Implementing Listening-As-Modelling in the classroom – a report on a 20-week experiment with my year 8 French classes and its impact on their aural comprehension skills

(Co-written with Steven Smith and Dylan Vinales)

1. Introduction

This post describes the broad lines of a twenty-week long instructional programme in listening skills (a pilot study) Gianfranco carried out with two Year 8 French classes as part of his performance management.

We’ll discuss its main findings and implications for teachers. We hope to include reference to this study in the book we are preparing about teaching listening called Breaking the Sound Barrier.

The study investigated whether extensive L.A.M. (Listening-as-modelling) instruction focusing on the enhancement of micro-listening skills would be more effective than the traditional textbook-like approach (i.e. doing topic-based listening comprehension tasks).

The preliminary findings of the study suggested that the L.A.M. programme was more effective than traditional instruction and significantly enhanced the students’ ability to:

  1. read non-words with correct pronunciation in isolation and in short sentences (including mastery of liaison)
  2. transcribe French texts under dictation conditions;
  3. identify parts of speech in aural input and
  4. comprehend aural texts.

One important caveat: being a pilot study, it was not designed and carried out with the rigour of a ‘proper’ research study. Also, its small scale and the opportunistic sampling (Gianfranco simply chose two classes he teaches) means that the findings are hardly conclusive and generalisable. You may find them interesting nevertheless and we’ll mention at the end some of their possible implications.

Also, it should be borne in mind that all the activities below were fully integrated with the other three skills and grammar teaching. So,for instance, the micro-listening and aural vocabulary activities were always staged prior to communicative oral tasks. This required considerable planning effort – another crucial caveat.

2. The rationale and aims of the study

The main rationale behind this study was the hypothesis that MFL teachers may be able to substantially improve their students’ listening comprehension skills through explicit training in the five skill sets below:

  1. Decoding skills, i.e. ability to match letters and letter patterns to L2 phonemes (also known as GPC: grapheme-phoneme correspondence).
  2. Speech-segmentation skills, i.e. the ability to identify word-boundaries.
  3. Lexical retrieval skills, i.e. the ability to recognize words and retrieve their meaning in real time.
  4. Parsing skills (ability to recognize patterns in aural input).
  5. Alertness to and localisation of sound.

To test this hypothesis a L.A.M. instructional programme was devised and implemented as laid out below.

3. Research and instructional methodology

3.1 Setting and participants

The study was carried out at Garden International School, with two mixed-ability groups of (mostly Asian) Year 8 students of French of more-or-less equivalent number, gender distribution and ability (as measured through a T-test performed on the results of the three baseline assessments). Group A (the experimental group) received the L.A.M. training, whilst Group B (the comparison group) was taught the traditional way.

3.2 Baseline assessment

At the beginning of the study, the students were assessed as follows:

  1. Listening comprehension test: the students listened to a text covering vocabulary studied in Year 7 and primary school and answered questions listed in increasing order of difficulty.
  2. Decoding test: students were required to read out nonsense words in isolation and short sentences containing instances of liaison.
  3. Dictation test.
  4. Questionnaire asking how confident and motivated the students were in listening.

3.3 Baseline assessment findings

Listening Comprehension. The students performed reasonably well at the listening comprehension tests in both groups (75 being the mean score for Group A and 74 for Group B).

Decoding test. Students’ ability to pronounce non-words was poor, less than 50 % of the words correct for both groups (Group A 45.4 %; Group B 43.1 %). No student could accurately perform liaison. Issues were identified with the pronunciation of word endings (e.g. es, et), diphthongs (ou, eu, u) and the ability to differentiate between  le/de and les/des, un and une and other words of similar spelling but different pronunciation.

Dictation. Accuracy below 50% for both group (Group A 47 %; Group A 44 %)

Questionnaire. Both groups displayed average levels of motivation and self-efficacy, most of the students choosing the mid-point of the five-point scale adopted.

T-tests were performed on the results of all the four measures the baseline assessments and showed the two groups were equivalent on all measures  (A T-test assesses whether the averages of two groups are statistically different from each other.)

Let’s look next at the tasks done with the LAM group (group A), then the ‘traditional’ group (Group B).

4. The Listening- As-Modelling Programme 

Here is the instructional approach carried out with the experimental group

4.1 Enhancing alertness and localisation of sounds. A student who is alert is more likely to pick up details in the aural input. This skill is often taken for granted by teachers. In this case students’ level of alertness was raised through:

4.1.1 Spot the intruder;

4.1.2 Track the sound / word;

4.1.3 Spot the mistake;

4.1.4 Faulty echo (teacher reading a sentence twice and asking students to spot the mistake deliberately made in the pronunciation the second time).

4.2 Improvement of speech-segmentation skills. Segmentation is the most important skill in listening comprehension. The programme aimed to enhance segmentation skills by:

4.2.1 Aural/writing synergy.  Presenting new language items always in short sentences through a combination of aural and written input. Hence, I chose substitution tables or sentence builders, in the belief that visually processing the gaps in between words in a sentence whilst hearing that sentence being uttered would enhance students’ sensitivity to segmentation. Other tasks through which I attempted to realise this synergy were: (the word synergy may confuse some teachers) Oral ping-pong (see here for explanation) Paired reading aloud sessions of short texts. Break the flow’ tasks – fairly long sentences written without spacing. (The teacher reads the sentence aloud and students separate the words)

4.2.2 Pattern recognition tasks – effective pattern recognition skills facilitate segmenting in that it is easier to assign roles to the words one recognises in the input. Sentence builders, substitution tables, sentence puzzles and other tasks described (here) were used to focus students on patterns.

4.2.3 Sound-discrimination tasks. This is another crucial skill, as often misinterpretation of speech signals stems from misunderstanding similar sounding words (e.g. the classic ‘ship’ vs ‘sheep’). Tasks used in this area were:

  • Minimal pairs. (students must discriminate between two similar sounding words)
  • Spot the rhyming words (students are given three words containing different sounds and must match them with the words the teacher utters)
  • Spot the mistake (students are given a text which the teacher reads aloud; they must spot the words the teacher is mispronouncing)

4.2.4. Parallel text dictations with L1 parallel text. A short text was dictated. To help students segment, the English version of text was provided.

4. 2.5. Gapped texts dictations. Note that gapped words were not chosen randomly; they contained problematic sounds.

4.3 Improvement of decoding skills. The teaching of decoding skills occurred in every lesson for about ten minutes and focused on the following:

  • Silent versus voiced word endings (weeks 1 to 4), through tasks such as ‘spot the silent letter’ or syllable completion tasks.
  • Discrimination between the sounds un, in, en, an (weeks 5 to 10) using ‘minimal pairs’ or ‘spot the rhyme’.
  • Discrimination between oi, ou, eu and u (weeks 11 to 15).
  • Discrimination between eu(x), é, è  and e (weeks 16 to 20).

As already mentioned above, the tasks typically used in this area were:

  • M.L.E. (Micro-Listening Enhancers) such as minimal pairs and syllable-completion tasks and others (see here ).
  • Critical listening tasks were also used before any production tasks (student A reads aloud whilst student B is briefed to pay selective attention to a target sound and provide feedback on its pronunciation).
  • Short dictations.
  • Highly structured oral Interactional activities which elicited production of the target phonemes (e.g. communicative oral drills and find-someone-who with cards ).

Please note: instruction followed my M.A.R.S. approach, i.e.: Modelling/Awareness-raising first, followed by receptive processing, then structured production. In other words, students only produce the sounds in structured tasks after much aural exposure to the target sounds.

4.4 Improvement of memory for words and speed of lexical retrieval. Four sets of aural tasks were used:

4.5 Improvement of parsing skills. Besides the tasks described in 2.2 above, listening tasks requiring the students to categorize words by word-class/part of speech or by tense were carried out in order to enhance their grammatical sensitivity through the aural medium.

5. The comparison group (Group B)

Besides the typical textbook-like listening comprehension activities non-treatment group the comparison group performed full dictation tasks with the same frequency as the experimental group, in order to ensure that they would not be at a disadvantage at post-test, where the four assessment tasks would include a dictation task.

6. Post-test (final assessment)

At the end of the experiment the students were assessed through the same tasks used at pre-test, plus an aural vocabulary matching task:

– Reading out non-words in isolation and in short sentences.

– Dictation.

– Listening comprehension (done how?)

– Vocabulary recognition (20 words uttered at 5 seconds interval to assess speed of retrieval).

7. Findings

  1. Enhanced alertness to sounds – the students have unanimously reported paying much more attention to sounds, putting it down to the baseline test and the micro-listening activities.
  2. Improvements in decoding skills and pronunciation. Statistically significant improvements in decoding skills were observed from pre-test to post-test (from 45.4 % to 83.4). The results were statistically significantly superior to the comparison group (83.4 % accuracy vs 60.5%). The students put them down to the masses of aural input, to the emphasis the lessons put on phonemes and the micro-listening activities, which they did not always enjoy but significantly enhanced their awareness of the differences between similar-sounding phonemes and words.
  3. Improvements in transcription skills under dictation. Statistically significant improvements were found for this task too. The ratio correct to incorrect words rising by nearly 40% for the experimental group (Group A) and only by 20 % for Group B.
  4. Listening comprehension task. The score of the listening comprehension task too showed a statistically significant advantage for the experimental group, although not as marked as in the other two tasks (group A’s mean score 89.2; group B’s 80.3)
  5. Word matching lexical task. The experimental group very significantly outperformed the comparison group (group A’s mean score 88.4; group B’s 69.7). Interestingly, some comparison group students reporting knowing most of the words they heard but being overwhelmed by the time pressure.
  6. Can-do attitude and motivation vis-à-vis listening. The questionnaire investigating this area has not been administered as yet (planning to do it next week). However, from my conversations with my students throughout the process it emerges clearly that they felt much more confident than at the beginning of the process and that they saw a causality between the L.A.M. activities and their enhanced can-do attitude.

Do note that the effect size was high (i.e. the gains were evenly distributed amongst the students in Group A).

Finally, it is worth pointing out that the students reported finding sentence-builders, substitution tables and sentence puzzles as the most useful activities. The activity they enjoyed the most were spot the intruder, spot the mistake oral repetition sparring and find-someone-who with cards. They did not enjoy dictations but they reported learning a lot from them.

8. Concluding remarks

The study concisely outlined in this post is a small-scale pilot with lots of limitations. Having used the same approach with various classes in the last three years, it does however confirm my observation that students do feel more confident and become substantially more competent at decoding L2 graphemes, transcribing and comprehending of L2 input as a result of this sort of training.

The observation can be made, of course, that students get better at what they practice, so it is little surprise that phonological and spelling skills improve when you spend more time on them. More significant, perhaps, is the fact that listening comprehension improved somewhat along with decoding and transcription. An interesting question to ask is: to what extent would comprehension alone improve if an experimental diet without any transcription was tried, i.e. in this case, if Group B had done no dictation?

However, it is entirely plausible that improving decoding and transcription skills does have the added pay-off discovered in this experiment. Even common sense might predict such an outcome.

Does this suggest that teachers should give more time to the types of tasks done with Group A above? Almost certainly. We would suggest that this is particularly the case with beginners and near-beginners. For intermediate and advanced students such tasks can still play a role, e.g. advanced level transcription, gap-filling and reading aloud. At higher levels, however, it is likely that most listening will focus to a greater extent on comprehension and the ability to participate in more sophisticated discussion.

Of note in the study was the fact that not all decoding skill tasks were enjoyed by the students for long periods of time. How significant is this? Well, on the one hand students vary and do not always like the same things. In addition, we would argue that if students become competent at a skill it increases their confidence (‘self-efficacy’) and makes them enjoy learning in the longer term.  Self-efficacy is a key determinant of success. On the other hand, it may make sense to acknowledge the potential limitations of some tasks by working on micro-listening skills in short burst of, say, 10 minutes, preferably incorporated within other meaningful, communicative, two-way activities so that lessons do not become overtly focused on the form of the language at the expense of meaning. There is a potential danger that too much ‘focus on form’ becomes tedious for some students. Games such as the ones described by Gianfranco here may be used as motivational tools to make the focus on decoding skills less tiresome.

One final point: we are aware that some of the skills described above, e.g. segmenting the sound stream, lexical knowledge and transcription skill are developed by other, non-listening tasks. This is a reminder that listening cannot be taught in isolation and depends on the effective development of speaking, reading and writing.


Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s