PERSONALITY ATTITUDES. AND
AFFECT AS PREDICTORSOF
SECONDLANGUAGE
COMMUNICATION
PETERD. MACINTYRE
University College of Cape Bretotl
CATHERINECIIAROS
Oxfutd.Unioereib
Numzrous stud,izs haoe established th2 irrwortarlce of dffectfue odliablcs, su.h as aftia.dae,
motiuation, percei.ted cornpetence,ctnd altsiet1| in pred'icting suecessin xeond' language
leatning and communication. Path arnJysb uas used'to inveetbgfu @rcldtiotue dntotug
impactontfu frequenc! ofsecond'lznglrye corrlrrr 'niaatiotu,
tl&eevaiables, toesntir&thei
dtuil to eaa,mine the role of global percotualibl trdits, Sienifiaatut pdths affectitr'g {he
frequenc! of communicdtiotu wele found, from uillir4tese to comtuttuicate itu the second'
langaage (L2), language learning matiuatinn, perceived L2 communicdtioe campetetbe,
dt d,thc opportunitxl for cotutaclwith L2 speakerc.Furthpr ftsults dernbnatldte thatglobdL
personality trai,ts and,l6neuage-reldted, dffectiue uoi.ables (su..h as attitudes end onebt!)
set tfu psychol,ogical contert for seard language cortunlttticdtiorl, Thzse resulls are
interpreted in teli8 of mod,elsof second langtdge ledrning dnd communicatioru.
Cornmunicationand secondlanguageacquisitionare closelytied
approach
together.On onehand,recenttrendstowarda conversational
to secondlanguagepedagogyreflectthe belief that onemust use the
languageto developproficiency,that is, one must talk to learn. On
the other hand, communicationis more than a means of facilitating
languageleaming, it is an important goal in itself. Whether the
orientationfor languageleaming stemsfrom the desireto meet new
people,travel, experience
otherculture8,or simplyto useit in one'6job
AI-IIHORS'NOTE: ?/r.is research was eupportcd by a postd'octoral fell'ouship ftottt the
Social Scietuceand Huma,nities Reeeatch Couruil of Canoda to thc firct duthor The
authors woul.d,lihe to etpress our grdtitud'e to Richdrd Cl'rrEtuL n. C. GardtueLdtud
the two dtuorrlrtuousreuiewert fot their inualuable commznte on previous d'rdfte of
this tTllt tuscript. Portions of thie stud,! uere presetuted'at thc dntuud corrferetuaeof the
Canad,ian Psychological Associo,tion in Charlottetown, PEI, Canadd, Jutue 1995,
Correspond.ence rcldted ta this article thould be addressed to thc fit't d1lthor clo
Ilnivereitl College ofCa.peBreton, PO Box 5300, St&et Noua Scotid, C6hada BlP 6LZ
or by E-nail to [email protected].
JOURNALOF LANGUAGEAND SOCTALPSYCHOLOGYVoL 15No. r. March 1996 3-26
Tnc.
o 1996SasePublications.
j{
JOURNAL OF LANGUAGEAND SOCIAL pSyCfiOLOGy / March 1996
(Cl6ment& IGuidenier,1986),the primary reasonfor languageleaming
oJtenis to use it to communicate.Unfortunately,rese;chlnto these
two topics,communicationand secondlanguageleaming, has developed
alongsomewhatdifferentlines.
This adicle draws upon models developedwithin these two
streams of researchin an attempt to predictthe frequencyof second
language communication in a bilingual context. predictoi variables
will be drawn from Gardner's(198b)socio-educational
morlelof language
learningand Maclntf're's(1994)modelof willingrressto communicate.
adaptedto refer to the secondlanguage.The major purposeof th;
presentstudy will be to test this hybrid modelusing path analyeis.
Becauseboth of the constituentmodelsdescribeindividualdifferences.
a secondpurposewill be to examinethe role ofthe globalpersonality
traite. Recentwork on the Five Factor Model (e.g.,Goldberg,1993)
suggests that it represents a taxonomy of global traits, and this
developmentwill be appliedto the secondlanguagedomain.
SOCIO-EDUCATIONALMODEL
OF SECOND LANGUAGE (L2) LEARNING
Individual differencesin secondlanguageleaming havebeenstudied
for severalyears. One of the most active and productiverese€rrch
programs has been undertaken by Gardner, Lambert. and their
associates(Gardner& Lambert, 1972).The emergingproductofthis
researchactivity is the socioeducational
model of secondlanguage
acquisition(Gardner,1985,1988).Whereasportionsofthe modelhave
been,and will continueto be, updatedto incorporatenew research
results,the basicmodelhas consistentlybeenreplicated(for a review
seeGardner,1985;Gardner& Maclntyr.e,1992,1999b).
The socioeducational
model(Gardner,1985)proposes
that two basic
attitudes-integrativeness andattitudestowardthe leaming gituatton_
arisefrom the leamers'socioculturalmilieu. Integrativenessrefersto the
desireto learn a secondlanguageto meet and communicatewith mem_
bers ofthe target languagecommunity.Attitudes toward the learmng
situation refer to the evaluationofthe languageteacherand the course.
Both integrativenessand attitudestowardthe learningsituation contribute to the learners'level of motivation.This tripartite cluster,shown
in Figure 1, has beenca-lledthe integrative motive (Gardner,19gb).
Integrative motivation, in turn, influencesthe activity level of the
learnerin learning situations,includingboth formal (e.g.,classroom)
andinformal(e.g.,onthestreet)typesoflearning.Muchofthe research
on the socioeducational
modelhas examinedthe role of motivationin
languageleaning and demonstratesits importance in producing individual differencesin variousforms of languagelearningachievement
(Gardner,1980,1985,1988;Gardner& Maclnt re, 1992,1999a).
()
HC\
ooI
F o o
c, lz
o
q
; b o
bn .i
J
Ctl
.9
.x
6
:
R
x
t
!; 9E. Es E8
: E s 5 ,x
JOIJRNAL OF LANGUAGEAND SOCIAL PSYCHOLOGY/ March 1996
Researchhas shown that in addition to attitudes and motivation,
has a signifrcanteffect
anxietyaboutsecondlanguagecommunication
on secondlanguageleaming (Honnitz,Horwitz & Cope,1986;Horwitz
& Young,1991;Maclntyre & Gartlneq1991b,1994b).Languagearxiety,
apprehensiongeneratedin secondlanwhich is the situation-specific
guageconte).ts,has beenshownto correlatenegativelywith second
Ianguagecoursegrades(Horwitz, 1986)and the ability to take in,
process,and output secondlanguageinforrnation(Maclntyre & Gardner,
it hasbeenshownthat
1994a,1994b).With respectto communicating,
speakingin the secondlanguagecanbe especiallyanxiety provoking
(Horwitz et a1.,1986;Koch & lbrrell, 1991;Maclntyre & Garalner,
1991c).Gardnerand Maclntlre (1993a)found that amongattitudes,
motivation, ard anxiety,measuresof languagearxiety showedthe
strongest correlations with several indexes of second language
achievement.
In a theoreticaldiscussionof the relationbetweenlanguageanxiety
and motivation,Gardnerand Maclntlre (1993b)proposedreciprocal
paths between motivation and language a-nxiety;high levels of motivation are likely to abate arxiety, and high levels of anxiety are likely
and
to inhibit motivation.Althoughthis relationhasbeensuggested,
the available data show a negative correlation between language
anxiety ard motivation,languageanxiety has not been consistently
model
included as a variable in tests of Gardner'ssocioeducational
(Maclntyre & Gartlner, 1991b).Having measuresof both language
anxiety and integrativemotivationin the presentstudy will allow for
an examinationofthe connectionbetweenthem.
WILLINGNESS TO COMMIJNICATE
Whereasthe study of anxiety in secondlanguagecommunication
is emerging as an important topic in the literature, native language
communicationapprehensionhas been widely studied for many
years (Richmond& McCroskey,1985).Communicationapprehension is conceptuallysimilar to languageanxiety; they both refer to
anxiety about communicating(seeDaly, 1991;Horwitz et al., 1986).
communicationapprehensionin the native languagehas
Researchon
shown its potentially dehimental effectson the frequencyand qualiff of
communication(Beatty,1988;McCroskey,L977,1984;McCroskey&
Richmond,1987).
Oneof the strongest,mostreliableeffectsof communicationapprehensionis a reduceddesireto communicate(Beatty,1987).Recently,
(Mccroskey,1992;McCroskey& Baer,1985;
McCroskeyand associates
McCroskey& McCroskey,1986;McCroskey& Richmond,1991)have
that capturesthe
proposeda construct,'\trillingnessto communicate,"
Maclatyre, Charos / L2 COMMITNICATION
m4jor irnplicationsthat conceptslike communicationapprehension,
introversion,reticence,and shynesshavefor communicative
behavior.
Willingness to communicateis defined as a stable predisposition
toward communicationwhen free to ctrooseto do so(McCroskey& Baer,
1985).A personmay be unwilling to communicatefor a variety of
reasons,such as anxiety,introversion,alienation, or a lack of communicative competence,for example.All of these variables have
shown significant correlations with willingness to communicate
(Burgoon, 1976;McCroskey& Richmond,1991;McCroskeyRichmond,&
McCroskey,
1987).
Maclntyre (1994)developeda path modelto predictwillingnessto
communicatein the native language(seeFigure 2).The modelpostulates
that higher levels of willingness to communicateare based on a
combinationof greater perceivedcommunicativecompetenceand a
relativelackof communication
apprehension.
The modelfurther shows
the influence of personality traits (see also McCroskey,1984;.
McCroskey& Baer, 1985).The globaltrait ofintrover€ion contributesro
both communicationapprehensionand the perceptionof communicative
competence,
and self-esteemwas found to play a role in developing
communicationapprehension.Finally, the modelincorporatesBurgoon,s
(1976) suggestionthat societal pressures,reflected in feelings of
alienation and anomie,play a role in generatingan unwillingnessto
communicate.One avenueavailable for further explorationis the
impact ofdifferentsituationalcontextsonthe model.Maclntyre(1994)
recommended
exploringthe interactionbetweenpersonalityand
specificsituationalcharacteristicsin their influenceon willingnessto
communicate(p. 140).Situationsin which a communicatoruseshis or
her secondla-nguage
representan opportunityto both test the model
and integrate it with existing languagelearning research.
Thusthe willingnessto communicate
modelwillbeappliedtosecond
languagecommunication.In fact,its two key variables,anxietyabout
communication
andthe perceptionof communicative
competence,
have
consistentlyappearedas correlatesof secondlanguageproficiency
(Cl6ment,Gardner,& Smythe, 1977,1980;Maclntyre, 1992).In his
contextualmodel, Cl6ment (1980, 1986)considersthem to form a
higher order construct,self-confrdence,
that providesincreasedmotivation for secondlanguageleaming and contact(communication)
with
the target languagecommunif. WhereasCl6ment'smodeltreats selfconfidence
as a unitary construct,Maclntyre's(1994)modelpostulates
a specifrcrelation betweenits major components,specificallythat
anxiety influencesthe perceptionof competence.
In support of this
hypothesis,Maclntyre and Noels( 1994)foundthat arxious language
leamerstendedto underestimate,and relaxedstudentsto overestimate,
their ability to speakand comprehend
the secondlalguage as evaluated by independent,bilingual raters.
o0
.E
=
F
irU
E> E3
.UE
co :9i
c
)': >,
= x
r)
{L)
a
+
(.)
(n
Maclatyre, Charos/ L2 COMMIJMCAfiON
GLOBAL PERSONALITY TRAITS
As noted above,the influenceofglobal personalitytraits (seeFunder,
1991)has beenexaminedin t}te researchonwillingnessto communicate.
AIso,someresearchhasbeenconductedonthe roleofpersonalitytraits
in secondlanguageleaming. Although personaliff haits are hypothesizedto be important in languageacquisition(Gardner,1991),past
researchin this area has producedinconsistentresults (Lalonde&
Gardner,1984;Skehan,1989).
Introversion-extroversionhasbeenthe trait of primary interest and,
acrossseveralinvestigations,hasproducedinconsistentrelationswith
avariety of measuresoflanguageachievement(Chastain,1975;Naiman,
Frolich, Stem, & Todesco,1978;Pritchard, 1952;Smart, Elton, &
Burnett, 1970;Swain& Burnaby,1976).In consideringthis ambiguity,
Skehan(1989)notesthat for academicachievementin general,introversionis usually the more desirableend ofthe trait dimension.But
for languageleaming, the desirableend may be either extroversionor
introverrion, dependingonthe leaming contextor instructional methods.
For example,languageleaming basedin a formal classroomsetting
emphasizingrote memoryfor vocabularyand grammar rules might
favor the introvert; laaguageleamingbasedon communication
would
likely favor the extrovert. In general,the correlationbetweenextroversion and languageachievementhas been nonsignificantor slightly
positive and the results seemto be tied to the particular measures
employedin the studies (Skehan,1989).Studies focusingon other
personalityvariableshavealsoyieldedmixedresults.
Lalondeand Gardner(1984)attemptedto incorporatepersonality
variablesinto the socioeducationa.l
model.Eighteenpersonalityvariables
were included,but they showedvery few conelationswith lalguage
achievernent,aptitude, or perceivedsecondlanguagecompetence.
However,integrativeness,motivation, and attitudes toward the learning situation did show significant correlationswith two post hoc
groupingsofpersonality traits that were tentatively labelledas arnlytic
orientatbn and senouszess.
Usedin a causalmodelpredictinglanguage
achievement,analytic orientationwasfoundto underlieintegrativeness,
and seriousnesswas related to attitudes toward the leaming situation.
Lalondeand Gardnerconcluded
thatbroadpersonalitytraits haveonly
an effectchanneled
an indirect effecton secondlaaguageachievement,
through language-relatedattitudes ard motivation.
The study ofpersonalityprovidesno shortageoftheoreticalmodels
from which to chooseor levelsat which to conceptualize
traits. The
strategy used by Lalonde and Gardner (1984) in grouping traits
based on factor analysisimplies that amongthe 18 traits that they
measured,morebasicpersonalitytraits exist.Asimilar rationalehas
been employedby researchersusing factor analysis to investigate
the Five Factor Model (or Big-Five; see Goldberg,1993).This
JOURNAL OF LANGUAGDAND SOCIAL PSYCHOLOGY/ March 1996
"modeL'isintendedto representa taxonomyof the mostbasic,global
personalitytraits. Whereasvariousversionsofthe Five FactorModel
have been reported (Digman,1980;McCrae& Costa,1989;Norman,
1963), their structure and interpretation are remarkably consistent
(Goldberg, 1993). This emerging consensuson the basic pe$onality
factorsmay renewinterest in the study of the role ofpersonalif factors
in languagelearning.
Using Goldbergis(1992)terminology for the Big-Five, the frrst trait,
'labelledintrouersinn-ertrcoer
sinn, conttastsresewed,quiet, andunassertive with outgoing,ta-lkative,arld active.The secondfactoris labelled
and.conttastscold,selfish,and distrustpleasantness
or agreeableness
Thethird tr t,conscientinus'
andgenerous.
ful witlr kind, cooperative,
nessor ilependabilily,contrastsdisorganized,careless'and lazy with
responsible,thorough,and hardworking.The fourth trait, etnotional
stability,contraststense,newous,and unstablewith calm,contented,
and unemotional.The frnal factor,intellector sophisticatinn,cor:trasts
analytical, imaginative, and creative with unintelligent, unreflective,
and uninquisitive. This final trait has also been labelled as culture ot
to etpeienre arld may bestbe seento reflectthe degreeof a
openness
search, tolerance, and appreciationfor the unfamiliar (Costa &
McCrae,1985).
The present study attempts to combinethe work on language
leaming, willingnessto communicate,and globalpersonalitytraits'
The prihary purposeof this study is to test the abilitv of a hvbrid of
modeland MaclntFe's (1994)willGarrlner's(1985)socioeducational
ingnessto communicatemodelto predict the frequencyof using the
secondlanguagein daily interactions.This combinedmodelwill be
testedusing path analysis.A secondpurposewill be to examinethe
influenceofglobaltraits that rrill be integratedinto the modelin ways
that are consistentwith previousresearch.
Theinitial modelis shownin Figure3.Them{or elenentsofGardner's
(1985)modeland their interrelationsare shown:Integrativenessand
attituales toward the learning situation are independent, and both
contributeto motivationfor languageleaming' The modelalsoshows
the relationsanong elementsof Maclntlre's (1994)model:Language
anxiety reducesperceivedcommunicativecompetence,and both of
thesevariables influencewillingness to communicate.Consistentwith
Cl6menf,s(1980,1986)contextualmodel,a path from willingnessto
communicateto motivationis proposed.Both willingnessto communicate and motivationfor languageleaming contributeto determining
the frequencyof secondlanguagecommunication.
The model also shows the influence of personality haits. It is
possibleto speculateon a large number of potentia.Ipaths involving
lhesevariables.Basedon Lalondeand Gardner's(1984)suggestions,
the influence of the Big-Five is expectedto operateindirectly, via
anxiety,and perceivedcompetenceAs
attitutles,motivation,la.nguage
Macl-utyre,Charor/ L2 COMMIJMCATION
11
describedbelow,there is some reason to expectthat each of the
and
Big-Fivetraits mayplay arole in secondlanguagecommunication,
specificpathswereproposedprior to the study (asshownin Figure 3).
For the sakeofparsimony,eachtrait wasassigrtedonlyonepredictedpath.
Languagestudente who are higher in intellect may perceivethempersonsand alsomorecompetentin the
selvesasmoreknowledgeable
secondlangrrage(seeCl6ment& Iftuidenier, 1986),thus a path from
is
intellect to perceivedsecondlanguagecommunicativecompetence
proposed.Extroverts'preferencefor socialactivity may lead them to
have lower levelsof languagearxiety about socialinteractionusing
the L2 (Maclntyre& Noels,1994;Skehan,1989),thereforea path from
Peoplewith higherlevels
extroversionto languageanxietyis proposed.
who are more pleasant,likely will anticipatemore
of agreeableness,
positiveinteraction with memberqofthe secondlanguagegroup(see
Cl6ment,1980).The mostlikely variable to be affectedby this seems
to
to be integlativeness,thus a path is proposedfrom agreeableness
integrativenesg.Individuals with lower emotionalstability may be
more prone to languageanxiety. However,a path from emotional
stability to languagea-nxietyis not explicitly proposedhere because
prior research has demonstrated that language anxiety is not
strongly related to general trait anxiety,which would be reflected
in a lack of emotionalstability.Apath is expectedbetweenemotional
stability and integrativeness.This is basedon Segalowitz's(1976)
finding that thosewho havelesspositiveattitudestoward the target
languagegoup alsofeeluncomfortablespeakingto a memberof that
group. Finally, conscientiousness
also may play a role in that those
who are better organizedrnayapproachlanguagelearningin a more
systematic manner and may 6how better cognitive processingof
language inshuction (Krashen, 1981; Lalonde & Gardner, 1984).
to attitudes toward the leamTherefore,a path from conscientiousness
ing situation is proposed.
Figure 3 showsoneadditionalelement.As notedby Cl6ment(1980,
1986),the sociolinguisticcontextplaysa potentiallyimportant role in
providing the opportunity for frequent and/or pleasant L2 contact.
Maclntyre ( 1994)alsoproposedthat studiesexaminethe influenceof
situationson willingnessto communicate,as did the work of Burgoon
(1976).Therefore,a measureof the languagecontext,as definedby a
self-repodofthe relative concentrationofLl and L2 at homeand at
variablein the model.Based
work, will be includedas ar exogenous
on Cl6ments (1980, 1986)model, context is expectedto influence
becauseit followsfrom Cl6mentsdescripwillingnessto communicate
tion that self-confidentindividualswill be morewilling to communicate in the secondlanguage.Further, a direct path from contextto
is proposedbecausethe numberofopporfrequencyof communication
tunities to communicatein the secondlanguageshouldinfluencethe
frequencyof doingso.
72
JOIJRNAI OF LA.NGUAGEAND SOCIAL PSYCHOLOGY/ Mardr 1996
hJ,cilncMatl
ttuent.r.hkd
Alallffi
Ebdioo.l Subiiiry
c.ui.nri6qs
P.rc.ivtdcmp.rr.
in de L:
ljsur.
Arjer,
IocgdivaK
AA: ftdd6
bwrd
de lming.i@id
(*bd &muF)
Sdtagee.
r-zwirriryEiso
cruicr!
Flqu.nrofusriu
soql jrnl!!!f
in t2
sp.:IiDg riMrirs
Moivdion for
L.luse lojn!
,t\gzle 3. Base path analyals uodel to bo tosted.
The purposeof the presentstudy is to examinethe relationswithin
and betweenthe languagelearning and communicationmodelscited
above,and to extendthe modelsby testing their ability to predict in
vivo languageuse. In addition,the somewhatneglectedissueof personality dispositionsthat favor secondlanguageacquisitionand use
will be examined.
METHOD
PARTICIPANTS
in Ottawa,whichis a large,bilingual(FrenchThis studywasconducted
English) city.Ninety-twoAaglophonestudents,taking introductorylevel conversationalFrench, were recruited from adult evening
classes offered by local schoolboards.Thirty of the padicipants
were male (meanage = 32.6,SD = 12.3)and 62 were female(mean
age = 33.8,SD = 12.0).All of the participantsspokeEnglish as their
Maclntyre, Charos / L2 COMMIJMCATION
only a minimal level of
native languageand, as a group, possessed
Frenchcompetence.
MATERIALS
The materialsrequiredfor this studyincludedself-reportmeasures
of the Big-Five personality traits, frequency of communication,
attitudes,motivation,
willingnessto communicate,perceivedcompetence,
and the amount of Frenchpresent in the work and home context.The
measuresctrogenhere were selectedto be brief but soundpsychometric
instruments. The limited time availablefor testing (30-45minutes) did
not permit the use of longerinstruments.The scaleswere presented
A descriptionof these scales,an
in one of eight random sequences.
alpha
coeffrcient(o) for the present
exampleitem, and the Cronbach
follows:
sanple are as
Measures of Language Learning Affect
Aseriesofbriefmeasuresofattitudes,motivation,aJIdanxietywere
administered.All ratings were made on a 7-point Likert-tlpe scale.
"Guilford-style"
Gardner and Maclntyre (1993a)have shownthat these
convergentand predictivevalidity.
items haveacceptable
1. Integrativeness (o,= .86). This measures the degree to which respondents
were learning French for the pu.rposeofinteracting and communicating
with Flancophones. This aggregate measure was composed of three
single-item measures of integrative orientation, attitude toward French
"IvIy
Canailians, and interest in foreign languages. An example item is
with
Flench
Canadians
in
order
to
interact
leaming
French
feelings about
are: Weak-Strong."
2. Motivation (cr = .65). Motivation was measureil with three single-item
measures of the desire to learn French, motivational intensity (elfort),
"If I were to
and attituile toward learning French' An example item is
rate how bard I work at learning Frenclr" I wor:ld characterize it as: Very
Much-Very Little."
3. Attituiles toward the learning situation (cr,= .89). This was measured by
two items, attitude toward the French teacher and attitude towaril the
"My attitude toward my French course is:
cowse. An example itern is
Unfavorable-Favorable."
.
4. Language anxieiy (a = .48).' This was also measured by two items, one
assessing Flenctrclassroom anxiety and the other measwing French use
"My anxiety wheu speaking French is:
anxiety. An example item is
Low-High."
Communication-Related Variables
The followingthree variableswere adaptedto refer to communication using French.Eachofthe measulespresents12 communication
JOIJRNAL OF LANGUAGEAND SOCIAL PSYCHOLOGY/ Mard 1996
contexts involving speakingto friends, acquaintances,and strangers
in four settings: dyads, small groups,formal meetings,and public
speakingsituations.
1. Perceived competence in French (cr,= .98). lvelve items, adapted from
McCroskey, Richmond, and McCroskey (1987), assessedthe average
percentage of time, ranging from lVo to 700Vo, Lhat respondents felt
competent in using French to speak in 12 situations. An example is "talk
in a small group offriends."
2, Frequency of communication in French (cl = .97). Items from the
perceived competence scale were adapted to measure the frequency
ofcommunicating in French for each of 12 situations, using a 7-point
scale with the anchors never and many n1,an! tim,es- The potential
range of scores was 0 to 72, An example item is "talk with an
acquaintance."
3. Willingness to communicate in Frendr (o = .97). Tbenlr items from
McOroskey & Baer (1985) assessed the average percentage of time,
ranging from 0Voto 7OO4o,that respondents would choose to communicate in French in various situations. In addition to the 12 speaking
contexts not€d in items number 1 and number 2 immediately above, 8
"filler" items were also included. Al example is "talk to a service station
attendant."
Personalit5r Measure
(1992)transparentbipolarscalewasusedto assessfiveglobal
Goldberg's
personality traits. Goldberg(1992)arguesthat the transparent bipolar
inventory is an acceptablesubstitute for a longer scaleof the Big-Five
(suchastheNEO-PI).Sevenitemsona g-pointsemanticdifferentialscale
were usedto measureeachof the followingpersonalitytraits:
l. Extroversion (vs. introversion) (cl = .85). An example item is silenttalkative.
2. Agreeableness (vs. disagreeableness)(o =.63). An example item is
cooperative-uncooperative.
3. Conscientiousness (vs, negligence) (d, = .74). An example item is disorganized----organized.
4. Emotional stability (vs. neuroticiem) (o = .56). Arr example item is
relaxed-tense.
5. Intellect (vs. unsophisticated) (u = .77). An example item is creativeuncreative.
Social Context
This measurewas written specificallyfor this investigation.
1. Context (cr = .60), The degree to which reepondents encounter Frendr in
their neighborhoods and in their workplace was assessed using single
items with 5-point scales.The Iabels aesignedto the 5 points were (l = oll
English, no French;2 = mostly English, somz French;3 = half English,
half French; 4 = some English, tnostlx French; and,5 = no English, all
Maclntyre, Charos/ L2 COMMUMCATION
French). The contqL variable is the sum of the ratinss of the amount of
French encountered at work and at home. The rating for the people who
indicateil that they did not work outsiile the home (2 = 17) was obtained
by doubling their rating ofthe home context on the assumption that these
inilividuals are, in effect, working at home.
PROCEDIJRE
The boards of education offering adult evening classes were con-
tacted about the proceduresfor this study.A list of 12 instructors of
introductoryJevelFrench-as-a-secondJanguage
courseswasprovided.
Theseinstructorswere then contactedby phoneand askedfor their
cooperationin contacting and testing students during class time.
Eight instructorsparticipated,3 couldnot be reachedby phonewithin
the time frameset for the study,and 1 was unableto relinquish class
time for testing.With the permissionofthe instructors,all classeswere
tested over a 2-weekperiod, near the end of their courses.As the
studentsarrivedfor class,they wereinformedthat a researchstudy of
secondlalguagecommunicahon
wasbeingconductedandthat partici
pation was voluntar;rThey werethen presentedwith a consentform
that summarizedthe proceduresof the study,describedthe types of
measuresbeing used, and noted that the data would be collected
anonymously.
Almost all studentsdecidedto participate(90 to 954o),
and respondentsweregivenas muchtime asrequiredto completethe
questionnaire,lessthan 45 minutesin all cases.
R"ESUI]TSAND DISCUSSION
Apath analysis,usinga maximumlikelihoodsolutionfrom LISREL
VII (J0reskog& SOrbom,1989),was conductedon the corelation
matrix shownin the appendk.'Thebasemodelshowsa reasonablefit
to the data, although the chi-squareis signifrcantQ2(46 = L87.2,
p < .001).Tbelve of the 15 path coefficientsin the initial modelwere
siglifrcant (r > 2.0).All ofthe pathsthat werederivedfromthe Gardner
(1985)and Maclntyre (1994)modelswere replicated,including the propoeedeffectson the frequencyofl,2 comrnunication.In addition,4 ofthe
5 predictedpatbs involving the globalpersonalityhaits were obtained,
as were both paths involving socialcontext.The only nonsignificalt
paths were from agreeableness
to integrativeness,integrativenessto
language anxiety, and willingness to communicateto motivation. It
shouldbe notedthat all 3 ofthese paths werebasedon speculations
about the relations among the variables and had not been tested
exptcitly in prior investigations.It would be inappropriate,however,to
concludethat thesevariablesarenot at all relatedto eachother,because
JOURNAL OF L{NGUAGE AND SOCIAL PSYCHOLOGy/ March 1996
that would anount to acceptingthe null hypothesis.Therefore,we will
not speculateon the reasonsfor the absenceofthesepaths.
The significant chi-squaretest for the base modelindicatesthat
additionalvariancecouldbe accountedfor if new paths were adderl.
Strictly speaking,addingadditionalpathsis an exploratoryproceclure
(MacCa-llum,
Roznowski,& Necowitz,1992),as opposedto the confirmatory approachadoptedup to this point. Such paths should be
regardedas"datadriven"andthereforetentative,pendingreplication.
Nevertheless,the major value of such results may be to provide
potentially interesting avenuesfor future investigations. Therefore,
additionalpathswereadded,oneatatime, until the chi-squarebecame
nonsignificant LISREL modificationindexes,which expressthe exrent
to which chi-squarewould be reducedby adding an additionalpath,
were employedin this procedure.At each step, the path with the
highestmodificationindexwasadded.Finally,the threenonsignificant
paths wereremovedbeforethe final modelwas evaluated.Ttre sequence
ofstepstakento revisethe modelis shownin Table1.Thestandardized
solutionfor the frnal modelis shownin Figure 4.
The last variablein the causalchainis frequencyofsecondlanguage
communication.
Four significant,positivepathswereobtainedfor this
variable:onesleadingfrom willingnessto communicate,motivation,
perceivedcommunicativecompetence,
and context.Ttleseresultsconfrrm
pathssuggested
by Gardner's(1985)socioeducational
modelaswell as
Maclntyre's (1994)model of willingness to communicate(see also
McCroskey,1992),as shownin Figure 3. In the caseof motivation,
studentswho havegreatermotivationfor languagelearningreport using
the languagemorefrequently In the caseof willingnessto communicate,
studentswho are morewilling to communicatevdll be more likely to
do so.Further,havingthe opportuniff to conversewith Francophones.
reflectedin the contextvariable,alsoplaysa role.The largestsingieeffeci
was obtainedfor perceivedcommunicativecompetence.This is reasonable becauseall of the respondentswere at a relatively low level of
actualcompetence.
This might suggestthat simplyperceivingthat one
has the ability to communicate,
regardlessof one'sactualproficiency,
can affect the rate of participation in L2 conversation.
Underlyingmotivationfor languagelearningin Gardner'ssocioeducational model are integrativenessand attitudes toward the learning
situation (Gardner,1985).The present results replicate Gardner,s
tripartite divisionofthe integrativemotive.Asignificantpath wasalso
found from integrativenessto attitudes toward the learning situahon.
Using different exogenousvariablesand measurementtechniques,
and in a unilingual context,Gardner,Lalonde,and pierson (1989)
showedthe oppositepath betweenthesetwo constructs.In the present
study, frequent contactwith the L2 group is a constantpossibiliby,
thereforeintegrativenesslikely wouldbe a muchmorestableattitude
than attitudes toward the leaming situation. This may explain why
Maclatge, Charos/ L2 COMMITNICATION
l7
Table 1
Step-Br-StepPlocad.urefor ReoisingtheMod,elto Add.D.tta-DriuenPaths
A4.
Gfl
GrI
RMSR x2td.f
.151
13?.85
.79
3.06
Revision 2r
Add a patJr from
perceived coEpeteace to
ftequenry of communication
77.t7
.82
1.75
Revision 3:
AiLl a path from contet t to
perceived competence
62.00
.84
1.44
Revision 4:
Adcl a path from
agreeableness to williagaess
to coDnutdcate
53.28
46
Base model
Revisiou 1:
Add a path from
int€Srativeness to attitudes
toward ttre leanirg situation
Final model:
Erase all uonsigaifrcant paths
.81
42
.93
45
.92
.061
Notei GJI = goodnessoffit inder Adj. GFI = adjusteil gpodnessoffrt iadex; RMSR = root
tuea! squar€ riesrdual.
the path from integrativeness to attitudes toward the learning situation emergedin the presentstudy.
The hypothesizedcausesof willingness to communicatewere
replicated(Maclnty'e, 1994).In this case,both languagearxiety and
perceivedcompetenceexerted a direct influence on willingness to
and the predictedeffectofanxietyon perceivedcommucommunicate,
was also supported.Also, the path from context
nicative competence
to willingnessto communicateindicatesthat increasedopportunities
for interaction directly affect one'swillingaess to communicatein the
L2. Thus we suggestthat the intention or willingrress to engagein L2
communicationis determinedby a combinationofthe student'spercephonof his or her secondlanguageproficiency,the opportunityto
usethe language,and a lack ofapprehensionaboutspeaking.
At the far left of Figure 4 are the exogenousvariables:the five
personalityfactorsand the measuteof socialcontext.With respectto
context,having more opportunities for interaction in French may lead
a greaterwillingnessto comto an increasein perceivedcompetence,
nunicate in French, and more frequent communication.The path
leading to perceivedcompetencesuggeststhat the opportunity to
communicateleadsto the development0f greateractual competence,
likely through a combinationof practice and the adjustment of learner
beliefs. Obviously,the learners' ability to communicatein the L2 will
improvewith practice.But this path might alsoindicatethat exposure
18
TOIJRNAI OF LANGUAGEAND SOCIAL PSYCHOLOGY/ March 1996
Rmt MS R.sidual= .067
t.;;
[email protected]
J
i
-. '
trt d
io rh. tl
t4g!.3.^dic.y
SNrdItu@8.
Cl]mdd.
sond l.ngug. io 12
slc,riu siurioc
(tupErirc
&t._n od)
Emcidrsobiriry
6 FleorrrE
d
rh. l6ios
siMior
(dd'd&.ffi)
Figure 4. F]lna'l moilel showiag all slguificaDt
paths.
to authentic oral input reducesthe formality (e.g.,proper grammar
rules) that some students experiencein the language classroom.
Perhapsthe perceptionof competence
increasesbecausethe level of
objectiueL2 competence
requiredfor daily communicationis shownto
be lower than expected.For example,studentsmay learn effective
verbal (e.g.,requestingsynonymsfor unfamiliar words,switchingto
Ll sometimes)andnonverbalstrategies(e.g.,usinggesturesandfacial
expressionsto conveymeaning)to avoid conversational
trouble spots
(seeOxford,1991).This path is data driven,and this proposedexplanation might make an interesting avenuefor future investigation.
The global personality traits eachshoweda significant path in the
present model.The measureof intellect was found to underlie perceivedcompetence,
as predicted.It would appearthat personswho
considerthemselvesto be moreintellectual,sophisticated,or opento
(synonymsfor this factoraccordingto Goldberg,1993),also
experience
seethemselvesas morecompetentin French.
Extroversion,as predicted,showeda negativeeffecton language
anxiety.In mostsituations,itis notnecessarilythe casethat extroverts
Maclatyre, Charos/ L2 COMMIJMCATION
shouldfeellessnervousthan introverts,becauseextroversionreflects
sociabilityrather than emotionalarousal(Goldberg,1993).Within the
languagedomain,both a significalt path from extroversionto language
anxiety coupledwith the lack of a path from emotionalstability are
consistentwith Maclntlre & Garrlner's(1989, 1991a)frndingsthat
general trait arxiety, as would be reflected in the emotional stability
factor, is not typically associatedwith languageanxiety.This supports
the assertionthat it is the socialand communicativedemandsof
L2 interaction,and not a predispositionto nervousness,
that drive
languagearxiety.
A significantpath wasalsoobtainedleadingftom agreeableness,
but
not to its predicteddestinationof integrativeness.Instead, a datato willingnessto communicate.
drivenpathwasfoundfrom agreeableness
themselves
wouldbemore
Peoplewho aremorepleasantandagreeable
likely to havepleasantcontactswith target languagegroupmembers,
and this appearsto be reflectedin their willingnessto communicate.
Cl6ment (1980,1986)suggeststhat frequent,pleasantcontactwith
membersof the target languagecommunitywill primarily determine
a constructvery similar to willingthe level of linguistic self-confidence,
nessto communicate.We might suggestthat whereasself-confidence
leadsto a motivationfor languageleaming, willingnessto communicate
for languageuse,asifsuffrcientcompetence
is basedon apreparedness
had alreadybeen achievedfor the communicativepurposeat hand.
Expressedin terms ofpath analysis,in a future study,we wouldexpect
to motivationfor languagelearning,but
a path from self-confrdence
not onefrom willingaessto commwricat€to motivation.This hypothesis
must be consideredtentativebut vrouldindicatea differencebetween
and willingnessto communicate.
self-confidence
As predicted,a significant path was obtajnedleadingfrom emotional
As notedabove,this factorreflectsa type
stability to integrativeness.
of trait anxiety,and peoplewho feel less arxious appearto be more
disposedto interactingwith memberrofthe secondlanguagecommunity.
Segalowitz(1976)foundthat peoplewho were morenervousspeaking
to a member of another languagegoup had less positive attitudes
toward them. This may be attributableto a processof self-perception
(Bem, 1972; Fazio, 1987)whereby individuals perceivetheir own
anxiety and decidethat there must be somethingaboutthe language
groupthat is makingthem nervous.Thereforethat groupis perceived
as less likable and this would be reflectedin a reducedlevel of
integrativenes
s.
The frnal exogenousvariable to show a significant path was
It washypothesized,
basedonLalondeandGardner
conscientiousness.
(1984)and lGashen(1981),that studentswho were better organized
might showmorepositiveattitudestowardthe learningsituation.We
in a structuredcourse,as opposedto language
speculatethat success
acquisition"on the street,"might be moredependenton the student's
JOIJRNAI OF I"ANGUAGEAND SOCIAL PSYCHOLOGY/ Mardr 1996
level of organizationand attention to detail. Thus studentswho are
more conscientiousand well organized,as comparedto less well
organizedones,may possessan advantagein study habits that leads
to a morepositiveattitude toward the languagecourse.
Consideringthe final modelas a whole,we havenot obtainedany
recursive paths, all of the relations are shorvn as one-way.This
reflectsour desireto developa rnodelthat predictsthe frequencyofL2
communicationat a given time. there is no doubt that, over time, as
studentscontinueto leam and usethe secondlanguage,changesin other
variables should be expected.For example,Cldments (1980, 1986)
contextualmodelclearlystatesthat frequentand pleasantinterethnic
contact will influence anxiety and the perception of competence.
Whereasthe modelshownin Figure 4 doesnot includepathsfrom 12
communicationbackwardto other variables,we canbe certainthat a
longitudinalstudy would showrecursiverelations.
CONCLUSIONS
This study attemptedto integratea numberofissuesthat arisein
The mqjor
referenceto secondlanguageleaming and communication.
purposeof the investigationwas an attempt to replicate relations
describedin a modelof languagelearning motivationand a modelof
willingness to communicate,and to assessinterrelations between
thosemodels.The majorelementsof Gardner's(1985)socioeducational
modelshowthe expectedrelationswith eachotherandwith frequency
of L2 communication.As well, the willingnessto communicateconstruct appearsto adaptwell to the secondlanguagecontextand may
representa prolitable addition to this literature. TWoof the three
expectedpathsbetweenthetwo modelswerenotsupportedin this data
set. The presentanalysissuggeststhat communicatingin a second
language,in a bilingual milieu, amongbeginninglanguagestudents,
appearsto be relatedto a willingnessto engagein L2 communication,
motivation for language leaming, the opportuni$ for contact, a-nd,
perhapsmost important, the perceptionof competence.
In turn, languageanxiety,intellect, and the socialcontextwere shownto influence
in this group.Ttrediversity ofinfluences
the perceptionof L2 competence
(social,personality,and affective)convergingon thesevariablesshows
the complexityinvolvedin L2 communication.
The secondpurposeof this study was to examinethe potentialrole
ofglobal personalitytraits, as identified in the Five FactorModel,in the
hybrid model.Basedonthe presentresults,itwould appearthat global
personality traits are implicated indirectly, via their influence on
language-relatedatbitudes,languagearxiety, perceivedL2 competence,
motivation for languagelearning, and willingrressto communicate.
The contribution of both personality traits and social context to
Macllrtyre, Charos/ L2 COMMITNICATION
predictingthe frequencyofl2 communication
is clearlyevidentin the
presentstudy
It shouldbe notedthat the modelemergingftom this investigation
containsboth exploratoryand replicatedpaths.One of the valuesof
the exploratorypaths identified in this modelis to offer suggestions
for future research.Suchinvestigationsmight examinethe relations
motivation,and self-confrdence,
amongwillingnessto communicate,
in
particular to considerwhetherwillingnessto communicatelacks the
motivational properties implied by the self-confidenceconstruct.
Perceivedcompetenceexerteda direct and strong influenceon the
frequencyof communicationin this group of beginning students.
Future researchmight considerwhetherthe effectis as strongamong
students with more languagetraining. The path from integrativeness
to attitudes toward the learning situation was reversed from that
obtainedin an earlier study by Gardner et al. (1983).If this path
replicates,future researchmight investigatewhetherthis is afunction
of the socialmilieu, the nature of the languagecoursein which the
subjectswere enrolled,or the obtainedlevel ofsecondlanguagecompetence.Finally, willingnessto communicatewas influencedby four
trait, and indirect
direct paths,includingonefrom the agreeableness
paths from extroversionand intellect. In light of these findings,
research on native language willingness to communicatemight
examinethe interactionof globaltraits in producinga willingnessto
communicate.
Futureresearchmight alsoexaminethe degleeto whichthe present
modelreplicateswithin variousgroupsand acrosssocialcontexts.The
meaning of bilingualism and secondlarguage communicationmay
dependon the relationsbetweenlanguagegroupsinvolved.Whether
the samemodelcar be appliedto both minority and m{ority groups
learning each others'languagewould be an interesting avenuefor
further study (seeCl6ment,1986).AIso,the potentialfor the modelto
show genderdifrerencesshouldbe explored.Finally it should be noted
was measuredusing self-report.
that the frequencyof communication
Although there is evidencethat willingness to communicateis significantly correlatedwith observedcommunicativebehavior(Zakahi &
McCroskey,1989),the link with overtbehaviormust bewell established.
In spite of the needfor future researchand the limitations ofthe
present study, this investigationhas demonstratedthat modelsof
with those
secondlanguageacquisitioncanbe integratedsuccessfully
from the communicationdomain.In addition,both globalpersonality
traits andsocialcontexthavean impactontheprocessofusingasecond
language.Theseareasare often isolatedfrom eachother,and their
synthesisshowsthat potentiallypowerfulmodelscanemergefrom the
of knowledgein variousdomains.In terms of predicting
convergence
the frequencyof secondlanguagecommunication,the presentstudy
clearlyshowsthat it is a complexprocessindeed.
q n
;
@ r a i
ccica
i
crr io an
qaotc!n
B
'j
[email protected]
ocalicorraoc\
<,o+ao3ts-16
q<qclnaqo?o?
X
E
X E
g
cocort(ocn@;i
cn .+ cl c! ro o c.t rr
t
t
t
t
t
t
t l
t
Frc!
Z B
1r!rrolitNroFcn
.'?.{aai:clclcl
. j i
{'o
.
'6
;
'jn']?
F
16<lroaocnoroo
r0 .o i cr c.) alr <i <n
i
cnF(OrlJo)C\!-r:f
i
Orq)<OIo@O)cOO
e1n1ac!.'?c
I
o)
v)
nqao?
r{cOrr
{rO<O+
i F ( o c n d F v F
ct)<,[email protected]@
'4qcQ
u?no?q-ln'{c
3
icolloicnF-ots-
-9>
' i - 6 6
P a:t
22
d
F
.i
c.i
E€ .
e: **
=f,a;;
Ef€E
>o:J <
c/ji6
>,.^
E tr
*€3
€ F " 5 ! . Es 3 . E
Eg F
B i g Fs - s i
E E F !E
.il
g
9 ^
n6
,d
3t!
6g 6
Fi oFlFl <r'l ()
Fodoi9i*g
o
a
Macl-nt1re, Charos / L2 COMMUNICAfiON
23
NOTES
1, Tlre coeihcient alpha for sooe ofthese measures, particularly language aDxietJ4is
low lhe alphas for motivation, agreeableness, enotional stabilit1t and the measure of
coutext also are somewhat lower thao desilable. Tttis may result ftoE attempting to
measure a bmacl clomain witl: a small number of scale items, Ar a result, some oftlte
path coefficients involving tJeesevariables Day be attenuated.
2. Path aaalysis was selected iDstead ofa latent variable aualysis because two ofthe
constructs, context and laDguage aDxietlt would have oolJr two observed indicator
variables. In cases where only two observecl variables define a latent coustruct, an
indeterminate solution may arise.
REFERENCES
Beatty, M. J (198?). Commu:eication appleheDsiou as a determinant of avoidance,
srithdrawal aoil performance atxiely. Commutuicd.tiorrQuarterlf , 35, 202-277 .
Beatty, M. J (1988), Situationa.l and predispositioual correlates of public speakiug
aDlD.iebit.
Conti.utuiad,tion Edu,catian, 37, 28-39.
Bem, D. J. ( 1972). Self-perception theory. In L. Eerkowitz (Ed.), Advarnes in exp*imental
eoci.al,psychalogr (Iol. 6, pp. 1-62). San Diego, CA: Academic Press.
Burgoon, J. K. (1976). ttle unwillingness to communicate scale: Development and
validslio\, Corrstuuni.ation Morrogaphs, 43, 60-69.
Chastaia, K. (1975). AJfective auil ability factors in second language acquisition.
Langtage l*arnitg, 25, 163-16 1.
Cl6ment, R. (1980). Etlaicity, contact aud communicative competence in a seconil
language. In H. Giles, W. P RobiDson, & P M. Smith (Eds.), Language: Socid,l
psXtchologicalperspectives (pp. 147 -754). Oxfor4 UK: Pergamon.
ClSment, R. (1986). Seconil lauguage profrciency and acculturation: An iovestigation of
the effe.cts of langlage atatus aud inclividual characteristics. Joumal of larrgudge
and. Said Ptychologg, 5, 27 L-290,
Cl6ment, R., Gard:rer, R. C., & Smythe, P C. (1977). Motivational variablss in second
laaguage acquisitioor A study ofFrarcophones Iearning Etglish. Canql,lan JournaJ
of Behz.uioural Science, 9, 723-133.
Cl6meut, R., Garduer, R. C., & Smythe, P C. (1980). Social ard inilividual factors in
seconcl language acquisitci,on. Canad,ia.n Journal of Behaoioutal Science, 12, 293902.
Cl6ment, R., & Kruideoie4 B. G. ( 1986). ?fu effect of cotutert on the conpo*itior. and, role
ofofietutd.tiorrs i, Eecond,lang.r.ageacquisition. Ottawa, Cauada: Iuteraational Centre
for Research on Bilingualism.
Costa, P T., & McOrae, R. R. (1985). ?tu IVEO Peworlolitr ltuoentot! Manual. Odessa,
FLr PsychologicalAeseesmeut R4sources,
Daly, J. (1991). Understanding communication apprehension: An iatroduction for
language eilucators. In E. K. Horwitz & D . J. Young (Eds. ), Longutge d.nxiety: Frorn
theory and, reeearch to classroom itnplicatioas (pp. 3-13). Englewood Cliffs, NJ:
Prentice Hall.
Digaan, J. M. (1980). Personality structr.re: Emergence ofthe five-fac+ottuodeLAntuudl
Review of Psxrchology,4 1, 4L'l -440.
Fazio, R. H. (1987). Self-perception theory: Acurrent perspective. Irr M. P Zanaa, J. M.
Olson, & C. P Hermau (Eds.), Socitl influence: The Ontario Slmposiutn (Yol.6,
pp. 129-150). Hillsdale, NJ: Lawrence Erlbaum.
24
JOIJRNAI
OF LANGUAGE AND SOCIAL PSyCHOLOGY/ Marctr 1996
Frrndea D. ( 1991). Global traits: A Neo-Allportian approach to persotoJ!,y. Pcychological
Scienre.2.31-39.
Gard:rer, R. C. ( 1980). On tlrc validity of a.frectivevariables i! secondlaaguage acquisitiou:
Conceptual, contextual anil statistical consiclerations. Ldngudge Led,ming, 32,
255-269.
Gardner, R. C. (1985). Soclol pcychologt and, second,langtage learninl: Thz tole of
attitudes 6nd, rnotiod.rioz. Irndon: Edward Arnold.
Gardner, R. C. (1988). The socio-educational model of second language learniugl
Assnmptions, fi.ndings and issues. -Lor4uqge I'edtuir.g, 38,10!-126.
Gardner, R. C. (1991). Second language learning in aclults: Cot?elateE of profrcienqr
Applied. LanguaAe Learning, 2, !-28
Gardaer, R. C., Lalonile, R., & Pierson, R, (1983).Ihe socio-educational moilel ofsecond
laaguage acquisition: An investigation using IISREL causal modellimg. Journal of
Lanepdge and. Social Psychology, 2,5l-65,
Gardner, R. C., & Lambert, W E. (1972\. Altihrdcs and, motintion in eeand, language
learning. Rowley, MAr Newbur5r House.
Gardner, R. C., & Maclntpe, P. D. (1992). A student's contributions to second language
learning: Part I. Cognili.ve iadablee. Language Tbachiag, 25, 271-220.
Gardner, R. C., & Madrrt}*e, P. D. (1993a). Oo the beasureoent of afrective variables
-194.
in seconil lansuage lea$r g.IntEaqAe ladmhg,43,757
Gardner, R. C., & Maclntyre, P D. (1993b). A student's contributioos to secondlanguage
learning: Part II. Affectfite variahlea.Innguage fbdchtna, 26, l-17.
Goldberg, L. R. (1992). The development of markers for tJ:e big-frve factor structure.
Pslch.ologicdl AeseEernetut,4, 2642.
Goldberg, L. R. (1993). Ihe structure of phenotypic personality lt it,s, Amaficdn
Peychologist, 48, 26-34.
Horritz, E. K. (1986). Preliminary eviclence for the reliability and validity of a foreiga
language anxiety scale, TESOL Qrnrterly , 22, 559-662.
Horwitz, E. I{., Horwitz, M. B., & Cope, J. (1986), ForeiS:r laDguage classroom anxiet5r
Modern Inn4tage Jouna\ 70, 125-732.
Horwitz, E. K., & Youug, D. J. (1991). Ionguage d.ntiat!: Frorrr theory d.bd,research to
classroom implicationr. Englewooil Clifrs, NJ: Prentice Hall.
Jareskog, K., & Stirbom, D. (1989). ,Lisrel 7 user's reference glade. Moorewille, IN:
Scientific Software.
Koch, A. S., & tbrrell, T. D. (1991). Affeetive reactions offoreign language stuileuts to
natural approac.h activities and teaching techniques. In E. K. Horqritz & D. J.
Yoin!,g (Edg.\, Ldtwtage ansiety: Frotn theory and research to classroom implicatians
(pp. 109-126). Erglewood Cli.ffs, NJ: Preotiee Hall.
Krashen, S. ( 1981). Second.langvage acquisition arul second.langtage Learning, Toton.ta:
Pergamon Press.
Lelonde, R. N, & Gardrcr, R C. (1984). IrvestiSatilg a causal motlel of secondlaaguage
acquisitiou: I{here cloes personality fit? Cazadian Journal of Behatioural Science,
16, 22/l-237.
MacCalluo, R. C., Roznowski, M., & Necovritz, L. B. (1992). Model modifrcations in
covaliaDce structure analysis: The probtremofcapitalization on chance. Prychalogical
Bulletin, 111, 490-504.
Maclntyre, P. D, (7992), An*iety 6tud.ld.hgudge led|nitug ftorrt d. etdges of prcce'sing
pelepectioe, Unpublished Doctoral Dissertation, University of Western Ontario,
Lonclon, Canaila.
Mactrt5ne, P D. (1994). Variables uncler\ring willingness to comnuaicate: A causal
arta\ysis. Commrnication Reeearch Reports, 11, 1?6-142.
Maclntyre, P D., & Gardner, R. C. ( 1989). A.nxietSrancl se.condlanguage learniog lbward
a theoretical clarfficatiot Langpage Learning, 39, 25L-27 6.
'
Maclotlre, Charos/ L COMMUMCATION
25
Maclntyre, P D., & Gardner, R. C. (1991a). Investigating laaguage claes arxietlr using
the focused essay tecbai qu.e,The Modon La.nguage Journz.l, 75, 296-304.
Maclatyre, P D., & Gardner, R. C. (1991b), Language anxiety: Its rslation to other
anxieties ajril to processing in oative aail secood languages. Lan4tuage l*arning,4l,
513-534.
Maclntyre, P D., & Gardaeq R. C. (1991c). Methoils and resu.lts i.n the study ofanxiety
in languag€ learning: A review of t):e literabre.I'atLgltdge Ledllrirtg,41,85-L77 ,
Maclatyre, P D., & Gardoer, R. C. (19944). The efrects ofiniluced aaxiety on cognitive
processing irr seconil lanSuage leaning, Studies in Second.Langaage Acquisition, 16,
r,r7.
Maclntyre, P D., & Gardrrer, R. C. (1994b). ltle subtle effects of language arxiety on
cogrdtive prccssing ia the second laaguage, Ianguage l*arning,44,283-305.
Mactrtyre, P. D., & Noels, I{- A. (June, 1994\, Communicatian appreheneion, perceir.ted
corry)etetuce,dnd dchtal conpetence in ct secotudlatuEuo8e. Presented at the anaual
conference of the Canadian Psychological Association, Penticton B.C.
McCrae, R. R., & Costa, P T. (1989). Mote reasons to adopt t,I:e five-factor model.
American P*chalogist' 44, 467-452.
McCmskey, J. C. ( 1977). Oral coomunication apprehension: A summary ofrecent theory
and researih. Ilnman Commutieation Reseatch, 4, 7 8-96.
Mc{roskey, J. C. (1984). The coEEunication apprehension pertpective. In J. A. Daly &
J. C. McCroskey (E& -), Aooid,ing communiaation (pp, 13-38). Newbury Park, CA:
Sage.
McCroskey, J. C. ( 1992). Reliability anil validity ofthe willingless to communicate scale.
Cor*nunicdtion Quarterly, 40, 16-25.
McOroskey,J. C., & Baer, J. E. (NovembeB 1986). WillinEless to communicate: The
coDstnrct and its Eeasurement. Paper presented at the amual convention ofttre Speeeh
Comnunication Associatioo. Deuver.
McCroskey, J. C,, & McCroskey L. L. (1986). Predictots ofwillingness to communicate:
Implications for screening and remediation. Paper preseoteil at tbe annual convention
of the Intonational CommunicatioaAssociation, Chicago, IL.
McOroskey, J. C., & Ridrmond, V P (1987). Willingress to communicate. In J. C.
McCroskey & J. A. Daly (Ns.), Personnlt0 and' Ituterpe/8onolCoturunhalion (pp l29'
156).Newbury Par\ CA: Sage.
McCroskey, J. C., & Riclmond, V P. (1991). Wilinsless to communicate: A cognitive
perspective. te M. Booth-Butterfielil (Ed.), Communication, cognitiotu, arLd dntiet!
(pp. 19-37).Newbury Park, CA: Sage.
McCroskey, J. C., Richmond, V P, & McCmskey, L. L. (1987). Correlates ofself-per€eived
competence. Paper preseoteil at the annual convention of the
coonulication
Iutemational Coromunication Association, Montreal.
Naioaa, N., Fmlich, M., Stern, H. H., & Toclesa, A. (\978). The eood' langudge leanter.
Research iu educatiou series, 7. Tbronto: OISE.
Norman, W. T. (1963). Tbward an ailequate taxonomy of personality attributes:
Replicatecl factor structure in peer oominated persouality r^trin.gv. Jourtuq,l of
Abtuornal and' Sociol Psychologt, 66 , 57 4-589
Oxfor4 R. (1991). langr age led.m'ing strategies: Whdt eoery teacher should know. New
York Newbury House.
Pritcharal, D. F. ( 1952). Ar1investigation into the relationship between persooality traits
aud ability in ooclern languages. Eritish Joumal of Ed'ucatiortrtl Psychologt, 22,
747 -748.
Richmoud, V P, & McCroskey, J. Q. (7986) Comnunicatinn: Apptehension, aooid'ance
and effectiuenesc, Scottsdale, AZi Gorsudl Scarisbrick.
Segalowitz, N. (19?6). Coomunicative incompetence aud the non-flueot bilingual.
Cdnad.ian Journ'al of Behaoioural Scbnce, 8, 122-\3L.
JOITINAI OF LANGUAGEAND SOCIAL PSYCHOLOGY/ March 1996
Skehan, P (1989). Indbid.ual differmces in eecond,ldrtgaage le.trning. Inndoni Edward
Arnold.
Smart, M., Elton, C. E, & Buraett, C. W: (19?0). Unilerachievers and overachievers in
iutermediate Fleuch . Modern Latguage Journal., 54 , 4L5-442,
Swain, M., & Burnaby, B. (1976). Persouality characteristics a.tcclse.coud language
learning in young c-hildren: A pilot study. Worhing Papere on BilinguaJiem, 11,
115.128.
Zakahi \]l1 R., & McCmskey, J. C. (1989). Willineless to cobmu.nicate: A potentially
confouuding variable in coDmu.Dication res eardt. Communication Repor-ts,2,96-104.