Task-based language learning and teaching: An action-research study

I Introduction

The growing prominence of tasks and the role they play in facilitating second language learning has been noted in recent task-based language teaching (TBLT) studies (Ellis, 2003; Foster, 2009; Samuda and Bygate, 2008; Van den Branden, Bygate and Norris, 2009; Willis and Willis, 2007). The consensus based on the growing empirical literature is that tasks have positive effects on second language (L2) learning, but less is known about the utility of such tasks by teachers in actual L2 classrooms. Therefore, the purpose of this article is to explore the ways in which TBLT research can be converted into classroom practice by adopting an action research framework.

Action research, first established in the 1940s by Kurt Lewin, involves staging research activities as the following: (1) identify a problem or question; (2) carry out an action; (3) observe and reflect on the outcome; and (4) plan another action (Lewin, 1946). The goal of action research for language teachers is to improve their pedagogical practices by deepening their understanding of students’ learning processes, experimenting with a variety of methodological options, and critically examining and reflecting on their lessons and activities with a view to taking initiatives. Ultimately, action research empowers teachers by giving them a voice to be heard about their classrooms as well as allowing their perspectives to be documented in the field. This approach, which is used here, serves as a productive way of facilitating collaboration among teachers themselves as well as with classroom researchers (Sowa, 2009; Wyatt, 2011).

II The problem examined

Despite the increasingly well-documented theoretical and pedagogical value of tasks, the use of language learning tasks remains a challenge for many teachers because the realities of using tasks in the classroom are not always straightforward. Textbook-provided activities do not always meet the criteria of a task, and tasks that are readily available may not be appropriate to meet learners’ goals and needs. Teachers may design their own tasks, but this poses several challenges in itself; chiefly:

How can a teacher who is inexperienced with tasks create ones that collectively engage learners, appropriately align to their level, and effectively meet shared learning goals?

The issue of setting a task at an appropriate level is a fundamental concern and can largely determine the efficacy of a task. In fact, a practical challenge facing teachers, especially those who are inexperienced, is students’ perception of tasks as excessively difficult (Skehan and Foster, 2001; Van den Branden, 2006). Essential for the success of task-based instruction is the ability of teachers to design and implement language learning tasks at an appropriate level of difficulty so that students can engage with and learn effectively from the materials provided.

This action-research project, namely the experiences of a teacher (the author) exploring tasks, will be discussed. The teacher had no direct instruction in TBLT or the use of tasks in her teaching. Nevertheless, she designed a task in an English for Occupational Purposes course (a course entitled ‘English for hotel and food service workers’), which targeted recent refugees and asylees looking to work in the hospitality and restaurant industries. Classes met for three hours twice a week over a period of 10 weeks. Since none of the learners had prior experience in the food service or hospitality industries, course topics covered workplace routines and etiquette, food safety, housekeeping duties, paychecks and schedules, using the telephone, locating job sources, and meeting job interview expectations.

Thirteen students between the ages of 25 and 56 years old enrolled in the course. Seven of these students were Nepali Bhutanese, three were Ethiopian, one was Eritrean, and two were Iranian. In terms of English language proficiency, educational background, and time spent in the USA, there were notable discrepancies among these learners. At one extreme were students with secondary and post-secondary education from their home countries with an intermediate level of English. At the opposite end in terms of educational experience and English level was a Nepali couple who were pre-literate in their own language, had had no previous formal classroom education, and exhibited extremely limited English skills even after several months of full-time instruction. The majority of students could be classified as high beginners, and most had some literacy issues as well. Even for learners with relatively strong educational backgrounds, none had a first language that was typically written in the Latin alphabet, meaning that decoding letters was problematic. All were enrolled simultaneously in basic English skills courses.

Given the pedagogical challenges described above, if the teacher wished to incorporate tasks into her teaching effectively, she would have to deal with the ecological realities of literacy issues, a multilevel group, and limited educational, occupational, and cultural experience, as well as the typical linguistic horizons that accompany lower level classes. In the following section, examples of how teachers can overcome such challenges by engaging in critical systematic examination and reflection of a task are provided.

III Addressing the problem: Teacher task evaluation as a solution

To address these concerns, the teacher decided to experiment with the development and evaluation of a task. An increasing number of studies evaluating the effectiveness of tasks and TBLT programs involve teachers in the development of tasks (Andon and Eckerth, 2009; Carless, 2004; Ilin, Inozu and Yumru, 2007; McDonough and Chaikitmongkol, 2007; Tavakoli, 2009). This step is deeply important if, as noted by Edwards and Willis (2005), tasks are expected to be widely used and accepted in the classroom. More crucially, opportunities for teachers to revise task materials may explain the relative success of one program over another (Ellis, 2003). Furthermore, retrospective evaluation can help teachers determine whether the materials can be used again, which activities work or not, and how the materials can be modified so as to increase task effectiveness in future lessons (Ellis, 1997). With this background in mind, the question becomes: How would a classroom teacher design and evaluate a task?

4 Evaluation of Task 1

The outcomes of Task 1 were tallied from the handouts collected immediately following the task cycle, compiled, and then analysed in comparison to the intended closed outcomes. The results are summarized in Table 1 below. All student names are pseudonyms.

OPEN IN VIEWER

Table 1.

Response-based results of Task 1.

Participant name	Participant level	Responses (numbers correspond to position)	Four duties correctly identified?	Duties divided correctly?	Ratio of correct responses to total given
Gagan (1A)	Low intermediate	2, 3, 4, 7, 10	no	no	1/5
Kebedech (1B)	High beginner	5, 8, 9, 11, 14	no	no	1/5
Bijay (2A)	High beginner	13, 12, 5, 14	no	no	1/4
Amita (2B)	High beginner	4, 6	no	no	0/2
Parto (3A)	High beginner	1, 4, 10	no	no	0/3
Kalavati (3B)	Beginner	9, 12, 2	no	no	0/3
Asnakech (3C)	Intermediate	9	no	no	0/1
Bhupen (4A)	High beginner	12, 14	yes	yes	2/2
Muluwork (4B)	Intermediate	10, 11	yes		2/2
			Total		7/27 (26%)

The success of the task outcomes was measured in two ways. First, were the students able to discern the four unfinished duties correctly given the information they had? Second, were the students able to correctly divide the unfinished duties between them without overlap? Table 1 shows that only one pair of students were able to complete the task accurately (i.e. determining the correct number and nature of unfinished duties and by dividing them). Discussion with the classroom assistant, an intern from a local high school who was present that day, revealed later that there was potentially some interference with the data for this exceptional pair. She describes her assistance to them by saying, ‘I explained it … twice, and you explained it … three times … I sat down with them and basically said, “This is what you would do and this is what you would do”.’ The data, however, remain in the analysis since it is impossible to determine precisely whether or not the nature of her assistance could be considered as providing the correct answers, or merely explaining and scaffolding, something which is a rather subtle distinction.

It should also be noted that, for similar reasons, data from the low-literacy pair of students were excluded entirely from this study. Answers were provided to them directly by the assistant and they merely copied the necessary words. For the other pairs of students, the few responses that matched their targets may also be interpreted as mere chance, since no pair was even able to identify the correct number of responses. In fact, the percentage of total correct responses for these other students did not exceed 25% in any single case. Allowing for the possibility that Pair 4 did in fact successfully complete this task of their own accord, a generous measure of the overall response-based success rate of this task would be 26%.

The questionnaire responses showed that for Task 1, three out of five comments displayed an overall negative attitude towards the task. Parto, for example, wrote, ‘I feel very hard. This activity must be approved’, by which it can be inferred that she meant ‘improved’. Of the remaining two students who submitted responses, one seemed to blame himself for his dissatisfaction with the task: ‘Becous you are good theacher If she study to me I am understand … Becous It is not hard or very hard when I am hard stadie it is very easy [sic].’ The only somewhat positive response was ‘the [father] is smileing Peace [sic].’ After careful examination of the data, the teacher’s and assistant’s initial conclusions were that the students struggled with the task and felt significant frustration. The difficulty of the task is illustrated in the following interaction between the teacher and her students:

T:

… No it’s not finished. Did you clean the bathtub?

Gagan:

Cleaned the bathtub.

T:

You did not clean the bathtub. [To other student] Did you clean the bathtub?

Kebedech:

Yes.

T:

No.

Kebedech:

No no no …

T:

Kebedech, if there’s a check, you did it …

In this interaction, the students’ willingness to engage is apparent, but the teacher and students are both frustrated by the students’ inability to understand. The teacher reverts to explaining a key concept (the function of a check on a checklist) again, realizing that this may be a barrier to effective task completion.

IV Critical examination and teacher reflection: Task 1

By all accounts, Task 1 was far from a success. Students were not able to accomplish the task set before them, reported negative attitudes towards the task, and even exhibited possible emotional barriers that may impede language learning. This latter issue is extremely common for this particular population. One could easily see how a teacher unfamiliar with TBLT, yet making a first foray into task design and implementation, might conclude that tasks are not ultimately appropriate for her students, or even that tasks in general are not an effective means of instruction. A less likely but even more lamentable result might be that a teacher could refuse to modify her teaching and dogmatically insist on the use of such tasks, leaving her students feeling frustrated and pushing them out of the learning process rather than inviting them into it. In the action research context, reflection and modification become essential, thus leading to more effective solutions.

In the process of reflection, the teacher examined the results, discussed them with more experienced colleagues at her school, and examined relevant materials such as Ellis (2003) and Willis and Willis (2007). An analysis of literature was also conducted into issues of task complexity (Foster, 2009; Pica et al., 2009; Robinson, 2001; Robinson and Gilabert, 2007; Skehan, 2001).

In light of this particular classroom context, the most relevant revelation is a consideration of various factors that influence attentional demands on the students. Skehan and Foster (2001) divide these factors into three main areas: (1) code complexity, which refers essentially to the difficulty of the language; (2) cognitive complexity, which relates to the cognitive demands placed on participants to understand and process the task input and to perform task requirements; and (3) communicative stress, which refers to features of how the task is expected to be carried out, such as within a time limit or without one, that make it more or less difficult to complete. The second factor (cognitive complexity) is of particular concern to us and thus features largely in the barriers described in Figure 1. That is to say, the amount of cognitive load in Task 1 made it difficult for the learners to focus on linguistic factors due to their limited attentional capacity.

OPEN IN VIEWER

Figure 1.

Perceived barriers to successful task completion.

From this analysis, research and reflection, five general categories of factors that posed barriers to effective task completion were outlined in Figure 1. These include code complexity, cultural and schematic barriers, students’ task familiarity and orientation, the teacher’s role in the task, and cognitive difficulties. Code complexity, i.e. the linguistic difficulty of the task, clearly cannot and should not be eliminated entirely or reduced excessively, or the purpose of the task will be defeated; however, there are certain aspects of the language that posed undue difficulties on the learners. The checklist, for instance, was 14 items, a number which is unnecessarily high for low literacy and beginning learners and which creates a rather untenable amount of vocabulary words to retain. Furthermore, additional practice with the language itself prior to beginning the task would have prepared students to better understand the vocabulary involved, thereby freeing their attention to devote to other aspects of the task.

In terms of cultural and schematic difficulties, it became apparent from reflection that students had struggled with the unfamiliar concept of a checklist. This might not be a problem with learners who have strong educational backgrounds. For the limited or interrupted education of these learners as well as their limited literacy, however, we cannot assume a square and a crooked line through it will not automatically be interpreted as ‘finished’. Furthermore, ‘to do’ lists were a concept that were also not likely to be familiar for low-literacy learners who, even if they were familiar with goal-setting practices, were not likely to write their goals down and mark them upon completion. Another cultural barrier that may have played a role was that learners were not necessarily predisposed towards assigning individual chores once the uncompleted chores were determined. Rather, they may have wished to simply work together on whatever needed to be done, allowing this process to take place in a more organic fashion.

A lack of task familiarity also appeared to be an issue and was one mentioned in the literature with far more educationally experienced learners in mind as well (see, for example, Ellis, 2003: Chapter 8). Learners in this situation, in which even the teacher had had limited experience with classroom tasks, clearly did not have a mental concept of how this type of classroom activity was meant to progress and conclude. Some learners, more familiar with exercises of this nature, began simply to read the checklists aloud. Of particular concern was the conceptual difficulty posed by the teacher’s use of dual-sided cards because they were intended to ensure that a gap was present (i.e. learners could not look at both sides of the card at the same time), but instead became a major source of confusion.

The teacher’s role in preparing students for the task and guiding them through it is an important one to examine. It was found that pre-task activities were insufficient and ineffective, leaving students with too little opportunity to learn both the cultural concepts and minimum language required for the task. The modeling activity, while useful in theory, was poorly executed since students were unable to see what the teacher was doing with the assistant and instead only listened to an interaction. This was likely to confuse rather than help them. Moreover, although the teacher designed the checklist in order to ensure a genuine information gap, prevent cheating, and afford opportunities for easy task repetition, this design decision made it more difficult for her to quickly assess the accuracy of students’ results because each pair’s checklist was different. Her role in facilitating the task was, therefore, less effective.

The last major barrier to task completion worth considering was cognitive difficulties. One difficulty that was already mentioned was that the number of items on the checklist was unnecessarily long; however, the fact that these items were unnumbered merely contributed to the difficulty of coping with this list. Furthermore, we found that for several reasons the design of the task increased the cognitive difficulty and communicative stress that these learners faced rather than moderating it. First, the dual-sided card was unfamiliar to the students, as mentioned earlier. Second, the students had to read each list twice – once on the card and once on the handout in front of them – and had to deal with multiple copies of the checklist (four for each pair). Third, the students’ job of transferring the information from the card to their handouts was a difficult task in and of itself, and one that was not explained to students clearly or even explicitly required of them. This excerpt from Task 1 demonstrates the difficulty in transferring this information, even when the choices were relatively limited, as they were at the heading of the handout.

This exchange reveals that even determining the codes used by the teacher that were intended to simplify the task (room numbers and ‘A’ and ‘B’ for each partner), merely complicated the task for these overtaxed students. In short, students were not able to assess their own achievement, and felt unsure if they were accomplishing the task correctly or not.

V Action: Modification of the original task

After critically examining and reflecting on the outcomes of Task 1, a number of modifications were made to address the difficulties described in Figure 1. The changes are summarized in Figure 2. For example, the additional scaffolding activities were included to (1) familiarize students with the cultural concepts of checklists and checks, (2) increase familiarity with the linguistic items on the checklists, and (3) provide students with more practice and preparation in describing finished and unfinished housekeeping duties as well as the process of negotiating and dividing these duties.

OPEN IN VIEWER

Figure 2.

Modifications of Task 1 made in Task 2.

To reduce these barriers, the ‘to do’ list was removed as it posed one more unnecessary cultural/linguistic element of the task to be understood. The students were also each given handouts that they were allowed to compare side by side if necessary. This essentially halved the amount of information that they were required to comprehend in order to understand the task and prevented errors in the process of transference. Although this may have resulted in less spoken language between participants, there are linguistic (reading, negotiating division of labor), cognitive (identifying, interpreting, comparing) and cultural (checklists, teamwork, polite negotiation of and/or volunteering for duties) challenges that learners still cannot circumvent even if the two papers are placed side by side. This modification also gives students greater control over the level of difficulty they are able to manage with the task; the more capable and ambitious students may not look at each other’s checklists in order to complete the task while less advanced students can benefit from having the option of doing so.

Another modification involves the decision to replace the modeling activity as a performance of the task between the teacher and her assistant with the use of a whole-class sample task completion. As described in the task cycle below, the teacher wrote one checklist on the chalkboard in large letters (‘Housekeeper A’s checklist), and guided students through actually doing the ‘finished’ chores before checking them off. She then distributed copies of ‘Housekeeper B’s checklist’ to the students and guided the students through a version of the task extremely similar to the paired information gap. Whereas students were passive listeners in Task 1, unable to connect what they were hearing to the small-sized papers before them, in Task 2 the idea was that they would be able to participate, experiment linguistically, or at the very least better observe the actual process of the task completion in order to better understand it. The advantages of this kind of activity were echoed by Prabhu (1987) amongst others as an important pre-task means of scaffolding the partner task for learners. Whole-class modeling is, in addition, a means of addressing the lack of task familiarity. Embedding the task in a realistic situation of ‘finished’ and ‘unfinished’ tasks (e.g. having students actually vacuum the floor, etc.) maintains a sense that some real meaning is being communicated and also reduces the likelihood that learners will merely treat the task as an exercise. Furthermore, it gives learners a better understanding of information gap activities as a means of language learning since they are not very likely to have prior experience with these.

2 Evaluation of Task 2

As shown in Table 2, Task 2 produced positive results in that the students were able to determine which chores had and had not been finished in each room. In every pair, even the group in which there was one significantly low-literacy student, all students were in fact able to locate the unfinished tasks. For the second layer of the task, it appears that only one pair of students (Pair 3) out of six failed to complete this portion of the task accurately. These results suggest that the task was aimed at a more appropriate level for these learners than Task 1, and may even have been easy. Combining the two aspects of the task into 12 total parts, two for each of the six pairs, we can consider the overall response-based success rate of this task to be 91.7%.

OPEN IN VIEWER

Table 2.

Response-based results of Task 2.

Participant name	Participant level	Responses (numbers correspond to position)	Four duties correctly identified?	Duties divided correctly?	Ratio of correct responses to total given
Gagan (1A)*	Low intermediate	6, 9	yes	yes	2/2
Asnakech (1B)*	Intermediate	2, 5	yes	yes	2/2
Bina (2A)	Beginner	5, 6	yes	yes	2/2
Bijay (2B)*	High beginner	2, 9	yes	Yes	2/2
Amita (3A)*	High beginner	2, 9	yes	no	2/2
Kalavati (3B)*	Beginner	7, 9	yes	no	2/2
Bhupen (4A)*	High beginner	2, 6	yes	yes	2/2
Ganesh (4B)*	True beginner and low literacy	5, 9	yes	yes	2/2
Parto (5A)*	High beginner	2, 6	yes	yes	2/2
Kebedech (5B)*	High beginner	5, 9	yes	yes	2/2
Afsaneh (6A)	Low intermediate	6, 9	yes	yes	2/2
Muluwork (6B)*	Intermediate	2, 5	yes	yes	2/2
				Total	24/24 (100%)

Note. Possible correct responses-same for all pairs: 2, 5, 6, 9; * Indicates student had been present for the original task.

In addition to the improved accuracy of responses, the student-based evaluation as measured through the brief questionnaires immediately following the task also demonstrated greater student perceptions of enjoyment and success for Task 2. From the teacher’s and assistant’s perspectives, this task seemed to go much more smoothly and the mood in the classroom seemed to remain lighter throughout. The students’ own words corroborate these views; all seven questionnaires submitted marked the task as either ‘good’ or ‘very good’ (note: the actual results are not included in this article due to space constraints, but are available upon request). Open-ended comments also demonstrated positive feelings and a subjective sense of learning; one student wrote, ‘I teach some new sentence and new [words] and I become brave for conversation … my class is very good because It isn’t bouring and It is very active and fun! [sic]’ Another wrote, ‘I’m happy For practice on the class … I am soon learning in the class … I think activity was good [sic].’ One of the more advanced students marked the task as ‘very good’ and compared it specifically to the first task: ‘[T]he second time experians and the team work is good [sic].’

VI Critical examination and reflection: Comparison of Task 1 and Task 2

The results from the response-based and student-based evaluations show that Task 2 was a more effective and more preferable task than Task 1 because (1) student responses in Task 2 showed at least 66.7 % greater accuracy, (2) the task was completed much more quickly (13 minutes as opposed to 32 minutes), and (3) the students as well as the instructor had much more positive perceptions of the second task than the first. These differences are summarized in Table 3. Using these factors as criteria, the concluding impression is that the modifications and additional activities implemented between Task 1 and Task 2 were effective in making this task more accessible and enjoyable for learners.

OPEN IN VIEWER

Table 3.

Comparison of results for Task 1 and Task 2.

	Time for Main Task completion	Response-based evaluation	Student-based evaluation
Task 1	32 minutes	26% accurate responses	3 out of 5 comments were negative; no positive comments
Task 2	13 minutes	91.7% accurate response	7 out of 7 comments were positive
Difference between Task 1 and Task 2	Task 2 was accomplished in 19 minutes less time	Task 2 produced 66.7% more accurate responses	Task 2 received far more positive student responses

While not all modifications can be directly attributed to this positive change, certain modifications showed evidence that the items in Task 2 were more effective. A good example is the modification that removed the double-sided cards and allowed the learners to compare their checklists side by side. At first the teacher was rather reluctant to make the change as she feared it could eliminate the required element in the information gap activity. However, this modification to the original task was possibly the only procedural element that made the task comprehensible for some students, as observed in this excerpt from Task 2:

It seems that without this ability to visually discern the differences in the checklists, the concepts behind the task may have been inscrutable to these students. While it is likely that a more advanced group of students might find double-sided cards motivating, this original design feature was found to be an inappropriate in this context.

Another modification that the teacher hesitated to make was adding numbers to the checklists because she feared that the students’ focus would shift away from the target vocabulary to the far easier vocabulary of numbers. After the numbers were included in Task 2, these fears were observed to be true in some cases, but not nearly in every circumstance. As proof, learners from three different pairs in the process of completing the modified task (Task 2) are quoted below.

The language used may not be grammatically or pragmatically correct, but learners are able to draw from a variety of linguistic resources to convey meaning and get the task done without having to resort to numbers. Furthermore, regardless of whether or not the linguistic forms used were limited to numbers in some cases (arguably still a useful set of lexical items for beginning learners), numbering was seen as essential in reducing the cognitive load that beginning learners face in completing the task. Therefore, this added feature in Task 2 was important in making the challenge manageable.

VII Conclusions

The creation, implementation, and evaluation of these two tasks (Task 1 and Task 2) bore significant benefits for both teachers and classroom researchers, namely providing a greater understanding of learners’ needs and limitations and how best to address those needs. The study also highlighted ways in which teachers can make tasks more accessible and enjoyable for students. Tasks can serve as a means of incorporating enjoyable, beneficial, and communicative activities in the L2 classroom, as Task 2 attests. Task 1, on the other hand, can serve as a caveat to teachers to carefully develop and test their tasks, and to consider all of the possible schematic and cognitive limitations faced by their learners.

The contrast between these two tasks allows broader conclusions to be rendered. First, if a single task does not appear to succeed with learners, teachers should not necessarily conclude that the task is an ineffective means of learning. This is because several changes in the task structure, input, and implementation may be required to better address the specific learner population involved. The process of modification exemplified in this action research and its documented success can also serve as inspiration to teachers who may not otherwise be inclined to create their own tasks or improve upon existing tasks, or who may need examples of what modifications are required.

While the process of creating and evaluating one’s own tasks may be overwhelming, it can also be valuable and rewarding for teachers and learners alike. This action research demonstrates these benefits by reporting how the teacher’s critical systematic examination and reflection enabled her to modify a task-based pedagogical activity to improve its effectiveness. The teacher’s successful implementation of the modified task also led to positive changes in how she viewed task-based teaching. To conclude, this article serves as an example of how teachers can create their own tasks and of the importance of evaluating them empirically. In so doing, it also highlights the importance of action research as a means by which language teachers can address problems that arise in a TBLT lesson and, more generally, develop their reflective skills.