Enhancing Oral Academic English as a Medium of Instruction (EMI) through self-regulated Listening to Authentic Educational Podcast

Research questions

The following questions were formulated to guide the inquiry:

• Does the implementation of self-regulated listening practices with authentic lecture podcasts significantly improve EMI teachers’ use of instructional discourse functions?

• Does listening practice have a significant effect on various discourse functions among EMI teachers or within the EMI classroom?

Educational context

In Ethiopia (where this study was carried out), English is taught as a foreign language from grade one. Learners’ L1 (Amharic in this specific setting) is used as a medium of instruction in grades 1-8. The use of EMI starts in 9th grade and continues to serve this role in higher education. Students across grades1-12 take 180-200 minutes of English lessons per week. The student population in the country has barely any exposure to the English language outside the school context. This lack of exposure is largely taken as a major flaw in the competence involving the English language (Fuad et al., 2019. Bartley, 2014. Ministry of Education, 2017 ). There is a widely shared consensus that the efforts made in EFL classrooms are not enough to train learners in using the language for the required academic ends. Accordingly, there is a pressing need to provide more exposure to target language use by the language proficiency of content area teachers ( Tafesse, 1998. World Bank, 2013 ).

Research design

This study employed a mixed research design in which a combination of experimental intervention and exploratory tools was used.

Participants

The participants in this study belong to a population of content area teachers in upper secondary schools (grade 11) of suburban communities in Ethiopia. Multi-phase cluster sampling was used to draw 40 teachers from six government schools. During the first phase, the schools were selected using simple random sampling. Then, the content area teacher population for eleventh grade (168) of the schools was classified into two discourse communities, namely natural sciences and social sciences. 28 teachers from each stratified group were selected for a total of 56. This selection was followed by further screening on the teachers’ competence by means of an observation scheme on bundles of English language skills for EMI purposes adapted from Flowerdew (1994) and James and Philip (2000). The selected teachers were observed in five instructional sessions (1 h each). Based on this assessment, 40 teachers (18 from social science and 22 from natural science) were found to be significantly deficient (below 30% of the expected performance). These teachers were divided into control and experimental groups. Afterwards, group equivalence between control and experimental groups was tested. As shown in Table 1 , the independent sample t-test results on the performance of the two groups indicated that they were at an equal level of competence (mean difference: 1.5%; t, 12, p, 0.683).

Selection of listening resources and piloting

The listening resources were drawn from MIT (Massachusetts Institute of Technology) courseware; they were a stock of lecture podcasts of several academic disciplines. Resource selection involved two stages. In the first stage, an assessment was made to select lecture podcasts with reasonable level of comprehensibility and richness in relevant academic language functions. Two lectures for each academic discipline (physics, biology, geography, and civics) were selected in this stage of screening. In the second stage, the comprehensibility of the listening resources was assessed (as a piloting process) using listening content comprehension tests. Finally, those with no comprehensibility barrier, yet rich in EMI language tools (introduction discourse functions, definition, explanation, hypothesizing, questioning, and discourse markers) were selected for the final listening practice.

Training on self-regulated listening

The experimental group participants were trained on self-regulated listening and received a brief description of the content of the podcasts. The training familiarized participants with detailed actions of the five phases of SRL: 1) task description, 2) goal setting and planning, 3) strategy enactment, 4) monitoring, and 5) metacognitive adaptation. They were also trained to have a selfregulated listening process log. The log consisted of a set of templates for the participants to record their listening activities and track their progress in attaining their goal (discovering and using the target language functions). Finally, all the actions of the process were modeled and demonstrated by the researcher to the experimental group participants. Once they received the training, they told to listen to it for three weeks, and their performance after the SRL was assessed through observation tools.

Data collection tools and procedures

The study employed an observation scheme and an in-depth interview as major tools for data generation. The observation scheme was used to assess the target language competence of teachers across control and treatment groups (both in the pre-treatment and post-treatment sessions). The observation scheme was constructed based on Flowerdew (1994) and James and Philip’s (2000) model for assessing oral competence regarding EMI. The scheme takes macro discourse functions (introduction, explanation, definition, hypothesizing, questioning, and discourse markers) along with realizing micro language tools as targets of the learning outcomes. Apart from these structured schemes of observation, videotaping was used to capture the instructional moves of teachers.

In scoring the teachers’ performance, each of the macro discourse functions (lesson introduction, explanation, questioning, etc.) was taken separately for assessment. The scoring primarily considered whether the macro discourse exists in the teacher repertoire. Furthermore, it rates from 1-100% whether the micro language tools are appropriate in word/phrase selection, combination of words/ phrases, grammatical accuracy, and overall fluency regarding the specific communicative purpose. The observation was piloted in the preliminary stages of the study, and an inter-rater score made by three EAP professional teacher educators was found to be reliable (r = 0.86).

Finally, an in-depth interview was held with the experimental group teachers on their self-regulated metacognitive journey of exploiting the listening resources. During the interview, the teachers were told to 1) reflect freely and extensively on what they did in the five stages of SRL, 2) describe their cognitive processes in the course of this process, and 3) describe what they attained or failed to attain. The interview was conducted in their L1 (Amharic) as well as audio-recorded.

Data analysis methods

An independent sample t-test was employed to analyze the quantitative data on the teachers’ performance throughout the observation scheme. This was cross-referenced with qualitative data on metacognitive processes related to SRL. These datasets were coded following the five moves of self-regulated listening practice and analyzed thematically.

Language gains in operative academic discourse functions

The operative domain of language use takes up the larger space of content instruction. It is the component in which teachers detail their instruction through definitions, elaborations, comparisons, etc. As such, it requires a number of discourse moves to fulfill the instructional goals depending on the nature of the lesson content and outcomes. In the observed lessons, the teachers took an informing stance as a way to organize their thoughts and to fulfill their instructional goals. Definition, elaboration, and hypothesizing were employed to this effect.

Table 3 shows the level of success at employing these discourse functions for the desired goal.

The results on operative instructional language use demonstrate that the intervention group teachers significantly benefited from the self-regulated listening practice across discourse functions regarding definition (t: 2.14, p < 0.05), elaboration (t: 2.78 , p < 0.05), and hypothesizing (t: 2.03, p < 0.05). The impact was also evidenced by a significant effect size across the three operative discourse functions (r > 0.06).

Defining

To mention illustrative evidence of this impact from the instructional episodes, the teachers in the post-intervention instruction sessions used definitions with syntactic sufficiency and variety.

… Land rehabilitation is a biological process of returning a damaged land [yetegodameret] to its former condition. (Geography, episode 1)

… a pulley is a wheel that carries a rope or cable to supports movements… (Physics, episode 1)

… Monosaccharaides are forms of carbohydrates that exist as independent or connected units of sugar. (Biology, episode 1)

These definitions produced by teachers after the listening practice constitute the syntactic components of scientific definitions (concept + category + specifier). While these components are visible, the teachers are also able to express the specifying features through varying linguistic tools involving relative clauses, adjectives, reduced relatives, and appropriate tense use.

It is important to note that the teachers made very limited code switching and judicious translation in defining the terms. Nevertheless, this is significantly different from similar definitions given in the pre-listening time of their instruction, where the teachers mostly defined terms in the L1 or with a frequent switch to L1, failing to show learners how to define concepts in full English sentences.

Elaborating

As an extension of the definitions, the teachers employed the discourse function of elaboration quite effectively in the post-listening period. This impact is generally evidenced by the teachers’ use of the target medium of instruction (English) to restate or re-present concepts, to provide more details through examples, and to make the elaborated element more precise by means of particularizing, summarizing, evaluating, etc.

Furthermore, as a way to ensure comprehension, the teachers used short declarative sentences with a reasonable degree of accuracy rather than switching to the L1. As the teachers extended their explanations towards task elaboration, they selectively used L1 equivalents of key meaningbearing words. In addition, they reread instructions from the textbook in order to highlight the content of the task and the directions to be followed in solving problems. This is a well-reasoned component of the instructional process largely targeting content comprehension and avoiding misunderstanding of key points in the tasks.

One of the teachers in a geography class explained:

[… Bezih ymelemameja project wust yemabrarut… beafrica ahgur wust… be tefetro habt atekakem ena bezelaki limat… kinwun mekakel yalewn kurgnt new]

Note the instructions of this project. In it, you are expected to show what is happening in Africa with regard to the use of natural resources and sustainable development. (Geography, episode 2)

Hypothesizing

Another set of operative language tools internalized by teachers is hypothesizing. These linguistic tools are lexical bundles and word starters (for example: let’s think/say/assume/imagine, (so) what would happen (if), what will happen if, what happens if) that help teachers express interpretations about a practical situation or condition through simple imaginary situations. The instructional repertoire of teachers in the post-listening period of their instruction largely demonstrated proper use of these tools. Particularly, unlike the pre-listening time of their instruction, the teachers employed these tools to make abstract and complex concepts accessible to learners.

In a physics lesson, while explaining the concept of acceleration (a rate of change in the speed of an object), a teacher employed hypothesizing as a simplifying tool:

. let us assume that a car travels from here (the school compound) to the church compound at 60 km/h for 5 minutes; from the church compound to the gas station at a speed of 70 km/h for 10 minutes; and from the gas station to the market at 100 km/h for 12 minutes… Now we can see that speed of the car changes three times between the school compound and the market… Now let’s calculate the acceleration from these changes. (Physics, episode 2)

Imagine the parliament makes a law that contradicts… emm… conflicts with the constitution? [kehigemengistugayemigachhigbiawota]… What happens in such conditions… Can you predict … [minyemifeteryimeslachihual]?… the house of federation will … eemm repeal it [… yiserzewal] … This is check-and-balance. (Physics, episode 3)

The instructional repertoire in these excerpts demonstrates that the teachers smoothly demonstrate hypothetical conditions to explain complex concepts, namely acceleration in physics and check-and-balance in constitutional principles. Of course, the teacher in the civic class switched to the learners’ L1 quite frequently. However, his ideas were largely repetitions of what he had already stated in English, and he did it as a judicious move to ensure comprehension.

Questioning

Questioning is one of the primary instructional competences that teachers are expected to possess in EMI. It is also a competence that teachers are expected to pass onto their students as a linguistic tool for inquiring about disciplinary knowledge. However. it is one of the EMI competences that remained undeveloped among the teachers participating in this study.

Table 4 shows that the teachers made only limited improvement with regard to questioning in the target language (t: 1.03, p > 0.05, r < 0.06) after the listening practice. A closer examination of their instruction processes showed that their teaching required different forms of questioning: information-seeking questions (asking for unknown information), rhetorical questions (where a response from the audience is not expected), display (seeking verification of student knowledge), and eliciting questions (asking for explanation and reasons).

However, these question types were not adequately employed in the instructional process. Some of the questions (for example, informationseeking and verification questions) lacked the appropriate question or auxiliary verbs.

…. What are *we say* about the characteristics of this material… em, em… in previous lesson? [In other instances, the auxiliary verb ‘did’ is missing.]

*You remember the formula …? How *we solve the problem in last lesson? (Physics, episode 4)

In eliciting questions, the teachers mostly started the question in English but quickly switched to the learners’ L1.

Why is … [lemindinewyihchigiryemifataraw] [why does this problem occur?] [pause]... Because the chemical decrease the air circulation in the soil…

How people…

[hiizbochyepoletikawukilinayemiagegnutendet new?] [how are people politically represented?] [pause] They are represented by election (Geography, episode 3)

In these elicitation questions, the teachers barely used English to state their commutative intent. Of course, switching to L1 in such questions may facilitate content comprehension, but the way in which the teachers switched does not constitute a move to overcome comprehension barriers. It is mainly due to their undeveloped grammatical proficiency at asking that kind of question.

Finally, it is worth highlighting that, unlike other types of question, the teachers demonstrated a better improvement in rhetorical questions:

… a lot of gas emission comes out of the industries… but the Americans government cannot stop the gas emission because that damage their economy. Is the effect bad or good for the world? [no pause] It is bad, isn’t it? (Geography, episode 3)

As we discussed in the previous lesson, democratic institutions have different functions… What are the democratic institutions that make laws? [no pause] These are law making councils such as the house of representatives and regional councils. (Geography, episode 4)

To summarize, apart from these linguistic features related to rhetorical questions, the teachers’ learning gain for the larger forms of EMI questioning was limited.

Using discourse markers

Discourse markers are organizational signals through which teachers demonstrate threads of relationships between units of ideas in their instructional talk. As such, they are linguistic tools that shape the information structure of instructional discourse and enhance learners’ content comprehensibility (Chaudron and Richards, 1986; Thomson, 1994; Smit, 2006 ). Teachers’ use of these tools in English across academic content classes also means giving learners the opportunity to discover these tools for organizing disciplinary thought in oral academic communication.

The participant teachers in this study demonstrated a marked improvement (t = 2.84, p < 0.05, r > 0.06) in using these tools following the self-regulated listening practice. Their gain in these tools was demonstrated both in their use of macro and micro discourse markers. Through macrodiscourse markers, the teachers prepared students for the upcoming discourse, demonstrated how the topic is related to other parts of the instruction, and used topic shifters with a smooth flow of thought and prosody. All these were done in English with reasonable accuracy and rare switching to the learners’ L1.

… You remember our discussion on causes of global warming. Today we will be discussing one of these causes… That is industrial waste… so much about industrial waste… Now, we come to another cause… (Geography, episode 5)

As for micro-markers, a marked improvement was demonstrated by the teachers in signaling temporal, causal, contrastive, and appositive relationships between ideas while maintaining the internal cohesion of their instructional discourse.

Metacognitive evidence of self-regulated learning

So far, the reported results have been largely generated from the observation scheme on teachers’ instructional performance as the result of the listening practice. To cross-validate the data, an in-depth interview was carried out on the teachers’ metacognitive processes. The data obtained substantiate the results, revealing the metacognitive moves leading to the learning gains. As such, the teachers’ listening practice was guided by the steps of a self-regulated learning process: forethought (task description, goal setting, and strategic planning), strategy enactment, monitoring, and adaptation (Greene and Azevedo, 2007).

Evidencing the first two steps, one of the participants stated the following:

... as the researcher told me about the listening practice and the nature of the texts, I decided to take useful language structures that helped me to communicate ideas in my teaching… Before I came to the listening, I had a lot of problems of English language use in teaching my subject from the beginning of a lesson to the end... so, I went into the listening with this in mind to improve myself… and I concentrated throughout on those useful phrases to take into my mind. (Participant 1)

This remark suggests that the participant teachers moved proactively to exploit the listening input in order to correct their deficiency in oral academic language. Particularly, it demonstrates the actions of the participant teacher in the first phase (forethought) of self-regulated learning. By looking into specific actions of the participant teacher in this phase, it could be observed that he made a task analysis based on the briefing from the researcher on the content of input. Once he knew the content of the input (an authentic lecture podcast), he set the goal of picking language tools which were helpful for his teaching practice. Furthermore, he devised strategies of focused attention and concentration in order to discover these tools. This remark also shows the drivers of self-regulated learning such as self-motivation and beliefs, which are reflected in the form of outcome expectations, intrinsic interest, and goal orientation, which sets the stage for selfregulation of cognitive processes in exploiting the listening input.

Further evidencing the role of meta-processes in the learning of the discourse functions, another participant teacher reflected:

… As I listen to the lecture, I try to discover phrases useful for introducing my lessons: those that help me to explain concepts in different ways, how to ask questions to students and how to restate ideas, how to move from on portion (point) to another, etc. I discover useful phrases for each of these purposes… once I find them, I silently try to use them and rehearse them repeatedly… Then, I listen to the lecture again and I check how much I have mastered the use of the phrases… (Participant 2)

This participant’s reflection indicates that she monitors her cognitive moves in processing the listening input with clear statements of what she seeks to achieve and what she has done as a way to attain this end. Furthermore, her monitoring actions continue with her attempts to track her progress through meta-tools of self-experimentation, selfjudgment, imagery, and self-instruction.

Yet another instance of self-regulated metaevidence related to selective attention and effort management is reported by another teacher participant:

I was selective in what I concentrate on in the listening. I decided this step-by-step. In the first listening, I simply survey the language tools in the listening text. I could see in the first listening that the lecture has language tools of introducing a lesson, explaining concepts at greater depth, asking questions, making transitions, and concluding a lesson, yet I see that not all of them are necessary for me. For example, I don’t have a problem in introducing a lesson. Also, I can conclude a lesson quite effectively. My problem areas are explaining and asking questions in English correctly. So, [during the listening], I focused on the latter language tools to improve my language for teaching. I did this in the next listening I made, and I could master these phrases. (Participant 3)

This reflection again shows that the participant teacher did not rush into goal setting before he went into the listening. He had to characterize the input through a survey in a first-round listening. Then, he set his goals (mastering discourse functions regarding explanation and questioning) and planning strategies to attain this. Selective attention was particularly employed to achieve this goal.