1. Introduction
The design and sophistication of help options in computer-based second language (L2) listening applications have evolved in line with the advances in computer usability and capability (Lynch, Reference Lynch2009). Initially, help options were text-based and they consisted of first language annotated transcripts and dictionary-like entries. With the addition of static and interactive images, help options incorporated visual elements as a way to complement texts and increase learner motivation. More recently, help options have been developed with links to online resources. Nowadays, they offer the listener alternative ways of exploiting aural texts through pronunciation cues, cultural notes, dictionaries and grammar modules (Robin, Reference Robin2007). Despite such advances, however, the conceptualization and design of help options in computer-based L2 listening remains an under-researched area.
In this paper, we put forward the Conceptualization and Design (CoDe) framework in an effort to spur the research and development of help options in computer-based L2 listening. We open our discussion with an examination of current classifications of help options in CALL. After setting out four empirical studies, we present conceptualization concepts and design guidelines. We conclude by exploring the integration of the conceptualization and design components of the framework and by providing suggestions for further research.
2. Situating the study
Earlier, we defined help options as embedded application resources that assist learners in performing computing operations and/or support language learning (Cárdenas-Claros & Gruba, Reference Cárdenas-Claros and Gruba2009). We acknowledge that the purpose of ‘Help’ toolbars and elements are to assist, but throughout our work we have argued that the actual use depends on each individual listener deciding whether, when and how to utilize the available options. Choice is fundamental.
Notably, our views differ from those introduced by Pujolà (Reference Pujolà2002) and Hegelheimer (Reference Hegelheimer2003). In Pujolà's work, for example, help facilities were divided into two major groups: assistance and guidance. Facilities for assistance provide learners with help for comprehension of learning materials; guidance facilities provide learners with help to act upon tasks (Table 1).
Table 1 Help facilities framework (Pujolà, Reference Pujolà2002)
In a related framing, Hegelheimer (Reference Hegelheimer2003) set out two functions: operational guidance that allows the learner to repair miscues caused by software; and task guidance, that allows learners to act upon language learning tasks. Such views of help options may require revision: First, the terms ‘assistance facilities’ and ‘task guidance’ do not fully acknowledge learner autonomy; secondly, though in-built resources have the potential to be assistive, their interpretation and subsequent use is dependent on a particular listener's context and perception. On a third note, current classifications are not sufficiently fine-grained to be able to conduct close analysis of empirical listening data. By simply grouping a variety of options in a single category (i.e., ‘assistance facilities’ or ‘task guidance’), the intended functions of differing options are obscured. For instance, when a listening transcript is accessed, the aural-to-visual resource is designed to assist comprehension. A cultural note, however, acts to enrich or supplement listener understanding and not ‘assist’ or ‘guide’. For these reasons, frameworks must be complex enough to be used in the sophisticated analysis of differing resources, purposes and listener motivations.
Although current classifications of help options are widely acknowledged in the CALL literature (see Levy & Stockwell, Reference Levy and Stockwell2006), such frameworks remain largely theoretical. To extend earlier work, here we include listeners’ voices as a way of providing a richer account of what features they deem relevant for comprehension in computer-based L2 listening.
3. Empirical studies
To lay the groundwork for our proposed CoDe framework, we review the literature in three main areas: 1) L2 listening and computer-based L2 listening; 2) help options use/non-use in CALL, interactive learning environments and help systems design and 3) user-centered design with a particular focus on participatory design practices. Thus, we consulted works that include:
3.1 Listening
• L2 listening: Flowerdew & Miller, Reference Flowerdew and Miller2005; Lynch, Reference Lynch2009; Rost, Reference Rost2002, Reference Rost2006, Reference Rost2007; Vandergrift, Reference Vandergrift2007.
• Computer-based L2 listening: Brett, Reference Brett1995, Reference Brett1996, Reference Brett1997, Reference Brett2000; Hubbard, Reference Hubbard2001, Reference Hubbard2004, Reference Hubbard2006; Ramirez & Alonso, Reference Ramirez and Alonso2007; Vanderplank, Reference Vanderplank2009; Weinberg, Reference Weinberg2002.
3.2 Help options use/non-use:
• Computer-based L2 listening: (Cárdenas-Claros, Reference Cárdenas-Claros2005; Chapelle, Reference Chapelle2003, Reference Chapelle2005, Reference Chapelle2009; Chun & Payne, Reference Chun and Payne2004; Chun & Plass, Reference Chun and Plass1996; Grgurović & Hegelheimer, Reference Grgurović and Hegelheimer2007; Hegelheimer, Reference Hegelheimer2003; Hegelheimer & Tower, Reference Hegelheimer and Tower2004; Hernández, Reference Hernández2004; Hoven, Reference Hoven1999, Reference Hoven2003; Hsu, Reference Hsu1994; Jones, Reference Jones2003, Reference Jones2006, Reference Jones2009; Liou, Reference Liou2000; Lin & Chen, Reference Lin and Chen2009; Pujolá, Reference Pujolà2002; Jones & Plass, Reference Jones and Plass2002; Rivens Mompean & Guichon, Reference Rivens Mompean and Guichon2009; Sun, Reference Sun2010).
• Interactive learning environments (ILE): Aleven, Stahl, Schworm, Fischer & Wallace, Reference Aleven, Stahl, Schworm, Fischer and Wallace2003; Aleven, McLaren, Roll, Koedinger, & Lester, Reference Aleven, McLaren, Roll, Koedinger and Lester2004; Bartholomé, Stahl & Bromme, Reference Bartholomé, Stahl and Bromme2004; Clarebout & Elen, Reference Clarebout and Elen2006; Reference Clarebout and Elen2009; Horz, Winters & Fries, Reference Horz, Winters and Fries2009; Manlove, Lazonder & De Jong, Reference Manlove, Lazonder and De Jong2009.
• Help systems design environments: Dworman and Rosembaun, Reference Dworman and Rosenbaum2004; Ellison, Reference Ellison2007; Hughes, Reference Hughes2007; Kelleher & Pausch, Reference Kelleher and Pausch2005; Murray & VanLehn, Reference Murray and VanLehn2005; Sampson, Reference Sampson2007.
3.3 User-centered design
• CALL design practices: Colpaert, Reference Colpaert2004; Cushion & Hemard, Reference Cushion and Hémard2003; Heift, Reference Heift2006; Hémard & Cushion, Reference Hémard and Cushion2001; Fotos & Browne, Reference Fotos and Browne2004.
• Participatory design: Müller, Reference Müller2003; Bromberg & Henderson, Reference Bromberg and Henderson1990; Bødker, & Iversen, Reference Bødker and Iversen2002; Farmer, Reference Farmer2008; Farmer & Gruba, Reference Farmer and Gruba2006; Könings et al., Reference Könings, Brand-Gruwelb and van Merrienboerb2010, Salman et al., Reference Salman, Ince, Kim, Cheng and Yildirim2009; Zaphiris & Zacharia, Reference Zaphiris and Zacharia2001.
Our review of the literature identified three main gaps regarding the lack of studies pertaining to: 1) the listener experience, 2) design features that stimulate use, and 3) learning tasks that stimulate use. Importantly, we note that much research into help options has been investigated through quantitative methods using non-disruptive technologies such as tracking systems and screen capturing devices. Although useful, we argue here that such measures do not allow for an opportunity to ask participants if, how, and when they resort to the use of help options; that is, a better understanding of listener reasons for use/non-use is not elicited. The questions that guided our overall investigation included:
1. What triggers participants to use help options in a computer-based L2 listening activity?
2. What inhibits participants from using help options in a computer-based L2 listening activity?
3. What design features of help options can contribute to their use?
To answer these questions, four independent but interrelated empirical studies were structured (Table 2). The first two studies (Study One and Study Two) sought to explore reasons for use/non-use of help options with 15 adult learners of English from Colombia who interacted with selected listening tasks from the Longman English Interactive© (LEI©) program in three sessions. A series of semi-structured interviews were audio recorded. Data were translated and coded and inter-translator reliability established (0.95). Cyclical coding and recoding of data resulted in the development of five themes: Relevance, Recovery, Challenge, Familiarity and Compatibility (Appendix A).
Table 2 Summary of research design and findings in Cárdenas-Claros, (Reference Cárdenas-Claros2011)
In the two complementary studies, Study Three and Study Four, participatory design sessions were held that consisted of eight language learners, two human-computer interaction specialists, two language teachers and a computer programmer. To minimize issues of cultural influence, study participants were predominantly from Colombia and all had been language learners themselves. Thus, participatory design sessions were held in Spanish and participants evaluated three researcher-generated paper prototypes, designed two paper prototypes (beginner and high-intermediate) and iterated the design on paper (Study Three) and on the computer screen (Study Four). Design outcomes and collaborative interaction between participants were analyzed and qualities of help options were identified across studies using software (Nvivo 8.0) to assist qualitative data analysis. Design outcomes were presented in light of five features of design: Type, Location, Sequence, Click-Through and Display (Appendix B). Analysis of interaction data across studies yielded four themes: Ease of use, Learner control, Guidance and Learning (Appendix C).Footnote 1 In this paper, we focus on elements that provide repeated and slowed audio text delivery (audio control buttons), transcribed texts, translations, cultural notes, dictionaries, glossed words, listening tips and feedback. Captions and subtitles, although one of the most common forms of listening assistance (Winke, Gass & Sydorenko, Reference Winke, Gass and Sydorenko2010) were not explored because participants primarily interacted with audio-only materials (on English Listening Lounge© and Randall's Cyber Listening Lab©) or with video clips that offered transcripts instead (Longman English Interactive©).
4. The CoDe framework
As noted previously, we based the construction of this proposed framing on three different sources: 1) results of previous research on help options in CALL, 2) findings from research investigating help options/users’ assistance in interactive learning environments, and 3) analysis of the four empirical studies conducted as part of a doctoral study that investigated the use/non-use and design features of help options in computer based L2 listening. Thus, our proposal is both theoretical and empirical and consists of two components: Conceptualization (Co) and Design (De).
4.1 The conceptualization component
Acknowledging the work of Pujolà (Reference Pujolà2002) and Hegelheimer (Reference Hegelheimer2003), we propose four components that set the boundaries of help options: Operational, Regulatory, Compensatory and Explanatory. In Table 3, we set out the conceptualization component of the framework. In the first and second columns we describe the types of help options and their functions, focusing on the activities or actions where learners may benefit from the use of a particular help option element. In the third column, we explain when and/or how each type of help option is used. In particular, we consider the moment when the option is used as it transforms original input in order to ease comprehension, facilitate language processing, and draw the learners’ attention to key linguistic features. In the final two columns, we set out the type of enhanced input as identified by Chapelle (Reference Chapelle2003) and how such input has been, and can be, operationalized in CALL.
Table 3 Conceptualization of help options in computer-based L2 listening
4.1.1 Component One: Operational help options
Operational help options are designed to assist listeners to master software and hardware functions in advance of potential problems. In general, operational help options encompass available advice on how to run, approach and address software operations; examples of such options include help menus, user manuals, training modules and interactive tutorials.
As we conducted Studies One and Two, we found that participants stressed the need to include components that directed them where to click when a problem arises. With a view to minimizing frustrations that interfere with learning, participants in Studies Three and Four recommended that operational help options be presented at an introductory level. At this level, participants noted, all other functions of the program could be explained both in the target language and in the first language of the listener. In this way, participants suggested, learners could gain familiarity with the software as well as benefit from exposure to the target language. In the introduction section of a computer-based listening resource, participants argued that listeners should be required to complete a module that explains each option before being allowed to access any further material.
4.1.2 Component Two: Regulatory help options
Regulatory help options, as we see them, are designed to influence and guide listeners on how to pace their learning such that they can adjust their behaviour in response to task demands. Sensitive to the moment of use, we set out two types of regulatory help options: in preparation for task demands and after task completion.
Listeners use regulatory help options in preparation for task demands as a way of approaching texts. Instances of regulatory help options used in preparation for task demands are listening tips, hints on strategy training and guidance on if, when, and how to use help options. Such options, for example, provide listeners with instructions on how to proceed with learning, how to approach a task and how to attend to focal language features. To illustrate, imagine that the goal of a lesson is to familiarize students with numbers from zero to nine. Here, the task would be to listen to a conversation and then fill in a chart with the correct phone numbers. In this instance, a regulatory option would be presented as a listening tip to suggest listening for specific information in the text (numbers). The listening tip would direct learners to employ strategies that would allow for the identification of key information.
Regulatory help options used after task completion promote the self-regulation of learning by raising the awareness of the power of self-assessment for both comprehension and misunderstanding. Such options go beyond traditional corrective feedback (where answers are marked as right or wrong) as they provide listeners with further opportunities for text comprehension, task completion and/or learning. Instances of such regulatory help options include explanatory feedback and messages that are displayed after a listener clicks on a check or submit button.
Ideally, regulatory help options should be accessible throughout the completion of a task. At present, we understand that regulatory options in the form of ‘interactive hints’, for example, may be designed for reading, vocabulary and/or grammar but at present the promise of (more) Intelligent CALL for listening is still in its infancy. Additionally, research would be needed to clarify interactional factors and pinpoint specific instances that require assistance. The construction of such sophisticated options would require a substantial investment as ‘hints’ would need to be tailor-made for each audio segment. We would argue that the mere availability of sophisticated options would not guarantee that learners would actually use them. In fact, throughout our studies participants noted that simultaneous interactions with help options and aural texts placed such cognitive demands on them that they were distracted, not assisted.
4.1.3 Component Three: Compensatory help options
Compensatory help options assist listeners to recover from breakdowns in understanding. Such breakdowns may result from the lack of background knowledge, frail confidence, or weak task and software familiarity as well as low working memory capacity. To classify compensatory help options, we argue, SLA principles of input and interaction from an interactionist perspective (Chapelle, Reference Chapelle2005b) are the most salient. In this view, input undergoes three types of modification: aural-to-visual, aural-to-aural, and visual-to-visual.
In the aural-to-visual modifications, we distinguish between pictorial (e.g., images, video clips) and orthographic (e.g., transcripts) elements. Visual elements in the video were used by most participants in Studies One and Two to complement their understanding of the audio track. Some participants, however, found the images distracting, and thus avoided watching the video. Importantly, participants noted that they needed to have a minimum grasp of the aural text to make sense of the video; otherwise, they thought, images could be misinterpreted and understanding negatively affected. The participants in Study Three and Study Four found that a transcript, one form of aural-to-visual modification, was crucial for language development and they were keen to use it more often.
The second type of modification, aural-to-aural modification, is performed as users replay, pause, rewind and/or forward specific or complete aural segments. Aural-to-aural modification makes input salient in the form of repetition, and this key feature of input is believed to push language development (Chapelle, Reference Chapelle2003). Such modification is operationalized as audio/video control buttons, media controller bars, and variable speed playback buttons. Through these tools, learners can pace the delivery of input in tune with their learning styles and working memory capacity.
In a third type of modification, visual-to-visual modification, input is presented through written texts in the target language of the user. Operationalized as native language translation options, learners can use these modifications to comprehend the target language, to confirm their own hypotheses and to check their understanding of individual lexical items.
4.1.4 Component Four: Explanatory help options
Explanatory help options offer a listener the ability to access additional information about the input rather than presenting the same input in a different modality. Two functions motivate the use of explanatory resources: to draw learner's attention to key words/linguistic features and to expose learners to enriched input. Hyperlinks to glossaries, definitions, and glossed words, for example, can draw learners’ attention to specific elements of the input.
The second function of explanatory help options is that of enriching original input to encourage learners to explore language beyond the interaction with original input. Instances of explanatory help options that enrich input include culture notes, online-concordances, grammar explanations, and dictionaries.
In the listening section of the LEI© program, explanatory help options are operationalized as culture notes and as an online monolingual dictionary. The cultural notes option is presented as a separate button in the listening comprehension activities and is available across each level of language proficiency. Participants in our studies, particularly beginners and high-beginners, neglected the culture notes option because they did not see an immediate relevance to make use of it for either text comprehension or task completion. Moreover, lower proficiency listeners neglected culture notes because they could not understand the input. Learners at the intermediate and high-intermediate proficiency levels, however, used cultural notes to extend their understanding to go beyond the initial information that was available in the aural input.
A software design flaw prevented the effective use of culture notes for all learners because scrolling caused the notes to be hidden and thus neglected. In our studies, it appeared that the benefit of accessing cultural notes was not apparent; participants in Study Three, for example, asked for cultural notes to be better highlighted to ensure better interaction. Listeners in Study Four recommended that the option be omitted in designs for low proficiency learners, as a way of freeing cognitive resources for other listening tasks.
Another explanatory component, the monolingual dictionary option, is offered in the LEI© program. Participants found it difficult to use as it could only be accessed through the homepage. Such poor design structure, we found, forces learners to abandon listening exercises in order to access the in-built dictionary. Difficulties in accessing and then trying to make use of a dictionary involved additional cognitive demands which listeners were often unable to meet. When they did use a dictionary, participants in our studies made an effort to exit the LEI© program and access an online translation tool.
4.2 The Design component
The Design component of the CoDe framework was primarily, but not exclusively, informed by the analysis of collaborative interactions between L2 learners, language teachers, computer programmers, and HCI designers. In this section, we propose four guidelines based on what we ourselves found and with other previous research and principles.
4.2.1 Design Guideline One: Strive for simple and intuitive design
Our first contention is that the design of help options in computer-based L2 listening should be easy to understand regardless of previous computer experience, language proficiency and the level of engagement with listening activities. Throughout our studies, we found that participants too often reported problems in finding available options because they were difficult to find and use. To illustrate, we recorded the actions of a listener who neglected key vocabulary options because of a confusion between links that were intended for complete units and those intended for specific listening tasks. In another example, we saw high proficiency listeners who were unable to access glossed words because the program first required them to open up one of the transcripts.
To improve design, the design session professionals who participated in our studies suggested that help options be: 1) grouped and located in a horizontal toolbar immediately above the questions; 2) named by the function they accomplish; 3) offered in a single click, and 4) be displayed continually on the screen. Ideally, they argued, the location of any given help option would emulate the design patterns that have been established by major companies (e.g., Google, Apple, Microsoft) and located horizontally along the screen. Customization by users would also be preferable. Further, the design professionals argued that options be given functional names so that learners could anticipate what they would do once they were accessed. Finally, learner effort and cognitive resources are likely to be wasted if options are not a single click away and constantly displayed (Lynch & Horton, Reference Lynch and Horton2009).
4.2.2 Design Guideline Two: Provide different routes for interaction
Help options need to be designed to accommodate different learner preferences and learning styles. At present, a common flaw in help option design forces learners to make use of a single path of interaction (Colpaert, Reference Colpaert2004; Grgurović & Hegelheimer, Reference Grgurović and Hegelheimer2007; Cárdenas-Claros & Gruba, Reference Cárdenas-Claros and Gruba2009). Current help option designs are not flexible enough to cater for the different preferences and learning styles of a wide variety of language learners.
Throughout Studies One and Two, we witnessed several listeners who missed a number of interaction opportunities, for example, with glossed words and translation options. Students were often unable to effectively transfer classroom strategies, such as vocabulary preview, to the computer-based environment. Confusion and frustration were readily apparent.
Design professionals in our work made a range of specific recommendations to foster ways to better accommodate individual preferences and learning styles. They suggested that glossed vocabulary, for example, be accessed both within the context of a transcript and as part of a stand-alone list. Similarly, translation options would be better presented if they were made available in tandem with an aural text, or placed alongside a transcript, or if possible be accessed through a combination of these two modalities.
4.2.3 Design Guideline Three: Provide fewer, rather than more, help options
Help options, we found, are best presented to listeners in terms of proficiency. Offering too many options, in the hope of providing assistance ‘just in time’, seems to unnecessarily confuse listeners and places additional cognitive demands on them. In our studies, we found that participants at the lower proficiencies were seemingly confronted with too many options and thus did not make effective use of them.
Design professionals recommended that ideally four or five options be made available at each level of proficiency. Help options that are not seen to be beneficial for learners at different stages can be removed or replaced by others. Translation options, for example, could be offered for beginners; for other levels, transcript and glossary options could be included in the program.
4.2.4 Design Guideline Four: Minimize potential distractions
The design of help options should help learners to stay on task. Help options that require learners to take a number of unnecessary steps in response to task demands creates frustration and signals the presence of poor design practices (Colpaert, Reference Colpaert2004; Hegelheimer & Tower, Reference Hegelheimer and Tower2004). In particular, we highlight the poor design of the dictionary option in the LEI© program that required listeners to exit interactive activities. As a result of the flaw, listeners opted to use an online translation tool because it was familiar, easy to access and could be constantly displayed. Going online, and thus outside the listening program, encouraged simultaneous interaction between the definitions and the listening activities.
Dictionary options, design professionals suggested, could be presented in pop-up windows and appear on the top-right-hand-side of a page. Although such a design may cause listeners to ‘momentarily abandon’ a task to work with the dictionary, such an interruption would not be as abrupt as exiting to work on an entirely different page; hopefully, learners could recover more easily to then resume task completion.
4.3 Integrating the conceptualization and design of help options
Integrated findings from four empirical studies show a relationship between the conceptualization and the design components in our proposed CoDe framework (Table 4). The first column of Table 4 lists the four components of the conceptual part; the second column lists the help options suggested by design professionals. In each of the remaining columns, we list a single design guideline of the CoDe framework. It is worth noting that, although all of the suggested help options are grounded in Guideline One: strive for simple and intuitive design, only a few of the options are informed by Guideline Two: offer different routes of interaction or Guideline Four: minimize potential distractions.
Table 4 Relationship between conceptualization and design of help options
We found that only two options, ‘glossed words’ and ‘translation’, follow Guideline Two and allow multiple routes for access. We recognize, however, that providing varying paths to access to help options such as audio control buttons would result in potential complexity and further distraction. Thus, simple and intuitive design would seemingly be sacrificed. Clearly, more research into effective help option design on this point is warranted.
Guideline Three, provide fewer, rather than more help options suggests that L2 listeners should be provided with the options they are more likely to use based on language proficiency. Design professionals acknowledged that some help options were more useful for learners at particular proficiencies than at others, as in the case of translation options. Therefore, they argued that option access should differ by language proficiency.
Although such a view addresses the concern to do with having too many options, this guideline should be interpreted cautiously because the listeners may use a program without training and for a short time. It is likely, though, that additional exposure to the learning environment along with training may result in a reduction of cognitive load without implying a further restriction of options.
With regard to Guideline Four, minimize potential distractions this is mostly apparent in the use of a dictionary option that is offered in a separate page of interaction. We found that listeners neglected dictionary options because they were required to set aside productive interactions. As a result, listeners were distracted away from the task at hand.
5. Conclusions
In this paper, we have proposed a re-conceptualization of help options in the CoDE framework. Such a framing leads to two insights relating to both the neglected and the indiscriminate use of help options. With regard to the former, researchers have consistently reported an underuse or neglect (Cárdenas-Claros, Reference Cárdenas-Claros2005; Grgurović & Hegelheimer, Reference Grgurović and Hegelheimer2007; Hsu, Reference Hsu1994; Hegelheimer & Tower, Reference Hegelheimer and Tower2004; Liou, Reference Liou1997; Rivens Mompean & Guichon, Reference Rivens Mompean and Guichon2009); here, we argue that in particular regulatory and explanatory help options tend to be the most commonly neglected. Further study, directed at regulatory help options specifically, should be conducted in light of their increased use throughout CALL materials.
Indiscriminate use of help options may well hinder the effective development of L2 listening skills; that is, beyond their utility for immediate task completion, only participants at the high levels of proficiency could make a link between computer-based learning and real-world application. If a poor transference between modalities persists, listeners run the risk of misusing help options. At present, it appears as if only regulatory help options prepare learners to work with real world materials. To improve the utility of help options to support autonomous learning, for example, we would argue that students need to be better directed to make links between assistance and future usages of a particular aspect of listening.
6. Pedagogical implications
In this section, we argue that our proposed CoDe framework could well serve as a roadmap in the CALL classroom to train L2 listeners to make better use of help options; in particular, we would like to encourage learners to use in-built computer resources for purposes beyond immediate task completion and/or text comprehension. Clearly, the potential of help elements for learning should be emphasized through systematic reflection on the part of learners as well as structured guidance from CALL instructors. A frequent discussion of mutual experiences, we believe, may lead to a shift in the misconceptions often associated with help option use/non-use.
We understand that our proposed design guidelines in the CoDe framework do not ensure that listeners will use the help options as intended, but we hope that their implementation may ease some of the usability issues that have been reported in the literature. We view the guidelines as an initial effort to reconcile the oft-recurring debate between ‘usability’ and ‘learnability’, and hope to encourage the classroom teacher to take a more central role in leading students towards the discovery of ways that may allow them to make better use of help options.
7. Limitations and avenues for further research
Our work in developing the CoDe framework was based on qualitative studies. We consciously sought to pursue an agenda that recognized the learner voice in use/non-use of help options. Specifically, our analyses were grounded in the relatively short interactions of fifteen adult Colombian learners of English with a particular piece of software. Through quantitative procedures, a next step in the development would be to refine and validate the proposed framework.
Once an empirical research design was constructed that identified key variables (e.g., time of access, purpose of access, frequency of access), interaction data with learners and help options could focus on the role of different native languages, different target second languages and different types of software packages. Longitudinal studies, perhaps embedded in a curriculum in which computer-based L2 listening played a big part, could offer opportunities to expand the CoDe framework. Extrapolating the four-part classification of help options to language skills other than listening would be another avenue worth exploring.
In our proposal, we sought to address aspects of visual design and interaction. Whether and how our suggested guidelines contribute to L2 listening development remains speculative. Can learning be made more effective, in measurable terms, on the basis of improved software design? We are cautious of making strong claims, and take note of Garrett (Reference Garrett2009):
“The availability of tools […] does not guarantee that students will, in fact, use them in the way or to the extent that developers intend; only careful structured assignments and follow up work can effectively promote such use” (op. cit.: 277).
To conclude, it is teacher and learner training that is paramount in promoting usable design and fostering meaningful contributions to language learning.
Appendix A
Definition of themes and factors
Appendix B
Study Three and Four: Summary of findings - design outcomes
Appendix C Appendix C
Study Three and Four: Summary of findings-Interaction data
