Do They Make a Difference? The Impact of English Language Programs on Second Language Students in Canadian Universities

This study refinesour understanding of the positive impact of ESL and EAP programs moder-on L2 university students’ academic and social engagement... Specifically, the study examined 1 t

Do They Make a Difference? The

Impact of English Language Programs

on Second Language Students in

pro-on questipro-onnaire resppro-onses of 641 L2 students studying in 36 Englishlanguage programs in 26 Canadian universities The researchers iden-tified programs by their activity emphasis as either English as a sec-ond language (ESL) or English for academic purposes (EAP).Activity emphasizing speaking, social interaction, and general lan-guage development was viewed as ESL, whereas activity that empha-sized academic reading, writing, and language development wasconsidered EAP The researchers used structural equation modelingprocedures to examine the network of relationships between lan-guage program emphasis and participants’ background characteristics

in influencing academic and social engagement A model of ated mediation (Wu & Zumbo, 2008) was confirmed; that is, languageprogram activities were found to account for variation in strategieswhich mediated academic and social engagement However, theimpact was moderated (lessened or strengthened) by three personalbackground factors: anxiety, stress, and motivation This study refinesour understanding of the positive impact of ESL and EAP programs

moder-on L2 university students’ academic and social engagement

doi: 10.1002/tesq.103

In response to the increasing numbers of second language (L2)learners studying in Canadian universities, many new language pro-grams have been developed These programs tend to share the com-mon goal of developing L2 learners’ English language in order tosupport their successful transition to and engagement with academicwork in university (Cheng & Fox, 2008; Fox, Cheng, Berman, Song,

& Myles, 2006), but they differ considerably in emphasis, ogy, and approach Surveys of such programs across Canada havedemonstrated little consistency (Berman, 2002) and less evidence oftheir specific contribution in supporting L2 students’ engagementwith the disciplinary work of the university Indeed, these surveys(Berman, 2002; Cheng & Fox, 2008) reveal a patchwork of programs,which may be roughly classified into two types: English as a secondlanguage (ESL; general English) or English for academic purposes(EAP) We use ESL rather than English for speakers of other lan-guages (ESOL) or English as an additional language (EAL), because

methodol-at the time of this study most Canadian universities continued tolabel their language support courses as ESL and/or EAP Previousresearch suggests that ESL programs tend to focus on overall lan-guage development and conversational English They tend to utilize

a range of teacher-made thematic or textbook units in order toboost students’ acquisition of idioms, grammar, vocabulary, meta-phor, and so forth, across a range of different communicative situa-tions (e.g., booking travel in Canada, managing a meeting) and texts(e.g., magazine/newspaper articles, radio or television news broad-casts, film) The goals of ESL programs are typically related toincreasing communicative capability (Savignon, 2002) and tend toemphasize speaking activities and social interaction (e.g., role-plays,conversation gaps, group work) while fostering a sense of belonging,community, and connection with the new (Canadian) context andculture

EAP programs, on the other hand, are explicitly concerned withthe skills and strategies in English that are “required for study pur-poses in formal education systems” (English-Teaching InformationCentre, cited in Jordan, 2004, p 1) They have typically been charac-terized by activities that simulate academic work; for example, read-ing academic texts and taking notes; writing research reports oressays; making formal seminar presentations; and acquiring strategies

to increase comprehension by scanning, skimming, and evaluatingacademic texts (Cheng, Myles, & Curtis, 2004) EAP programs aim

to develop a “general academic English register, incorporating a mal, academic style with proficiency in the language use” (Jordan,

for-2004, p 5)

The impact of such ESL and EAP language programs has remainedunderexamined, however At present, in the face of economic pres-sures, cutbacks, funding shortages, and accountability agendas, manyEnglish language programs are being confronted with the need toprovide evidence of their effectiveness (Fox, 2009; Swales, Barks, Oster-mann, & Simpson, 2001) However, generally such programs have nocoherent (or convincing) research to fall back on for support Further,over the past decade, the pedagogical value of EAP approaches hasbeen questioned by researchers (e.g., Freedman, 1999; Hansen, 2000)who argue that language is embedded in discipline-specific academicand professional contexts These researchers maintain that languagecourses which focus on general academic knowledge, skills, or strate-gies may be of little value to students, who must meet the narrowerdemands of specific disciplinary academic cultures Further, some haveargued that English language teaching has not been sufficiently critical

of covert political and economic influences that have motivated trends

in the field (Pennycook, 1996) “They point out, for example, thatwhat is considered plagiarism by EAP teachers and the dominantacademic culture they represent devalues L2 students’ academic andprofessional experiences in their native countries” (Cheng & Fox,

2008, p 312)

The interaction of students’ personal background characteristicswith program emphasis and activity has also been underexamined(Berman, 2002; Cheng & Fox, 2008; Fox, 2004; Fox et al., 2006) Thusthere remain many unanswered questions For example, do both ESLand EAP activities have the same impact on academic engagement?

Do ESL and EAP programs increase L2 students’ satisfaction with thequality of life of their new academic study cultures? What is the rela-tionship between program type, individual students’ background char-acteristics, and academic and social engagement? This study begins toaddress these questions by investigating the impact of English lan-guage support programs on L2 students studying in Canadian universi-ties Specifically, the study examined (1) the relationship betweenlanguage program activity and student background characteristics and(2) the impact of that relationship on academic engagement and satisfac-tion with life in Canada, operationalized in the study as social engage-ment The results of the study highlight differences between variablesthat mediated the impact of language program activity on studentengagement (i.e., directly linked program emphasis and outcomes toimpact) and those that moderated engagement (i.e., lessened orstrengthened program impact) Before considering the study itself, inthe following section we provide an overview of literature that hasexamined factors that affect L2 students’ academic and social engage-ment in other research contexts

FACTORS THAT IMPACT THE ACADEMIC AND SOCIAL ENGAGEMENT OF L2 STUDENTS

Many factors that may potentially impact L2 student learning havebeen examined in the research literature Some studies have focused

on language proficiency as a predictor of academic achievement, oftenindicated by grade point averages (GPAs) at the end of a course or atprogram completion, or by the perceptions of faculty and students.However, as Christopher (1993) points out, GPAs do not take intoaccount students who may have dropped out, nor do they allow fordisciplinary differences Other research suggests that language acquisi-tion variables such as time in an English-dominant country, type ofhigh school attended, and previous university experience are goodpredictors of future academic achievement—and better than sociode-mographic factors (i.e., gender, parents’ education, or ethnicity) orstandardized test scores (see, e.g., Fox, 2009; Shih & Brown, 2000).Some research (e.g., Johnson, 1988) has found that standardized profi-ciency tests (e.g., Test of English as a Foreign Language, TOEFL) cor-relate with outcomes such as academic credits earned and programcompletion; however, such studies have been questioned (Cotton &Conrow, 1998; Fox, 2004; Kerstjens & Nery, 2000), because so manyother factors potentially affect academic outcomes (e.g., motivation,educational background, field of study, family pressures, support).Macrae (1997) argued that test scores need to be systematically exam-ined in relation to such factors as age, motivation, language, and edu-cational and cultural background Further, Leki (1995) foundevidence that new cultural contexts—both academic and social—impact student performance (for a review of these studies and factorswhich impact L2 students in English-medium universities, see Cheng

& Fox, 2008; Fox et al., 2006)

Al-Sharideh and Goe (1998) identify social connections and/orfriendship as a key factor in the academic engagement process, find-ing that L2 students needed to find a balance between their friend-ships with students from their own L2 cultural groups with localstudent friendships Their research highlights the importance of asocial network in academic retention and success Their findings aresupported by Scanlon, Rowling, and Weber’s research (2007), whichreports on transition to university as a “loss experience” (p 223), argu-ing that new university students often undergo a type of identity dis-continuity They found that new students in transition to university,regardless of language or cultural background, typically drew onknowledge of past learning contexts in their attempts to negotiate newidentities as university students, but that this was generally unhelpful

Their findings suggest the importance of social interaction in making

an effective transition to university: “Identity results from situatedinteractions in which students pick up cues regarding the horizons ofpossibility for identity formation in university transition” (p 223).They argue that “it is the nexus of situated [social] interactions withlecturers and other students that is the context and process of identityformation” (p 223) These studies suggest that any examination of stu-dents’ academic engagement process must be related to considerations

of their social engagement, and whether or not effective or satisfyingsocial connections are being made A number of large-scale surveyshave thus investigated student perceptions of the relative importance

of language, academic, and social skills (Leki, 2001) For example, man and Cheng (2001) reported that the most difficult academic lan-guage skills for L2 students in a Canadian university were those thatrequired either oral communication (e.g., group work, presentations,class discussions) or writing (e.g., research essays, reports) Further, anumber of studies have examined individual L2 students’ written andspoken accounts of their transition to university-level study (e.g., Ivan-i
c, 1998, 2006; Johns, 1997; Leki, 1995) and documented how studentidentities were constructed over time These studies have providedvaluable insights regarding the role of language program support

Ber-In addition, in recent years there has been an increasing number

of longitudinal and large-scale studies of EAP program support Fox(2004, 2005) provided evidence of the role of EAP in L2 students’engagement with university in longitudinal tracking studies that com-pared two cohorts of L2 students (one that had EAP support andone that did not), with their native-English-speaking counterparts Inboth studies, L2 students with language support had a much greaterlikelihood of academic success Banerjee and Wall (2006) reported

on a survey of U.K institutions’ presessional EAP courses and therole of tests in acceptance and placement They discussed challengesfaced by EAP programs in determining what constitutes successful com-pletion and concomitant readiness for academic study, and provided achecklist of criteria to consider in assessing the academic readiness

of presessional EAP students Green (2007) examined the impact ofthe International English Language Testing System (IELTS) on theprovision of EAP courses and the relationship between proficiency,test preparation, and the development of academic literacy, findingevidence that many presessional programs have felt pressure to mea-sure gains in proficiency using IELTS as the outcome He alsoreported on a disconnect between the content of the test and theobjectives of most presessional EAP courses, which rendered suchexpectations for gains in proficiency unrealistic Moore and Morton(2005) also noted that in many Australian institutions presessional

EAP courses have a dual purpose: to prepare students both for versity study and for the IELTS.

uni-The studies cited here suggest that a range of academic, social, andprogram factors may impact an L2 student’s engagement with univer-sity study They also suggest a growing awareness of the complexity oflanguage learning and a burgeoning interest in differing approaches

to language program support

HYPOTHETICAL MODEL OF IMPACT

The present study addressed the need for evidence of the impact oflanguage support on L2 students’ academic and social engagement inCanadian universities It was informed by the results of a qualitativestudy (Cheng & Fox, 2008; Fox et al., 2006) that reported on semi-structured interviews with 56 students attending three Canadian uni-versities, who were asked about their experiences in ESL, EAP, anddisciplinary courses In accounting for their transition to the academicdemands of university study in Canada, the participants demonstrated,

as Morita (2000) points out, that transition is “complex, locally ated [and] involves dynamic negotiations of expertise and iden-tity” (p 304) However dynamic and situated this process may be,Cheng and Fox (2008) identify recurring patterns of response acrossthe theoretical samples (Charmaz, 2006) drawn from the threeresearch sites These were reduced to seven major themes (subcatego-ries) and three categories (Figure 1):

situ-1 Academic and social engagement (e.g., using/developing ing strategies, accessing social/academic resources, adapting toindividual field of study)

learn-2 Student characteristics (e.g., anxiety, stress, motivation, guage and academic background)

lan-3 Language program characteristics (e.g., ESL, EAP)

These results suggest that language program support activities (ESL

or EAP) may be linked to academic engagement and satisfaction withlife in Canada, and that the academic acculturation process is medi-ated by course outcomes (e.g., increased English language use, strategyuse, goal setting) At the same time, the results suggest that the impact

of language support is moderated (i.e., strengthened or lessened) as aresult of a student’s personal characteristics (e.g., motivation, stressdue to external pressures, individual [internal] feelings of anxiety)

We hypothesized that the relationship between language supportprogram characteristics and academic or social engagement might best

be described as one of moderated mediation Models of moderatedmediation are well documented in the research literature (cf Muller,Judd, & Yzerbyt, 2005; Wu & Zumbo, 2008) and are used in a variety

of disciplines (e.g., physics, medicine, business, psychology) to explaincomplex interactions involving multiple variables:

The moderated mediation model hypothesizes that the mediationeffect, which is responsible for producing the effect of the independentvariable [i.e., language program characteristics] on the dependent vari-ables [i.e., academic and social engagement] depends on the value orlevel of the moderator [i.e., student characteristics]

(Wu & Zumbo, 2008, p 384)Whereas the mediating variables might explain why and how therelationship occurred (in this case, language support brought aboutchanges in students’ language use, strategy use, and goal-directed prac-tices, and those changes in turn affected students’ academic and socialengagement), the moderating variables might explain when or for whom(Wu & Zumbo, 2008), based on students’ levels of anxiety, stress, andmotivation

Figure 2 provides an illustration of the hypothesized moderatedmediation model accounting for the impact of ESL or EAP programcharacteristics on academic and social engagement In the diagram, thedirect effect of English language program on academic and socialengagement is illustrated by path c The role of course outcomes inmediating academic and social engagement is in bold The mediationeffect (illustrated by paths a and b) is moderated by student characteris-tics, which either lessen or strengthen program impact Moderatingeffects on a and b are demonstrated by the arrows with shading

Themes Categories

1 Using or developing learning strategies

2 Accessing social/academic resources Academic and Social Engagement

3 Adapting to individual fields of study

4 Academic motivation

5 Language and academic background Student Characteristics

6 Personal affect (anxiety or emotional state)

7 EAP and ESL program support Language Program Characteristics

FIGURE 1 Themes (subcategories) and categories in L2 students’ accounts of their

academic acculturation processes (N = 56) (Cheng & Fox, 2008).

Note The information in Figure 1 is reprinted with permission from The Canadian Modern Language Review/La Revue Canadienne des Langues Vivants, 65, 275–305 [doi:10.3138/cmlr.65.

2.307] © 2008 University of Toronto Press.

In examining the hypothesized theoretical relationships in Figure 2,

we specifically sought to answer the following three questions:

1 Is there evidence that ESL and EAP programs have a directeffect on L2 students’ academic and social engagement?

2 Do course outcomes (defined in this study as goal-setting andother goal-directed practices, English language use, and strategyuse) mediate academic and social engagement? If so, does media-tion differ in relation to program emphasis?

3 Do student characteristics (i.e., levels of anxiety; stress caused bypressure from parents, finances, etc.; and motivation) moderatecourse outcomes? Does moderation differ in relation to pro-gram characteristics (ESL or EAP)? If so, how does it differ?

METHOD

Participants

A questionnaire (see the Appendix) was distributed to ium language support programs in universities across Canada Alluniversity language programs which were members of official organiza-tions (e.g., TESL Canada, TESL Ontario, Alberta-TESL) and/or werementioned on official university websites were included in the survey.FIGURE 2 Hypothesized theoretical relationship of moderated mediation effects of language program characteristics on academic and social engagement.

English-med-Only English-medium universities in Canada were included in thisstudy.

It is important to point out that many Canadian universities have anumber of separate and independent programs to support L2 stu-dents, which are housed within different faculties, departments, orunits Further, in larger Canadian universities, there are many ad hoclanguage support programs which have arisen in response to localneeds within units, departments, or faculties These ad hoc programswere not included in the present survey

Fully completed questionnaires were received from 641 L2 studentsstudying in 36 language support programs (ESL and/or EAP) in 26Canadian universities Of the 641 participants, 80% (n = 513) wereinternational students, with the remaining 20% (n = 128) permanentresidents of Canada or new immigrants They spoke 44 different firstlanguages, with 80% speaking only one additional language (English).The 5 most often reported first languages in order of frequency wereChinese, Korean, Japanese, Arabic, and French Most respondentswere 20–22 years old at the time of the study (n = 269 or 42%),although 2 participants indicated they were 17, and 10 indicated theywere 40 or older The mean age of the respondents was 22.3 years(SD = 4.9); 359 (56%) indicated that their study in Canada was sup-ported by their family; 51 (7%) reported they were fully supported by

a scholarship; another 52 (8%) students indicated they had receivedloans to support their study; and 43 (6%) reported that they wereworking in order to support their study in Canada The remaining 136(21%) indicated that they received some combination of supportfrom, for example, family and scholarship, family and work, work andloan, and so forth Eighty-six percent (n = 550) indicated that theirparents had some form of tertiary-level education (college or univer-sity) The largest number of respondents, 325 (50.7%), were studying

in EAP programs at the time of the study; 255 (39.8%) were studyingEnglish in pre-university ESL intensive programs, and 27 (4.2%) hadrecently completed EAP and were newly enrolled in undergraduatestudy; 34 (5.3%) did not indicate the type of language support theyhad or were receiving

Instruments

A questionnaire (see the Appendix) was designed to investigate therelationships illustrated in Figure 2 In the first section of the ques-tionnaire, Items 1–7 elicited demographic information (e.g., age, gen-der, first language [L1], intended program of study, grades, parents’education); in Section 2, Items 8–19 asked students to identify the

relative importance of factors influencing their choice of university(e.g., career goals, friends, parents) and to provide additional back-ground information (e.g., about financial support, schooling, workexperience) The third section, consisting of 36 Likert-scale items, elic-ited responses regarding motivation, stress, anxiety, strategy use, lan-guage use (L1 vs L2), goals, and satisfaction with academic and sociallife in Canada In the fourth section participants responded to 10 Lik-ert-scale items regarding activity in their L2 English support courseand to 6 open-ended items, which elicited additional informationabout program, plans, and the usefulness of their language support.These 6 items are the focus of a separate study (see Cheng & Fox,2008).

The Likert-scale items in Sections 3 and 4 were of particular tance in testing the validity of the hypothesized model (Figure 2).They were, in effect, operational definitions of each of the variables inthe model For example, language use was operationalized by itemssuch as 43 (I spend most of my time speaking English) and 45 (Ispend most of my time speaking my native language) Anxiety, a stu-dent’s reported uneasiness about English (or internal state of concern

impor-or wimpor-orry), was operationalized by items such as 31 (I get nervous abouthow to say things in English) and 32 (I worry that my English is notgood enough) Stress, a student’s perception of external pressures, wasoperationalized by items such as 25 (I study more because my familywants me to) and 26 (I feel stressed because of my family’s expecta-tions) Motivation was operationalized by items such as 20 (I am seriousabout my academic studies) and 21 (I am serious about my Englishlanguage studies) The questionnaire allowed us to collect data neces-sary to confirm or reject our hypothesized model of the impact of lan-guage program type, ESL or EAP, on student engagement (academicand social)

Analysis

In order to test the relationships in the hypothesized model ure 2), we applied structural equation modeling (SEM) procedures,which allowed us to simultaneously assess multiple paths in the model,

(Fig-to explicitly incorporate error in the analysis, and (Fig-to consider overallestimates of model fit to the data as part of the analytical process (Wu

& Zumbo, 2008) As D€ornyei (2007) points out, the “big advantage” of

an SEM modeling approach is that it “also includes directional pathsbetween the variables and not just information about how the variableshang together” (p 238) In other words, SEM allowed us to testcause-and-effect relationships among the variables, and because it is a

confirmatory technique it also allowed us to confirm the adequacy ofthe hypothesized model in accounting for the data elicited by thequestionnaire.

There are a number of statistical programs available for conductingSEM analysis, for example, LISREL, AMOS, and EQS (Byrne, 1998)

In the present study, we used LISREL version 8.54 to refine our standing of the relationships identified in Figure 1 and to investigatethe validity of the hypothesized model illustrated in Figure 2 SEManalysis provides goodness-of-fit measures which can be used to judgehow well a model accounts for the data One measure is provided by achi-square test, which is reported along with degrees of freedom Ifthe model fits the data, the chi-square test is not significant, and theratio between the chi-square and degrees of freedom is low However,

under-“the sensitivity of the [chi-square] to sample size is widely known”(Byrne, 1994, p 54), and as a result other indices of model fit are alsoreported The root mean square error of approximation (RMSEA) is

“one of the most informative criteria in covariance structure ing” (Byrne, 1998, p 112) and a key indicator of model fit (Loehlin,2004) The RMSEA measures the differences on average of residuals

model-in the observed data and the proposed model In general, values ofless than 08 are deemed to be evidence of good model fit and thosegreater than 10 are not Yet another indication of good model fit isthe number of iterations required before a solution is found (i.e., con-vergence) “The best scenario is a situation where only a few iterationsare needed to reach convergence” (Byrne, 1994, p 55)

Prior to undertaking the SEM analysis, we conducted reliability ysis (i.e., Cronbach’s alpha) for both the overall questionnaire and foreach of the variables in turn In all cases reliability met or exceeded.80, which is considered adequate (D€ornyei, 2007) In some cases,items that undermined reliability and were not considered theoreti-cally substantive were trimmed to improve the clarity and interpretabil-ity of the analysis Then, we applied factor analysis to the Likert-scaleitems drawn from Sections 3 and 4 of the questionnaire in order toexamine the structure of the data in accounting for the constructs(variables) in the hypothesized model

anal-Both the Kaiser-Meyer-Olkin Measure of Sampling Adequacy (KMO)and Bartlett’s Test of Sphericity indicated that the data were suitablefor factor analysis (i.e., KMO = + 8 and Bartlett = 000) Maximumlikelihood with orthogonal rotation was used to confirm the viabilityand interpretability of the factors and further specify the model Sub-sequently, SEM analysis was used to investigate each path in the model

in relation to direct (mediated) and indirect (moderated) effects.Thus, six models were investigated across the two program types, ESLand EAP (Table 1)

In order to examine the moderating effects of students’ personalcharacteristics, the questionnaire responses were divided into twogroups based on high or low levels of reported anxiety, stress, andmotivation Holding the statistical parameters constant, we could thenexamine whether there were significant differences based on groupmembership through the application of a chi-squared difference test.

A significant difference indicated that the variable was acting like a

“dimmer switch in lighting” (Wu & Zumbo, 2008, p 383), moderatingthe effect by lessening or strengthening its intensity

RESULTS AND DISCUSSION

The results of the SEM analysis are discussed in relation to thethree research questions

Research Question 1: Is there evidence that ESL and EAP courses have adirect effect on L2 students’ academic and social engagement?

Analysis of the data provided evidence that both EAP and ESLcourses had a direct effect on the study’s L2 participants’ academicand social engagement Table 2 provides an overview of t-values forthe direct effect of EAP and ESL programs on L2 students’ engage-ment, both academic and social According to this analysis the directeffects of EAP programs on academic engagement are significant,strong, and positive, with t-values ranging from 6.7 to 7.2 (with 0indicating a null effect); EAP programs have a significant but some-what weaker effect on social engagement, with t-values ranging from3.9 to 5.9 The results for ESL programs are similar, with direct

Moderating variables

Impact on engagement

Goal directed practices Motivation

Goal directed practices Motivation

Goal directed practices Motivation

Goal directed practices Motivation

effects on academic engagement ranging from t-values of 7.5 to 7.8and direct effects on social engagement ranging from t-values of 4.2

to 6.0 Overall, these findings attest to the important impact of bothESL and EAP language support on L2 participants’ reported aca-demic and social engagement

Research Question 2: Do course outcomes (defined in this study as directed practices, English language use, and strategy use) mediate aca-demic and social engagement? If so, does mediation differ in relation

plan-• Language use (outcomes that emphasize increased use of English

in spoken communicative/interactive activities, which are moreprevalent in ESL programs)

• Strategy use (outcomes that emphasize the development of demic skills and strategies, which are more prevalent in EAPprograms)

aca-Interestingly, the findings suggest that, regardless of programemphasis, only strategy use appears to have a mediating impact on aca-demic engagement This was the case for both ESL and EAP languageprograms

TABLE 2

Overview of Direct Effects of EAP and ESL Program Characteristics on Academic and Social Engagement by Causal Path (N = 641)

Language program characteristics Direct effect (t-values) Impact on engagement

In the case of the relationship between EAP?Strategy Use?AcademicEngagement, although the chi-square was significant, indicating a poten-tially poor fit of the model to the data, the RMSEA = 067 indicatedgood model fit (Byrne, 1994, 1998; Loehlin, 2004).

The results of the ESL?Strategy Use?Academic Engagement analysiswere similar The model converged in two iterations, and although thechi-square was significant, the RMSEA was 057, indicating a goodmodel fit

This finding clarifies to a degree the nature of the relationshipbetween language program characteristics (ESL and EAP) and aca-demic engagement It suggests that these language support programshave a direct effect on strategies used by L2 students and that this, inturn, affects their degree of academic engagement It lends support tothose who argue that the primary role of ESL and EAP teachers in thecontext of language support programs at the university level is one ofsupporting the development of skills and strategies that help students

Do course outcome variables

mediate impact? [yes or no]

Direct effect?

[yes or no]

Do these variables moderate (lessen or strengthen) impact? [yes or no]

Goal-directed

practices

Language use Strategy use

Anxiety (internal state)

Stress (external

Goal-directed

practices

Language use Strategy use

Anxiety (internal state)

Stress (external