{"id":16196,"date":"2021-01-16T23:57:37","date_gmt":"2021-01-16T14:57:37","guid":{"rendered":"http:\/\/www.tesl-ej.org\/wordpress\/?page_id=16196"},"modified":"2021-02-16T07:44:48","modified_gmt":"2021-02-15T22:44:48","slug":"ej96a3","status":"publish","type":"page","link":"https:\/\/tesl-ej.org\/wordpress\/issues\/volume24\/ej96\/ej96a3\/","title":{"rendered":"Language Task Engagement: An Evidence-Based Model"},"content":{"rendered":"

February 2021 – Volume 24, Number 4<\/h2>\n\n\n\n\n\n
Joy L. Egbert<\/strong>
\nWashington State University
\n<jegbert\"atmark\"wsu.edu><\/td>\n		Reima Abobaker<\/strong>
\nIndependent Researcher
\n<abobakerreima\"atmark\"gmail.com><\/td>\n		Mira Bekar<\/strong>
\nSS. Cyril and Methodius University
\n<m.bekar\"atmark\"flf.ukim.edu.mk><\/td>\n<\/tr>\n
Seyed Abdollah Shahrokni<\/strong>
\nWestern Oregon University
\n<shahroknis\"atmark\"wou.edu><\/td>\n		Pruksapan Bantawtook<\/strong>
\nWashington State University
\n<p.bantawtook\"atmark\"wsu.edu><\/td>\n		Mary F. Roe<\/strong>
\nWashington State University
\n<mary.roe\"atmark\"aggiemail.usu.edu><\/td>\n<\/tr>\n
Xue (Snowy) Zhang<\/strong>
\nThe University of Hong Kong
\n<dodosnowy\"atmark\"163.com><\/td>\n		Haixia He<\/strong>
\nWashington State University
\n<haixia.he\"atmark\"wsu.edu><\/td>\n		Keun Huh<\/strong>
\nHannam University
\n<keun\"atmark\"hnu.kr><\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n
\n

Abstract<\/h3>\n

Recent research points to the need for a specific research focus on language task engagement because task engagement can lead to increased motivation, persistence, satisfaction, and learner achievement (Early, Rogge, & Deci, 2014; Henri, Halverson, & Graham, 2015; Reeve & Lee, 2014); a major gap in the research in this area is the lack of a unifying model. This study responds to this gap in order to move understandings of language task engagement forward. To meet this purpose, the present study applies both descriptive and statistical data to develop and validate a model of language task engagement. The article describes the exploration of language task engagement from two main sources: 1) the large body of literature around engagement, which was used as one source of data for model-building, and 2) online surveys of student, teacher, and researcher perspectives that were collected and analyzed from multiple sources and contexts. To explain the model, the article first presents a brief justification for exploring the engagement construct, differentiating task engagement from related concepts. The paper then outlines the study methodology, presents and describes the model based on the literature and other data, and provides conclusions and recommendations.<\/p><\/blockquote>\n

There is good reason to address the construct of language task engagement; one reason is that not enough is known about engagement in general. The construct is embedded in what scholars describe as different types of student engagement; these include disciplinary (e.g., Wang, Fredricks, Ye, Hofkens, & Linn, 2016), situational (e.g., Inkinen et al., 2019), and class engagement (e.g., N\u00fa\u00f1ez & Le\u00f3n, 2019). Most often mentioned in the literature is school <\/em>engagement (also called contextual engagement), which is students\u2019 general commitment to participate in schooling (Jimerson, Campos, & Greif, 2003). As Moreira, Cunha, and Inman (2019) assert, there is much that is still unknown about the multiple dimensions of the engagement construct, and that is one reason to study it.<\/p>\n

A more specific way to look at engagement is in the notion of task engagement. In this study, \u201ctask\u201d is defined as a classroom activity or exercise with<\/p>\n

    \n
  1. one or more clear goals,<\/li>\n
  2. a sequence or process for meeting those goals, and<\/li>\n
  3. a specific beginning and end.<\/li>\n<\/ol>\n

    Originally defined as time-on-task, that is, the time a learner spends on a classroom activity (van Gog, 2013), task engagement has also been considered as effort quality and amount of student interaction (Alsawaier, 2018). More recently, task engagement has come to refer to a more complex interplay of concepts. For example, Csikszentmih\u00e1lyi (1990) and his colleagues encapsulated many of these concepts in the construct of \u201cflow\u201d (e.g., Almetev, 2018; Cavanagh, 2014; Gardiner, 2017; Ghani & Deshpande, 1994). Studies of flow in language learning (see, e.g., Amini, Ayari, & Amini, 2016; Egbert, 2003; Liu, Wang, & Tai, 2016; Markovic, 2020; Mirlohi, Egbert, & Ghonsooly, 2011; Tardy & Snyder, 2004) have provided evidence that deep task involvement can be experienced in language classrooms, is engendered by a variety of factors, is fluid, and can lead to willingness to communicate and greater language learning. However, while the construct of flow addresses optimal experience, some researchers assert that it does not address the overall complexity and range of task engagement (see, e.g., Christenson, Reschly, & Wylie, 2012; Taylor, 2016). Vasalampi, et al. (2016) define task engagement as showing \u201ccommitment and involvement in a learning task\u201d (p. 46), but they and many others note a lack of deep understanding of language task engagement (see, for example, Mercer, 2019). The potential benefits of task engagement to language learning are additional reasons to study it.<\/p>\n

    Oga-Baldwin (2019) asserts that engagement is \u201cperhaps one of the most crucial steps in predicting how students succeed at languages in formal education settings\u201d (p. 4). More research on task engagement is clearly needed, particularly in the area of language learning (Aubrey, King, & Almukhaild, 2020; Olga-Baldwin, 2019). The gap that this work fills, then, is the creation of a model that integrates available and collected information about language task engagement. It attempts to eliminate the competition between various definitions and related attributes and offer a cohesive model that includes relevant pieces of evidence-based importance. To do so, it reviews both the broader literature around engagement and the specific data collected from language teachers, learners, and researchers. Because the literature was considered part of the data for the study, the literature \u201creview\u201d is included in the findings section.<\/p>\n

    Methodology<\/h3>\n

    Research Team<\/h4>\n

    Nine researchers participated in this project. All team members were involved in every aspect of the study, from literature review and data collection to data coding and model development.<\/p>\n

    At the university in the US Pacific Northwest where the project was centered, one professor emeritus, one full professor (project lead), one adjunct instructor, and two doctoral students participated. Other team members included faculty members from Hong Kong, Korea, the southeastern US, and the Republic of North Macedonia.<\/p>\n

    Data Sources and Analysis<\/h4>\n

    Along with evidence from previous studies conducted by members of the research team (e.g., Egbert, 2003; Egbert, & Abobaker, 2018; Egbert & Borysenko, 2018), documents and descriptive evidence from researchers\u2019 classrooms and on-going studies contributed to the creation of the model. For this study, the research team also reviewed the current literature on engagement and constructed and implemented a set of surveys of English language teachers and English as second\/foreign language (ESL\/EFL) students. These latter two data sources and their analyses are described below.<\/p>\n

    Literature. <\/strong>To explore the evidence around task engagement, we first conducted a literature review of existing studies, instruments (e.g., Wigfield & Guthrie\u2019s 1997 MRQ), teacher anecdotes, and theories to discern the most salient elements of language task engagement. Also included were articles from peer-reviewed journals dating from 1987-2020, including current in-press empirical studies and theoretical\/ conceptual essays from sources such as ResearchGate<\/em> (researchgate.net). All of the items included engagement as a keyword either in the title or the text of the item and were included regardless of methodological framework.<\/p>\n

    Each team member chose articles to read. Ultimately, the team reviewed more than 200 items. Team members created summaries in a shared spreadsheet of each article that included five items: (1) instruments and measures used; (2) proposed engagement elements, (3) conditions for engagement, (4) precursors, facilitators, and antecedents for engagement, and (5) outcomes of engagement. These and other data directly relevant to task engagement were included in a shared spreadsheet until the information became highly redundant.  These summaries were read by the team members before meeting to discuss the model in order to uncover repeated themes and other relevant ideas, and during model-building discussions the ideas in the summaries (e.g., mentions of \u201cinterest\u201d) were aggregated using the \u201csearch\u201d feature.
    \n<\/a>
    \nFormal student and teacher surveys. <\/strong>Two formal, anonymous online surveys were employed to gather data about facilitators of task engagement, one for teachers and one for students. The student survey was composed of six demographic questions (four multiple choice and two open-ended), followed by four open-ended focus questions (found in     Appendix A<\/a>). The demographic\/ background questions inquired about the respondents’ place of birth, current place of study, first language, academic level, language proficiency level, and gender. The open-ended questions targeted the respondents\u2019 favorite activities in general (and the reason why they were preferred), activities they would rather do than what they currently did in class, and what they preferred a teacher do in class. Based on the published literature, including previous studies by team members, the focus questions provided respondents with the opportunity to make both broad and specific statements and to add to the data regarding what students say engages them.
    \n    <\/a>
    \nThe online teacher survey was composed of six demographic and two open-ended focus questions (in     Appendix B<\/a>). The demographic questions tapped the respondents\u2019 place of birth, current teaching location, first language, teaching level, duration of practice, and gender. The open-ended items inquired about how the respondents engaged their students and when they believed students were engaged. These data added additional perspectives on what might engage students and why they become (or do not become) engaged. This study complied with standards of research involving humans as subjects, including that all participants responded voluntarily and no identifying information was collected.<\/p>\n

    In addition to being provided in English, the surveys were translated into 10 languages (Arabic, Chinese simplified, Chinese Traditional, Indonesian, Korean, Persian\/Farsi, Russian, Spanish, Ukrainian, Turkish) by team members and colleagues in other countries who volunteered to participate; they were then back-translated and created using Qualtrics web-based survey-development software (Qualtrics, 2018). The introduction page of the survey asked respondents to complete the survey in the language in which they felt most comfortable.<\/p>\n

    The online surveys were provided to faculty members and students in the field through personal messages, professional list-servs, and face-to-face meetings. These colleagues then passed on the survey link to others and a snowball effect helped to obtain additional responses. Table 1 shows the overall number of responses in each language from students and teachers. However, the language used for the survey does not imply any specific country of either origin or study; in addition to students with the 10 language backgrounds in which the survey was offered, students from other language backgrounds also completed the survey. These included, for example, a Japanese student who completed the survey in Chinese, one student from India and one from Germany in English, and one from Cyprus who took the survey in Russian. Further, 38 Macedonian students whose L1 is Macedonian completed the survey in English, as did five students who were born and live in Macedonia but are Albanians (with Albanian as L1 and Macedonian as L2) and one student born in Macedonia but of Turkish heritage.<\/p>\n

    Table 1. Student and Teacher Responses and Survey Language.<\/strong><\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
    Language<\/strong><\/td>\nStudent Responses<\/strong><\/td>\nTeacher Responses<\/strong><\/td>\n<\/tr>\n
    Arabic<\/td>\n 18<\/td>\n 4<\/td>\n<\/tr>\n
    Chinese simplified<\/td>\n396<\/td>\n 54<\/td>\n<\/tr>\n
    Chinese Traditional<\/td>\n0<\/td>\n 2<\/td>\n<\/tr>\n
    English<\/td>\n82<\/td>\n84<\/td>\n<\/tr>\n
    Indonesian<\/td>\n 0<\/td>\n 0<\/td>\n<\/tr>\n
    Korean<\/td>\n 58<\/td>\n4<\/td>\n<\/tr>\n
    Persian\/Farsi<\/td>\n40<\/td>\n8<\/td>\n<\/tr>\n
    Russian<\/td>\n 4<\/td>\n 0<\/td>\n<\/tr>\n
    Spanish<\/td>\n1<\/td>\n 0<\/td>\n<\/tr>\n
    Ukrainian<\/td>\n 8<\/td>\n4<\/td>\n<\/tr>\n
    Turkish<\/td>\n1<\/td>\n 0<\/td>\n<\/tr>\n
    TOTALS:<\/strong><\/td>\n608<\/td>\n160<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

    The survey responses completed in various languages were translated into English and then checked another time for accuracy by native speakers of the language of the survey. Sets of surveys were coded by pairs of researchers using a coding scheme that was created based on subsets of the data. The coding scheme was revised and retested with four researchers three times. A final revision was normed with the whole research team and supported with written explanations of details such as how to decide how to discern meaning units. The scheme consisted of nine category codes (e.g., Tools, Strategies, Authenticity, etc.) and 84 sub-codes (e.g., self-efficacy, connection to language, real-world, current and future academics, interactions with peers, teachers, and\/or experts). Coders then assigned one code to each meaning unit in the data (see Dao, 2019, for an explanation). Data with no sub-code fit were coded in the broader category.<\/p>\n

    Each set of researchers coded the data individually and then met to reconcile differences. The results of initial inter-coder reliability showed a high rate of agreement between the coders (from 78% to 100%, with an average of 95%). All data were then reconciled through discussion to 100%. The 3,995 individual codes (designated \u201ccomments\u201d in the rest of this article) were entered into a spreadsheet. Finally, descriptive statistics were conducted for each language group and for teacher and learner groups separately.<\/p>\n

    Further, confirmatory factor analysis (CFA) techniques were used to examine the engagement facilitators coded from responses to each teacher and student question in the formal surveys. A dichotomous indicator was calculated for each engagement facilitator to represent the presence or absence of the facilitator for a single question. This produced 54 dichotomous indicators (nine for each question). CFA models were then built using the indicators for each question and estimated using MPLUS 8.3 (Muth\u00e9n & Muth\u00e9n, 1998-2017). Model fit was assessed using Chi-square, CFI, RMSEA, and SRMR. If the model fit was inadequate according to established fit thresholds (i.e., CFI > .95, RMSEA < .06, SRMR < .08; Hu & Bentler, 1999) then the indicator with the lowest R-square was eliminated and the model re-estimated. Figure 1 shows an example of the CFA model for Teacher question 7. Panel A of the Figure shows the initial model estimated with all nine indicators which would not converge to produce estimates. Panel B shows the reduced model with the best fit (c2<\/sup> = 8.90 p<\/em> = .45, CFI = 1.0 RMSEA = 0.0 (0.0, 0.09) SRMR = .08).<\/p>\n


    \nFigure 1. Example of the Full CFA Model (Panel A) and Final CFA Model (Panel B) for Teacher Question 7.<\/strong><\/p>\n

    Table 2 shows the fit statistics for the final models for all six student and teacher questions. Only one model (Student Q9) produced a significant Chi-square and one model produced a CFI well below the .95 threshold (Student Q8), indicating a lack of fit for those models. This makes sense because of the differences in the content of each question asked. Therefore, overall fit statistics, as well as factors including parsimony, response count, and lack of improvement in models with fewer terms, led us to accept the model of the engagement facilitators that were the most meaningful in answering the teacher and student questions.<\/p>\n

    Table 2. Final Model Fit Statistics.<\/strong><\/p>\n\n\n\n\n\n\n\n\n\n
    <\/td>\nQuestion<\/td>\nCFI<\/td>\nRMSEA<\/td>\nRMSEA 95% CI<\/td>\nSRMR<\/td>\nChi-square<\/td>\np<\/em>-value<\/td>\n<\/tr>\n
    Teacher<\/td>\n7<\/td>\n1.000<\/td>\n0.000<\/td>\n(0.000, 0.089)<\/td>\n0.083<\/td>\n8.904<\/td>\n0.446<\/td>\n<\/tr>\n
    8<\/td>\n1.000<\/td>\n0.000<\/td>\n(0.000, 0.096)<\/td>\n0.085<\/td>\n3.770<\/td>\n0.583<\/td>\n<\/tr>\n
    Student<\/td>\n7<\/td>\n0.996<\/td>\n0.016<\/td>\n(0.000, 0.086)<\/td>\n0.030<\/td>\n2.297<\/td>\n0.317<\/td>\n<\/tr>\n
    8<\/td>\n0.916<\/td>\n0.039<\/td>\n(0.000, 0.067)<\/td>\n0.087<\/td>\n16.819<\/td>\n0.052<\/td>\n<\/tr>\n
    9<\/td>\n0.945<\/td>\n0.043<\/td>\n(0.014, 0.071)<\/td>\n0.082<\/td>\n18.669<\/td>\n0.028<\/td>\n<\/tr>\n
    10<\/td>\n0.957<\/td>\n0.022<\/td>\n(0.000, 0.048)<\/td>\n0.074<\/td>\n17.817<\/td>\n0.215<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

    Table 3 illustrates the engagement facilitators that were significant elements in the final factor models for each teacher and student question. Cells with significant loadings only (p<\/em> < .05) are shown in the Table, denoting which engagement facilitators were significant indicators for the final factor models estimated for each question. Blank cells denote indicators that did not load onto the final factor model for each particular question. For example, the final factor model for teacher question 7 shows significant loadings for the engagement facilitators authenticity, autonomy, learning support, and interest, while the social interaction and challenge were non-significant. Standardized loadings are shown for the teacher questions, while unstandardized are shown for the student loadings due to estimation problems encountered with the student models.<\/p>\n

    Table 3. Results of Confirmatory Factor Modeling of Engagement Facilitators by Question.<\/strong><\/p>\n\n\n\n\n\n\n\n\n\n\n
    <\/td>\n<\/td>\nTeacher
    \n(Standardized)<\/td>\n    		Student
    \n(Unstandardized)<\/td>\n<\/tr>\n
    Number<\/td>\nEngagement Facilitator<\/td>\nQ7<\/td>\nQ8<\/td>\nQ7<\/td>\nQ8<\/td>\nQ9<\/td>\nQ10<\/td>\n<\/tr>\n
    1.<\/td>\nAuthenticity<\/td>\n-0.535<\/td>\n0.433<\/td>\n—<\/td>\n5.219<\/td>\n-0.287<\/td>\n1.431<\/td>\n<\/tr>\n
    2.<\/td>\nSocial interaction<\/td>\n—<\/td>\n—<\/td>\n0.477<\/td>\n2.609<\/td>\n1.051<\/td>\n4.715<\/td>\n<\/tr>\n
    3.<\/td>\nChallenge<\/td>\n—<\/td>\n—<\/td>\n—<\/td>\n1.956<\/td>\n—<\/td>\n—<\/td>\n<\/tr>\n
    4.<\/td>\nAutonomy<\/td>\n-0.813<\/td>\n—<\/td>\n—<\/td>\n1.898<\/td>\n—<\/td>\n1.197<\/td>\n<\/tr>\n
    5.<\/td>\nLearning Support<\/td>\n-0.311<\/td>\n0.463<\/td>\n—<\/td>\n—<\/td>\n—<\/td>\n2.742<\/td>\n<\/tr>\n
    6.<\/td>\nInterest<\/td>\n-0.190<\/td>\n-0.470<\/td>\n—<\/td>\n5.802<\/td>\n-0.680<\/td>\n0.525<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

    Loadings are shown for each engagement facilitator that was a significant (p<\/em> < .05) indicator for the final factor model estimated for each question.<\/p>\n

    A major limitation of the formal surveys analysis is the interpretation of results given that responses were coded by a rater after obtaining written responses from participants. This leads to questions while interpreting individual loadings and the resultant factors themselves within the context of each survey question; however, a content analysis shows that this is a reasonable outcome based on the framing of the survey questions.<\/p>\n

    Informal surveys. <\/strong>In order to gain additional data, an informal digital survey using PollEverywhere software was composed of two open-ended questions in English that asked attendees at sessions at five ESL\/EFL-focused conferences (three in the U.S., one in Taiwan, and one in the Republic of North Macedonia) what adjectives they would use to describe the best and worst language teaching and\/or learning activities they had taught or experienced. The anonymous surveys were conducted at sessions that focused on task engagement, and the voluntary data were used as part of the conference presentation to demonstrate concepts of task engagement to the audience. The reason for the open-ended questions was to see whether and how the respondents addressed engagement elements and what other aspects of the activities were salient to them. The survey was created using Google Forms and was completed by the respondents on their own devices during the conference sessions. One hundred and seventy-seven students and teachers completed the informal survey; no return rate was calculated. The informal surveys, which were conducted while the study was in progress, were analyzed using the same coding scheme and process as the formal surveys. The total number of codes assigned was 606. Initial inter-rater agreement was 89%, and then all codes were reconciled through discussion to 100%. The data were input into a spreadsheet and descriptive statistics conducted.<\/p>\n

    While it may be true, as Oga-Baldwin (2019) asserts, that language learners who “are getting high quality input, producing regular output, repeating their practice, and being instructed on the form of the language” (p. 22) are likely to be engaged, the study data from language learners and educators and the broader literature provide both a wider and deeper view of both the definition and components of task engagement. The results contribute to an evidence-based model based in theory of language task engagement, which was created based on a series of discussions by the research team of the overall patterns in all of the data.<\/p>\n

    Findings and Interpretations<\/h3>\n

    The findings from the literature and survey data support the set of facilitators and indicators in the model in Figure 2. The literature contributed relatively more to the task elements and outcomes constructs in the model, while all of the data were considered equally for the facilitators and indicators. This section explains and describes each part of the model, based on the literature, the data analysis, or both, and it includes the relative contribution of each. All student and teacher comments were used verbatim from the formal surveys.<\/p>\n


    \nFigure 2. Model of Language Task Engagement.<\/strong><\/p>\n

    Language Task Elements<\/h4>\n

    Based on a synthesis of descriptions in the literature review, elements of a task include all aspects of a task that a teacher can design, from choice of topic and content to the product the students produce. Future research can show the elements where task engagement makes the biggest difference for language learners.<\/p>\n

    Task Engagement Facilitators<\/h4>\n

    Based on the literature review and survey data, the research team identified six primary facilitators that may enhance language task engagement when integrated into task components and perceived by learners. The collected data does not suggest to what extent or in how many task components these facilitators should be integrated; this appears to be context-based and remains a topic for future research. Table 4 provides frequencies for the six facilitators gleaned from the formal survey data. A description of each facilitator with additional findings follows.<\/p>\n

    Table 4. Frequencies of Comments about Facilitators in the Formal Survey Data.<\/strong><\/p>\n\n\n\n\n\n\n\n\n\n\n
    Frequencies of Comments on the Facilitators <\/strong><\/td>\n<\/tr>\n
     Facilitator<\/strong><\/td>\nTeachers<\/strong><\/td>\nStudents<\/strong><\/td>\nTotal<\/strong><\/td>\n<\/tr>\n
    Authenticity<\/td>\n141<\/td>\n670<\/td>\n811<\/td>\n<\/tr>\n
    Social interaction<\/td>\n120<\/td>\n560<\/td>\n680<\/td>\n<\/tr>\n
    Learning support<\/td>\n81<\/td>\n400<\/td>\n481<\/td>\n<\/tr>\n
    Interest<\/td>\n61<\/td>\n388<\/td>\n449<\/td>\n<\/tr>\n
    Autonomy<\/td>\n46<\/td>\n171<\/td>\n217<\/td>\n<\/tr>\n
    Challenge<\/td>\n10<\/td>\n112<\/td>\n122<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

    Authenticity.<\/strong> Overall, learner perceptions of authenticity in the formal surveys (811 comments) support the literature that proposes it as a core facilitator of language task engagement (present in about a third of the literature reviewed). The informal surveys also support this finding, with authenticity as the most-mentioned facilitator, with 83 overall comments. Authenticity has been defined in the literature in various ways. This includes the genuineness of materials and cultural practices exemplifying the native speaker community (Widdowson, 1996), the use of real-life langue and language in conveying a real message (Benson & Voller, 1997), and meaningful real social\/classroom communication, negotiation, and interaction (Guariento & Morely, 2001). Svalberg (2018) posits the notion of \u201cengagement with language\u201d (p. 21) in instructed language contexts, basing it on the construct of meaningfulness, which includes the concepts of purposefulness, utility, and enjoyment.<\/p>\n

    A synthesis of the literature reviewed for this study further provides a broad definition of authenticity as learner perceptions of relevance and value of a task, which may be embodied in different task components in different ways. Our findings underscored three aspects of authenticity: connections, self-efficacy, and meeting learning needs\/goals. First, learners appear to perceive authenticity in a task when it makes personal connections. These include connections to learners\u2019 lives outside of class (i.e., their backgrounds, cultures, interests, etc.), to their previous learning, to their future learning, and\/or to the task topics and tools. As one student said, if the students \u201cdon\u2019t see a point, learning won\u2019t begin.\u201d Another added that teachers should use topics from the \u201cdaily lives of students.\u201d A teacher commented that students are more likely to be engaged when \u201cThey are valued for who they are, their learning and growth, and their contributions to their peers\u2019 and teacher\u2019s development.\u201d<\/p>\n

    Further, the surveys indicated that when students feel confident\/ self-efficacious the task can help them to successfully achieve their personal learning goals and they are more apt to actively participate in and accomplish the tasks. Third, learners noted that they find a task authentic when they perceive it as addressing their language, life, and academic goals. For example, one student stated that activities should help them with \u201cself-improvement not just the language,\u201d while another remarked that teachers \u201chave the responsibility to also teach humanism, what is the basic right and wrong…stimulate [students\u2019] minds and help them develop [as individuals].\u201d<\/p>\n

    Social interaction.<\/strong> As noted in the literature review, a large body of research across various perspectives, i.e., cognitive (e.g., Ahn, 2016; Coyle, 2007; Robinson & Ellis, 2008), sociocultural (e.g., Duff, 2007; Norton, 2000; Peterson, 2012), and sociocognitive (e.g., Atkinson, 2014; Batstone, 2010; Han & Hyland, 2019), discusses the essential role of interaction in support of learners\u2019 engagement in language learning, and ultimately in their language achievement. The cognitive perspective, emphasizing processing language input and producing output, sees interaction as participatory experiences in social contexts where learners can negotiate meaning or form (Robinson & Ellis, 2008). The sociocultural perspective considers interaction as a way to acquire language and establish new identities (Norton, 2000). The sociocognitive perspective, which combines the above two perspectives, underlines the importance of both language use and interaction (Batstone, 2010).<\/p>\n

    However, the survey respondents perceived social interaction more pragmatically as communication between teacher and learner or learner and learner. According to the formal survey, types of interaction opportunities favored by students and teachers include interaction with peers (127 mentions), teachers (100 mentions), specific others such as native speakers and experts (25 mentions), and other non-specified social interaction such as discussion and conversation (428 mentions). As one student explained, \u201cI\u2019d like to be involved in a discussion, because I can organize my own views and understand the thoughts of various people. This is the activity that you will learn best about the subject.\u201d Another student noted that social interaction \u201chelps students relax and form bonds with each other.\u201d Some of the Macedonian students suggested ideas like, \u201cIt would be interesting and good if we had an opportunity to spend our lessons with a native speaker.\u201d In addition, effective communication with teachers was highlighted in the survey responses. One student suggested that teachers \u201ccommunicating with students is a key factor that either makes or breaks the atmosphere,\u201d while another stated that teachers should be \u201ccollaborative but not [too] much friendly.\u201d<\/p>\n

    With 44 comments, social interaction was the third most mentioned facilitator in the informal surveys, which supports both the literature and the formal surveys around the perceived importance of social interaction to task engagement (Dao, 2020). However, a few dissenting voices from learners noted that they preferred to work individually, asserting that, \u201c…in classes, not everyone likes to collaborate or participate.\u201d<\/p>\n

    Learning support. <\/strong>A large body of literature shows that certain types of learning support from teachers can assist student engagement in learning activities: building positive teacher-student relationships; showing personal interest in or concern for students; promoting students\u2019 needs, interests, and goals; supporting students\u2019 need for autonomy; employing relevant learning activities; presenting various types of performance feedback; providing scaffolds and structure to learning; and explicitly directing students\u2019 learning in structured tasks (see, e.g., N\u00fa\u00f1ez & Le\u00f3n, 2019; Reeve, 2013; Shernoff, Ruzek, & Sinha, 2017; Tian & Zhou, 2020; van Uden, Ritzen, & Pieters, 2014). Additionally, other specific types of learning support, such as using humor, rewards, and visuals, have been found effective to support language learning (e.g., Bell & Pomerantz, 2015; Curry & Lillis, 2004, Kiss & Weninger, 2017).<\/p>\n

    The formal survey respondents also identified many of these literature-based practices.  They commented 481 times on 18 distinct areas of learning support that they perceived as important. Among them, individualized feedback, equal support of all students, and clear instructions were among the most often mentioned. Two of the participants who responded in Arabic indicated that they needed \u201cto be given the opportunity to ask about anything in the lecture such as the translation of unknown words and [for the teacher] to not refuse to translate some words assuming that students are supposed to be familiar with them.\u201d A Macedonian student suggested that,<\/p>\n

    Generally, students are not motivated to do academic work and extracurricular activities on their own (it is responsibility of both parties). A good teacher needs to know in which direction the class is moving and plan activities for mixed-ability classes. Then to engage all students in activities, usually the same ones discuss issues in class, the shy ones are silent most of the time.<\/p>\n

    In the informal surveys, 40 comments mentioned learning support as part of the \u201cbest\u201d language learning activity, while 19 items stated that a lack of learning support was part of their worst learning activity. The findings not only provide evidence that students perceive the teacher as responsible for supporting their learning, but also that they believe that teachers should help students support each other as much as possible; this makes learning support an essential element in the model.<\/p>\n

    Interest. <\/strong>The reviewed literature suggests that interest is one of the most crucial facilitators of task engagement; however, the type of interest that matters most remains unclear.  Situational interest, for example, includes aspects such as novelty, exploration, challenge, and instant enjoyment of the subject matter (Ainley, 2012; Janna, 2019). Ainley suggests that when personal interest aligns with the content taught, it can promote situational interest, allowing students with higher initial interest to persist longer in tasks. The author contends that personal interest, on the other hand, involves learners\u2019 relatively permanent individual preferences based on real-life interests and values that may or may not be reflected and\/or maintained in the learning environment.<\/p>\n

    Scholars and participants noted interest as an important facilitator of task engagement. In the formal surveys, 449 comments addressed interest.  In the informal surveys, 81 comments mentioned interest as an important aspect of engagement in the best language learning activities and 110 comments addressed the idea that lack of interest was part of their worst activities. One Arabic student called for more personally interesting tasks for \u201cdeveloping the curriculum to include practical activities rather than teaching theories.\u201d Another student expressed the desire to include in the curriculum \u201cstories and novels. I want to read them because they are enjoyable, and they impress\/ interest me and grab my attention.\u201d Many of the surveyed students mentioned their positive engagement in online activities and games and stated that those tasks captivated their interest, were most efficient for learning, and made them want more.<\/p>\n

    Autonomy.<\/strong> Autonomy refers to learners\u2019 ability to exercise control over their learning; it is characterized by choice and responsibility (Benson & Voller, 1997). Existing scholarship sheds much light on the specific benefits of autonomy. Specifically, the impetus of language learning comes intrinsically from the learner, which puts learner autonomy at the heart of the process of language learning (Little, 2007; Nunez & Leon, 2019). In addition, the right amount of autonomy, measured through student input, teacher observation, and trial and error in different tasks, can support learners\u2019 engagement and ultimately enhance learning outcomes (e.g., Nakamura, Phung, & Reinders, 2020; N\u00fa\u00f1ez & Le\u00f3n, 2019; Reeve, 2013; Shernoff et al., 2017). When learners are allowed to make significant decisions about what is to be learned, along with how, when, where, and why to learn, they are more likely to be motivated and engaged in the learning task (Mozgalina, 2015). Further, when learners take responsibility for their own learning, such as monitoring their learning process and evaluating their learning outcomes, they are more likely to experience task engagement (Jang, Reeve, & Deci, 2010).<\/p>\n

    Respondents to the formal surveys mentioned autonomy far less than other facilitators, with a total of 217 mentions out of 2,760 comments (see Table 4). Further, it appears that culture\/ educational background may play a role in perceptions of autonomy. The data do not explain this phenomenon; however, some Macedonian students did address their desire for control during tasks such as discussion, stating that teachers should be \u201cgiving students the freedom to talk to each other. If the teacher wants to give their own opinion on the matter, they should do so in short and clear sentences, instead of taking over the whole discussion.\u201d In other words, while some of the respondents did not mention autonomy or choice as something they perceived or wanted, others within the same language group or of the same educational background did. In the informal surveys, autonomy was mentioned only five times as being important to the \u201cbest\u201d language learning activity. While autonomy may be of more importance to some students than others, it remains a facilitator of task engagement in general.<\/p>\n

    Challenge.<\/strong> According to flow theory (Csikszentmih\u00e1lyi, 1990), when learners\u2019 skill levels align with the level of task difficulty, learners optimize their learning by experiencing flow (i.e., full involvement and enjoyment in learning). Second language acquisition research has also found that task engagement can be facilitated when learners\u2019 skill levels match the challenges of a language learning task (e.g., Czimmermann & Piniel, 2016; Egbert, 2003). This indicates that language tasks should provide challenge at an optimal level so that the task pushes learners to think and invest effort rather than being bored or disengaged (Aubrey, King, & Almukhaild, 2020; Shernoff, 2013). Although found in the language learning literature as a salient aspect of student engagement (Pawlak, et al., 2020), participants mentioned challenge the least number of times on the formal surveys; the teachers mentioned it 10 times, whereas students made 112 comments that focused on this facilitator. Some anecdotes suggest that this relatively low number might stem from students\u2019 inability to consider that levels of challenge in their educational systems can change, or they might consider other aspects more important in considering engagement. Keeping with this same pattern, only 11 comments in the informal surveys indicated that sufficient challenge was part of their best learning activity.<\/p>\n

    Overall, students and teachers recognized that engagement supports learning, and teachers recognized this same idea, noting that engaged students do better work. The survey data uncovered strong patterns to support existing scholarship around the facilitators of learning support, social interaction, and authenticity. Additional but weaker patterns arose for interest, autonomy, and challenge. No other facilitators rose to the level of \u201cpattern,\u201d i.e., had enough support in the data or the literature, to be considered central to language task engagement, but future research may discover others.<\/p>\n

    Level of Task Engagement (Indicators)<\/h4>\n

    The model posits that, when students are engaged by the facilitators integrated into their language tasks, they should exhibit signs of their level of engagement. These signs, or indicators, synthesized from existing scholarship, fall into five main categories: behavioral, cognitive, emotional, agentive, and social.<\/p>\n

    Behavioral. <\/strong>Teachers can tap into students\u2019 behavior as an indicator of language task involvement (Assor, Kaplan & Roth, 2002; Henrie, Halverson, & Graham, 2015; Law, Chung, Leung, & Wong, 2017). As one teacher responded, \u201cI can discern from their body language whether they are engaged or not,\u201d and another noted, \u201cI can tell they are engaged when I have to struggle to interrupt them (and they are working).\u201d The literature suggested that teachers can observe how often and how much students participate and in what language, the focus and length of their attention, the effort they appear to put toward the task, their paralinguistic behaviors such as speaking louder, their persistence in the face of challenges, time limits, and other potential distractors. The literature provides examples such as when a student constantly sighs loudly and leans away from his or her desk, this can be taken as a sign of frustration or even disengagement. It also suggests that across contexts and cultures, not all students will outwardly display the same engagement behaviors. For example, one student might look around the classroom, seemingly disengaged but actually thinking hard about the task, while another might stare at the assigned text but not think about it at all. When teachers know their students well, they might obtain a better idea about what their students\u2019 behaviors indicate. However, the literature indicates that teachers might find it more effective during instruction to use indicators in addition to these behavioral ones to evaluate how engaged their students are.<\/p>\n

    Cognitive. <\/strong>Cognitive engagement can be challenging to detect since it entails mental processing, such as students’ understanding of and focus on content (Assor, Kaplan, & Roth, 2002; Henry & Thorsen, 2018). Various scholars have attempted a number of assessment tools to identify it, for example: self-report reflections (Henrie, Halverson, & Graham, 2015); use of strategies (Pintrich & De Groot,1990); introspective interviews (see van Uden, Ritzen, & Pieters, 2014), and language output (Dao, 2019). The literature review indicated that cognitive indicators were the most difficult to assess.<\/p>\n

    Emotional. <\/strong>Emotion\/affect can also be a major indicator of level of engagement (Han & Hyland, 2015; Sato, 2017). Positive affect is seen as part of deeper engagement, while displays of negative affect can indicate less engagement or complete disengagement. Studies support the use of physical emotion indicators such as laughter, signs of interest, positive reactions to tasks, and peer and faculty relationships to ascertain whether and to what level students are engaged; others posit that student self-reports can be used, such as an emotion or enjoyment survey or questionnaire items (for examples of studies that address emotional indicators of engagement, see Dao, 2019; Early et. al, 2014; Gunuc & Kuzu, 2015). While instruments do exist that purport to measure emotional indicators of engagement (see, for example, Kim, Park, Cozart, & Lee, 2015), they have not been integrated into an overall measure of task engagement and have not been used in the field of language learning.<\/p>\n

    Agentive. <\/strong>Levels of student agency have recently gained focus in the literature as an indicator of engagement (see, e.g., D\u00f6rnyei, 2008; Henry & Thorsen, 2019; N\u00fa\u00f1ez & Le\u00f3n, 2019; York, 2020). While cognitive indicators address mental processes, this indicator focuses on student action upon the learning activities, or, as Henry and Thorsen (2019) indicate, students \u201cnot only react to learning activities, but also proact upon them\u201d (p. 5). Agentive indicators can be seen when tasks are designed to allow students a sense of belonging and self-expression. This sense of control is embodied in the autonomy facilitator. Several teachers commented on student agency, noting that engagement occurred when the teacher \u201c[gives] them every opportunity to be leaders in and out of the classroom,\u201d and when they \u201chave the opportunity to create their own [research] questions.\u201d Future research can provide both measures of and evidence around this indicator.<\/p>\n

    Social. <\/strong>While some scholars categorize willingness to communicate as a behavioral indicator of engagement, others have assigned it to the category of social indicators (Almetev, 2018; Mercer, 2019). Teachers noted specific contexts in which this indicator was present; for example, \u201cif they have a particular role within the group, this makes them feel that their contribution matters, no matter how small it is.\u201d This distinction is an area for future research.<\/p>\n

    Overall, if students are experiencing some level of language task engagement, it can be measured in part by observing their actions and affect. However, this is a simplistic view of measuring a complex construct, and other measures of cognitive, agentive, and social indicators may provide a more thorough understanding of the nature and extent of language task engagement.<\/p>\n

    Outcomes<\/h4>\n

    As noted previously, scholars basically agree that task engagement can lead to achievement (see Hiver, et al., 2021, for example), but they vary widely in what outcomes they address. For example, while some studies emphasize social outcomes (e.g., Ainley, 2012), others investigate performance and assessment (e.g., Craig, 2016), while still others explore confidence and achievement (e.g., Crick & Goldspink, 2014). Researchers and teachers can use the model described in this article to examine all of these outcomes and more specific achievement in language and content learning. The results of such studies can help to refine the model and add to the literature base on language task engagement.<\/p>\n

    Implications<\/h3>\n

    Overall, the model presented in this study indicates that task engagement facilitators can be integrated into one or more task elements to increase student engagement and support their learning. In turn, teachers can monitor students\u2019 levels of engagement (by observation, survey, or other means) and revise instruction as appropriate. Some of the implications of this study\u2019s outcomes for teaching and research are discussed below.<\/p>\n

    Teaching<\/h4>\n

    Using the proposed model as a reminder of how and in what aspects language teachers can design engaging tasks may help them to gain a general proclivity toward engaging their language learners. Once teachers know the interests, needs, and abilities of their students, they can try out different combinations of facilitators and task elements from the model as they design tasks to reach course goals. Because each classroom context is unique, it is not possible to prescribe what teachers should do to engage their students, only what they may do based on the model.<\/p>\n

    For example:<\/p>\n