Causal Relationships Between Communication Confidence, Beliefs About Group Work, and Willingness to Communicate in Foreign Language Group Work

Causal Relationships BetweenCommunication Confidence, Beliefs About Group Work, and Willingness to Communicate in Foreign Language doi: 10.5054/tq.2010.235993 G roup work including pair

Causal Relationships Between

Communication Confidence, Beliefs

About Group Work, and Willingness to Communicate in Foreign Language

doi: 10.5054/tq.2010.235993

G roup work (including pair work) has been increasingly used inEnglish as a second language (ESL) or English as a foreignlanguage (EFL) classrooms as the instructional focus has shifted fromteaching discreet aspects of language, such as grammar and vocabulary,

to developing students’ communicative competence in English In group work, students can have ample opportunities to interact with eachother in their second language (L2) in natural ways, which is likely todevelop their communicative competence, that is, various components

small-of their language competence, such as organizational competence,pragmatic competence, and strategic competence (Bachman, 1990) Use

of small-group work is also recommended by cooperative learning

positively interdependent manner, with each group membercontributing his or her share of the work In these circumstances,students benefit from recognizing the overall success of the group andfrom observing the success of its individual members (Johnson &Johnson, 1999) They claim that cooperative small-group work canenhance learning outcomes, communication skills, learning motivation,and psychological health (e.g., Cohen, 1994; Jacobs, Power, & Loh, 2002;Johnson & Johnson, 1999; Kagan, 1994; Sharan & Sharan, 1992).Unlike teacher-fronted instruction, where the teacher controlsinteractional patterns, in small-group learning, how students participate

in group work is rather unpredictable The teacher cannot forcestudents to participate in groups in the way the teacher wants or expectsthem to Instead, students are free to choose whether or not to workactively in groups, while nonlinguistic and noncognitive contextualfeatures, such as classroom layout, students’ interpretation of their tasks,and peer relationships, may influence how students act in groups (cf.Coughlan & Duff, 1994; Do¨ rnyei & Kormos, 2000; Jacob, 1999).Participation in group activity includes students’ willingness to commu-nicate (WTC) with peers in the target language Several factors mayimpact on such willingness

Factors related to overall L2 WTC have been researched andidentified (e.g., Cle´ment, Baker, & MacIntyre, 2003; Yashima, 2002;Yashima, Zenuk-Nishide, & Shimizu, 2004) However, factors specific toWTC in L2 group work have not been identified Likewise, there is a lack

of understanding of the relationships among factors that may affectstudents’ willingness to work in L2 groups

BELIEFS ABOUT L2 GROUP WORK

One possible influence on students’ WTC with group members in theL2 is what they believe about the value of group work It is felt thatstudents’ beliefs about L2 group work are formed at least in part based

on their experiences in social contexts, including language learningsituations Such beliefs are not only cognitive but also social constructsformed in social contexts in which learners accumulate experience andencounter problems (Barcelos, 2006) Their beliefs are relatively stablebut also changeable (Sakui & Gaies, 1999; Woods, 2006); they alsoinfluence students’ L2 learning behavior (Horwitz, 1985, 1987, 1988;Woods, 2006) In the study I discuss here, I operated on the assumptionthat such beliefs about L2 group work are strongly related to willingness

to participate in L2 group work

Assuming that beliefs about L2 group work that students hold vary,for example, ‘‘L2 group work is useful for L2 development’’ and ‘‘L2

group work is of intrinsic value in and of itself,’’ as shown in my previousstudy (Fushino, 2006a), it was hypothesized that beliefs about groupwork consist of five components discussed in the cooperative learningliterature and based on my own experience as an English teacher whooften uses group work In a subsequent study (Fushino, 2006b), Icategorized them as beliefs concerning (a) the value of cooperation, (b)the efficiency of group work, (c) knowledge co-construction/peerscaffolding in group work, (d) relationships with other group members,and (e) teacher–student roles in college English classrooms However, inconducting a factor analysis of those components, only three belieffactors were identified: positive beliefs about the value of group work,beliefs about group work usefulness, and negative traditional instructionorientation (Fushino, 2008).

Some forms of group activities require relatively high level nication skills and the ability to engage in spontaneous interaction.Dwyer (1998) investigated the relationship between communicationapprehension and learning style preferences of undergraduate students

commu-in a public speakcommu-ing course at a U.S university and found that femalestudents with high communication apprehension often do not havecommunication skills adequate to help them feel competent in an active,undirected learning environment It is not difficult to imagine thatindividuals with a high degree of communication apprehension will notparticipate in group work actively Keaten, Kelly, and Pribyl (1997)measured levels of self-reported communication apprehension inelementary and secondary school students in Japan, with the PersonalReport of Communication Fear questionnaire (McCroskey, Andersen,Richmond, & Wheeless, 1981) Keaten et al reported that the Japaneseschool children in their study showed increasing levels of first language(L1) communication apprehension in the classroom in general as theygrew older (from elementary school to junior high school)

It has also been reported that Japanese college students showedstronger communication apprehension, both in Japanese and English,

than their U.S counterparts (McCroskey, Gudykunst, & Nishida, 1985;Pribyl, Keaten, Sakamoto, & Koshikawa, 1998) In the study byMcCroskey et al., the group of respondents (Japanese universitystudents) was split into two, and the first half reported their L2communication apprehension while the second half reported their L1communication apprehension on the short version of the PersonalReport of Communication Apprehension (McCroskey, 1978) TheJapanese students showed statistically significant higher L1 and L2communication apprehension than their U.S counterparts speakingEnglish, and the Japanese students’ levels of L1 and L2 communicationapprehension were almost the same Pribyl et al (1998) investigatedJapanese university students’ levels of communication apprehensionwith the Personal Report of Communication Apprehension-24(McCroskey, Beatty, Kearney, & Plax, 1985) They compared the resultswith normative means taken from U.S samples published in Richmondand McCroskey (1995) and reported that Japanese college students hadhigher communication apprehension than their U.S counterparts.However, those studies were focused on trait-like communicationapprehension and self-perceived communication competence, andthose constructs were not directly related to L2 group work In thecooperative learning and small-group-work literature, no researchershave conducted focused studies on communication apprehension andself-perceived communication competence.

In L2 research, MacIntyre, Cle´ment, Do¨ rnyei, and Noels (1998),based on the original proposal by Cle´ment (1980), posited that L2communication confidence consists of L2 communication apprehensionand perceived L2 communicative competence Perceived L2 communicativecompetence is students’ self-perception of their ability to communicate

in an L2 In the study described here, I adapted their idea and viewedcommunication confidence in L2 group work as being made up ofcommunication apprehension in L2 group work and self-perceivedcommunicative competence in L2 group work

MacIntyre, Noels, and Cle´ment (1997) conducted a study in aFrench-as-a-second-language course at a Canadian university in order

to investigate how language anxiety, perceived and actual L2competence, were related The participants responded to a ques-tionnaire on language anxiety and a can-do questionnaire Then theyperformed tasks that corresponded to items on the can-do ques-tionnaire MacIntyre et al reported that self-perceived competencewas significantly correlated with anxiety and actual competence Theyalso suggested that less anxious students who perceived themselves asmore competent chose to communicate more willingly Therefore, itmay be that, even if students have low communication apprehensionbut view their communicative competence as being high (i.e., high

communication confidence in L2 group work), they will more likely

be ready to participate in L2 group work Moreover, both nication apprehension in L2 group work and self-perceived commu-nicative competence in L2 group work can change as studentsexperience L2 group work If students become more confident in L2group work, their increased confidence will presumably make themmore willing to engage in L2 group work

commu-WTC IN AN L2

The notion of WTC in an L2 has recently begun to attract interest insecond language research WTC in an L2, which has been defined as ‘‘areadiness to enter into discourse at a particular time with a specificperson or persons, using an L2’’ (MacIntyre et al., 1998, p 547), isconsidered to be the direct antecedent of students’ actual engagement

in L2 communication

MacIntyre et al (1998) proposed a well-known six-layered heuristicmodel of variables that influence L2 WTC The first layer is actual L2use, which is directly influenced by L2 WTC (Layer 2) Layer 3 has twoconstructs: desire to communicate with a specific person and degree ofcommunicative self-confidence Factors in Layer 3 directly influence L2WTC These three layers are considered situation-specific influences.Layers 4, 5, and 6 are considered stable, long-lasting influences on L2WTC Layer 4 represents motivational propensities and consists of threeconstructs: interpersonal motivation, intergroup motivation, and self-confidence Layer 5, which represents the affective-cognitive context,includes intergroup attitudes, social situation, and communicativecompetence The bottom layer, Layer 6, is the social and individualcontext that consists of intergroup climate and personality Layersfurther from L2 WTC (Layer 2) are felt to have less direct influence onWTC

MacIntyre and Charos (1996) conducted a study in a Canadianbilingual (English and French) context in order to investigate astructure of L2 WTC by testing a path model A questionnaire wasadministered to introductory-level adult learners of French It was foundthat frequency of actual use of L2 was predicted by motivation and WTC,and WTC was predicted by perceived communication competence andcommunication anxiety No relationship between WTC and motivationwas found in this bilingual setting

MacIntyre and Cle´ment (1996) reported a questionnaire studyconducted in a Canadian monolingual university with Anglophonestudents learning French as a second language The study was a

replication of MacIntyre and Charos’s study (1996) MacIntyre andCle´ment reported that perceived L2 communicative competence, L2anxiety, and motivation are directly related to L2 WTC in themonolingual (English-speaking) setting They also reported thatintegrativeness and attitudes toward the learning situation are related

to L2 WTC through motivation

Recently, efforts have been made to situate the heuristic model ofMacIntyre et al (1998) in culture-specific contexts (e.g., Wen & Cle´ment,2003; Yashima, 2002; Yashima et al., 2004) Yashima and Yashima et al.applied the above-mentioned model presented in MacIntyre and Cle´ment(1996) to the Japanese context First-year Japanese university studentsmajoring in information science, in Yashima’s study, and first-yearJapanese high school students receiving intensive English languageinstruction in their high school, in the study by Yashima et al., responded

to a set of questionnaires made up of attitudinal–motivational measuresand L2 WTC scales Yashima and Yashima et al hypothesized that L2 WTC

in the Japanese context is directly influenced by (a) students’ confidence

in L2 communication, composed of Communication Apprehension andPerceived Communicative Confidence in the L2, (b) Motivation, and (c)International Posture, which consists of ‘‘interest in foreign or interna-tional affairs,’’ ‘‘willingness to go overseas to stay or work,’’ and a

‘‘readiness to interact with intercultural partners’’ (Yashima et al., 2004,

p 125) If a student has low L2 Communication Apprehension and high(self) Perceived Communicative Competence in the L2, the person isconsidered to have high L2 communication confidence Yashima andYashima et al tested their model using structural equation modeling andfound that L2 WTC is directly influenced by International Posture andCommunication Confidence, but not by L2 Learning Motivation L2motivation is an intervening variable that indirectly influences L2 WTC byinfluencing L2 communication confidence

PURPOSE OF THIS STUDY

As explained above, L2 communication confidence, beliefs about L2group work, and L2 WTC may all impact on students’ participation ingroup work Thus far in this article these components of WTC have beenexplored separately Expectancy-value theory (Wigfield, 1994; Wigfield &Eccles, 2000) attempts to link these three constructs in L2 group worksettings According to this theory, people are motivated to engage intasks that they are confident that they can accomplish and that theyvalue Applying this theory to the L2 group work context, it can behypothesized that if students are confident in L2 communication and

have strong, positive beliefs about L2 group work, they will willinglycommunicate in L2 group work.

However, the causal relationships among these three factors have notbeen investigated empirically In other words, we do not know howcommunication confidence in L2 group work—an aggregated constructthat consists of communication apprehension in L2 group work and self-perceived communicative competence in L2 group work (cf Cle´ment

et al., 2003; Yashima, 2002)—and beliefs about L2 group work togetheraffect WTC in L2 group work Yashima investigated factors that affectWTC in an L2 in four different situations with three differentinterlocutors, but in the study described here, communication occurred

in a much more limited situation, that is, in L2 group work As such,direct application of her L2 WTC model to the L2 group work situationmight not work Therefore, the purpose of this study was to investigatethe causal relationships among (a) Communication Confidence in L2Group Work, (b) Beliefs About L2 Group Work, and (c) WTC in L2Group Work while taking the models of Yashima (2002) and MacIntyre

et al (1998) into consideration The hypothesized model is shown inFigure 1 The belief underlying this study was that, if causal relationshipsamong Communication Confidence in L2 Group Work, Beliefs AboutL2 Group Work, and WTC in L2 Group Work are identified, classroomteachers will be better equipped to tailor their instruction in ways thatwill boost students’ engagement in classroom tasks The study wasdesigned to investigate the following research question: How docommunication confidence in L2 group work and beliefs about L2group work affect students’ WTC in L2 group work? A structuralequation modeling (SEM) approach was employed to answer thisresearch question

METHOD

Participants and the Research Site

Seven hundred and twenty-nine first-year students (with a mean age

of 18.65 years, standard deviation [SD] 5 0.86) from various

FIGURE 1 Second language (L2) group work (GW) model (WTC 5 willingness to

communicate).

departments at a co-ed university in the Tokyo area participated in thisstudy Among them, 329 were males, 392 were females, and 8 did notindicate their gender The students had at least 6 years of formalEnglish education at their junior and high schools prior to enteringthis university The precise levels of the students’ English proficiencywere not measured; however, because the university is acknowledged to

be a prestigious university, their proficiency levels were in the higherrange, as defined in Japan Students took a placement test at thebeginning of the semester and were placed in classes based on theresults of the test Two courses were prepared for them: aCommunicative Course and a Language and Culture Course In eachsemester, three kinds of classes (four 90-minute lessons per week intotal) were offered In the Communicative Course, the oral commu-nication class (twice a week), the reading and writing class (once aweek), and the listening class (once a week) were provided, and in theLanguage and Culture Course, the listening and reading class (twice aweek), the extensive reading class or learning with video class (once aweek), and intercultural communication class (once a week) wereoffered Within the same course, different textbooks were used forclasses with higher proficiency students and for those with lowerproficiency The English program at the university employed anEnglish-only policy, so teachers were told to use English for instruction

in the two courses, regardless of the students’ English proficiency.Japanese was avoided as much as possible even in the lower proficiencyclasses

Material

The data used in this study were solely obtained from a questionnaire.The questionnaire (in Japanese) was constructed for a larger researchproject in which this study was included Response items were writtenbased on the theories of cooperative learning, my observation ofstudents’ group work, communication apprehension (McCroskey et al.,

1985, in particular), and self-perceived communication competenceliterature as well as WTC literature I also sought advice from acooperative learning specialist The draft version of the questionnairewas shown to one of my previous students (a second-year universitystudent at that time) to determine if there were any difficulties inunderstanding and responding to the items The student filled out thequestionnaire and timed how long it took her to complete it Shereported that it took her about 10 minutes to complete thequestionnaire and that no item was unclear to her Originally, thequestionnaire had 98 items, consisting of 8 parts: (a) communicationapprehension in L2 group work, (b) self-perceived communicativecompetence in L2 group work, (c) the value of cooperation, (d) the

efficiency of group work, (e) knowledge co-construction/peer ing in group work, (f) relationships with other group members, (g)teacher–student roles in college English classrooms, and (h) WTC in L2group work However, after a pilot administration and a factor analysis(Fushino, 2008), only six factors were extracted, and only high-loadeditems were retained in the final version of the questionnaire Thequestionnaire used in this study consisted of 64 five-point Likert-scaleitems (1 5 strongly disagree; 2 5 disagree; 3 5 neutral; 4 5 agree; 5 5strongly agree) in six sections (see the Appendix for an Englishtranslation) The results of the questionnaire were submitted to factoranalysis, and six factors were identified, as expected The names andreliability indices of the questionnaire sections follow:

scaffold-Communication Apprehension in L2 Group Work (10 items, a 5 0.88): Theitems were written in consultation with the Personal Report ofCommunication Apprehension (PRCA-24; McCroskey & Richmond,1982) to measure students’ degree of communication apprehension inL2 group work A sample item is ‘‘I feel nervous when I work in a group

in English class.’’ Item 1 was reverse coded

Self-Perceived Communicative Competence in L2 Group Work (10 items, a 50.94): Considering what verbal actions students were likely to take in L2group work, 10 can-do-type items were constructed in order to measurehow students perceived their level of communication competence in L2group work A sample item is ‘‘If my group members ask me questions inEnglish, I can answer them in English.’’

Positive Beliefs About the Value of Group Work (6 items, a 5 0.93):Students were asked to indicate to what degree they agreed withstatements regarding beliefs about the value of group work, because itwas considered that students who believe in the value of group workwere likely to participate in group work actively A sample item is ‘‘Groupwork is important for human growth.’’

Negative Traditional Instruction Orientation (8 items, a 5 0.89): Studentswere asked to show what attitudes they had toward traditional instructionbased on the idea that students with a higher negative traditionalinstruction orientation were considered to be more positive toward moreinnovative instruction, such as group work, and thus were likely toengage in group work more actively Because negative wording wasavoided in order to prevent student confusion (Brown, 2001; Do¨ rnyei,2003), all the items were written in affirmative sentences, such as ‘‘I learnwell in a teacher-led class that has no group work.’’ Therefore, all theitems were reverse coded in order to make the higher score indicatehigher negative traditional instruction orientation

Beliefs of Group Work Usefulness (20 items, a 5 0.95): Students whobelieved that group work was useful for them and for their learning wereconsidered to be more likely to join group work, and the items in thissection were constructed to measure the degree to which studentsconsidered group work useful A sample item is, ‘‘During group work Ilearn various opinions and ideas from group members.’’

WTC in L2 Group Work (10 items, a 5 0.96): The items in this sectionasked how willing students were to communicate in the 10 situationsdescribed All the situations were the same as those for the items writtenfor Self-Perceived Communicative Competence in L2 Group Work Asample item is, ‘‘If my group members ask me questions in English, I amwilling to answer them in English.’’

Procedure

The questionnaire described above was administered anonymously atthe end of the spring semester in July, 2005, by eight colleagues of mine.Selection of the participants was based on convenience At this university,teachers could not choose which class to teach These eight teachers wereassigned to one of the three classes that the students took Hence it seemedunlikely that sampling by convenience would distort the results of the study.Analyses

As stated above, the purpose of this study was to explore therelationship between Communication Confidence in L2 Group Work,Beliefs About L2 Group Work, and WTC in L2 Group Work by usingSEM Communication Confidence in L2 Group Work and Beliefs AboutL2 Group Work are psychological constructs that cannot be measureddirectly Therefore, there was a need to operationally define thesevariables Communication Confidence in L2 Group Work and BeliefsAbout L2 Group Work were operationally defined as follows:

Confidence 5 (M of Apprehension Reversed + M of Confidence)/2Beliefs 5 (M of Value + M of Instruction + M of Usefulness)/3where Confidence 5 Communication Confidence in L2 Group Work;Apprehension Reversed 5 reverse-coded Communication Apprehension

in L2 Group Work; Competence 5 Self-Perceived CommunicativeCompetence in L2 Group Work; Beliefs 5 Beliefs About L2 GroupWork; Value 5 Positive Beliefs About the Value of Group Work;Instruction 5 Negative Traditional Instruction Orientation; andUsefulness 5 Beliefs of Group Work Usefulness The possible maximumscore was 5 The means and intercorrelations of these variables areshown in Table 1 After data screening and assumption checks, astructural equation model was made and tested using AMOS 5.1J(statistical software, Arbuckle, 2004) with 592 cases

RESULTS AND DISCUSSION

In order to investigate causal relationships among CommunicationConfidence in L2 Group Work, Beliefs About L2 Group Work, and WTC

in L2 Group Work, an SEM was tested SEM processes consist of a series

of regressions and confirmatory factor analyses (Byrne, 2001) In thatrespect, SEM is useful for investigating how well a theoretical modelexplains the interrelationships among a set of variables (Hu & Bentler,1999) That is, by using structural equation modeling, we can determinewhether a theory-driven model fits the observed data to a statisticallysignificant degree (Sasaki, 1993), and we can obtain information on howplausible the model is in terms of revealing the existence of cause–effectrelationships (Murayama, 1998) Although it is true that, in order tomeaningfully suggest any causation, there is a need for a great deal oftheory and research, as Kline (2005) explains, the pictorial aspect ofstructural equation modeling enables us to see visually the influence ofone variable on another variable It was this visual aspect of SEM that wasappropriate for the purposes of this study

The Hypothesized Model

The heuristic model of WTC in an L2 by MacIntyre et al (1998)indicates that WTC (Layer 2) is directly affected by State Communicative

of Group Work Usefulness; Beliefs 5 Beliefs About L2 Group Work; WTC 5 Willingness to Communicate in L2 Group Work.

All correlations are significant at p , 0.01 (1-tailed).

Self-Confidence (consisting of ‘‘perceived competence’’ and ‘‘a lack ofanxiety,’’ p 549) and Desire to Communicate with a Specific Person(Layer 3) However, their model was designed to account for L2 WTC inthe Canadian ESL context, which is very different from the context ofthis study, where English is a foreign language that is seldom usedoutside of English classrooms Besides, the primary purpose of the studywas to investigate students’ WTC in L2 Group Work in English classesand not their willingness to use the L2 in communicative situationsoutside classrooms Because of these differences, as Wen and Cle´ment(2003) noted in their conceptualization of WTC in the Chinese context,

it was not feasible to directly apply the model of MacIntyre et al to thecurrent situation However, their heuristic model became a usefulconceptual tool in constructing my model

Yashima (2002) investigated an L2 communication model in theJapanese EFL context by testing via SEM and found that WTC in an L2within the Japanese EFL framework was directly affected by two latentvariables: L2 Communication Confidence, as indicated by ‘‘communica-tion anxiety in the L2’’ and ‘‘perceived communication competence in theL2,’’ and International Posture Latent variables are ‘‘theoreticalconstructs that cannot be observed directly’’ (Byrne, 2001, p 4).Observed variables are those that can be directly measured by observation,such as self-report responses and test scores (Byrne, 2001) However, WTC

in the L2 in Yashima’s study was considered to be WTC outside of Englishclassrooms In the present study, the focus was on WTC in L2 GroupWork, in particular; therefore, Beliefs About L2 Group Work, rather thanInternational Posture, was hypothesized to affect WTC in L2 Group Work

I also view these beliefs to be relatively stable, not changing from class toclass On the other hand, beliefs can change over time; thus they are not acomplete trait Therefore, Beliefs About L2 Group Work should belocated in Layer 4 in the model of MacIntyre et al (1998) In thisconceptualization, Beliefs About L2 Group Work indirectly affects WTC inL2 Group Work via Communication Confidence in L2 Group Work,which in turn, is affected by Beliefs About L2 Group Work

Based on this idea, I developed a hypothetical model The headed arrows in Figure 2 illustrate the hypothesis that Beliefs About L2Group Work (a latent variable) indirectly affects WTC in L2 Group Work(an observed variable) via Communication Confidence in L2 Group Work(a latent variable), which has two indicator variables, CommunicationApprehension in L2 Group Work and Self-Perceived CommunicativeCompetence in L2 Group Work Beliefs About L2 Group Work has threeindicator variables: Positive Beliefs About the Value of Group Work,Negative Traditional Instruction Orientation, and Beliefs About GroupWork Usefulness Each indicator variable was specified by aggregating thevalues of all the items used in the measurement variable Ovals in the