1. Introduction
Technology has long played a role as the central medium for L2 listening, teaching and learning (see Jones Reference Jones2008 for an historical overview). However, it has been the advent of digital technology primarily in the shape of multimedia CALL that has opened up new avenues and opportunities for L2 listening pedagogy and associated research. The major focus of this research has thus far been on investigating language learning through computer-based listening facilitated by help options. Cárdenas-Claros & Gruba (Reference Cárdenas-Claros and Gruba2009: 69) define help options as ‘embedded application resources that assist learners in performing computing operations and/or support language learning’. This kind of learner assistance may be presented in written/verbal/visual modes in the guise of pictures and videos, vocabulary annotations, captions, subtitles, transcripts, translations, control functions (e.g. pause and rewind), listening advice and feedback, dictionaries, or cultural notes.
1.1 Associated theory
The principal theoretical perspective underlying the use of help options in computer-based L2 learning is an interactionist one (Cárdenas-Claros & Gruba Reference Cárdenas-Claros and Gruba2009), and a number of CALL researchers exploring how help options for L2 listening can assist learning have explicitly positioned their studies in accordance with interactionist views (e.g., Hegelheimer & Tower Reference Hegelheimer and Tower2004; Grgurović & Hegelheimer Reference Grgurović and Hegelheimer2007; Cárdenas-Claros & Gruba Reference Cárdenas-Claros and Gruba2014). In brief, interactionist theory in CALL (Chapelle Reference Chapelle1998, Reference Chapelle and Kaplan2002, Reference Chapelle, Egbert and Petrie2005a) extends the notion of interaction as the impetus for language acquisition beyond interpersonal activity and intrapersonal activity (Ellis Reference Ellis1999) to include learner-computer activity, with all three types of activity possible in CALL tasks. The crucial point about such activities is that the process of interaction draws learners’ attention to language as they engage in meaning-making (Chapelle Reference Chapelle and Kaplan2002, Reference Chapelle, Egbert and Petrie2005a). This process is hypothesized to benefit language acquisition through negotiation of meaning, obtaining enhanced input, and mental processing of linguistic input (Chapelle Reference Chapelle2003, Reference Chapelle, Egbert and Petrie2005a). Help options are one way in which learners can access input enhancement in learner-computer activity. Regarding help options for L2 listening, input enhancement involves making adjustments to the language data to increase listening comprehension. This may be accomplished by salience (e.g., highlighting words in transcripts, repeated listening using control functions), modification of input (e.g., using images, written/visual annotations), and elaboration (e.g., cultural notes) (see Chapelle Reference Chapelle2003 and Cárdenas-Claros & Gruba Reference Cárdenas-Claros and Gruba2009 for details). In addition, to better theorise multimodal input-based L2 learning, Plass & Jones (Reference Plass, Jones and Mayer2005: 471) introduced a model for multimedia which brings interactionist views on L2 learning together with generative theory of multimedia learning (Mayer Reference Mayer1997, Reference Mayer2001, Reference Mayer2002) and dual-coding theory (Paivio Reference Paivio1971, Reference Paivio1986), and aspects of this model have informed understanding of help-option use in L2 listening CALL environments in several studies (e.g., see Jones Reference Jones2006).
CALL researchers have also drawn on a sociocultural theory (SCT) perspective. Again in brief, based on the work of Vygotsky (Reference Vygotsky, Cole, John-Steiner, Scribner and Souberman1978, Reference Vygotsky and Wertsch1981, Reference Vygotsky1996), SCT informs CALL according to three aspects – mediation, social learning, and genetic analysis (Warschauer Reference Warschauer, Egbert and Petrie2005). Mediation refers to the notion that external, culturally constructed physical (e.g., computers) and symbolic (e.g., language) tools, activities and concepts empower humans to self-regulate higher forms of thinking (e.g., problem-solving, learning) (Lantolf Reference Lantolf2006, Reference Lantolf, Gass and Mackey2012). Social learning reflects the concept that, through the mechanism of mediation, these higher mental abilities first emerge in socially situated activity and may subsequently be internalised by participating individuals (Lantolf & Thorne Reference Lantolf and Thorne2006). Genetic analysis encapsulates a methodology that aims to reveal over time the genesis and cause of thinking processes, and is explanatory rather than descriptive in nature (Lantolf & Thorne Reference Lantolf and Thorne2006; Lantolf & Poehner Reference Lantolf and Poehner2014). SCT-oriented research may investigate learners’ mental processes by examining interpersonal activity (e.g., computer-mediated communication), intrapersonal activity (e.g., private speech), and/or learner-computer activity (e.g., help-option use). Related research has primarily applied SCT in relation to computer-mediated communication (Warschauer Reference Warschauer, Egbert and Petrie2005), with only a few investigations considering it in terms of collaboration between learners interacting at, and with, a computer, for example, in a rare SCT-informed study of help options for L2 listening by Jones (Reference Jones2006) (see also Gánem-Gutiérrez Reference Gánem-Gutiérrez2006; McDonough & Sunitham Reference McDonough and Sunitham2009). Hoven (Reference Hoven1999) has also outlined a CALL-relevant instructional design model based on SCT principles, though this model has rarely been referred to in help-option research.
1.2 Empirical research
Turning to empirical research, from a broad review of the associated literature for this article, it is evident that after an initial surge in interest among listening researchers in the fairly early days of CALL with respect to examining help options, fascination waned from around 2007 and we have seen only a smattering of related studies published since then, apart from research into captions, which has experienced ongoing interest (e.g., Winke, Gass & Sydorenko Reference Winke, Gass and Sydorenko2010; Hsu Reference Hsu2013; Montero Perez et al. Reference Montero Perez, Peters, Clarebout and Desmet2014). Also as a result of this review, a number of general research design issues surfaced. These are important to highlight for they inform aspects of the approach to the various tasks I propose for generating further research, about which more robust claims for language learning and listening pedagogy can be made. Firstly, there are currently very few longitudinal studies apparent in published material. Language learning in CALL contexts requires exploring at regular intervals over time with sufficiently dense data collected for analysis rather than the short-term ‘snapshot’ approach which has often been used (Blin et al. Reference Blin, Caws, Hamel, Heift, Schulze, Smith and Barr2013). Also, few studies have employed a mixed-methods design, an aspect which is important for investigating L2 listening (Cross & Vandergrift Reference Cross and Vandergrift2015) in conjunction with the increasingly complex and dynamic medium of computer-based learning from multiple perspectives (Cárdenas-Claros & Gruba Reference Cárdenas-Claros and Gruba2009; Blin et al. Reference Blin, Caws, Hamel, Heift, Schulze, Smith and Barr2013). Another issue is that learning outcomes have not often been examined using a range of process-oriented and product-oriented measures, which is crucial for advancing arguments about learners’ interaction with technology and its relationship to performance both within, and outside of, a CALL task (Chapelle Reference Chapelle and Kaplan2002, Reference Chapelle2004, Reference Chapelle and Chapelle2012). Moreover, Cárdenas-Claros & Gruba (Reference Cárdenas-Claros and Gruba2009) identify three other factors which are points of concern regarding research design in previous help-option studies. Firstly, low participant numbers has been problematic. For example, the majority of help-option interaction-focused studies had between 10 and 20 participants, which restricted data interpretations and made it difficult for researchers to establish reliable patterns of use. Also, there is often a lack of diversity in data collection instruments, meaning triangulation of data is impeded. Thirdly, researchers have imposed restrictive time constraints on learners’ help-option interactions. This can reinforce the task-completion focus of learners (Chapelle Reference Chapelle2005b), which may limit their efforts to access and exploit help options fully for learning and lead to unreliable data being collected. A further issue is that there is still no concrete L2 listening pedagogy drawn from empirical research which teachers can refer to for exploiting help options or training learners in their use, which may be one reason why listening in CALL contexts has yet to become fully accepted and exploited in mainstream listening pedagogy (Hubbard Reference Hubbard and Chapelle2012). Indeed, very few help-option studies draw on, are underpinned by, or show links to established L2 listening theory and practice in literature reviews or discussions of findings.
In sum, given the rather dated and stagnant nature of associated empirical research and the research design problems identified, it seems timely to propose an agenda for investigating the effective use of help options during listening in CALL environments, which may promulgate renewed/further interest in this area. The approach to the tasks I suggest also seeks to address many of the concerns raised.
2. Research agenda
In my review of the literature concerning help options in CALL listening-related studies, I ascertained three main research themes: learner perceptions and experiences; comparisons of different conditions; and learner variables. These themes serve as a useful means for structuring the research agenda I am proposing. In addition, several studies have pointed to the importance of providing instruction to learners to maximise help-option benefits for comprehension. Thus, learner training also seems to be a fruitful line of enquiry and is a fourth theme considered. For each theme, I will briefly review relevant research in order to identify, and then outline the approach to, associated research tasks that may be undertaken to examine and further our understanding of L2 listening comprehension mediated by help options in CALL, as well as stimulate the move towards establishing an associated pedagogy which is more integrated with mainstream L2 listening pedagogy.
2.1 Learner perceptions and experiences
Several studies (e.g., Liou Reference Liou1997; Pujolà Reference Pujolà2002; Jones Reference Jones2003; Grgurović & Hegelheimer Reference Grgurović and Hegelheimer2007; Cárdenas-Claros & Gruba Reference Cárdenas-Claros and Gruba2014) have reported on learners’ attitudes to various help options and their associated interactions with this kind of learning support. Broadly speaking, these studies established that there was great variability in listeners’ preference for, and use of, different help options. For various reasons, e.g., previous experience or easing task completion, learners either rarely relied on help options, used some but not others, or were compulsive users. Of the studies indicated, I will focus in more detail on the Cárdenas-Claros & Gruba article here due to its use of a range of qualitative data collection methods immediate to the listening and help-option experience.
Cárdenas-Claros & Gruba (Reference Cárdenas-Claros and Gruba2014) were interested in understanding more about why listeners often do not access or interact with available CALL help options to facilitate learning or use them ineffectively (e.g., see Hegelheimer & Tower Reference Hegelheimer and Tower2004; Grgurović & Hegelheimer Reference Grgurović and Hegelheimer2007). The authors investigated fifteen Colombian English as a second language (four) and English as a foreign language (eleven) learners’ thoughts on, and use of, nine different help options according to five themes that appeared in their analysis of data collected through verbal reports (stimulated recalls), direct observation and semi-structured interviews over three one-hour sessions. In brief, Cárdenas-Claros & Gruba (Reference Cárdenas-Claros and Gruba2014) found that learners’ were positive about help options relevant to task management and completion, and the use of help options was broadly influenced by learners’ views about language learning, comprehension failures, familiarity with help-option use, and ease of help option interaction. The authors acknowledged that their study was exploratory in nature, with few participants, each of whom exhibited idiosyncratic behaviour regarding help-option usage/non-usage that made it problematic to establish particular patterns of use. As such, further studies would be useful in order to examine help-option perceptions and usage with respect to larger and different learner groups, levels of proficiency, listening texts and tasks and/or target languages over longer time periods (Grgurović & Hegelheimer Reference Grgurović and Hegelheimer2007; Cárdenas-Claros & Gruba Reference Cárdenas-Claros and Gruba2014).
Research task 1
Extend Cárdenas-Claros & Gruba's ( Reference Cárdenas-Claros and Gruba 2014 ) study to investigate listeners’ attitudes to, interactions with, and related learning outcomes for various help options using a mixed-methods approach.
The scope of Cárdenas-Claros & Gruba (Reference Cárdenas-Claros and Gruba2014) could be extended to exploit the quantitative and qualitative data from screen capture software and eye tracking systems, respectively, which can be employed to record learners’ interactions with a computer. CALL L2 listening studies (e.g., Pujolà Reference Pujolà2002; Grgurović & Hegelheimer Reference Grgurović and Hegelheimer2007; Rivens Mompean & Guichon Reference Rivens Mompean and Guichon2009) have usefully exploited screen capture software to gather quantitative data on help-option interactions. Cárdenas-Claros & Gruba (Reference Cárdenas-Claros and Gruba2014) also used screen capture software to provide prompts for stimulated recalls only, but decided not to exploit this numerical data on participant's help-option usage in their analysis due to concerns regarding its utility in providing insights into learners’ cognitive processes (in contrast to the other three data collection methods they used). Eye tracking systems may overcome this issue since this technology can be used to make moment-by-moment inferences about cognitive processing and corresponding levels of complexity through monitoring eye movements (Blin et al. Reference Blin, Caws, Hamel, Heift, Schulze, Smith and Barr2013). Quantitative and qualitative data captured in real time and less intrusively (compared to introspective methods such as verbal reports) by screen capture software and eye tracking systems would be of value alongside qualitative data collected with the other instruments Cárdenas-Claros & Gruba (Reference Cárdenas-Claros and Gruba2014) originally employed. Such technology no longer requires a high level of technical expertise to use, with systems now enabling easy data storage in a variety of formats and supported by automated data analysis and presentation of results (e.g., see Winke, Gass & Sydorenko Reference Winke, Gass and Sydorenko2013). Furthermore, it would also be useful to establish the relationship of help-option preferences and interactions to listeners’ language acquisition and text comprehension (e.g., as in Liou Reference Liou1997 and Hegelheimer & Tower Reference Hegelheimer and Tower2004) across such a study either through qualitative and/or quantitative measures. Reflecting on the issue that participants were rarely using some help options to facilitate their learning in her study, Liou (Reference Liou1997) raised the matter of determining whether listeners are being provided with more help options than are necessary or whether they have optimised their help-option use down to a few key choices. Furthermore, Cárdenas-Claros & Gruba's (Reference Cárdenas-Claros and Gruba2013) proposed framework for conceptualising L2 listening help options and associated design guidelines is in need of further examination and could usefully form the theoretical basis and guiding principles for this kind of research task. Lastly, Cárdenas-Claros & Gruba (Reference Cárdenas-Claros and Gruba2014) used Longman English Interactive (Rost & Fuchs Reference Rost and Fuchs2004, LEI© 2008), an off-the-shelf product designed for language learners with built-in help options. Given the comparatively larger amount of authentic streaming video available, it is important to examine the manner in which learners interact with such material using current media players. However, while subtitles, transcripts, and linked dictionaries are available for some authentic material (e.g., see CLILstore at http://multidict.net/clilstore/) or such help options can be readily added (e.g., see Ayamel at http://ayamel.byu.edu/), for other authentic streaming videos, researchers will need to consider how best to embed selected help options with streaming videos in a CALL environment. Overall, this mixed-methods approach for investigating learners’ cognitive and functional behaviours would be beneficial to addressing concerns mentioned and deepening our understanding of the which, why, and how of listeners’ help-option use. Moreover, understanding listeners’ interactions with help options can be seen as an important initial step for informing the subsequent design of learner training (see Research task 6) which reflects the effective interaction with help options by a given group of learners and addresses identified issues with misappropriate help option use.
2.2 Different conditions
Determining the relative impact of various help options (as opposed to one, for example, just captions) on facilitating both listening comprehension and vocabulary learning in comparison to no help options being available has been the focus of a few studies by Jones (e.g., Reference Jones2003, Reference Jones2006) and a colleague (Jones & Plass Reference Jones and Plass2002). These investigations focused on L2 French learners and various treatments providing pictorial/written annotations or no annotations in order to identify which annotation treatment was superior in aiding listening comprehension of, and vocabulary learning from, a single aural text. Furthermore, these studies investigated whether the use of these annotations as a help option was more effective than not having vocabulary annotations. Vocabulary learning was measured in the three studies with an immediate and delayed multiple-choice test, while listening comprehension was assessed through immediate and delayed written recall tests in the form of a summary of propositions contained in the given text. In each study, the learners who had access to both written or verbal and pictorial annotations showed the highest performance in the immediate and delayed vocabulary and listening tests. These findings suggest that a combination of written and pictorial support from help options are most effective for comprehension recall and language learning. In their concluding remarks, Jones & Plass (Reference Jones and Plass2002) and Jones (Reference Jones2006) point out that further research of this kind is important for it can offer support (or otherwise) for Mayer's (Reference Mayer1997, Reference Mayer2001, Reference Mayer2002) generative theory of multimedia learning (and, I would suggest, Paivio's Reference Paivio1971, Reference Paivio1986 dual-coding theory) by illustrating whether mental models of text content are strengthened when learners integrate multimodal representations of that text and create referential connections between them in working memory. Accordingly, follow-on research is warranted, given that there are other types of help options beyond just vocabulary annotations, none of which – apart from captions – have been investigated either individually or in written/pictorial combinations with respect to facilitating comprehension of a listening text (audio-only or audiovisual) and language learning in comparison to one another and a control group which cannot access help options. This would be valuable in terms of adding to the minimal (and rather dated) knowledge base which currently exists regarding combinations of help options and their utility in promoting language development and assisting understanding of listening texts, as well as inform theory as previous studies have suggested.
Research task 2
Investigate the impact of different help options in various permutations on promoting language learning and facilitating listening comprehension versus no help options.
To approach this research task, it seems useful to follow the research design employed by Jones & Plass (Reference Jones and Plass2002) and Jones (Reference Jones2003, Reference Jones2006) and investigate a number of treatment groups which have different combinations of help options providing written and/or pictorial support along with a control group that cannot access any help options. The language features targeted via the help options could be grammatical or lexical. A word of caution is necessary regarding the use of captions, subtitles or transcriptions for support in this kind of study, for rather than language and text recall deriving from listening, such recall may likely be due to a reading effect (e.g., see Winke, Gass & Sydorenko Reference Winke, Gass and Sydorenko2013), which is an unwanted confound if claims are to be made regarding aural input as the mode of learning. In addition to a language pre-test (like the vocabulary one used in the original studies), a baseline pre-test of listening comprehension is warranted to determine if participants’ listening levels are homogenous. Moreover, as in the studies mentioned, we might use written and vocabulary recognition and recall tests immediately following the treatments, followed by delayed versions. Xu (Reference Xu2010) also recommends testing learners’ ability to produce language rather than merely recognise and recall it to establish knowledge gains. Moreover, the same non-authentic text (a recorded written narration) was used in all of the studies, so it is important to determine whether findings for comprehension recall/language learning hold true for a range of text types, particularly authentic texts. Jones & Plass (Reference Jones and Plass2002) did not monitor whether participants actually accessed the annotations in the different treatments, and recommend collecting data on look up patterns and time spent accessing various types of support, as well as assessing the related role individual differences play in this behaviour.
In the introduction to this article, I indicated that Jones (Reference Jones2006) had conducted help-option research situated within a sociocultural theory perspective. Jones drew on this theory (along with interactionist accounts) in her overview of the notion of collaborative learning, an aspect which she investigated in unison with annotations in the given study. As in Jones & Plass (Reference Jones and Plass2002) and Jones (Reference Jones2003), this later study involved different treatment groups: no annotations and no collaboration; annotations only; collaboration only; and annotations with collaboration. The collaboration between learners was achieved through writing notes to one another in their native language while working together at a shared computer. The results from a vocabulary recognition and recall test showed that this form of collaboration had little effect on vocabulary learning. In contrast, the propositional recall of the annotations with collaboration group in a listening test was significantly superior to other groups. Jones proposes that opportunities for meaning clarification, hypothesis testing, and peer support along with motivational effects were possible reasons for this finding, but this is mere conjecture given no related data was collected. At the end of her article, Jones offers several recommendations for future research, with one recommendation being that related studies should more thoroughly investigate the interaction between annotations and collaboration and its effect on language retention and text recall. Moreover, Jones recommended collaboration through aural communication should be explored, as it may lead to better language learning and text comprehension than collaboration through written communication.
Research task 3
Investigate the impact of annotations with collaboration through aural communication on promoting language learning and facilitating listening comprehension.
With respect to associated theory underpinning the design and data analysis for this research task, rather than bringing in both sociocultural theory (SCT) and interactionist perspectives as Jones (Reference Jones2006) did, I believe it would be of value to frame the task totally in terms of SCT. This theory readily informs interpersonal, intrapersonal, and learner–computer activity in CALL contexts (as I mentioned earlier), all of which can be examined in the task I suggest, and which no other study about CALL with help options in L2 listening has attempted. This would be a valuable addition to the very limited body of SCT-inspired L2 listening research, and would also potentially contribute to variety in second language acquisition theory building, which Swain & Deters (Reference Swain and Deters2007: 871) state is necessary ‘to explore the complexity of the L2 acquisition process’. A solid number of SCT-informed studies on language learning through collaborative dialogue, i.e., the interactional talk between two or more learners in which knowledge is co-constructed in problem-solving activity (Swain Reference Swain and Lantolf2000), are available for reference which can offer theoretical insights into this type of research (e.g., Swain Reference Swain and Lantolf2000, Reference Swain and Byrnes2006; Swain, Brooks & Tocalli-Bellar Reference Swain, Brooks and Tocalli-Beller2002; Swain & Watanabe Reference Swain, Watanabe and Chapelle2012). A rare study by Gánem-Gutiérrez (Reference Gánem-Gutiérrez2006) of collaboration during CALL tasks illustrates the potential for this type of SCT-informed research for investigating the combination of collaborative dialogue and help options in mediating language learning and listening comprehension. L2 listening lessons typically encompass dialogic interaction between learners in pre-, while, and post-listening stages, with this interaction seen to promote the construction and co-construction of metacognitive awareness of L2 listening processes (see Cross Reference Cross2010). Examining collaborative dialogue and its effect on listeners’ metacognitive awareness as they engage in the processes of planning, problem-solving, monitoring and evaluating their listening during various computer-based listening tasks supported by different help options would contribute to developing L2 listening pedagogy in this regard.
Turning to research design, this task could be carried out using the same treatment groups as in Jones (Reference Jones2006) and incorporating the suggestions described for the previous research task (see Research task 2). Gánem-Gutiérrez's (Reference Gánem-Gutiérrez2006) study is also informative regarding possible research design features. In Jones (Reference Jones2006), participants collaborated through their first language (L1) (see also Gánem-Gutiérrez Reference Gánem-Gutiérrez2006), but it would also be of value to investigate the use of the L2 for communication, reflecting those learning contexts where this condition is often mandated. Ideally, technology would be used to record learners’ dialogue (audio or video) and help-option interactions (e.g., screen capture software) simultaneously. This would provide multifaceted data regarding annotation use with collaboration through oral communication, language acquisition, and listening comprehension, which would then enable the researcher to offer data-based insights into collaboration effects. Researchers may also consider examining grammatical forms rather than lexis, or other individual/combinations of help options, with collaboration in the manner I have outlined for this research task to broaden understanding.
While promoting language acquisition through and facilitating comprehension of a single text have shown immediate gains in language and text recall, for CALL with help options to gain greater acceptance in mainstream listening pedagogy, we need empirical support from longitudinal research examining the effectiveness of this technology on language and listening ability in comparison to more traditional modes of listening input. Brett (Reference Brett1997) has compared learners’ comprehension of six texts presented in audio-only, video-only, and multimedia (video plus an immediate feedback help option). Findings showed that language recall and comprehension task success rates were generally higher for the multimedia condition, with Brett suggesting this was probably a result of the instant feedback being used by learners to shape, direct and verify their comprehension, which made the input comprehensible and focused attention on linguistic features. Therefore, as indicated by Brett, a natural extension to his study (see Vandergrift & Cross Reference Vandergrift and Cross2015, for proposed replications of Brett's Reference Brett1997 study) would be longitudinally to examine, using different texts, the relative effects of the three treatments (audio-only, video-only, and multimedia) on developing listening skills and promoting language acquisition to establish if multimedia (video plus help options) is a potentially superior mode of delivery in these regards.
Research task 4
Investigate the development of learners’ listening ability and promotion of language acquisition through multimedia (plus help options) compared to audio- and video-only conditions.
This task would be longitudinal and experimental or quasi-experimental in design with a pre-test (to establish homogeneity, as suggested by Brett) and post-test and, ideally, a delayed post-test. The test employed should be a standardised test with listening and language-focused sections. Pertinent research design features for listening comprehension studies (e.g., ensuring equal amounts of listening in each group, controlling for listening practice outside the study) indicated by Cross & Vandergrift (Reference Cross and Vandergrift2015) should also be respected. There is plenty of scope to explore any of the current range of help options in various combinations. In his Reference Brett1997 study, Brett decided only to include an immediate feedback help option for multimedia treatment to keep the (language) input conditions constant across the three treatments. In the kind of longitudinal, experimental, or quasi-experimental study with a standardised test which I am suggesting, this condition need not be respected. However, it is wise to explore help options which support listeners’ text understanding directly (e.g., subtitles, transcripts) separately from those that offer guidance to listeners regarding their approach to achieving text comprehension (e.g., response feedback, listening advice) in order to determine their relative effectiveness. The quantitative data from testing could also be supported by qualitative data from observations, interviews, verbal reports, screen capture software, and/or eye tracking systems where suitable in a mixed-methods approach to tease out and establish how the different delivery modes influence listeners’ comprehension and language learning approach.
2.3 Learner variables
Essentially, proficiency level has been the only learner variable which has been investigated with respect to its influence on help-option use/non-use. Related studies by Grgurović & Hegelheimer (Reference Grgurović and Hegelheimer2007), Hoven (Reference Hoven2003), Hegelheimer & Tower (Reference Hegelheimer and Tower2004), Liou (Reference Liou1997), and Rivens Mompean & Guichon (Reference Rivens Mompean and Guichon2009) have produced diverse findings which make it difficult to draw any real conclusions. For example, some studies reported that less-proficient learners preferred different help options to more-proficient learners and were more likely not to use any help options at all, while other research found no correlation between help-option use and proficiency.
As Pujolà (Reference Pujolà2002) and Grgurović & Hegelheimer (Reference Grgurović and Hegelheimer2007) indicate, in addition to proficiency, there are other learner variables which may influence help-option use in CALL L2 listening research. The amount of hands-on computer-based listening experience a learner has had has been found to be an influence (Cárdenas-Claros & Gruba Reference Cárdenas-Claros and Gruba2014); for example, it may affect their confidence and ability to navigate help options and exploit control functions to maximise comprehension. Motivation is also a learner variable that may impact on learners’ engagement with help options and task performance (Grgurović & Hegelheimer Reference Grgurović and Hegelheimer2007; Cárdenas-Claros & Gruba Reference Cárdenas-Claros and Gruba2014). Moreover, other learner variables considered to impact on L2 listening comprehension in general identified by Vandergrift & Baker (Reference Vandergrift and Baker2015) which may also influence help-option selection and use include L2 vocabulary knowledge (e.g., the need to access glossaries) and metacognitive awareness of listening (e.g., selecting help options purposefully). If we can capture and cluster learner similarities and differences at different ability levels, we will have a clearer view of the alternatives for instruction that can be instigated (Blin et al. Reference Blin, Caws, Hamel, Heift, Schulze, Smith and Barr2013).
Research task 5
Investigate the influence of key learner variables such as listening proficiency, motivation, metacognitive awareness, and L2 vocabulary knowledge on help-option selection and use and task performance.
Given the range of variables which can be examined and the different permutations possible, this research task may best be actualised as a series of interwoven sub-tasks according to the particular variable/s of interest to the researcher. However, this does not imply each variable should just be examined independently of other variables, for the interplay between the various factors is also important to identify. For example, the development and application of metacognitive awareness is determined by a learner's motivation to engage in this behaviour (Ushioda Reference Ushioda, Lasagabaster, Doiz and Sierra2014). Potential approaches to data collection regarding some of the key variables are outlined below, each illustrating the need to investigate the given variable using instruments which gather both broad and help-option-use-specific data.
2.3.1 Listening proficiency
Data regarding this variable may be collected through the use of a standardised computer-based listening test to tap listening ability in a CALL setting since listening processes in that context may differ to other listening environments. This research sub-task would also benefit from understanding listener proficiency through a researcher-designed test which reflects the kind of listening material with help options the users will experience and targets the same listening task objectives. Such quantitative data can be supplemented by data from qualitative instruments (e.g., retrospective verbal reports after each question) designed to capture the strategic competence of learners, which is a key facet of L2 listening proficiency (Field Reference Field, Geranpayeh and Taylor2013) and highly likely to shape listeners’ help-option use.
2.3.2 Motivation
Collecting data regarding motivation is largely done through developing and administering questionnaire surveys, or modifications thereof, to cater to contextual factors and the group of respondents, and there is no reason why this should not be so to capture general learning motivations for this sub-task. However, attention should also be given to measuring learners’ motivations (and demotivations) regarding learning to listen through technology and aspects related to help options as listening support in the given context (see Dörnyei Reference Dörnyei2010 for creating, administering and processing such questionnaires). Subsequent face-to-face interviews regarding responses would help to expand understanding and build a more robust picture of the nature of technology-related motivational factors and their influence.
2.3.3 Metacognitive awareness
The Metacognitive Awareness Listening Questionnaire (MALQ) (Vandergrift et al. Reference Vandergrift, Goh, Mareschal and Tafaghodtari2006) is a frequently used instrument to gather information regarding learners’ metacognitive awareness of L2 listening immediately following a listening task (e.g. see Vandergrift & Baker Reference Vandergrift and Baker2015). This data can also be enhanced with verbal and/or written prompts which elicit reflections regarding person, task, and strategy knowledge and use which the MALQ provides in quantitative form. The kind of data provided by the MALQ is task-related but rather generic. To narrow in on metacognitive behaviour in relation to the selection and use of various help options in CALL contexts, a bespoke questionnaire which attempts to measure person, task and strategy factors following a listening task in which learners have access to help options would be a valuable accompaniment.
2.3.4 L2 vocabulary knowledge
A broad indication of the aural vocabulary knowledge of L2 listeners can be established using the Listening Vocabulary Levels Test (LVLT) (McLean, Kramer & Beglar Reference McLean, Kramer and Beglar2015). In addition, establishing the depth of knowledge of lexical items is likely to be informative for understanding the influence of the selection and frequency of use of various help options (particularly glossaries and dictionaries). Lexical items included in the listening material accompanied by help options should be targeted, so this will probably mean the contents of off-the-shelf tests for depth of L2 vocabulary knowledge (for reading comprehension) such as the Vocabulary Knowledge Scale (Wesche & Paribakht Reference Wesche and Paribakht1996) or the Depth of Vocabulary Knowledge (see Qian & Schedl Reference Qian and Schedl2004) measure will not necessarily be appropriate to the particular context and content being investigated, but can serve as models of the format, question types and length for a similar instrument that can be developed.
In addition to collecting data on the variable(s) of interest, recording the type and frequency of help options consulted for the given listening task using, for example, screen capture software (e.g., see Grgurović & Hegelheimer Reference Grgurović and Hegelheimer2007) is necessary. Examining the statistical relationship between the learner variables could be achieved through the use of multivariate techniques (e.g., see Skehan Reference Skehan2014), with this type of analysis effective for constructing models illustrating how different variables interact (e.g., see Vandergrift & Baker Reference Vandergrift and Baker2015). Analysis of the qualitative data gathered would provide a more complete picture of the manner in which the variables shape listeners’ choice and utilisation of help options in task completion. Lastly, the researcher has the option of collecting data on these variables through a ‘one-shot’ approach in immediate conjunction with a computerised listening task supported by help options, or track variables and their relationships to each other, and to help-option selection and use in CALL listening environments longitudinally (which also would bring in data collection through listening diaries) in order to reveal how such variables change across the learning experience.
2.4 Learner training
In addition to a range of learner variables possibly being an influence on the use and non-use of various help options, it may also be that a lack of relevant help-option instruction is a potential factor, particularly for less-skilled listeners. Recognising this, some authors (e.g., Liou Reference Liou1997; Grgurović & Hegelheimer Reference Grgurović and Hegelheimer2007; Jones Reference Jones2008; Cárdenas-Claros & Gruba Reference Cárdenas-Claros and Gruba2009, Reference Cárdenas-Claros and Gruba2014) have stated that learner training would be advantageous and is worth implementing. Findings from the notably limited CALL research into learner training for listening (see Romeo & Hubbard Reference Romeo, Hubbard, Levy, Blin, Siskin and Takeuchi2010; Hubbard & Romeo Reference Hubbard, Romeo and Stockwell2012) and other skills support this view (see O'Bryan Reference O'Bryan2008), though none of these studies examined if the training led to language learning enhancements. In discussing the need for learner training in CALL, Hubbard (Reference Hubbard2013) offers four standpoints to vindicate why it is important: well-designed tasks and technology do not inevitably result in learner success; learners do not intuitively evolve into effective technology users over time; even ‘digital natives’ (Prensky Reference Prensky2001) require training; and both teachers and learners need specialised training. Given the urgings of the help-option researchers mentioned, the positive findings from the few learner training studies thus far, and the points Hubbard makes, there certainly appears to be a case for investigating the impact of learner training on help-option interactions and determining if there are associated performance gains, particularly for learners at different levels of listening proficiency who have been shown to vary in their approach to help-option use (e.g., see Grgurović & Hegelheimer Reference Grgurović and Hegelheimer2007). Indeed, researching this aspect will directly begin to address the lack of L2 listening pedagogy related to help options I mentioned in the introduction to this article.
Research task 6
Investigate the effect of learner training on help-option use and task performance for groups of learners of differing listening ability.
The research design for this task could broadly follow that employed by O'Bryan (Reference O'Bryan2008), i.e., an experimental, mixed-methods approach. This design should also include a pre- and post-test aspect, given the importance of establishing whether the training intervention affects performance is a key outcome of this research task. This testing should encompass a standardised computer-based listening test as a minimum, though it may usefully be supplemented by a researcher-designed test targeting particular listening skills or language attributes of interest to the researcher.
In more detail, a preliminary step for this task would be for the researcher to decide whether the focus is on the impact of training in help-option use on listening comprehension ability, or on the use of help options for learning language (e.g., vocabulary) through listening for they are very likely to require training in different types and combinations of help-option use. For example, captions and transcripts can draw listeners’ attention to language features being targeted, while answer feedback which directs the listener to a related area of the listening text may promote comprehension processes such as problem-solving and monitoring. Subsequently, and as I alluded to at the end of my commentary for Research task 1, it is important to collect data to establish how the given learners at beginner, intermediate, and advanced levels employ help options for task achievement across a range of tasks, for this will inform the nature of the training. This data collection need not be as exhaustive as for Research task 1 for it should reflect what listening teachers can put into practice in classroom contexts. As such, observation notes, interviews and/or learner journals could be utilised. A further consideration is regarding which help options to choose for inclusion in the training. Cárdenas-Claros & Gruba's (Reference Cárdenas-Claros and Gruba2013) conceptualisation of a typology of L2 listening help options and general guidelines indicating the number and combination of help options are important considerations at different levels of language (and listening) proficiency. Therefore, studies which explore learner training in terms of the relative utility of the help-option types suggested by Cárdenas-Claros & Gruba (Reference Cárdenas-Claros and Gruba2013), and how many and which paths of use for help options are preferable at different levels of ability is one potential focus for researchers.
With respect to the actual approach to the learner training itself subsequent to the initial steps outlined above, the investigations by O'Bryan (Reference O'Bryan2008), Romeo & Hubbard (Reference Romeo, Hubbard, Levy, Blin, Siskin and Takeuchi2010), Hubbard & Romeo (Reference Hubbard, Romeo and Stockwell2012) drew on a set of five associated principles proposed by Hubbard (Reference Hubbard, Fotos and Browne2004), and this pedagogical framework is suggested for this research task. These five principles are: give teachers first-hand experience of CALL; give learners teacher training; use a cyclic approach; use collaborative debriefings; and teach general exploitation strategies (see Hubbard Reference Hubbard, Fotos and Browne2004: 51–56 for details). Hubbard & Romeo (Reference Hubbard, Romeo and Stockwell2012) and Romeo & Hubbard (Reference Romeo, Hubbard, Levy, Blin, Siskin and Takeuchi2010) embedded these principles in a conceptual framework targeting three domains of learner training: how (technical training), what (strategic training), and why (pedagogical training). Regarding listening help options, while training in metatechnical skills for today's media players is typically fairly straightforward, strategic and pedagogical training could be underpinned by referral to publications covering metacognition in L2 listening (e.g., Vandergrift & Goh Reference Vandergrift and Goh2012). Such metacognition embraces knowledge of the listening task, knowledge of appropriate listening strategies, and learners’ knowledge about themselves as listeners (Vandergrift & Goh Reference Vandergrift and Goh2012). Training targeting these knowledge forms and integrating help-option use, e.g., instructing learners on task appropriate help options, linking sequences of listening strategies with help-option strategies, and selecting help options in line with their individual beliefs about learning, can usefully be explored in this research task. Alternatively, the researcher may examine training in help options which facilitates task completion and promotes learning within a framework of the metacognitive listening processes of planning, problem-solving, monitoring, and evaluation. Furthermore, earlier in this section I mentioned Cárdenas-Claros & Gruba's (Reference Cárdenas-Claros and Gruba2013) proposed conceptualisation of listening help-option types and general guidelines. As I stated in the introduction, help-option researchers rarely integrate mainstream L2 listening theory (e.g., metacognition, listening strategies, skills-based listening) into their research, instead focusing on interactionist theories, for instance. Effective learner training for developing listening skills aided by help options needs to bring together frameworks such as Cárdenas-Claros & Gruba's (Reference Cárdenas-Claros and Gruba2013) and the work of, for example, Vandergrift & Goh (Reference Vandergrift and Goh2012), and this approach provides another area for researchers to explore with different levels of learners.
The principle that learner training involves a cyclic approach suggests the need for this research task to be longitudinal, such as throughout an entire course (O'Bryan Reference O'Bryan2008), though the length of the study will probably depend on the level of the learners being investigated, i.e., beginner-level listeners are going to need more extensive learner training over time to develop their metacognitive knowledge and effective listening and help-option skills. This also points to the need to establish how often and for how long listeners at different levels (and for varieties of listening materials and task types) should be given training to maximise their listening and language development supported by help options.
Regarding data collection during the training, quantitative data from listening tests and questionnaires, e.g., the MALQ (Vandergrift et al. Reference Vandergrift, Goh, Mareschal and Tafaghodtari2006) could be used to measure progress at selected points to examine changes in metacognitive awareness, with help-option use data coming from an online tracking system (e.g., see O'Bryan Reference O'Bryan2008). Qualitative data can come from collaborative debriefings as part of the training cycle as well as from any of the usual range of instruments such as listening diaries and interviews.
4. Conclusion
In this article, I have presented an agenda for researching help options for L2 listening in CALL contexts. Each of the tasks within the four themes I have covered offers interested researchers direction for advancing various aspects of a field that seems to have fallen rather by the wayside in comparison to other strands of L2 listening research. I have attempted to recommend task approaches that address issues I pointed out at the start of this article and have thus endorsed a mixed-method and longitudinal design, a diversity of instruments, and a focus on learning outcomes where pertinent; I also looked to promote theory building and encourage concern for informing an associated pedagogy.
Much work still remains to be done to build a robust body of research regarding language learning through computer-based listening facilitated by help options. I hope that the research agenda I have proposed will encourage L2 listening researchers to respond to this challenge by reinvigorating interest in conducting related studies.
Jeremy Cross is an Associate Professor in the Graduate School of Humanities at Nagoya University, Japan. His primary research interest is L2 listening comprehension, and he has published related articles in various international journals.
