At the beginning of term, my line manager announced that as he deemed me a more experienced teacher I would no longer have to be observed teaching. ‘Great’ I told myself, ‘I avoid the sack for a while yet.’ However, I then realised that, ‘great’ as it was for me, it was less so for the students forced to undergo my lessons, and thus began a search for alternatives to standard classroom observations that would, nonetheless, allow me to improve my teaching.

Audio-recording of lessons followed by transcribing my teacher talk seemed the most valid alternative and this also coincided with my increasing interest in discourse analysis, notably that of teacher talk.

Persuading my colleagues of the merits of this undertaking was, however, more arduous. A dislike of their own voice, prior knowledge of their poor teacher talk habits, unpleasant memories of their teacher training recordings and metaphors involving dirty laundry were some of the pretexts put forward to explain this reticence.

Yet, the reasons for recording ourselves teaching are numerous. For one, it’s never been cheaper or more banal, and recording through the use of widely available portable devices (Dictaphones, iPods, USB keys) and recording apps on our smartphones is only a swipe or click away.

Another is that teachers’ personal recordings can be unannounced and discreet, then listened to in private and finally deleted before anyone hears of them, whilst the benefits of having listened to oneself and set up an autonomous improvement programme are long term.

A third reason is that our initial discomfort felt at listening to ourselves wanes and over the course of a year, we do notice our own progress.

“A dislike of their own voice, prior knowledge of their poor teacher talk habits, unpleasant memories of their teacher training recordings and metaphors involving dirty laundry were some of the pretexts put forward to explain this reticence.”


With that in mind, my colleagues and I embarked on a series of classroom recordings of ourselves, which we transcribed and analysed. Below are the points of observation that we noted after listening to ourselves and reading our transcripts.

  • Tone of speech. Does the voice sound stressed, relaxed, aggressive, friendly, distant, condescending, intimidated, unnatural?
  • Pronunciation. Is there accent reduction to conform to some perceived standard (common among native-speaker teachers whose accent is unlike Received Pronunciation)? Or over-pronunciation of certain syllables to make oneself clearer? Does a ‘teaching voice’ appear?
  • Repetitions of utterances. Which words and sentences are repeated? At which moments? Is there a pattern?
  • Sentence length and complexity. Do utterances tend to be long or short? Do they contain mostly simple or complex clauses?
  • False starts (starting then restarting utterances). Which utterances, how often and when?  
  • Verbal tics (good, now, right, eh, ahm, well, like). Are they overused? (If your students mock you for them either in front of you or behind you, the answer is yes.)
  • Speech tempo. How many syllables a second in the utterances?
  • Instruction-giving. How lengthy are instructions? How often do you repeat them?
  • Question types. Which frequency of display (answers known by the teacher), referential (answers unknown by the teacher) and higher-order (answers demand more thinking time)?
  • Wait time. How much time between posing a question and inviting a response? Longer wait time gives more thinking time, encourages more students to answer and increases the quality of the response.
  • Teacher talking time (TTT) and student talking time (STT). What is the ratio? Is it consistent with lesson aims?  
  • Teaching time and classroom management time. What is the target?
  • Students’ and teachers’ L1 and L2 use. What is the ratio and what are the targets?
  • Praise. Is it for ability or efforts? How often is it used and what is the effect on students?
  • Names. Which students’ names are used the most and the least? Are they boys’ or girls’? Advanced or lower levels? Badly- or well-behaved students?
  • Classroom interaction patterns. Do they always follow the standard teacher’s Initiation, then student’s Response followed by teacher’s Feedback/Evaluation (IRF and IRE)? How can classroom interaction be varied to increase student participation?

On listening to your recording, you will add to this list your own observations of your teacher talk. Once it’s established, it pays to reflect on the causes of the observed features as addressing them allows us to remedy them more thoroughly.  

In my case, I’ve addressed my stressed tone, which is a reflection of how I feel notably at the beginning of class, by getting into class earlier. This allows me to better welcome the students as they arrive and to sound more relaxed throughout the lesson. Repetition in instruction-giving has been reduced by getting silence before speaking, then by slowing my speech tempo. Wait time for answers to questions has been lengthened by ignoring the first hands up and then counting to five. I intend to ask more higher-order questions, although I can never think of any and my excessive L2 use is owing to laziness, which I’ve done little about.

So now you know what to listen for, how do you go about setting up a recording? Here are some practical suggestions that we’ve used throughout the year.  

Firstly, choose a challenging class. It will provide you with the largest scope for self-betterment.

Second, prepare your smartphone/device to avoid technical hitches. Charge the battery, insert a micro SD to increase memory capacity if necessary and place the phone in an unseen pocket with the microphone towards your mouth so that it picks up your voice.

Third, switch on the recording before class. It’ll give you time to forget you’re recording and you’ll speak more naturally.

Fourth, listen to your recording after class and note your first impressions before transcribing. These are important because frequent listening can make the text too familiar and we then become indifferent to its features.  

Lastly, determine the foci of your analysis, followed by an improvement plan, then another recording later in the year to monitor progress.

Although they can’t replace human observers, recording devices do replay our speech and provide us with access to data, and voice and speech patterns that is otherwise difficult to obtain. Their discreet nature also allows teachers to improve on their teaching talk and they benefit novice and experienced teachers alike.

Bibliography and references


Bauer L & Trudgill P (Eds) (1988) Language Myths. London: Penguin Books.

Coulthard M (1975) An Introduction to Discourse Analysis. London: Longman.

Crystal D (1988) The English Language. London: Penguin Books.

Kelly Hall J (2002) Teaching and Researching Language and Culture. Harlow: Longman.

Lemov D (2015) How to Teach Like a Champion. San Francisco, CA: Jossey-Bass.

Rex LA & Schiller L (2009) Using Discourse Analysis to Improve Classroom Interaction. New York: Routledge.

McCarthy M (1991) Discourse Analysis for Language Teachers. Cambridge: CUP.

Thornbury S (1996) Teachers Research Teacher Talk. International House magazine, International House.

Websites (for average speech tempo) (speech tempo) (for wait time) (increasing wait time) (referential questions) (display questions) (known & unknown self) (for alternatives to Initiation Response Feedback interaction patterns)

Stephen Reilly works at the British Council Paris. A pick-and-mix teacher, he ignores textbook themes, grammar and exam preparation in favour of debates, drama and stand-up comedy. He’s immensely popular with students till they remember they have exams. He’s immensely unpopular with colleagues when they inherit his classes of ‘semi-literate D-grade wannabe performance artists’. He’s permanently unpopular with managers. Contact