In the current digital age, practitioners and additional stakeholders in second language (L2) teaching have been integrating technology into classrooms to empower learners in a world in which communication via technology is critical for their academic and professional success (González-Lloret & Rock, Reference González-Lloret, Rock, Ziegler and González-Lloret2022). In particular, the surge in online instruction and the use of innovative AI-generated tools have ignited great interest in the application of technology within educational contexts. This study focuses on task-based language teaching (TBLT)—an independent academic domain dedicated to researching and teaching additional languages through tasks (https://www.iatblt.org/). Since González-Lloret and Ortega’s (Reference González-Lloret, Ortega, González-Lloret and Ortega2014b) introduction to the framework of technology-mediated TBLT—a concept that integrates tasks and technology to enhance language learning—there has been an increasing interest in the intersection of educational technology and TBLT (Ziegler, Reference Ziegler2016). However, to date, there remains limited knowledge regarding the methodological and substantive features of published empirical studies exploring the intersection of tasks and technology. Thus, this paper systematically reviews empirical studies on technology-mediated TBLT published between 2000 and 2022. It focuses on methodological trends in this domain and provides directions for future research.
Literature review
Technology-mediated TBLT: The intersection between TBLT, (instructed) SLA, and CALL
Various definitions of tasks have been introduced over the last two decades (e.g., Ellis, Reference Ellis2003; Long, Reference Long, Hyltenstam and Peinemann1985). A commonality among them is the need for tasks to “focus on meaning, be goal-oriented, and have an outcome apart from merely practicing the language” (González-Lloret & Rock, Reference González-Lloret, Rock, Ziegler and González-Lloret2022, p. 38). Despite critiques of TBLT, several review papers (Ellis, Reference Ellis2017; Long, Reference Long2016) and meta-analyses (Bryfonski & McKay, Reference Bryfonski and McKay2019) have emphasized the abundant empirical support for the benefits of using tasks for L2 instruction.
González-Lloret and Ortega (Reference González-Lloret, Ortega, González-Lloret and Ortega2014b) first introduced the concept of “technology-mediated TBLT” as the integration of tasks and technology for language learning. The incorporation of technology in task-based curricula provides learners with opportunities to improve not only their L2 learning but also their digital literacy and skills in utilizing technological tools, which in turn facilitates their engagement with real-world tasks in the digital world. Compared to more traditional definitions of a task (e.g., Ellis, Reference Ellis2003), the definition of a technology-mediated task should be more encompassing since learners’ digital literacy and technological proficiency are necessary for successful task completion and forming social relationships with others in technology-mediated settings (González-Lloret & Ortega, Reference González-Lloret, Ortega, González-Lloret and Ortega2014b; González-Lloret & Rock, Reference González-Lloret, Rock, Ziegler and González-Lloret2022; Ziegler, Reference Ziegler2016).
Technology-mediated TBLT is primarily shaped by three influential fields: TBLT, instructed second language acquisition (ISLA), and computer-assisted language learning (CALL). This section briefly discusses each of the three domains and then explores the intersections of their theories. First, TBLT, one of the most extensively researched L2 pedagogical approaches, has various conceptualizations within the literature. However, the key characteristics that define TBLT is that it is “an approach to course design, implementation, and evaluation intended to meet the communicative needs of diverse groups of learners” (Long, Reference Long2015, p. 5). In contrast to other pedagogical approaches, such as the structural approach (which focuses on teaching discrete grammatical structures), TBLT prioritizes meaning while still addressing form. TBLT emphasizes the significance of engaging learners’ natural abilities by promoting incidental language learning through the performance of tasks that draw learners’ attention toward the target language rather than solely on the isolated linguistic forms (Ellis et al., Reference Ellis, Skehan, Li, Shintani and Lambert2020). Second, ISLA, known as a sub-field of second language acquisition (SLA), is a research field that aims to “understand how the systematic manipulation of the mechanisms of learning and/or the conditions under which they occur enable or facilitate the development and acquisition of an additional language” (Loewen, Reference Loewen2020, pp. 2–3). The majority of ISLA research is motivated by theories of second language learning such as the interaction approach to language learning (Gass & Mackey, Reference Gass, Mackey, VanPatten, Keating and Wulff2020), skill acquisition theory (DeKeyser, Reference DeKeyser, VanPatten, Keating and Wulff2020), and sociocultural theory (Lantolf et al., Reference Lantolf, Poehner, Throne, VanPatten, Keating and Wulff2020). Finally, CALL focuses on the use of digital technology in language learning and teaching. CALL originated as a discussion of professional issues surrounding the use of technology for language instruction by a small group of researchers. However, CALL has since emerged as a prominent focus within applied linguistics, as the growing presence of technology in various aspects of L2 practice—including L2 use, teaching, and teacher education—has consistently driven the development of pedagogical approaches that utilize technology (Chapelle, Reference Chapelle and Hinkel2005).
Scholars have discussed the relationships among these three areas, namely between CALL and SLA (Chapelle, Reference Chapelle2009; Plonsky & Ziegler, Reference Plonsky and Ziegler2016), TBLT and ISLA (Loewen & Sato, Reference Loewen and Sato2021), and CALL and TBLT (Ziegler, Reference Ziegler2016). Accordingly, researchers have suggested the need for reciprocal collaboration between these research domains at both the theoretical and methodological levels. For instance, Chapelle (Reference Chapelle1997, Reference Chapelle2009) emphasizes the need to firmly ground CALL in SLA theory and methodology since CALL has become an independent research domain. Similarly, Loewen and Sato (Reference Loewen and Sato2021) claim that both TBLT and ISLA are concerned with L2 learning in instructed settings, and that the ultimate goal of these two fields is to find effective instructional conditions and learning mechanisms. The authors further highlight that they complement each other at the theoretical, empirical, and practical levels and identify the use of technology as a shared future direction for research and practice that can benefit from the field of CALL. Furthermore, tasks have become essential to all three disciplines, serving as both the main unit of instruction and as research tools. Consequently, scholars have highlighted the ways in which CALL research benefits from TBLT, namely providing a theoretical framework to “design more pedagogically effective computer-based activities” (Ziegler, Reference Ziegler2016, p. 137). The integration of technology, especially motivated by CALL research, can thus be argued to enrich TBLT as a pedagogy in the modern era where digital skills and tools are indispensable. Theories and empirical evidence of ISLA can also provide foundations for pedagogical interventions. In sum, technology-mediated TBLT can bring these three domains together, facilitating evidence-based language pedagogy using context-appropriate technologies in the current digital age. To better understand the research domain of technology-mediated TBLT, it is imperative to survey the types of technologies that have been utilized when designing tasks and how these tasks have been implemented. Particular attention should be paid to whether technologies are an essential part of task design or mere add-ons to traditional versions of the target tasks.
Research synthesis on technologies and digital space for L2 learning
To date, research syntheses have explored prevalent themes in CALL (Akiyama & Cunningham, Reference Akiyama and Cunningham2018; Dehghanzadeh et al., Reference Dehghanzadeh, Fardanesh, Hatami, Talaee and Noroozi2021; Dixon et al., Reference Dixon, Dixon and Jordan2022; Di Zou & Xie, Reference Di Zou and Xie2021; Gillespie, Reference Gillespie2020; Golonka et al., Reference Golonka, Bowles, Frank, Richardson and Freynik2014; Zhang & Zou, Reference Zhang and Zou2022). Specifically, some syntheses have reported the methodological features of CALL empirical studies examining a single type of technology, such as synchronous computer-mediated communication (SCMC) tools (Akiyama & Cunningham, Reference Akiyama and Cunningham2018) and digital games (Dehghanzadeh et al., Reference Dehghanzadeh, Fardanesh, Hatami, Talaee and Noroozi2021; Dixon et al., Reference Dixon, Dixon and Jordan2022; Di Zou & Xie, Reference Di Zou and Xie2021). For example, Akiyama and Cunningham (Reference Akiyama and Cunningham2018) reviewed 55 telecollaboration studies that utilized SCMC tools in L2 classrooms. The study investigated features including learner demographics, SCMC types, use of asynchronous computer-mediated communication (CMC) tools, and interaction setups. Similarly, Di Zou and Xie (Reference Di Zou and Xie2021) reviewed studies focusing on digital game-based vocabulary learning with a particular interest in the types of digital games and theoretical frameworks adopted in addition to their main research foci (e.g., learning outcomes, motivation, learner behavior), findings, and main implications. Additionally, Dixon et al. (Reference Dixon, Dixon and Jordan2022) investigated the extent to which digital gaming influenced L2 learning outcomes in 26 empirical studies. Other variables examined in their review included the game developers’ intended purpose of the game, outcome measures, and game design features (e.g., type of player interaction).
Few reviews have examined the diverse range of technologies used in CALL research from a more comprehensive perspective, instead of focusing on the methodological features of empirical studies that use a single type of technology (e.g., SCMC, digital games). For instance, Golonka et al. (Reference Golonka, Bowles, Frank, Richardson and Freynik2014) investigated types of technology and their effectiveness in 350 empirical studies focusing on foreign language learning. The authors categorized technology types as schoolhouse- or classroom-based technologies (e.g., course management system, interactive whiteboard), individual tools (e.g., electronic dictionary, grammar checker, automatic speech recognition), network-based social computing (e.g., virtual worlds, chat platforms, social networking, blog, Wiki), and mobile devices (e.g., tablets, personal computers, cell phones). More recently, Zhang and Zou (Reference Zhang and Zou2022) examined 51 CALL studies for the main types, purposes, and effectiveness of technologies used to enhance second and foreign language learning. The five primary uses of technologies were (a) mobile-assisted language learning, (b) multimedia language learning, (c) socialized language learning, (d) speech-to-text recognition and text-to-speech recognition-assisted language learning, and (e) gamified language learning.
Finally, syntheses have also focused on research topics that have been investigated in CALL research. For example, Gillespie (Reference Gillespie2020) conducted a comprehensive review of CALL research published in three CALL journals: ReCALL, CALICO Journal, and Computer Assisted Language Learning, examining their research topics and methods. His review highlights the relatively small array of topics explored in previous research (e.g., the four language skills, vocabulary, grammar, CMC), leaving cultural content and contexts under-explored. He further noted that most studies were small-scale in terms of study duration, session frequency, sample size, learners’ proficiency levels (i.e., beginner or intermediate levels), and number of institutions. Studies also predominantly examined English as the target language.
There have been several syntheses and reviews on technology-mediated TBLT research (Chong & Reinders, Reference Chong and Reinders2020; González-Lloret, Reference González-Lloret2022; Lai & Li, Reference Lai and Li2011; Ziegler, Reference Ziegler2016). To date, a few research syntheses have focused on how technology deepens our understanding of TBLT features using representative studies. For example, Lai and Li (Reference Lai and Li2011) demonstrated that empirical studies have used diverse technological affordances—such as text-based CMC, digital games, blogging, telecollaboration, and emails—to demonstrate that technology can both enhance L2 learning using tasks and enrich our understanding of TBLT features. The review further reported that TBLT serves as a pedagogical framework for advancing the field of technology-mediated language learning. However, the review also pointed to challenges such as learners’ need to develop technological skills for task completion, greater need for teacher involvement, and difficulties in researching complex constructs (e.g., learner agency, digital literacy). Ziegler (Reference Ziegler2016) also reviewed technology-mediated TBLT studies to understand how technology can support L2 development in task-based settings and how it contributes to our knowledge of TBLT and L2 learning processes. The findings revealed a growing body of studies supporting the positive effects of technology-mediated tasks. Tools such as multiplayer games, virtual worlds, online collaborations, and social networking were identified as not only supporting L2 acquisition but also positively influencing L2 learners’ attitudes toward technology. Ziegler emphasizes a need for more research on emerging multimodal and immersive environments as opposed to further investigations of written text chats.
In a later study, Chong and Reinders (Reference Chong and Reinders2020) adopted a grounded theory approach to synthesize the findings of 16 qualitative task-based studies. The synthesis identified the characteristics, affordances, limitations, and factors impacting the effectiveness of technology-mediated TBLT in naturalistic, classroom-based studies. The results showed that technology-mediated TBLT facilitates collaboration, interaction, and communication, thereby cultivating positive effects toward language learning, facilitating student-centered learning, and developing linguistic and nonlinguistic skills. The study also shared limitations, such as teachers’ difficulties in implementing technological tasks and learners’ concerns regarding lack of explicit grammar instruction and heavy workload.
Syntheses have also investigated the general trends in technologies used in task-based settings over time. For instance, focusing on L2 pragmatic competence, González-Lloret (Reference González-Lloret2022) presented a historical overview of the types of technologies used in empirical task-based studies. She reported that studies involving pragmatics follow the general trend in CALL research of examining text CMC and, more recently, oral CMC. These studies often involve telecollaboration between two remotely located institutions. Moving toward more innovative technologies, studies have used multimodal affordances such as video scenario-based computer simulations, games, synthetic environments, and social networks. González-Lloret (Reference González-Lloret2023) chronicled the history of technology use in language education based on publications in System from 1970 to 2023, focusing on the emergence of technologies. She emphasizes that “technology has been ‘normalized’ in language education research, and its place in academia is now as ubiquitous as it is in our lives” (p. 8).
Despite the increase in research syntheses on the use of technology in language learning, no systematic review paper has yet focused on technology-mediated TBLT as a whole, emphasizing the following features: empirical research design, various task-related focal constructs, and types of technology implemented in task design. Thus, there is a need to review the methodological and substantive features of previous empirical technology-mediated TBLT studies to understand the trend of research foci and to guide future research directions more comprehensively.
The current study
The overarching goal of the current review paper is three-fold: (a) to survey the research contexts, learner demographics, and methods in technology-mediated TBLT studies; (b) to examine the main research constructs that have been investigated in previous technology-mediated TBLT research and the measurements used; and (c) to explore the types of technology or digital spaces utilized in previous research and to evaluate their essentiality in task performance. The research questions guiding this study are as follows:
(1) What are the characteristics of previous technology-mediated TBLT research in terms of research contexts, learner demographics, and research methods?
(2) What research foci have been examined in previous technology-mediated TBLT research? What measurements have been used to examine these focal constructs?
(3) What types of technology or digital spaces have been used in previous technology-mediated TBLT studies? How essential was technology to task performance in these studies?
Methodology
Inclusion criteria and search techniques
The first stage of data collection involved a comprehensive search using three databases: Linguistics and Language Behavior Abstracts, Education Resources Information Center, and Google Scholar. There are various ways to implement tasks in syllabus design, and different terminologies have been introduced to describe how tasks are used in curriculum design. For instance, tasks can be the main unit of instruction in TBLT or a supplementary affordance in task-supported language teaching (task-based vs. task-supported; Samuda & Bygate, Reference Samuda and Bygate2008). Notably, the term “task-based” has been “loosely applied as an umbrella term to refer to any context in which tasks are used” (Samuda & Bygate, Reference Samuda and Bygate2008, p. 57). As the purpose of this study is not to examine research at the curricular level, we adopted this looser conceptualization of “task-based.” In other words, we used “task-based” to refer to any instructional or research context in which tasks were used for language learning purposes. Using the advanced search function, we searched for the combinations of the keywords “task-based” OR “TBLT” AND “technology.” We conducted another search using “technology-mediated TBLT.” This study only included peer-reviewed journal articles and book chapters. In addition to the keywords listed above, we screened studies using the following criteria:
(1) published between 2000 and 2022;
(2) referred to their instructional material as a “task” throughout the paper;
(3) utilized technology, such as computers, mobile phones, or tablets (studies utilizing analog tools, such as audio tapes or CD-ROMs were excluded from the dataset);
(4) reported empirical data related to technology-mediated tasks (studies that focused on the evaluation of a task-based language program, teacher training, L2 assessment or assessment tools, and the foundation of a task-based curriculum starting from a needs analysis were excluded);
(5) conducted in L2 contexts; and
(6) written in English.
Studies that utilized the same data but investigated different research questions or goals were included separately (e.g., Abe & Roever, Reference Abe and Roever2019, Reference Abe and Roever2020). Moreover, corpus-based studies that referred to their instructional materials for data elicitation as “tasks” were included (e.g., Black & Barron, Reference Black and Barron2018).
The second stage of data collection involved a manual search of studies in four CALL-specific journals (CALICO Journal, Computer Assisted Language Learning, Language Learning & Technology, ReCALL) and two edited books (Task-based Language Learning and Teaching with Technology edited by Thomas & Reinders (Reference Thomas and Reinders2010) and Technology-mediated TBLT: Researching Technology and Tasks edited by González-Lloret & Ortega (Reference González-Lloret and Ortega2014a)). A total of 254 empirical studies were included in the dataset (the full list of studies is provided in IRIS [instruments and data for research in language studies]). Figure 1 illustrates the increase in the number of studies from 2000 to 2022.
Studies of technology-mediated TBLT from 2000 to 2022.
Data coding
A coding scheme was developed to extract information relevant to the research questions of this study (see Table 1). To evaluate the essentiality of technology in performing tasks, we developed a coding scheme to classify each study based on the extent to which technology was essential for learners to complete the task. The purpose of evaluating the essentiality of technology was to investigate whether research has been effectively integrating new technologies with language tasks in an organic and mutually informative way, considering the reciprocal relationship between TBLT and technology (González-Lloret & Ortega, Reference González-Lloret, Ortega, González-Lloret and Ortega2014b). Table 2 presents the definitions of each value for technology essentiality (i.e., technology-optional, technology-facilitated, technology-essential).
Coding scheme
Note: All items were coded numerically (e.g., foreign language = 1, second language = 2, not specified = 3) unless marked with * to indicate items with an open-ended response. Also, if the author(s) provided the Common European Framework of Reference for Languages (CEFR) level, proficiency was categorized accordingly based on the CEFR level description.
Criteria for essentiality of technology
To ensure the reliability of coding, two raters coded selected features in the coding scheme. The first rater was the second author of this study, and the second rater was a Ph.D. student in Applied Linguistics. Both raters received formal training in TBLT through graduate-level courses and are familiar with the TBLT literature. The raters evaluated 32% of the empirical studies included in the dataset (n = 82) independently on the following features: research methodologies, statistical analyses, essentiality of technology (see Table 2), research foci, measurements (learner perception, learning outcome), and target interactional and linguistic features. The exact agreement for coding these features reached 90.03%. Discrepancies were negotiated through multiple discussion sessions with the first author, and all the remaining data, including other discrete items, such as sample size, were coded by the first rater.
Data analysis
The 254 primary studies were coded in Excel for an array of methodological and substantive features. The frequencies of each feature were then counted, and the percentages relevant to each research question were calculated. The percentages were rounded to the nearest whole number.
Results
To answer the first research question, we surveyed the studies focusing on their language learning context, educational setting, research setting, target language, and learners’ target language proficiency. As demonstrated in Table 3, previous technology-mediated TBLT research has noticeably favored foreign language contexts (80%), with university students (82%) learning English as the target language (66%) in classroom contexts (65%). Regarding learner proficiency level, intermediate-level learners (60%) have been most investigated, though 16% of the studies did not report learners’ proficiency.
Research contexts and learner demographics
Note: The percentages were calculated by dividing k by 254. The percentages for educational setting, target language, and proficiency do not add up to 100%, as some studies investigated more than one item.
* Other educational settings included online recruitment (k = 5), refugees (k = 1), and adult learners living in the UK (k = 1).
** Other target languages included Arabic, Croatian, Irish, and Māori (k = 1 for each target language).
Regarding the characteristics of their research design, technology-mediated TBLT studies have involved approximately 45 participants on average, though a large variance in sample size (SD = 82.74) was observed, ranging from 2 to 1,150 participants. Furthermore, previous research has used 3.68 tasks on average (SD = 4.93), ranging from 1 to 48 tasks. However, 43 studies (17%) out of the 254 studies did not state the number of tasks used in their research.
With respect to research methodology, Table 4 shows that technology-mediated TBLT research has favored a mixed-methods approach (63%). An examination of the statistical tests used in the studies adopting a quantitative method or a mixed-methods approach revealed that virtually all of them utilized descriptive statistics (99%) to report their findings. Furthermore, such research has used statistical tests, including t-tests (25%), ANOVA (23%), and nonparametric tests (11%), to analyze data. Of the quantitative-only and mixed-methods studies that used inferential statistics (k = 135), less than one-fifth of the studies (16%, k = 22) incorporated post hoc tests, and less than half (46%, k = 62) reported effect sizes.
Research methodologies and statistical analyses
Note: We considered descriptive statistics as a quantitative method. Thus, a qualitative study that included frequency/percentages was counted as a mixed-methods study. The percentages of statistical analyses were calculated by dividing k by the number of studies adopting a quantitative method or a mixed-methods approach (N = 221). The total percentage of statistical analyses does not add up to 100% as studies often used more than one statistical test.
* Other statistical tests included MANCOVA (k = 1) and factor analysis (k = 2).
The second research question pertained to the research foci of previous technology-mediated TBLT research (see Table 5). Over half of the studies investigated learner perception (56%), whereas less than one-third of the studies explored learning outcomes (30%). Additionally, out of the 254 studies, 38 (15%) investigated learner perspectives as dependent variables.
Research foci
Note: The percentages for each descriptor were calculated by dividing k by 254. The total percentage of the studies does not add up to 100%, as some had multiple research foci. The percentages of task performance components were calculated by dividing k by 187, and the percentages of learner perspective components were calculated by dividing k by 38.
When analyzing learners’ task performance, interactional features were extensively investigated (67%) in this research domain. As shown in Table 6, a closer look at these interactional features revealed that feedback (31%) received the greatest attention, followed by negotiation of meaning (25%) and language-related talk (24%) commonly operationalized as language-related episodes (LREs).
Interactional features examined
Note: The percentages were calculated by dividing k by 126. The total percentage does not add up to 100%, as some studies investigated multiple interactional features.
Furthermore, we analyzed the language features that were examined in previous research (see Table 7). Vocabulary (42%) and grammar (38%) were most investigated. In addition, CAF measures were predominantly examined (51%). In particular, complexity (25%; written: k = 20; oral: k = 7) and accuracy (17%; written: k = 18; oral: k = 7) were examined more often than fluency (9%; written: k = 10; oral: k = 5).
Language features examined
Note: The percentages were calculated by dividing k by 187, as 67 studies out of 254 did not examine language features but investigated only learner perception, learner perspective, and/or interactional features with no linguistic foci (e.g., gestures). Also, in our analysis, if a study investigated LREs on grammar or vocabulary, we coded the target linguistic features as grammar and vocabulary, respectively. The total percentage of the studies does not add up to 100%, as some investigated multiple language features.
Among the research foci identified, learner perception and learning outcomes were observed with various operationalizations. Thus, we delved deeply into the measurements used to operationalize each construct (see Table 8). Studies examining learner perception predominantly used surveys/questionnaires (71%) followed by interviews (48%). Notably, we found that less than half of the studies (k = 61, 43%) triangulated the data to investigate learner perception by utilizing multiple measurements. Furthermore, studies investigating learning outcomes mostly used receptive tests (35%), oral productive tests (30%), and written productive tests (30%). We also observed that 30 out of the 77 studies (i.e., 39%) examining learning outcomes adopted multiple tests to examine different knowledge types.
Measurements of learner perception and learning outcome
Note: The percentages for learner perception and learning outcome measurements were calculated by dividing k by the number of studies that investigated learner perception (N = 143) and learning outcomes (N = 77), respectively. The total percentage of the studies does not add up to 100%, as some used multiple measurements.
Finally, the third research question addressed different types of technology or digital spaces that have been examined in previous research. As shown in Table 9, studies have mostly focused on text-based SCMC (34%), video-based SCMC (20%), and Web 2.0 tools (16%), followed by asynchronous CMC (15%). We also observed that 14% (k = 35) of the 254 studies focused on tandem learning and telecollaboration. In these studies, learners performed tasks with native or more proficient speakers of the target language using digital technologies. Upon closer examination of the 35 studies out of the 254 that focused on tandem learning and telecollaboration, it was found that 13 of them utilized video-based SCMC (37%), 10 utilized text-based SCMC (29%), 7 utilized asynchronous CMC (20%), and only 1 utilized audio-based SCMC (3%). A few studies also used social media tools, such as Facebook (k = 3, 9%), and Web 2.0 tools, such as Blogger (k = 1, 3%), to investigate tandem learning and telecollaboration. Additionally, previous studies have used diverse platforms when incorporating technologies with tasks. For example, regarding SCMC studies, recent research has commonly used software such as Skype, Moodle, Zoom, Facebook Messenger, and WebEx, while the earlier studies frequently used platforms such as ChatNet, WebCT, mIRC, and iChat. As for Web 2.0 tools, studies have used platforms such as Google Docs, Facebook, and Wikis.
Types of technology or digital space
Note: Two studies were excluded from the coding of the type of technology, as they did not provide sufficient information about what technology or digital space was used. Thus, the percentages were calculated by dividing k by 252. Also, the percentages do not add up to 100%, as some studies focused on multiple types of technology.
* Other types included an online dictionary (k = 1) and a concordancer (k = 1).
Using open-source websites available on the Internet, some of the more recent studies developed their own tasks in creative ways instead of utilizing those already developed in earlier research to facilitate negotiated interaction (e.g., jigsaw, spot-the-difference, decision-making). For example, Timpe-Laughlin and Dombi (Reference Timpe-Laughlin and Dombi2020) developed a fully automated, interactive oral task to examine L2 learners’ request-making strategies using HALEF (Help Assistant–Language-Enabled and Free; http://halef.org), an open-source, web-based framework for designing the spoken dialogue system tasks (Ramanarayanan et al., Reference Ramanarayanan, Suendermann-Oeft, Lange, Mundkowsky, Ivanov, Yu, Evanini and Dahl2017). The tasks required learners to call a fictitious supervisor and to make two requests (e.g., schedule a meeting, ask for a review of documents). Another example of a creatively constructed task is from Cornillie et al. (Reference Cornillie, Buendgens-Kosten, Sauro and Van der Veken2021), which utilized Twine (http://twinery.org), an open-source tool for telling interactive stories. Twine was used to examine L2 learners’ writing of interactive fanfiction based on a digital game series. Additionally, researchers have also developed their own software, platforms, or games using programming skills. For example, Taguchi et al. (Reference Taguchi, Dixon, Qin and Chen2022) developed a game using Python to examine L2 learners’ acquisition of request-making forms.
The third research question further addressed whether the technologies utilized in the studies were chosen and incorporated into task design in recognition of the mutual relationship between TBLT and technology (González-Lloret & Ortega, Reference González-Lloret, Ortega, González-Lloret and Ortega2014b). Table 10 shows that most studies used technology as an essential part of task performance (68%). Also, Figure 2 presents how technology-essentiality has changed from 2000 to 2022. The findings show that technology-essential studies have sharply increased since 2007, while technology-facilitated studies demonstrated a smaller increase over time.
Trend of technology-essentiality from 2000 to 2022.
Evaluation of the essentiality of technology
Note: Two studies were excluded from the dataset when coding the essentiality of technology, as they did not provide sufficient information about their technology. Thus, the percentages were calculated by dividing k by 252. The percentages add up to over 100%, as the percentages were rounded to the nearest whole number.
Discussion
The overarching goal of the current study was to understand the methodological and substantive trends of technology-mediated TBLT research published between 2000 and 2022. Some of the findings are encouraging, but others are cause for concern in that they highlight neglected areas of research. In this section, we discuss the findings and provide suggestions for future research.
In terms of the first research question, the findings revealed that research contexts and learner backgrounds are skewed, which is in line with previous systematic review papers in applied linguistics. For instance, Plonsky’s (Reference Plonsky2023) synthetic analysis of 308 applied linguistics research articles found that university students (39%), intermediate level learners (31.65%), and adult learners (41.94%) are the most widely researched groups in terms of sampling-related features. The recent growing interest in research syntheses has allowed us to notice the trends of current sampling practices in addition to researchers’ shared concerns regarding their implications at the ethical, theoretical, and practical levels (Andringa & Godfroid, Reference Andringa and Godfroid2020; Ortega, Reference Ortega2005; Plonsky, Reference Plonsky2023). Our findings were not an exception, and we agree that this is the result of accessibility and convenience sampling (Plonsky, Reference Plonsky2023). With the explicit efforts for multilingual turns in SLA coupled with diversity and inclusion initiatives in educational research, we hope that future research will involve more diverse populations and instructional contexts in this research domain.
Regarding the treatment of research methods, we used a rather lenient standard for classifying research as mixed-method (i.e., reported both quantitative and qualitative data), although they might not have followed protocols of mixed-methods (Creswell & Clark, Reference Creswell and Clark2017). This resulted in a notable presence of mixed-method studies that were mainly qualitative in nature but incorporated descriptive statistics to report frequencies of target observations (k = 85; e.g., Ziegler & Phung, Reference Ziegler and Phung2019). For the types of statistical tests used in these studies, the results showed a wider variety of statistical tests used in the technology-mediated TBLT literature compared to previous CALL or TBLT synthesis results (e.g., Plonsky & Kim, Reference Plonsky and Kim2016). However, similar to Plonsky and Kim’s (Reference Plonsky and Kim2016) report, most studies used tests of mean differences between or among groups, such as ANOVA, t-tests, and nonparametric tests. Since the importance of reporting effect sizes in L2 research has been increasingly addressed, many journals now require effect sizes (e.g., Language Learning, Studies in Second Language Acquisition). Interestingly, less than half of the studies (46%) reviewed reported effect sizes, despite this becoming a norm in quantitative research. We would like to note that this relatively low percentage can be attributed to the wide range of time covered by our dataset, spanning from 2000 to 2022, during which the reporting standards may have continued to evolve.
With respect to the second research question, the findings of our review showed that studies have extensively investigated learner perception. Given the relatively innovative nature of learning a language via technology in contrast to traditional paper-based approaches, we believe that researchers have mainly focused on examining learners’ acceptance of technology and their perceptions regarding its effectiveness. Also, a few studies in our dataset examined learners’ learning outcomes through perception data (perceived learning outcomes; e.g., Batardière, Reference Batardière, Fowley, English and Thouësny2013). As new technologies may result in positive “novelty” effects (i.e., a more positive perception due to an instrument’s newness or uniqueness), the interpretation of learner perception data, especially in a short-term project, necessitates caution. Furthermore, more longitudinal projects are warranted to understand learner perception dynamics toward technology-mediated TBLT over time (e.g., Kim et al., Reference Kim, Jung and Tracy-Ventura2017).
Although the definition of “task” involves a clear, tangible outcome, surprisingly, our analysis revealed that only 18% of the studies focusing on task performance examined the quality of task outcomes systematically by utilizing a rubric or qualitatively describing the task outcomes. As shown in Table 7, the linguistic performance of task outcomes has been widely examined, partly due to the interests associated with CAF measures. Although such linguistic measures could provide insights on potential language development through task performance, the holistic quality of task output, in addition to language quality, needs further attention, especially from task-based assessment perspectives. Furthermore, despite the integration of tasks with technology, there was limited investigation into the actual use of technology during task performance (11%), such as scrolling, clicking, and using online resources. To better understand how technology can be effectively integrated into language-learning tasks, future research is warranted regarding learners’ use of technology. Also, while the field of SLA has emphasized the significance of individual factors (Li et al., Reference Li, Hiver, Papi, S., Hiver and Papi2022), there has been relatively less focus on analyzing learner perspective data in technology-mediated TBLT research (15%) compared to analyses of learner perception (56%), learning outcome (30%), and task performance (74%). Out of the various learner perspective studies, however, we were able to observe a recent increase in studies investigating motivation (Canals, Reference Canals2020) and engagement (Dao et al., Reference Dao, Nguyen, Duong and Tran-Thanh2021).
In terms of the linguistic features that were investigated, the trends in technology-mediated TBLT research included in this synthesis aligned with those found in Plonsky and Kim (Reference Plonsky and Kim2016). The more traditional aspects of language, such as vocabulary, grammar, and CAF measures, received primary attention compared to other features, such as pragmatics/interactional competence and pronunciation/phonology. In terms of the CAF model, written output was examined more often than spoken task performance, particularly focusing on accuracy and on complexity at the syntactic and lexical levels. Thus, we would like to highlight the need to expand the research domain by including other dimensions of task performance, such as functional adequacy (i.e., how successful a learner’s task performance is in achieving task goal efficiently). Furthermore, previous research has tended to focus more on productive skills rather than receptive skills, overlooking input-based tasks. As highlighted in Gillespie’s (Reference Gillespie2020) synthesis of CALL research, technology-mediated TBLT research would also benefit from expanding the scope of research topics and focusing on higher level critical thinking skills by moving beyond the examination of linguistic performance during tasks. Regarding interactional features, a wider array was investigated than those found in Plonsky and Kim (Reference Plonsky and Kim2016), probably due to the ever-expanding technological affordances accessible in technologies (e.g., nonverbal communication during videoconferencing, alignment during text-chats) available since that synthesis was conducted.
The last goal of the current review was to delve into the types of technologies used in technology-mediated TBLT research and their implementation. The findings revealed that SCMC (either text-based or video-based) and Web 2.0 tools have been most widely used in technology-mediated TBLT research. This trend has been observable over the last two decades in SLA’s research agenda. For instance, there has been a surge in research comparing face-to-face and SCMC interactions in terms of learner noticing (Gurzynski-Weiss & Baralt, Reference Gurzynski-Weiss and Baralt2015), feedback (Rassaei, Reference Rassaei2017), and learner engagement (Baralt et al., Reference Baralt, Gurzynski-Weiss, Kim, Sato and Ballinger2016). Additionally, a growing interest in technology-mediated collaborative writing tasks is reflected in the high percentage of research articles using Web 2.0 tools (Abrams, Reference Abrams2019). Moreover, multimodal approaches to conceptualizing and teaching L2 writing are on the rise, and the use of digital multimodal composing tasks has been increasingly studied in the field of TBLT (Kim et al., Reference Kim, Kang, Nam and Skalicky2022). Furthermore, studies have increasingly investigated L2 learning via simulations in virtual worlds, using tools such as Second Life (Chen & Kent, Reference Chen and Kent2020) and Mondly VR (Tai, Reference Tai2022). Overall, it seems clear that novel technologies were introduced not only due to the diversifying functions of technologies but also to tap into different theoretical motivations of research associated with such technologies.
Moving forward, future technology-mediated TBLT research should diversify the types of technologies investigated, as CMC tools have been a primary focus (approximately 70%). It is important to note that our dataset did not include studies published in 2023, a year that has seen a marked increase in interest related to the use of AI technologies. Thus, we expect more research using AI-assisted tools in the near future given the recent advances in AI technologies and their applications to language education (Godwin-Jones, Reference Godwin-Jones2023). Furthermore, we would like to highlight the significance of collaboration between TBLT researchers and instructional technology and CALL experts, as oftentimes the observed lack of up-to-date technology use in TBLT could be due to the boundaries between different disciplines. For instance, immersive VR environments and metaverse platforms have been increasingly implemented in instructional designs in the field of instructional technology (e.g., Lee et al., Reference Lee, Yang and Wu2023). Such technology-mediated language learning platforms offer ideal settings for designing authentic tasks because they can create immersive learning experiences that closely resemble real-world situations for L2 learners, particularly in foreign language learning contexts. Instructional technology and CALL researchers could create more theoretically and pedagogically sound technology-mediated tasks through the adoption of a TBLT framework. Conversely, through collaboration with scholars from instructional technology and CALL, TBLT researchers can gain valuable insights on more diverse and innovative instructional technologies. By overcoming disciplinary boundaries, scholars in CALL, ISLA, and TBLT can engage in reciprocally beneficial relationships.
González-Lloret and Ortega (Reference González-Lloret, Ortega, González-Lloret and Ortega2014b) highlighted that technology-mediated TBLT introduces a new understanding of technology and task integration in that technology should be “yoked with real tasks rather than being chosen as mere translations or extensions of exercises and activities of various kinds into computer platforms” (p. 5). As the role of technology is critical in technology-mediated TBLT, we examined how different technologies were used in task design and implementation and proposed three categories: technology-optional, technology-facilitated, and technology-essential. We find it encouraging that the technologies represented in this synthesis were used organically as a part of tasks and facilitated authentic language use and interactions. Technology-essential tasks included writing academic papers with peer feedback on Moodle (Payant & Zuniga, Reference Payant and Zuniga2022), collaborative writing with synchronous teacher feedback on Google Docs, (Shintani, Reference Shintani2016), and real-world simulated interactive tasks using avatars (Chen & Kent, Reference Chen and Kent2020).
Limitations of the synthesis
One of the main limitations of the present study is that it did not investigate the extent to which the tasks in the reviewed studies adhere to the definition of tasks from TBLT perspectives (González-Lloret & Ortega, Reference González-Lloret, Ortega, González-Lloret and Ortega2014b). Although technology might have been organically integrated into task design in previous research, as shown in this review, there needs to be a standardization of the operationalization of tasks in the technology-mediated TBLT literature so that there is a consensus on what tasks are across the three research domains (TBLT, instructed SLA, CALL). Also, new technologies have transformed what counts as “real-world” within the traditional TBLT perspective (Ortega & González-Lloret, Reference Ortega, González-Lloret and Bygate2015), and this requires us to rethink what constitutes a task in the digital age. Thus, the next step to take within this line of research is to investigate whether the tasks in technology-mediated TBLT studies meet the criteria of a task, and particularly what types of technology-mediated task features promote learning opportunities and subsequent learning. Furthermore, the current review only incorporated studies published in English, which is a common constraint in synthetic research across the field of applied linguistics. As a result, the findings in this paper should not be considered a comprehensive representation of all technology-mediated TBLT studies, as there are many studies written in other languages that have examined the incorporation of tasks with technology.
Conclusion
This study systematically reviewed 254 technology-mediated TBLT studies published between 2000 and 2022. We observed a large number of studies reporting both quantitative and qualitative data. Additionally, a great number of the studies were technology-essential, meaning they organically integrated tasks with technology. Previous research has also investigated a wide variety of interactional features during task performance due to the availability of diverse technological affordances. A concern, however, is that studies in this area have primarily focused on a limited range of educational contexts, target languages, and proficiency levels. Similarly, little attention has been given to assessing the quality of task outcomes and learner perspectives compared to learner perceptions and learning outcomes. Furthermore, the target language features investigated are also rather limited, overlooking higher-level language skills and cultural aspects. Finally, the types of technology were greatly skewed toward CMC, and we call for more studies on different types of technology to be conducted so as to better represent the diversification of authentic target language domains in the digital age. Although efforts were made to include many studies that integrate technology and tasks, we would like to acknowledge that the dataset included in this study should not be considered an exhaustive list of technology-mediated TBLT studies. Despite the limitations, our findings demonstrate the need for increased dialogues and research collaborations among CALL, ISLA, and TBLT researchers. Such reciprocity could produce theoretically and pedagogically sound research with carefully designed technology-essential tasks, which in turn could provide valuable insights that inform language pedagogy in the current digital era.
Acknowledgements
We are grateful for Dr. Alison Mackey and anonymous reviewers who provided helpful comments during the revision process. We would like to thank Nicole C. De Los Reyes who helped with data coding. The earlier version of this paper was presented at AAAL 2024 in Houston. Any remaining errors are our own.
Open access
