english as a second language students and english language

ESL Students and English Language Anxiety 327

BN. ELENI PAPPAMIHIEL

The Florida State University

English as a Second Language Studentsand English Language Anxiety:Issues in the Mainstream Classroom

RESEARCH IN THE TEACHING OF ENGLISH • VOLUME 36 • FEBRUARY 2002 327

This study addresses issues of English language anxiety in two settings: English as a secondlanguage and mainstream classrooms. Participants were 178 middle-school Mexican immi-grant students attending school in the U.S. Participants were given the English LanguageAnxiety Scale, which was analyzed with three statistical tests: paired t-tests to analyze broadlevels of anxiety between ESL and mainstream classes; ANCOVAs to identify how levels ofanxiety correlated with the specific factors of years in the U.S., levels of academic achieve-ment, listening and speaking skills, reading and writing skills, and gender; and an explor-atory factor analysis to identify additional factors contributing to anxiety. A second data set,focus group transcripts, was analyzed to identify additional factors and coping strategies.Results showed several related but different types of English language anxiety and a signifi-cant gender difference. The focus groups revealed that interaction with Chicano studentsraised anxiety levels and that such strategies as avoidance were used to reduce anxiety. Thepaper concludes with recommendations for teaching and research that recognize the complex-ity of anxiety for English language learners.

IntroductionIt is now the end of April, and still Pablois not participating in class. He oftentalks to his friends, and his teacher haseven had to conference with him abouthis inattention during classroom dis-cussions. His English skills had enabledhim to pass his classes and meet mini-mum requirements on the standardizedtest used by the district, thus allowinghim to be mainstreamed. Yet even withsupport from the English as a second

language (ESL) teacher, he is not doingas well in the mainstream class as he hadin ESL classes, remaining shut off fromthe native English-speaking students.Even when he is placed in a group withother U.S.-born students who are flu-ent in English and Spanish, he with-draws from contact with them.

Unfortunately, Pablo’s situation isnot rare among mainstreamed Englishlanguage learners (ELLs). In fact,preservice teachers who go out into the

SElson

Copyright © 2002 by the National Council of Teachers of English. All rights reserved.

328 RESEARCH IN THE TEACHING OF ENGLISH • VOLUME 36 • FEBRUARY 2002

schools often write in their journalsthat ELL students often interact less inthe mainstream classroom, choosinginstead to remain as far away as possiblefrom the action of the classroom. Whilethere are many reasons why ELL stu-dents have difficulty in the mainstreamclassroom, including a lack of teacherengagement (Verplaste, 1998) and lim-ited cognitive skills in English(Cummins, 1984), one source of diffi-culty that is often overlooked is theaffective environment of the main-stream classroom. While educators andadministrators concentrate mostly onEnglish proficiency, factors such asanxiety are often overlooked inmainstreaming decisions.

In order to begin to address thiscomplex issue, this study focuses on theEnglish language anxiety of 178 middleschool Spanish-speaking ELL studentswho were mainstreamed part of the day.More specifically, in this study I ana-lyzed students’ responses to the EnglishLanguage Anxiety Scale (ELAS) in orderto answer the following questions:

1. Is there a difference between levelsof English language anxiety in ESLand mainstream classes?

2. If differences occur, is there arelationship between levels ofEnglish language anxiety in ELL andmainstream classrooms and timespent in the U.S.; achievement (asnoted by grades in their ESL classes);listening and speaking skills; readingand writing skills; and/or gender?

To investigate factors not anticipatedthrough the original research questions,

I posed more exploratory questionsbased on both the ELAS and focusgroup data:

3. If differences occur, is there arelationship between levels ofEnglish language anxiety in ELL andmainstream classrooms and factorsthat emerge during a factor analysisof the ELAS?

4. If differences occur, what corrobo-rating and additional factors emergethrough focus group discussionswith research participants?

5. What coping strategies do highlyanxious participants use to reducelevels of anxiety?

Theoretical BackgroundIn light of the increasing diversity in U.S.classrooms, mainstream classroom teach-ers are seeing more ELL students intheir classes. Furthermore, recent legis-lative changes such as Proposition 227in California (which limits the amountof time an ELL student can spend in asheltered English class) increase theneed for extensive language accommo-dations for these students since theyenter the mainstream at earlier points intheir language acquisition. As ELL stu-dents enter classrooms with fewer lan-guage skills, they have more to benervous about, increasing the chancesthat anxiety will affect the learningprocess, as has happened in foreignlanguage classrooms (Daley, Onwueg-buzie, & Bailey, 1997; Ehrman & Ox-ford, 1995; Gardner, 1985; MacIntyre &Gardner, 1993, 1994). Consequently, itis imperative that mainstream teachersknow more about the anxiety levels of


their ELL students and, more impor-tantly, how to reduce their apprehension.

Theories of AnxietyI next outline specific theories ofanxiety to illustrate the development ofanxiety theory, moving from general-ized, all-encompassing theor ies(Bandura, 1991; Pekrun, 1992) to moresituation-specific theories of languagelearning anxiety (MacIntyre & Gardner,1994) and then to theories that focuson contextual levels of anxiety withinindividuals (Pappamihiel, 1999). In keep-ing with MacIntyre and Gardner’s (1991)view that “while the instruments usedto measure language anxiety should bespecific to the language area, theoreticallinks to the more general anxiety litera-ture can be strengthened” (p. 43), Iinclude not only frameworks devel-oped specifically for language learningbut also models derived from psychol-ogy.

Anxiety: Self-efficacy and AppraisalGeneral theories of anxiety can beconceptualized using two models:Pekrun’s (1992) Expectancy-ValueTheory of Anxiety (EVTA) andBandura’s (1991) theory of self-efficacy.Each of these models uses differenttypes of appraisals to explain and pre-dict anxiety reactions in individuals.According to these models, concepts ofworry and distraction relate first toappraisals of situations as threatening ornot and then to learners’ determina-tions of their efficacy in dealing withthese situations. Pekrun’s (1992) EVTAmodel combines situation-outcomeexpectancies (appraisals of a situation as

being threatening or not) with action-control expectancies (appraisals aboutone’s ability to initiate and carry out aneffective solution). Foreseeing negative,potentially harmful events in whichindividuals cannot see themselves aseffective mediators often produces anxi-ety. Similarly, Bandura’s (1991) theoryof self-efficacy posits that when asituation is perceived as threatening, theresultant anxiety is dependent on anindividual’s perception of his/her abil-ity to deal positively with that threat.Bandura additionally argues that self-esteem can act as a mitigating factor inanxiety-producing circumstances.

When learners see situations asthreatening, there can be an adverseaffect on learning. Because highly anx-ious individuals are often in a state ofdivided attentional resources (Eysenck,1979), their ability to concentrate andbe successful at learning tasks is ham-pered. In other words, when studentsare constantly preoccupied with thethreat a learning situation poses, theycannot fully concentrate on that task. Atthe early stages of learning, students areusing many attentional resources toaccomplish basic tasks that they havenot yet learned how to complete auto-matically (Schallert, 1991). Highly anx-ious students are not able to automatizeactions as effectively since theirattentional resources are divertedthrough task-irrelevant processingbrought about by high levels of anxiety.

In addition to challenges withresource allocation, learners sometimesengage in self-deprecating (“I’m stu-pid,” “I can’t do this”) and self-focusedthoughts that interfere with feelings of


self-efficacy. These negative thoughtsadversely affect a student’s ability totake advantage of learning opportuni-ties, affecting students’ ability to seethemselves as successful learners (Gib-bons, 1991; Hass & Eisenstadt, 1991;Sarason, 1972). Anxiety is a complexconcept, dependent upon not onlyone’s feelings of self-efficacy but alsoappraisals concerning the potential andperceived threats inherent in certainsituations. These many appraisals coupledwith the influence of task-irrelevantprocessing can negatively affect thelearning process, often in ways thatstudents are not even aware of (Tobias,1986).

State, Trait, and Situational AnxietyPekrun (1992) argues that in instancesof high anxiety, habitualized reactionscan cause individuals who have experi-enced many threatening situations inthe past to be more likely to perceivefuture situations as threatening. Simi-larly, Vasey and Daleiden (1996) notethat highly anxious individuals mayhave a lower threshold of threat recog-nition, seeing generally ambiguous situ-ations as potentially threatening moreoften than moderately anxious persons.Because of the possibility that someindividuals are more prone to anxietythan others, it is necessary to differenti-ate between individuals who are oftenanxious and those who are not.Spielberger (1983) describes this differ-entiation as the state/trait dichotomy.

Individuals who are more anxiousand more likely to become anxiousregardless of situation are referred to ashaving trait anxiety; that is, anxiety is a

part of their character or an aspect of amore serious disorder. However, thosewho are able to appraise situationsaccurately as being threatening or notwithin reasonable limits are said to havestate anxiety, a social type of anxiety thatoccurs under certain conditions. Forexample, a person may not ordinarilybe anxious but becomes so when askedto make a public address. This differen-tiation is critical in the study of anxietybecause it allows the separation ofindividuals who are likely to be anxiousin any variety of situations from thosewho would not normally be anxious.

Some researchers further differen-tiate the concept of anxiety by distin-guishing between cognitive (worry) andemotional (affective) components ofanxiety (Deffenbacher, 1980; Schwarzer,1986). According to Deffenbacher, anxi-ety related to cognitive interference(e.g., learning challenges) is due to ex-treme instances of worry, not the arousalelement of anxiety. Thus the cognitivetype of anxiety associated with class-room learning is rarely facilitative.

In the field of second languageacquisition, MacIntyre and Gardner(1991) draw upon the work done bySpielberger (1983) to make an addi-tional distinction, situation-specific anxi-ety. Individuals who suffer fromsituation-specific anxiety may appraisecertain events as anxiety-producingonly when certain factors are present.For example, a student may be anxiety-free when writing an essay in English.However, when asked to write a similaressay in French, a second language, thesame student may then feel higherlevels of anxiety. .


Horwitz, Horwitz, and Cope (1986)argue that people who feel competentin their native language can feel re-duced to a childlike state when asked touse their second language. Additionally,learners of a foreign language are oftensubjected to threats to their self-per-ception in the foreign language class-room setting. They conclude that foreignlanguage anxiety can be associated withthree factors: a fear of negative evalua-tion, test anxiety, and communicationapprehension.

Horwitz et al. (1986) conceptual-ize foreign language anxiety as “adistinct complex of self-perceptions,beliefs, feelings, and behaviors related toclassroom language learning arising fromthe uniqueness of the language learningprocess” (p.31). They distinguish for-eign language anxiety from communi-cation apprehension, as outlined byMcCroskey (1987), in that individualswho show no communication appre-hension in their native language willoften exhibit foreign language anxiety(Applebaum & Trotter, 1986), whichgoes beyond communication appre-hension to include a fear of negativeevaluation and test anxiety. Addition-ally, because of the differences betweena foreign language learning situationand an ESL learning situation, in whichstudents are expected to create a secondlanguage persona, a stronger link toidentity factors needs to be considered.

In terms of self-concept and iden-tity, Guiora (1983) argues that languagelearning can be extremely traumatic forsome students because it threatens theirsense of self and worldview. In fact, theriskiness of the language learning situa-

tion may lead learners to fossilize or halttheir language acquisition because therisk to the learner’s national identitymay be seen as greater than the per-ceived benefits of acquiring better,more proficient second language skills(Beebe, 1983). Additionally, Schumann(1997), while currently investigatingphysiological aspects of language anxi-ety, maintains his pidginization hypoth-esis in which he posits a strongrelationship between an individual’swillingness to acquire a language andhis or her relationship with the targetlanguage group. In other words, themore psychological and social distanceperceived, the less likely it will be forthat individual to achieve high levels oflanguage skill in the target language.

Hence, anxiety in general can beassociated with threats to self-efficacyand appraisals of situations as threaten-ing. In a specific situation such aslanguage learning, a fear of negativeevaluation, test anxiety, communicationapprehension, and threats to one’s senseof self can reduce feelings of self-efficacy and increase the chances that asecond language situation will be seenas threatening. Additionally, the socialdistance felt by many Mexicans in theU.S. can exacerbate these factors, result-ing in differing levels of languageanxiety based on the context of thesecond language situation and the so-cial distance perceived in each.

Situational Anxieties in ELL StudentsAssuming an individual with state anxi-ety rather than trait anxiety, concepts ofself-efficacy are tied to past successes,vicarious experiences, and social per-


suasion (Bandura, 1991). Yet MexicanELL students in the U.S. are often at riskbecause, in many cases, they lack vicari-ous experiences and social persuasionthat would provide successful modelsand encouragement from others(Cummins, 1996; Zambrana & Silva-Palacios, 1989). Also, ELL students canbecome more susceptible to high levelsof anxiety related to language learningbecause of the language shock oftenexperienced by ELLs (Olsen, 1997).

Hispanic Americans experienceextremely high drop out rates, withonly 63% of all 18-24-year-old Hispan-ics completing some sort of high schooleducation, either through the attain-ment of a traditional diploma or GED(Kaufman, Kwon, & Klein, 1999).Hence, they are limited in the numberof academically successful models andother vicarious experiences that wouldincrease their levels of self-efficacy.Additionally, because of the margin-alization of Hispanic American groupsin the U.S., Mexican ELL childrenoften internalize negative stereotypesthat reduce their feelings of self-effi-cacy by limiting the social persuasion(encouragement from others) thatwould ordinarily tell them that they cansucceed (Zambrana & Silva-Palacios,1989).

Often ELLs who lack proficiencyin English are subject not only tojudgments about their language abilitybut also about their significance asindividuals (Cummins, 1996). Thesejudgments can be aggravated by teach-ers’ misconceptions about languagelearning. The difference between socialEnglish and academic English profi-

ciency can create false impressionsabout how much English an ELLstudent should learn in a certain amountof time. Social English, referred to byCummins (1984, 2000) as basic inter-personal communication skills, can beacquired in one to two years. However,the cognitive/academic language profi-ciency that is needed to be successfulacademically can take anywhere fromfive to eight years to acquire (Collier,1987; Cummins, 2000). When ELLstudents are mainstreamed, both teach-ers and students can have the implicitidea that they are fully prepared to workwith academic English, and when theyneed further accommodations, students’feelings of self-efficacy can be reduced.

Although there have been fewinvestigations on the effect of Englishlanguage anxiety on the acquisition ofEnglish, there have been several studiesthat have concluded that foreign lan-guage anxiety does have an adverseeffect on learning (Daley et al., 1997;Ehrman & Oxford, 1995; Gardner,1985; MacIntyre & Gardner, 1993,1994). These same factors that adverselyaffect learning in the foreign languageclassroom are present in the ESL class-room but at a more intense level.Moreover, for Mexican ELL studentsthese challenges can be exacerbated bydifficulty in developing a positive En-glish language identity. An ELL student’sability to make effective situationalappraisals can be hampered not only bylinguistic differences but also by cul-tural differences.

In sum, general issues of self-efficacy and expectancy-value theorycan be influenced by a fear of negative


evaluation, test anxiety, communicationapprehension, and identity factors. How-ever, these influences on anxiety do notoccur in a vacuum. English languageanxiety can be described as a socialanxiety, dependent upon interactionswith others. Therefore, the implica-tions of English language anxiety froma socioconstructivist perspective mustbe considered. According to Vygotsky(1978) and others, learning is depen-dent upon the social interactions thatoccur in the classroom. Withdrawalfrom this interaction as a result of highlevels of English language anxiety isperhaps the most harmful effect ofEnglish language anxiety.

It is imperative that teachers andadministrators be able to identify highlyanxious students within their schoolsand classrooms, ideally before the main-streaming process takes place. Whenteachers and administrators can identifyhighly anxious students more efficiently,efficacy beliefs (Bandura, 1995) andexpectancy-value-oriented preventions(Pekrun, 1992) can be initiated. In pro-viding this affective support for ELLstudents in mainstream classes, educatorscan increase their chances for academicsuccess beyond minimum standards.

MethodSettingSituated in a medium-sized schooldistrict in a major city in Texas, theschools involved in this study weremostly in the poorer neighborhoodswhere most of the ELL children in thisdistrict were schooled. In four of theseven sites, ESL classes were conductedin mobile units outside the permanent

buildings. The ELL density of the sitesranged from 6% of total school popula-tion to 27% of the total population.School size ranged from 509 to 1201total students; the largest schools, how-ever, did not have the largest number ofELLs.

ParticipantsThe study included 178 Mexican-bornmiddle school students (grades 6-8)enrolled in ESL programs. These stu-dents had all been in the U.S. for at leastone year and were mainstreamed for aportion of their school day, many foralmost the entire day. The range of timespent in the U.S. ranged from 1-12years, with a mean of 2.47 years in theU.S. All students spoke English at theintermediate level or above as noted bytheir ESL teachers. Achievement levelsfor the students ranged from extremelysuccessful to struggling. The students’semester-end grades averaged 83.8, witha range between 50-100, on the 100-point scale used within the school(with 100 equal to an A and less than 60equal to an F). Although many studentsspoke only Spanish at home, approxi-mately 20% of the participants spokesome English at home, mostly withsiblings or cousins. Almost all partici-pants were identified as low SES basedon their participation in the district’sfree and reduced lunch program.

All ESL teachers involved in thisstudy had several years of experienceteaching ESL. Additionally, all had somelevel of proficiency in Spanish, with fiveteachers being native speakers of Span-ish. Five of the participating teacherswere female and two were male. .


Data CollectionEnglish Language Anxiety ScaleAfter participants were identified, theELAS (see Appendix A) was adminis-tered during class time. The ELAS is a20-item Likert-type scale probing par-ticipants’ level of agreement or dis-agreement (1-5 with a neutral option)with statements. Since each statementprobes anxiety levels in two differentenvironments (within ESL and main-stream classes), the participant respondsto 40 prompts. Students were free to askquestions and request clarification onitems.

The ELAS is modeled after atheoretically similar Likert-type anxi-ety assessment, the 33-item ForeignLanguage Classroom Anxiety Scale(FLCAS) designed by Horwitz et al.(1986). Validated by Horwitz (1991),the FLCAS is often used in the field ofsecond language acquisition. The ELASand FLCAS show statistical similarities.According to Cronbach’s alpha test, theELAS shows an internal consistencyreliability of .89 with the FLCAShaving an internal consistency reliabil-ity of .93.

While both the ELAS and FLCASstem from a similar theoretical base,there are significant differences. TheFLCAS was developed for a foreignlanguage population. In other words,the target population included studentswho were studying another languagefrom the perspective of their nativelanguage. The ELAS, on the other hand,was developed for ELLs, learners whoare attempting to acquire a new lan-guage within the context of that lan-

guage, separated to a certain extentfrom their own native language andculture. Hence, adaptation of the FLCASfor use with ELLs cannot be consideredthe ideal fit. Because of this less thanperfect situation, focus groups providedadditional data.

Since ELL students function in atleast two different learning environ-ments, within the ESL class and withinthe mainstream classroom with nativeEnglish speakers, the ELAS (unlike theFLCAS) probes each environment sepa-rately, and statements are presented inboth English and Spanish. Each state-ment in the ELAS is presented from bothperspectives, as in the example below:

1. In ESL classes, I forget how to saythings I know.En clases de ESL, Me siento tannervioso (a) que se me olvida cómo decircosas que ya sé.

2. In regular classes, I forget how to saythings I know.En clases regulares, Me siento tannervioso (a) que se me olvida cómo decircosas que ya sé.

Accordingly, the ELAS is based onthe self-efficacy and situational apprais-als that are described by Bandura (1995)and Pekrun (1992). More specifically, itincludes the three factors outlined byHorwitz et al. (1986) (a fear of negativeevaluation, test anxiety, and communi-cation apprehension) and an identityfactor. Since the FLCAS does notinclude an identity factor, two itemsthat probe identity were added to theELAS (items 7 and 13). .


Focus GroupsSince it can be difficult to give voice toquantitative data in terms of offeringexplanation (Flores & Alonso, 1995),the focus groups were chosen as aqualitative data-gathering method.Krueger (1994) recommends focusgroups for the triangulation of quanti-tative data. Focus groups are furtherfound to help with investigations ofemotions such as revelations of anxiety(Morgan, 1993), although they havenot been used previously in secondlanguage anxiety research. The focusgroups both identified new sources ofanxiety and provided corroboratingillustrations of types of anxiety identi-fied by the statistical analyses.

While focus groups have been usedmostly in the field of marketing andother business specialties, over the pastfew decades they have been used in thesocial sciences as supplemental datagathering instruments and, more re-cently, as the sole method of gatheringdata in qualitative studies. There areseveral benefits that make their use inthis study appropriate. The main benefitof focus group data is their ability touncover information that would notnormally come out in a one-on-oneinterview or would be difficult to seethrough observation.

Some participants have difficultyverbalizing their language anxiety. Al-though they report feeling anxiety, theyhave difficulty giving voice to thesefeelings and remembering specific in-stances. Focus groups are often helpfulin aiding participants in articulatingtheir feelings and reactions because ofthe group synergy generated in these

discussions. Since focus groups rely ongroup interaction more than individualreports, often individuals have the op-portunity to compare their experiencesto those of the other participants, andnew information or different perspec-tives may be sparked by this interaction(Hoppe, Wells, Morrison, Gillmore, &Wilsdon, 1995). According to Morgan(1997),

The basic argument in favor of self-containedfocus groups is that they reveal aspects of ex-periences and perspectives that would not beas accessible without group interaction. (p. 20)

The focus groups used in this studywere homogenous in terms of ELASscore (high anxiety) and gender, basedon Krueger’s (1998) recommendationthat focus groups be as homogenous aspossible while still maintaining theopportunity for alternate points ofview. When dealing with children andadolescents, several researchers suggestbuilding same gender groups to preventpeacock effects. For example, male partici-pants may structure their answers toplease or have a positive effect on femaleparticipants and vice versa (Gillmore, &Wilsdon, 1995; Hoppe et al., 1995;Krueger, 1998). Hoppe et al. also suggestthat when conducting focus groups withchildren that the groups include partici-pants who are already familiar with eachother because they “seemed to feel saferand were more willing to express theiropinions in a group of children theyalready knew” (p. 106). Some focusgroups can have a duration of more thantwo hours; however, Hoppe et al. recom-mend that children’s focus groups last nolonger than one hour in order to maxi-mize children’s attentional resources..


Focus group units were assembledso students of the same age range andgender were in the same group (Hoppeet al., 1995; Krueger, 1998). Hence,there were ten groups in all, four malegroups and six female groups withmembers between the ages of 11 and15. The group sessions were videotapedand conducted in the language mostcomfortable for the students, eitherSpanish or English. While there wasoccasional code switching between lan-guages, all groups chose to speak pri-marily in Spanish. Students who wereidentified as highly anxious were in-vited to participate in focus groupsduring class time. While most studentsagreed to participate in the groups,absences prevented every highly anx-ious student from being involved. Sincestudents themselves did not know theexact date of focus groups, these ab-sences did not appear to be attempts toavoid the group.

From the literature cited and priorexperiences with preliminary focusgroups, the following focus group pa-rameters were set: (a) Group partici-pants were of the same gender, (b)Group participants were familiar witheach other, (c) Group participants scoredwithin the highly anxious range on theELAS (1 SD above the mean), (d)Group sessions were no more than onehour, and (e) Ground rules in eachsession included respecting each other’sopinions, not making put-downs, andletting everyone have a chance to talk.

The focus groups in this studywere taped using both audio and videoequipment so individuals could beeasily recognized in the transcription

phase. The presence of the recordingequipment did not seem to interferewith students’ responses and studentsquickly seemed to forget about therecorder and camera.

The focus groups were used afterthe administration of the ELAS in orderto interpret the numerical data gath-ered. All focus groups were conductedduring class time at the research site.Only the participants and I were al-lowed in the room where the focusgroups were taking place. As partici-pants came in the room, they wereasked to put themselves at ease andwere offered refreshments, which theycould eat and drink while the focusgroup was going on. I reintroducedmyself and explained the purpose of thefocus group, with special attention paidto the fact that their input was the mostimportant aspect of the group discus-sion. Also, participants had the oppor-tunity to ask questions. Many questionscentered on my ethnic origin, such as,“How come you speak Spanish ifyou’re Greek?” After participants hadthe opportunity to ask questions andsnack they were given an informationsheet with the main focus group ques-tions (Appendix B). They were thenasked to look at the questions and writedown any thoughts they had on thetopics listed. Preliminary answers werewritten in order to reduce the influenceother participants would have on eachparticipant’s responses during the actualdiscussion.

After all participants had gatheredand had time to get something to eat ordrink and write answers to the ques-tions, I called the meeting together and


began the warm-up phase. Participantswere asked a warm-up question, suchas, “Do you like speaking English?” or“Do you speak English often?” Usually,this question prompted some studentsto comment that often they did not likespeaking English. This comment oftenled the group into the actual data-collection questions.

Because of the dynamics of groupinteraction, groups did not always fol-low this list of questions in their exactorder. Additionally, I sometimes had toredirect the discussion. On occasion thediscussion became more of a groupinterview with my asking a questionand students answering in turn, as theywould in a class. When this patternoccurred, I withdrew from the interac-tion, usually by sitting silently, until thestudents began to talk among them-selves. The last question asked in focusgroups probed student advice abouthow to reduce anxiety (“Is there anyadvice you would give to teachers sothat students aren’t nervous?”) and wasused as a wrap-up question, allowingevery student to participate and add anyadditional information.

Focus group data can be complexin that not only must one search forideas and concepts repeated in similarwords, but also the researcher mustinterpret individual responses that maysound different but apply to the sameconcept (Krueger, 1998). Hence, theprimary goal of focus group analysis isnot necessarily to comment on indi-vidual responses but to search for trendsand patterns that emerge across groups.Also, discussions are evolutionary innature, as participants influence each

other and possibly change each other’sminds. In order to offset this possibility,I often repeated or summarized ques-tions in order to confirm responses.Additionally, because of the fluidity offocus groups, spontaneous questionsoccurred in each group.

Demographic DataDemographic data (see Appendix C forinstrument) collected on each partici-pant include information regarding thetotal number of years they had been inthe U.S., the language(s) spoken in thehome, and self-report data that wereused in the statistical analyses. In theself-reports participants were asked torate their proficiency in subcategoriesof English skills. These subcategoriesincluded listening, speaking, reading,and writing skills. Participants ratedthemselves as “not good,” “okay,” “good,”and “very good.”

Data AnalysisEnglish Language Anxiety ScaleSeveral statistical tests were used toanalyze the ELAS:

Paired T-tests were used to analyzebroad levels of anxiety between ESLand mainstream classes.

ANCOVAs were run on both ESL andmainstream classes to identify howlevels of anxiety correlated with thespecific factors of years in the U.S.,levels of academic achievement,listening and speaking skills, readingand writing skills, and gender.

A factor analysis was run on both ESLand mainstream classes to analyzehow levels of anxiety correlated


with various factors. Because thefactor analysis was exploratory,factors were not identified a prioribut rather through the analysis basedon items with an eigen value ofmore than one.

The decision to perform two sepa-rate ANCOVAs followed from thedetermination of a significant differ-ence between the English languageanxiety in the ESL class and the Englishlanguage anxiety felt in the mainstreamclassroom. To explore these differences,I treated them as separate instances ofpotential anxiety. Covariates in theANCOVAs included the continuousvariables of achievement and numberof years in the U.S. I identified achieve-ment as a factor because past research inthe field of foreign language acquisitionhas shown a significant relationshipbetween achievement and foreign lan-guage anxiety (Aida, 1994; Gardner,1985; Gardner & MacIntyre, 1993;LaLonde & Gardner, 1985; MacIntyre& Gardner, 1994). I chose time in theU.S. as a factor since many participantsin the field test phase expressed a beliefthat their anxiety would decrease asthey spent more time in the U.S.

Fixed factors included participants’self-reports of their listening/speakingskills (8 levels), their reading/writingskills (7 levels), and their gender. In theANCOVA analyses the subcategoriesof self-reports were paired so that thesocial skills (listening and speaking)were distinguished from the more aca-demic skills (reading and writing). Thisseparation was determined to be usefulbased on prior student interviews dur-

ing the field test phase in which partici-pants reported feeling more comfort-able with either academic or interper-sonal skills.

Because this investigation is ex-ploratory in nature, promax rotationwas used with the factor analysis. Onlyfactors with an eigen value of morethan one were retained for analysis.Preliminary analyses showed that themajor factors might be related; how-ever, a higher order factor analysis didnot show any good fit among theinterrelated factors.

Focus GroupsThe analysis of focus group data re-quired several steps, including partici-pant verification and the coding of data.Data were coded and grouped with thehelp of an analytical program(NUD*IST). In this procedure, as top-ics were repeated and reinforced ingroup transcriptions, they were orga-nized together to make trends moreevident. Responses were then color-coded under several broad categoriesincluding responses pertaining to aca-demic and social aspects of anxiety.Responses that fell into each categorywere then cross-referenced with spe-cific focus group questions relating tothe interactions with teachers, Anglostudents, and Chicano students in ESLand mainstream classrooms. Hence, theexcerpts I report are representative ofthe cross-referenced categories. In manycases a student would make a remarkwith others either voicing assent ornodding in agreement. Excerpts re-ported from these trends are directquotes but also reflect the overall spirit


of the focus groups. Hence, while theactual words of different participantsvaried slightly, quotes that directly andclearly described a situation or feelingwere used to represent the entire par-ticipant pool. Excerpts fall into twobroad coding categories: interpersonaland academic English language anxiety.These categories were then dividedinto subcategories including Englishlanguage anxiety related to teachers,Chicano students, Anglo students, andfellow ESL students.

As a confirmation of accuracy,responses were clarified through par-ticipant verification procedures. Thisverification involves the confirmationthat the moderator has understood theintent of the participants. Here, confir-mation was accomplished, as thor-oughly as possible, through the use ofwritten responses immediately prior tothe focus group session and clarifica-tion checks made throughout the focusgroup time. During clarification checksthe moderator would restate or re-phrase an idea to allow participants theopportunity to clarify or correct. .

ResultsThe presentation of results follows theorganization of the research questions,

drawing from both statistical tests andqualitative data from the focus groups forevidence for the study’s major claims.

Is There a Difference between Levelsof English Language Anxiety in ESLand Mainstream Classes?The first research question asked simplywhether there were differences be-tween anxiety in the ESL and main-stream classrooms. The paired t-testshowed a significant difference be-tween ELAS score within mainstreamand ESL classes (see Table 1). The resultssuggest that when many ELL studentsgo to their mainstream classes, theiroverall English language anxiety levelincreases.

When asked about their anxiety inthe focus groups, participants offered avariety of reasons and clarifications thatalmost always dealt with anxiety associ-ated with social and interpersonal in-teraction, as in the following excerpts:

RESEARCHER: Are you more nervous inESL or regular (mainstream) classes?

PARTICIPANT A: Regular . . . because I feellike when I say something the otherstudents are going to laugh at me.

RESEARCHER: Are you more nervous inESL or regular classes?

TABLE 1Paired t-test

Paired Differences

STD. ERROR MEAN STED. DEVIATION MEAN T DF SIG. (2-TAILED)

Pair 1 ESL-MAIN -7.9888 12.4356 .9321 -8.571 177 <.001


PARTICIPANT B: In the regular classes . . .the ones who know English andSpanish get impatient. They don’twant to work with them (ELLs)because they don’t know how to doit.

RESEARCHER: Why don’t you like talk-ing with them (non-ELLs)?

PARTICIPANT D: Because they knowmore English than me, and if I saysomething wrong, they laugh.

Interactions with English-speaking stu-dents tended to be strained and avoided,as evidenced by this student’s comment:“There [in the mainstream classroom], Italk a little bit, with friends of minewho don’t speak Spanish; with them Idon’t like talking English.”

When asked about interactionswith teachers in the mainstream class-room, many students commented thatthey were not so nervous speaking withteachers who used at least some Spanishin the classroom. One participant, forinstance, said that “I feel okay withthose teachers, the ones who speakSpanish.” This comment suggests thatcomfort levels for at least some studentsincreased when teachers sanctionedstudents’ mother tongue in class.

What Accounts for This Difference?The ANCOVA analyses (see Table 2)helped to provide texture to the broadfinding that students were more anx-ious in mainstream than ESL classes. Toinvestigate the second research ques-tion, the ANCOVAs tested the ELASresponses for relationships between anxi-ety in the two settings and factorspreviously identified by researchers as

contributing to anxiety: years in school,levels of achievement, listening andspeaking skills, reading and writingskills, and gender. Of these factors,three—achievement, reading and writ-ing skills, and gender—were found tobe significantly related to anxiety.

AchievementIn ESL classes a significant relationship(.002 at p<.05) between achievementand English language anxiety appeared.Further investigation in the direction ofthis relationship indicated that as ESLachievement increased, English languageanxiety decreased (beta weight=-.383).This finding is consistent with otherswho have found similar results in for-eign language classes (Aida, 1994;MacIntyre & Gardner, 1993, 1994). Yet,the same relationship was not apparentwhen ELL students were in mainstreamclasses (achievement=.183). Therefore,in mainstream classes students whowere high achievers in their ESL classesseemed just as likely to suffer from highlevels of English language anxiety aslow ESL achievers. This result suggeststhat it would be difficult for teachersand administrators to rely on ESLachievement as an indicator of Englishlanguage anxiety in mainstream classes.

Reading and Writing SkillsANCOVA analysis revealed a slightmain effect in the mainstream class-room with reading/writing skills (.042).Pearson’s product moment correlationindicated a negative relationship (-.251)in which ELL students in mainstreamclasses who believed they had good orvery good reading and writing skills


TABLE 2ANCOVA for ESL and Mainstream Classes

Tests of Between-Subjects Effects

DEPENDENT TYPE IV SUM MEANSOURCE VARIABLE OF SQUARES DF SQUARE F SIG.

Corrected Model ESL classes 7095.443a 53 322.556 2.130 <.001

Mainstream classes 13207.819b 53 249.204 1.617 .017

Intercept ESL classes 12434.467 1 12434.467 82.121 <.001

Mainstream classes 10598.737 1 10598.737 68.755 <.001

YEARS ESL classes 473.003 1 473.003 3.124 .080

Mainstream classes 3.294 1 3.294 .021 .884

ACHIEVEMENT ESL classes 1530.297 1 1530.297 10.107 .002

Mainstream classes 276.915 1 276.915 1.796 .183

LISTENING & ESL classes 1163.595c 7 166.228 1.098 .369

SPEAKING Mainstream classes 1221.986c 7 174.569 1.132 .348

READING & ESL classes 1902.621c 6 317.104 2.094 .059

WRITING Mainstream classes 2090.096c 6 348.349 2.260 .042

GENDER ESL classes 344.661c 1 344.661 2.276 .134

Mainstream classes 848.434c 1 848.434 5.504 .021

LS * RW ESL classes 3868.606c 19 203.611 1.345 .170


LS * GENDER ESL classes 772.431c 6 128.738 .850 .534


RW * GENDER ESL classes 462.877c 5 92.575 .611 .691


LS * RW * GENDER ESL classes 736.323 6 122.721 .810 .564

Mainstream classes 240.965 6 40.161 .261 .954

Error ESL classes 17412.841 115 151.416

Mainstream classes 17727.589 115 154.153

Total ESL classes 747104.000 169

Mainstream classes 932997.000 169

Corrected Total ESL classes 34508.284 168

Mainstream classes 30935.408 168

a. R Squared = .495 (Adjusted R Squared = .263b. R Squared = .427 (Adjusted R Squared = .163c. The Type IV testable hypothesis is not unique.


were less anxious than ELL studentswho believed their academic skills werepoor. This finding is consistent withresults of others (Bailey, Onwuegbuzie,& Daley, 1999) who found that studentswith high self-perceived skills were lessanxious than students who estimatedtheir own skills as low.

GenderANCOVA results also showed a signifi-cant main effect for gender (.021) inmainstream classes. Many of the ex-cerpts used to illustrate the findings ofthis study involved girls. When highlyanxious students were separated out toform the groups, there were many girlsidentified as highly anxious when usingthe English language. This main effectis consistent with other studies showingthat girls tend to be more anxious thanboys (Bernstein, Garfinkel, & Hober-man, 1989; Gierl & Rogers, 1996;Padilla et al., 1988; Plancherel & Bolog-nini, 1995). However, this finding istempered by reports that males are lesslikely to admit anxiety than females(Williams, 1996).

Yet heightened anxiety among girlswas not evident in ESL classes, a resultthat the statistical tests cannot explainconclusively. Two sources of evidence—the focus group responses and thelikelihood that anxiety in the main-stream classroom is related to peerinteraction and performance—suggestthat the higher levels of English lan-guage anxiety were related to changingsocial relationships in adolescence. Girlsoften have close interpersonal relation-ships with teachers (Bracken & Crain,1994) that could mitigate high levels of

English language anxiety in the ESLclassroom where tension is more re-lated to academic achievement. How-ever, in mainstream classrooms whereEnglish language anxiety seemed to bemore closely related to performancetypes of anxiety, girls did not seem tohave adequate coping strategies to helpthem save face in front of their nativeEnglish-speaking or Chicana peers.

Additionally, new relationships inmainstream classrooms must cross morecultural and linguistic boundaries, cre-ating additional challenges for girls tonavigate in order to form safe peergroup interactions. This conclusion isconsistent with Schumann’s views ofsocial distance that mark the impor-tance of the relationships formed be-tween groups (1997). Social factorsseem to mark many of the most anxi-ety-producing situations within themainstream environment.

Is There a Relationship betweenLevels of English Language Anxietyin ESL and Mainstream Classroomsand Factors That Emerge during aFactor Analysis of the ELAS?The factor analysis (see Table 3) wasconducted on the ELAS to answer thethird research question, which did notinvestigate a priori factors, as did theANCOVA analyses, but rather wasexploratory, seeking to identify addi-tional or corroborating factors impli-cated in students’ anxiety. The factoranalysis revealed the presence of anacademic anxiety in ESL classroomsthat was different from the interactionalanxiety evident in mainstream classes.Performance anxiety included up to


24% of the total variance and academicanxiety only 5% of the total variance inthe mainstream classroom. This differ-ence may be attributable to the lowlevel of meaningful interaction be-tween teachers and ELLs in the main-stream classes. English interactions inthe mainstream classroom primarilyseemed to involve student-to-studentinteraction.

Several students in focus groupsconfirmed that ESL classrooms weresites of teacher-interaction anxiety:

RESEARCHER: Are you more comfort-able talking with teachers or stu-dents?

PARTICIPANT C: Students . . . becauseteachers are correcting me all thetime.

PARTICIPANT D: I’m nervous becausewhen you answer to the teachers,you had to do it very quickly. I feelnervous when I talk to the teachersand they correct me.

RESEARCHER: Are you more comfort-able talking with teachers or studentsin the ESL class?

PARTICIPANT D: With the students be-cause if I say something wrong, theydon’t know it either.

However, academic anxiety relatedto student-teacher interactions also sur-faced in the mainstream classroom.Several students commented that theydid not like interacting with teachers aswell as students and expressed frustra-tion. One student, for instance, said thatthe teachers “know English well, I feel

TABLE 3Factor Analysis for ESL and Mainstream Classes

Total Variance Explained: ELAS Score within Mainstream Classes

INITIAL EIGENVALUES

FACTOR TOTAL % OF VARIANCE CUMULATIVE %

Performance Anxiety 5.483 23.841 23.841Teacher-interaction Anxiety 1.392 6.052 29.893

Extraction Method: Principal Axis Factoring. a. When factors are correlated, sums of squared loadings cannot be added to obtain a total variance.

INITIAL EIGENVALUES

FACTOR TOTAL % OF VARIANCE CUMULATIVE %

ESL Achievement Anxiety 6.248 26.034 26.034Identity Anxiety 1.362 5.673 31.707Teacher-interaction Anxiety 1.215 5.062 36.769

Extraction Method: Principal Axis Factoring. a. When factors are correlated, sums of squared loadings cannot be added to obtain a total variance.

Total Variance Explained: ELAS Score within ESL Classes


like they’re not going to tell—well, Idon’t know what they’re saying to me!”Hence, while there did seem to be someacademic anxiety associated with themainstream classes, it appeared to beovershadowed by the interactional anxi-ety.

The factor analysis results suggestthat there were different types of En-glish language anxiety present in thedifferent environments. In the ESLclasses most of English language anxi-ety seemed to stem more from aca-demic sources, accounting for approxi-mately 26% of the total variance.

In ESL classes the ANCOVAshowed a significant negative relation-ship between ESL achievement andEnglish language anxiety. This result,along with the factor analysis thatshows a more achievement-related typeof anxiety in ESL classes, suggests thatin ESL classes students tended to bemore concerned about their ownachievement in English and how wellthey were learning the language thanthey were with social issues. However,the factor analysis shows that in main-stream classes students experienced so-cially-oriented English language anxiety.

What Corroborating and AdditionalFactors Emerge through Focus GroupDiscussions?The fourth research question relied ona systematic analysis of focus group datato identify additional or corroboratingfactors related to students’ anxiety lev-els. Focus group participants reported asubstantial amount of English languageanxiety associated with the Chicanostudents. This type of anxiety did not

emerge in the factor analysis, whichidentified teacher-student anxiety butnot specifically student-student anxiety.The ELAS is an instrument modeledafter the FLCAS, an instrument prima-rily designed to measure foreign lan-guage classroom anxiety. Hence, theimplied emphasis is on classroom inter-action, typically thought of as teacher-student interaction.

Student-student anxiety, however,became a significant topic in the focusgroups. In focus groups student-studentinteraction was a direct question andparticipants had more opportunity todiscuss the sources of their anxiety.

To some these two groups—Mexi-can-born ESL students and Chicanos—may seem to be similar in language andculture, but there are significant differ-ences that often put the two groups atodds with each other. Chicanos repre-sent a group of students who were bornin the U.S. and are of Mexican descent.These students may or may not speakSpanish. In this study Chicano studentslooked down upon and teased studentscoming from Mexico. Teachers andadministrators in the schools where thestudy was conducted commented onthe strained relations between the twogroups. This finding seemed to be oneof the most important issues related toEnglish language anxiety among thestudent participants and warrants fur-ther research.

This tension is evidenced in com-ments made by focus group partici-pants. One student, for instance, saidthat she was more comfortable withMexican students “because, well, theymake mistakes, too, like us. They make


mistakes, too, but the Chicanos, yes,them, they criticize a lot, our English.”In groups with male participants, theboys involved reported fighting quite abit with Chicano students because theywere often referred to as “mojados”(wetbacks). These sorts of conflicts andtensions appear to have contributed tothe kinds of anxiety experienced byMexican students in mainstream class-rooms.

What Coping Strategies Do HighlyAnxious Participants Use to ReduceLevels of Anxiety?The fifth research question relied onfocus group data to identify copingstrategies identified by the students toalleviate anxiety. According to Prins(1986) and Bailey et al. (1999), one ofthe most common coping strategies foradolescents is avoidance, and focusgroup participants reported using thisstrategy most often. When asked whatthey did to avoid being anxious in theirmainstream classroom, most respondedthat they do not speak in class. Onestudent commented, “I just sit there,silent.” Additionally, when I asked whattheir teachers could do to reduce theiranxiety, they said they wanted theirteachers to leave them alone and notrequire them to speak in front of theclass or when they are really nervous.Several students came up with sugges-tions, such as “When they ask ussomething, they should give us thechance to answer in Spanish or for thepeople that can’t speak English, theyshould help them, so the others don’ttease them.”

While avoidance was the most

common coping strategy, some stu-dents, especially the female students,had elaborate methods to participateindirectly. In some cases girls wouldform friendships with Mexican girlswho spoke English well and use them asintermediaries in class. For example, ifMaria (who was highly anxious) knewthe answer but did not want to speakout in class, she would tell the answer toJuana (who spoke English well), andJuana would raise her hand and re-spond. This strategy seemed to workwell for the girls as long as the teachertolerated student-to-student interac-tions during class.

Another coping strategy used bystudents with high levels of Englishlanguage anxiety was to pretend thatthere was no one else around andignore the class and teacher whenspeaking. One participant commented,“I imagine that there’s no one there.”Other students found relief in classeswhere teachers would allow them torespond in Spanish to questions. An-other participant in the same focusgroup was asked what she would do ifshe didn’t know a word in English. Shecommented, “I wouldn’t say it, or Iwould say it in Spanish.” Some weremade more comfortable by being al-lowed to express themselves more inwritten and nonverbal formats thanoral assignments.

Interestingly, students’ coping strat-egies were often in line with recom-mendations from teachers and otherresearchers (Young, 1992, 1993). Manyof the most common anxiety reductiontechniques were mentioned, including:(a) allowing for a reasonable silent


period (a stage of language acquisitioncharacterized by considerable languageprocessing but little language produc-tion), (b) giving ELL students extra waittime, (c) allowing the use of the stu-dents’ native language with classmates,and (d) not demanding that studentstalk in front of the class.

One strategy not mentioned by theteachers but put forth by the studentswas the avoidance of Chicano students.As noted, students often mentioned thetense relations they had with theChicanos. When asked with whom hewould rather speak English, one stu-dent responded, “They [the Chicanos]feel like they know more than us.”Avoiding Chicano students, then, ap-peared to be a coping strategy forreducing anxiety in mainstream classes.

DiscussionResults from this study suggest thatEnglish language anxiety is multidi-mensional, affecting ELL students dif-ferently depending on the context ofthe situation. Because of the apparentdynamic nature of English languageanxiety, it is possible that it will affectELLs at many different levels, engen-dering a need for a variety of copingstrategies.

As previously mentioned, ELL stu-dents often have difficulty in the transi-tion from ESL to mainstream classrooms.While some of these difficulties can betraced to the fact that these students areoften mainstreamed before they havereached the academic proficiencyneeded to be successful in the main-stream, there are also affective consider-ations that need to be dealt with. .

In the case of the ESL classroom,what was not anxiety producing mightbecome so in the mainstream class-room. In other words, where ELLstudents may have not been anxious inthe ESL classroom because they hadabove average achievement, they mayvery well suffer from high levels ofEnglish language anxiety in the main-stream classroom because of dimin-ished feelings of self-efficacy whenconfronted with higher demands inlistening and speaking skills and socialrelationships.

Furthermore, female ELLs whohave stable interpersonal relationshipsin their ESL classes may struggle toestablish safe relationships in the main-stream classroom and withdraw fromclassroom interactions. This type ofwithdrawal can be especially harmfulsince many coping strategies seem todepend on some type of student-to-student interaction. If female studentsare not able to form stable relationshipsin their mainstream classes, feelings ofself-efficacy could be reduced evenmore, creating an English languageanxiety cycle in which anxiety in-creases as coping opportunities de-crease.

In sum, changes in anxiety levelscan stem from many sources includingfactors that affect fears of negativeevaluation, test anxiety, communicationapprehension, and identity creation ten-sion, but all threaten an ELL student’ssense of self-efficacy, increasing anxietybeyond normal limits. Additionally, asPekrun’s (1992) EVTA would suggest,when ELL students enter into newenvironments, they may have difficulty


judging situations appropriately asthreatening or non-threatening sincethe mainstream classroom appears simi-lar to the ESL classroom yet can be verydifferent due to the presence of nativespeakers of English. Hence, ELL stu-dents might find it difficult to maintainhigh levels of self-efficacy and makepositive situational appraisals because ofthe mitigating factors associated withsocial distance, fear of negative evalua-tion, test anxiety, communication ap-prehension, and threats to their sensesof self. These factors tend to be verysituational, based on the environmentof the encounter.

Faced with this situation, somecontinue to suffer from or begin toexperience high levels of anxiety thatcan result in withdrawal from thelearning experience. Some of thesehighly anxious students begin to seeany English language situation as threat-ening and cannot see themselves asbeing able to overcome the perceivedthreats from their fears of negativeevaluation, test anxiety, communicationapprehension, and identity issues. Yetothers find ways to deal with thesefeelings of inadequacy and helplessness.

Educational ImplicationsSince the situational aspects of Englishlanguage anxiety are complementary tothe theories of Pekrun (1992) andBandura (1991), perhaps the most im-portant educational implication formainstream teachers lies in thesocioconstructivist view of learningthat takes into consideration the inter-personal construction of knowledge.Language and social interactions are the

major mediation tools in constructingknowledge (Vygotsky, 1978; Wertsch,1985). More specifically put, learningcan be optimized through the use ofsuch authentic interactions as peercollaboration (Tudge, 1990). When ELLstudents withdraw from interactionswith peers and/or teachers because ofanxiety, their English development canbe negatively altered.

Upon entering the mainstreamclassroom, ELL students often makenegative situation-outcome appraisals(the expected outcome of a potentiallythreatening situation) when confrontedwith situations in which their usualESL classroom coping strategies are nolonger valid or adequate. When theycannot see any successful course ofaction (action-outcome expectancies),high levels of anxiety involving the useof English result. Helping ELL studentsto view potentially threatening situa-tions in a different light can circumventthese negative appraisals. Also, by in-creasing feelings of self-efficacy, ELLstudents can feel more in control andbetter prepared to deal with negativeoutcomes.

Implications for Future ResearchThe differences that have been dis-cussed thus far may be attributable tothe distinction alluded to by Deffen-bacher (1980). The academic Englishlanguage anxiety felt by students in ESLclasses may be more associated withworry aspects of anxiety, and the En-glish language performance anxiety feltin mainstream classes may be moresimilar to the affective anxiety associ-ated with the more emotional aspects


of anxiety. Additionally, the differencesmay be characteristic of a variety ofsecond language situational anxietiesthat are currently being explored byother researchers (Bailey et al., 1999;Cheng, 1999; Saito, Horwitz, & Garza,1999).

Currently, in the field of foreignlanguage research, such researchers areinvestigating anxieties related to spe-cific skills such as reading and writing.In light of the findings in this study,such research should be done withELLs. Additionally, this study is a repre-sentation of a point in time. Futurestudies of English language anxietyshould be more longitudinal in naturein order to further explore changes inanxiety levels as students move fromsheltered classes to more mainstreamenvironments.

Finally, findings in this study implya deeper relationship among Englishlanguage anxiety, identity development,and interethnic interactions. Future stud-ies should focus on the interface amongthese factors. As more and more ELLsare mainstreamed at earlier points intheir English language acquisition, re-searchers should also focus on the socialfactors involved in this process of aca-demic, cultural, and linguistic accul-turation. It will not be sufficient tosingle out one factor, such as languageproficiency, to determine a student’sprobability of success in the mainstreamclassroom.

ConclusionsI have argued that the affective environ-ment of a classroom can affect learningto such a degree that teachers must take

an active role in reducing the Englishlanguage anxiety that ELL studentsoften experience in mainstream class-rooms. The research questions firstposed have been answered, but theanswers are not simple.

Although my data establish thatthere were differences between levels ofEnglish language anxiety in ESL andmainstream classes, this finding seemsto muddy the anxiety waters instead ofclearing them. There appear to bedifferent types of English languageanxiety at work, and the type of anxietyseems to be dependent upon the con-text of the interaction and how fears ofnegative evaluation, test anxiety, com-munication apprehensions, and identityissues affect the self-efficacy and ap-praisals of the ELL students involved.

My data also point to other factorsthat can affect the interactions withinthe different environments. In main-stream classes girls tended to be moreanxious than boys were. Students weremore stressed about the social aspects ofinteractions with peers in the main-stream classroom and more anxiousabout their academic performance inESL classes.

Finally, my focus group data pointedto a number of strategies students usedto reduce anxiety. Although avoidancewas the most common strategy, stu-dents used other, more elaborate strate-gies as well, including enlisting friendsto act as intermediaries, responding inSpanish, and using writing to expressthemselves. An unexpected coping strat-egy was avoiding Chicano studentswho were perceived as being highlycritical of my participants’ oral English.


Many teacher-training programscontinue to treat ESL issues as marginaland graduates from these programs feelless than competent when faced withan ELL student in their classes (Henley& Young, 1989). Furthermore, manyteachers do not understand the situ-ational anxiety that is the result ofreduced feelings of self-efficacy andnegative appraisals since they have rarelybeen in a similar situation (Diffey,1990). While it is never easy to putoneself in another person’s position,teachers of ELL students must strive tounderstand not only the linguistic chal-lenges that these students face but alsothe affective factors that can affectlearning. ELL students are not justlearning another language, but anotherlife.

This research provides teachers andadministrators with a starting point. Ifeducators can begin discussions with

students who are potentially sufferingfrom high levels of anxiety, they canwork to reduce those tensions beforethey become habitualized and result inacademic and emotional strain for ELLstudents. In ESL classes, where Englishlanguage anxiety seems to be moreassociated with academic tasks, thisanxiety can be treated by the teacher byusing less-anxiety causing activities andby acknowledging English languageanxiety. However, in the mainstreamclassroom, where English language anxi-ety is more closely related to interper-sonal anxieties, the teacher’s role isoften de-emphasized in favor of theimportance of peer relationships. Inthese situations the teacher can take onthe role of mediator between the ELLstudents and their native English-speak-ing peers, helping each to understandthe other. .

Author Note

Please contact the author at The Florida State University, Department of Curriculum and In-struction, Multilingual/Multicultural Education Program, 209 Carothers Hall, Tallahassee, FL32306; email: [email protected]

References

AIDA, Y. (1994). Examination of Horwitz, Horwitz and Cope’s construction of foreign languageanxiety: The case of students of Japanese. The Modern Language Journal, 78, 155-168.

APPLEBAUM, R. L., & TROTTER, R. T. II (1986, May). Communication apprehension and Hispanics: Anexploration of communication apprehension among Mexican Americans. Paper presented at theannual convention of the Speech Communication Association, Chicago, IL.

BAILEY, P., ONWUEGBUZIE, A. J., & DALEY, C. E. (1999). Foreign language anxiety and learningstyle. Foreign Language Annals, 32(1), 63-76.

BANDURA, A. (1991). Self-efficacy conception of anxiety. In R. Schwarzer & R. A. Wicklund(Eds.), Anxiety and self-focused attention (pp. 89-110). London: Harwood Academic Publishers.


BANDURA, A. (1995). Exercise of personal and collective efficacy in changing societies. In A. Bandura(Ed.), Self-efficacy in changing societies (pp.1-45). New York: Cambridge University Press.

BEEBE, L. M. (1983). Risk-taking and the language learner. In H. W. Selinger & M. H. Long(Eds.), Classroom-oriented research in second language acquisition (pp. 39-66). Rowley, MA:Newbury House.

BERNSTEIN, G. A., GARFINKEL, B. D., & HOBERMAN, H. M. (1989). Self-reported anxiety inadolescents. American Journal of Psychiatry, 146, 384-386.

BRACKEN, B. A., & CRAIN, R. M. (1994). Children’s and adolescent’s interpersonal relations: Doage, race and gender define normalcy? Journal of Psychoeducational Assessment, 12(1), 14-32.

CHENG, Y. S. (1999). Language anxiety: Differentiating writing and speaking components.Language Learning, 49, 417-446.

COLLIER, V. (1987). Age and rate of acquisition of second language for academic purposes.TESOL Quarterly, 21, 617-641.

CUMMINS, J. (1984). Bilingualism and special education: Issues in assessment and pedagogy. San Diego:College-Hill.

CUMMINS, J. (1996). Negotiating identities: Education for empowerment in a diverse society. Ontario,CA: California Association of Bilingual Education.

CUMMINS, J. (2000). Language, power and pedagogy: Bilingual children in the crossfire. Toronto:Multilingual Matters.

DALEY, C. E., ONWUEGBUZIE, A. J., & BAILEY, P. (1997, November). Predicting achievement in collegelevel foreign language courses. Paper presented at the meeting of Midsouth EducationalResearch Association, Memphis, TN.

DEFFENBACHER, J. L. (1980). Worry and emotionality in test anxiety. In I. G. Sarason (Ed.), Testanxiety: Theory, research and applications (pp. 131-159). Hillsdale, NJ: Erlbaum.

DIFFEY, N. (1990). The other side of the desk: Experiencing learning a new language. TESLCanada Journal, 8, 27-36.

EHRMAN, M., & OXFORD, R. (1995). Adult learning styles and strategies in an intensive trainingsetting. Modern Language Journal, 74, 311-327.

EYSENCK, M. W. (1979). Anxiety, learning and memory: A reconceptualization. Journal of Researchin Personality, 13, 363-385.

FLORES, J. G., & ALONSO, C. G. (1995). Using focus groups in educational research: Exploringteachers’ perspectives on educational change. Evaluation Review, 19(1), 84-101.

GARDNER, R. C. (1985). Social psychology and second language learning: The role of attitudes andmotivation. London: Edward Arnold.

GARDNER, R. C., & MACINTYRE, P. D. (1993). On the measurement of affective variables insecond language learning. Language Learning, 43, 157-194.

GIBBONS, F. X. (1991). Self-evaluation and self-perception: The role of attention in the experi-ence of anxiety. In R. Schwarzer & R. A. Wicklund (Eds.), Anxiety and self-focused attention(pp. 15-26). London: Harwood Academic Publishers.


GIERL, M. J., & ROGERS, W. T. (1996). A confirmatory factor analysis of the test anxiety inventoryusing Canadian high school students. Educational and Psychological Measurement, 56, 315-324.

GUIORA, A. Z. (1983). Introduction: An epistemology for the language sciences. LanguageLearning, 33, 6-11.

HASS, R. G., & EISENSTADT, D. (1991). The effects of self-focused attention on perspective-takingand anxiety. In R. Schwarzer & R. A. Wicklund (Eds.), Anxiety and self-focused attention (pp.55-66). London: Harwood Academic Publishers.

HENLEY, R., & YOUNG, J. (1989). Multicultural teacher education Part 4: Revitalizing faculties ofeducation. Multiculturalism, 12(3), 40-41.

HOPPE, M. J., WELLS, E. A., MORRISON, D. M., GILLMORE, M. R., & WILSDON, A. (1995). Usingfocus groups to discuss sensitive topics with children. Evaluation Review, 19(1), 102-114.

HORWITZ, E. K. (1991). Preliminary evidence for the reliability and validity of a foreignlanguage anxiety scale. In E. K. Horwitz & D. J. Young (Eds.), Language anxiety: From theoryand research to classroom implications (pp. 37-39). Englewood Cliffs, NJ: Prentice-Hall.

HORWITZ, M. B., HORWITZ, E. K., & COPE, J. A. (1986). Foreign language classroom anxiety. TheModern Language Journal, 70(2), 125-132.

KAUFMAN, P., KWON, J. Y., & KLEIN, S. (1999). Dropout rates in the United States-1999. Washington,DC: U.S. Department of Research and Improvement, National Center for EducationStatistics.

KRUEGER, R. A. (1998). Analyzing and reporting focus group results. Thousand Oaks, CA: Sage.

LALONDE, R. N., & GARDNER, R. C. (1985). On the predictive validity of the Attitude/Motivation Test Battery. Journal of Multilingual and Multicultural Development, 5, 403-412.

MACINTYRE, P. D., & GARDNER, R. C. (1991). Methods and results in the study of anxiety andlanguage learning: A review of the literature. Language Learning, 41(1), 85-117.

MACINTYRE, P. D., & GARDNER, R. C. (1993). The effects of induced anxiety on three stages ofcognitive processing in computerized vocabulary learning. Studies in Second LanguageAcquisition, 16, 1-17.

MACINTYRE, P. D., & GARDNER, R. C. (1994). The subtle effects of language anxiety oncognitive processing in the second language. Language Learning, 44, 283-305.

MCCROSKEY, J. C. (1987). Willingness to communicate. In J. C. McCroskey & J. A. Daly (Eds.),Personality and interpersonal communication (pp. 129-156). Thousand Oaks, CA: Sage.

MORGAN, D. L. (1993). Successful focus groups: Advancing the state of the art. Thousand Oaks, CA: Sage.

MORGAN, D. L. (1997). Focus groups as qualitative research, Vol. 16. (2nd ed.). Thousand Oaks, CA:Sage.

OLSEN, L. (1997). Made in America: Immigrant students in our public schools. New York: The New Press.

PADILLA, A. M., CERVANTES, R. C., MALDONADO, M., & GARCIA, R. E. (1988).Coping responsesto psychosocial stressors among Mexican and Central American immigrants. Journal ofCommunity Psychology, 16, 418-427.


PAPPAMIHIEL, N. E. (1999). The development of an English language anxiety assessment instrument forMexican middle school English language learners. Unpublished doctoral dissertation, University ofTexas at Austin.

PEKRUN, R. (1992). Expectancy-value theory of anxiety: Overview and implications. In D.Forgays & T. Sosnowski (Eds.), Anxiety: Recent developments in cognitive, psychophysiological, andhealth research (pp. 23-39). Washington, DC: Hemisphere.

PLANCHEREL, B., & BOLOGNINI, N. (1995). Coping and mental health in early adolescence.Journal of Adololescence, 18, 459-474.

PRINS, P. J. (1986). Children’s self speech and self-regulation during a fear-provoking behavioraltest. Behaviour, Research and Therapy, 24, 181-191.

SAITO, Y., HORWITZ, E. K., & GARZA, T. J. (1999). Foreign language reading anxiety. ModernLanguage Journal, 83, 202-218.

SARASON, I. G. (1972). Experimental approaches to test anxiety: Attention and the uses ofinformation. In C. D. Spielberger (Ed.), Anxiety: Current trends in theory and research, Vol. II(pp. 382-405). New York: Academic Press.

SCHALLERT, D. L. (1991). The contribution of psychology to teaching the language arts. In J.Flood, J. M. Jensen, D. Lapp, & J. R. Squire (Eds.), Handbook of research on teaching the Englishlanguage arts (pp. 30-39). New York: Macmillan.

SCHUMANN, J. H. (1997). The neurobiology of affect in language. Malden, MA: Blackwell.

SCHWARZER, R. (1986). Self-related cognition in anxiety and motivation: An introduction. In R.Schwarzer (Ed.), Self-related cognition in anxiety and motivation (pp. 1-17). Hillsdale, NJ: Erlbaum.

SPIELBERGER, C. D. (1983). Manual for the State-Trait Anxiety Inventory (Form Y). Palo Alto, CA:Consulting Psychologists Press.

TOBIAS, S. (1986). Anxiety and cognitive processing of instruction. In R. Schwarzer (Ed.), Self-related cognition in anxiety and motivation (pp. 35-54). Hillsdale, NJ: Erlbaum.

TUDGE, J. (1990) Vygotsky, the zone of proximal development and peer collaboration: Implica-tions for classroom practice. In L. Moll (Ed.), Vygotsky and education: Instructional implicationsand applications of sociohistorical psychology (pp. 155-172). New York: Cambridge UniversityPress.

VASEY, M. W., & DALEIDEN, E. L. (1996). Information processing pathways to cognitive interfer-ence in childhood. In I. G. Sarason, G. R. Pierce, & B. R. Sarason (Eds.), Cognitive interferencetheories: Methods and findings (pp. 117-138). Hillsdale, NJ: Erlbaum.

VERPLAETSE, L. S. (1998). How content teachers interact with English language learners. TESOLJournal, 7(5), 24-28.

VYGOTSKY, L. S. (1978). Mind in society: The development of higher psychological processes. (M. Cole, V.John-Steiner, S. Scribner, & E. Souberman, Eds.). Cambridge, MA: Harvard University Press.

WERTSCH, J. V. (Ed.)(1985). Culture, communication and cognition: Vygotskian perspectives. New York:Cambridge University Press.

WILLIAMS, J. E. (1996). Gender related worry and emotionality test anxiety for high-achievingstudents. Psychology in the Schools, 33, 159-162.


YOUNG, D. J. (1992). Language anxiety from the foreign language specialist’s perspective:Interviews with Krashen, Omaggio, Hadley, Terrell and Rardin. Foreign Language Annals, 25,157-172.

YOUNG, D. J. (1993). Creating a low-anxiety classroom environment: What does the languageanxiety research suggest? Modern Language Journal, 75, 426-439.

ZAMBRANA, R. E., & SILVA-PALACIOS, V. (1989). Gender differences in stress among Mexicanimmigrant adolescents in Los Angeles, California. Journal of Adolescent Research, 4, 426-442.

APPENDIX A: ENGLISH LANGUAGE ANXIETY SCALE

1. In ESL classes, I forget how to say things I know.En clases de ESL, Me siento tan nervioso (a) que se me olvida cómo decir cosas que ya sé.

2. In regular classes, I forget how to say things I know.En clases regulares, Me siento tan nervioso (a) que se me olvida cómo decir cosas que ya sé.

3. In ESL classes, I tremble when I know I’m going to have to speak in English.En clases de ESL, tiemblo cuando sé que voy a tener que hablar en inglés.

4. In regular classes, I tremble when I know I’m going to have to speak in English.En clases regulares, tiemblo cuando sé que voy a tener que hablar en inglés.

5. In ESL classes, I start to panic when I have to speak English without preparation.En clases de ESL, empiezo a sentir pánico cuando tengo que hablar inglés sin preparación.

6. In regular classes, I start to panic when I have to speak English without preparation.En clases regulares, empiezo a sentir pánico cuando tengo que hablar inglés sin preparación.

7. In ESL classes, when I speak English, I feel like a different person.En clases de ESL, cuando hablo inglés, me siento como una persona diferente.

8. In regular classes, when I speak English, I feel like a different person.En clases regulares, cuando hablo inglés, me siento como una persona diferente.

9. In ESL classes, even when I’m prepared to speak English, I get nervous.En clases de ESL, aún cuando estoy preparado(a) para hablar en inglés, me pongo nervioso(a).

10. In regular classes, even when I’m prepared to speak English, I get nervous.En clases regulares, aún cuando estoy preparado(a) para hablar en inglés, me pongo nervioso(a).

11. In ESL classes, I’m afraid that my teachers are ready to correct every mistake I make.En clases de ESL, me da miedo pensar que mis maestros están listos para corregir cada error que cometa eninglés.

12. In regular classes, I’m afraid that my teachers are ready to correct every mistake I make.En clases regulares, me da miedo pensar que mis maestros están listos para corregir cada error que cometa eninglés.


13. In ESL classes, sometimes I can’t express my true feelings in English and this makes meuncomfortable.En clases de ESL, hay veces que no puedo expresar mis verdaderos sentimientos in inglés y esto me incomoda.

14. In regular classes, I can’t express my true feelings in English and this makes me uncomfortable.En clases regulares, hay veces que no puedo expresar mis verdaderos sentimientos in inglés y esto me incomoda.

15. In regular classes, I feel very self-conscious about speaking English in front of native speakingstudents.En clases regulares, pienso demasiado cuando hablo inglés ante los estudiantes cuyo primer idioma es el inglés.

16. In ESL classes, I get nervous and confused when I’m speaking English.En clases de ESL, me siento nervioso(a) y lleno(a) de confusión cuando estoy hablando inglés.

17. In regular classes, I get nervous and confused when I’m speaking English.En clases regulares, me siento nervioso(a) y lleno(a) de confusión cuando estoy hablando inglés.

18. In regular classes, there are so many rules in English, I feel like I can’t learn them all.En clases regularles, hay tantas reglas en inglés que siento que nunca las voy a aprender todas.

19. In ESL classes, there are so many rules in English, I feel like I can’t learn them all.En clases de ESL, hay tantas reglas en inglés que siento que nunca las voy a aprender todas.

20. In ESL classes, I’m afraid that native English speakers will laugh at me when I speak English.En clases de ESL, tengo miedo que las personas cuyo primer idioma es el inglés se burlarán de mí cuando hableinglés.

APPENDIX B: FOCUS GROUP QUESTIONS

1. In what situations do you feel uncomfortable speaking English?

2. Are you more comfortable speaking English in ESL classes or regular classes?

3. Are you more comfortable speaking English with teachers or other students?

4. Are you more comfortable speaking English with Hispanic students or Anglo students?

5. How does someone get over being nervous when he or she speaks English?

6. When does someone become an American?

7. When you speak English, do you feel Mexican or like a different person?

8.Is there any advice you would give to teachers so that students aren’t nervous?


APPENDIX C: ELAS INFORMATION SHEET

Nombre: _______________________________________

Número de identificación: _________________________

Edad: _______ Sexo: _________ Grado: ______________

¿Cuántos años tiene Ud. en Los Estados Unidos? __________________

¿Cuántos años tenía Ud. cuando vinó a Los Estados Unidos? _______________

A veces, habla Ud. inglés en su casa? ____________

¿Con quién? __________________________________________________________

¿Antes de vivir en Los Estados Unidos, estudió inglés en Mexíco? _________________

En las siguientes preguntas, hace Ud. un círculo sobre su respuesta.

¿Qué nivel escolar terminaron sus padres?Madre: escuela primaria escuela secundaria universidadPadre: escuela primaria escuela secundaria universidad

Por lo general, ¿cómo se siente cuando habla inglés?

…en las clases de ESL?Confortable poco nervioso(a) muy nervioso(a)

…en las clases regulares?Confortable poco nervioso(a) muy nervioso(a)

How well do you speak English?Very well good okay not good

How well do you read English?Very well good okay not good

How well do you understand other people speaking to you in English?Very well good okay not good

How well do you write English?Very well good okay not good

Si estaría Ud. disponible para participar en grupo de discusion, déme su número de teléfonoaquí: ________________

english as a second language students and english language

Documents