that there is no proof that by practicing speaking a foreign language, one
becomes more proficient. I t may be that the extent of second language
input is the most important factor. In a classroom, for example, extraverts
may encourage more input from the teacher, but introverts and extraverts
alike may benefit equally from this input (Larsen-Freeman 1980, personal
communication). Seliger (1977), however, in his study did not find that
Low Input Generators benefited equally from the input generated by the
High Input Generators. Instead, he found that HIGs progressed in the
learning of English as a second language at a faster rate than the LIGS. He
also states that there is little reason to believe that LlGs would compensate
for their disadvantage in the classroom by seeking language contact outside
of the classroom, which he found to be the case with all of the LIGs in his
study.
Still another explanation for the lack of significant results could be that
introversion-extraversion in itself is not enough to account for much of the
variance in English proficiency scores. Rather, a combination of certain
factors would be more likely to influence a person’s success or failure in
learning a second language. In the field of psychology, “trait theory” (the
theory that stable response dispositions such as introversion-extraversion
are enduring causes of behavior, Mischel 1968) has undergone much
criticism. Rather than ascribing behavior to individual traits, proponents
130 Language Learning Vol. 32, No. I

of “situation theory” say that behavior is a function of the situation, not

personality characteristics. Still others feel that behavior is a combination
of the two. While it is beyond the scope of this paper to outline all of the
arguments in the controversy (see Ekehammer 1974), it does seem feasible
that situation together with the combination of personality characteristics
plays an important role in determining a person’s behavior. The implica-
tion for this study is that if extraversion is measured as an isolated
component, little may be learned about a person’s propensity for learninga
foreign language. Rather, introversion-extraversion along with other
personality factors must be examined in the context of the situation under
which the foreign language is learned. Whether a student is introverted or
extraverted may be specific to a given setting. For students who rely mostly
on formal instruction (i.e., the classroom) to learn a foreign language, the
situation may be a crucial factor. In Japan, role behavior is conditioned to
a strong degree. The classroom situation requires that students exhibit
introverted behavior out of respect for the teacher; thus, the situation itself
may prevent students from obtaining as much input in the foreign language
as they would in a different situation. As a result, introversion-extraversion
would have very little t o d o with language proficiency in an EFL situation
unless the classroom learning situation were changed.
The situation factor may also explain why introversion-extraversion was
a significant factor for males, but not females, in attaining oral proficiency.
Personality studies have shown that there are no basic, significant
differences in personality between Japanese men and women (DeVos and
Wagatsuma 1970). However, role behavior allows men to be more
dominant and assertive, while women have to find a means of self-
expression indirectly within the structure of the female role. The male role
may permit men to exhibit extraverted behavior in the classroom or
elsewhere. Therefore, extraverted behavior may be instrumental in assist-
ing males to learn to speak a foreign language. Women, on the other hand,
may have to find other ways of learning to speak a language within the
confines of the female role. Extraverted behavior may not be a salient
factor for women in learning to speak a foreign language because of
culturally determined roles in Japan.
The results of this study indicate that, when measured as a stable
personality construct, introversion-extraversion does not seem to have
strong correlations with EFL proficiency. Perhaps direct observation of
behavior in relevant settings is a more reliable way of determining the
relationship between personality characteristics and second language
 Busch 131

learning, if there is such a relationship. Further research would help to shed

more light on the relationship between the situation/ behavioral aspects of
introversion-extraversion and second language proficiency.

REFERENCES
Brislin, R.W. 1970. Back-translation for cross cultural research. Journal of Cross Cultural
Psychofogy 1:185-216.
Brown, H.D. 1973. Affective variables in second language acquisition. Language Learning
23:23 1-244.
Cattell. R.B., and M.D. Cattell. 1969. Handbook f o r the Junior-Senior High School Per-
sonality Questionnaire. Champaign, III.:I.P.A.T.
Chastain, K. 1975. Affective and ability factors in second language acquisition. Language
Learning 25: 153- 16 I .
DeVos, G., and H. Wagatsuma. 1970. Status and role behavior in changing Japan. In Seward,
G.H., and R.C. Williamson (eds.), Sex Roles in Changing Society. New Yoik: Random
House.
Ekehammar. R. 1974. Interactionism in personality from a historical perspective. Psycho-
logical Bullerin 8 I :I 0 2 6 1048.
Entwistle. N.J. 1972. Personality and academic attainment. British Journal oJ Educational
Psychofogy 42: 137-1 5 I .
Eysenck, H.J., and S.B. Eysenck. 1968. Manual for the Eysenck Personality Inventory. San
Diego. Calif.: Educational and Industrial Testing Service.
Farhady, H. 1978. The differential performance of foreign students o n discrete point and
integrative tests. [Link]. UCLA.
Imamura, S. 1978. Japanese still not proficient in English. Article from clippings file. Japan:
Mainichi Daily News.
Iwawaki, S., S.B.J. Eysenck, and H.J. Eysenck. 1977. Differences in personality between
Japanese and English. The Journal of Social Psychology 102:27-33.
Japanese YMCA Language Education Standardization Committee(Gogaku Kyoiku Suishin
Jinka). 1979. The YMCA Tesr of English Proficiency. Osaka, Japan: Osaka YMCA.
Kelly, R.V., Jr. 1973. Instructional objectives, learner personality, and prediction ofacademic
achievement. Doctoral Dissertation, Purdue University, 1972. Dissertation Abstracts
International 34.186A (University microfilms no. 73-15.821).
Larsen-Freeman, D. 1980. Personal communication.
Lynn, R., and S.L. Hampson. 1975. National differences in extraversion and neuroticism.
British Journal of Social and Clinical Psychology I4:223-240.
Mischel, W. 1968. Personality and Assessment. New York: Wiley and Sons.
Morris, W.M. 1979. Exrraversion andlntroversion: An Inreractiunal Perspeclive. New York:
Halstead Press.
Mullen, K.A. 1978. Rater reliability and oral proficiency evaluators. Occasional Papers in
Lingusitics, no. 3, Proceedings of the International Conference on Frontiers in Longuage
Proficiency and Dominance Testing, Southern Illinois University at Carbondale 3: 133- 142.
Naiman. N., M. Frohlich, and H.H. Stern. 1975. The GoodLanguage Learner. Toronto: The
Ontario Institute for Studies in Education.
Naiman. N., M. Frohlich. H.H. Stern, and A . Todesco. 1978. The GoodLanguage fearner.
Research in Education Series No. 7. Toronto: The Ontario Institute for Studies in Educa-
tion.
132 Language Learning Vol. 32, No. I

Nie, H., C.H. Hull, J.G. Jenkins, K. [Link] D.H. Brent. 1975. Staristical Pac,kage
f i r the Suciul Sciences. New York: McCraw Hill.
Power, C.N. 1977. Effects of student characteristics and level of student-teacher interaction
on achievement and attitudes. Cunremporary Educational Psychology 2:265-274.
Rossier, J.C. 1975. Extraversion-introversion as a significant variable in the learning of
English as a second language. Doctoral Dissertation, University of Southern California.
Dissertation Abstracts International 36:7308A-7309A.
Rubin, J . 1975. What the ‘good language learner’can teach us. TESOL Quarrer/y9:41-51.
Seliger, H.W. 1977. A study of interaction patterns and L2 competence. Condensed version of
oral presentation. Los Angeles: Los Angeles Second Language Research Forum, UCLA.
Spolsky, B. 1969. Attitudinal aspects of second language learning. Lunguuge Learning
19~27 1-283.