We start by considering which elements of speech can be evaluated, and which cannot - or should not. We then think about what form evaluation can take and how the different ways to evaluate work. In the third part of this article we will analyse a common formal speaking test: IELTS.
What can we evaluate?
Speakers need to be able to produce the phonological features of speech well enough to be understood, so it is fundamental that these are included in evaluation in some way. Things such as the individual sounds, stressed and weak sounds in words and speech, rhythm and intonation patterns are easy to elicit and identify. We can then measure them against a standard based on whether we can understand them or not, or perhaps more accurately, whether a typical listener could.
Rules of language
Speakers need to be able to understand and obey the rules of structure, lexis and discourse when they speak. Again, this is easy to evaluate through observation, although we need to start thinking now about providing the speaker with suitable tasks and a suitable context, e.g. in interaction with other speakers. Observing the speaker, we can ask questions such as:
- Is the speaker choosing the right vocabulary? Does it make sense? Is it formal enough? Does it collocate with other words correctly?
- Is the speaker following rules about grammatical structure?
- Is the speaker connecting together what they say and connecting this with what other speakers say effectively? In other words, is it coherent and cohesive?
It is clear that speakers need to be able to understand and use paralinguistic devices, as they are an essential part of communication, and competence includes the ability to manage these devices. We can convey an enormous amount with use of eye contact and facial expression, for example, and gestures work in the same way as linguistic communication. However, there are problems. For example, it is difficult to evaluate many of these features explicitly. The speaker's use of gestures, expressions and verbal tools such as noises can be observed, but can we establish a standard, correct use? Eye contact, movement of the body and head, and posture all send powerful messages but how do we describe them in a framework for evaluation? How do we elicit them in a controlled form? In addition to this, we may feel that it is not appropriate to evaluate this area at all as part of spoken language testing, and that the best way to address these may be under another heading, such as intercultural communicative competence. This then means evaluating them separately, using very different techniques.
Speakers need to be able to recognise, understand and use the communicative functions of speech. This means what speakers actually communicate with their choices of vocabulary and grammar, intonation and stress, changes in volume and tone etc. These features can be evaluated through observation of the speaker's performance and comparison against a standard. As we are evaluating communicative functions, it is relevant to evaluate a speaker from this perspective, for example by asking:
- Does the speaker use intonation and stress effectively to support their message?
- Does the speaker use the right functional language (exponents) to express their message?
- Does the speaker manage volume and tone appropriately to support communication?
- Does the speaker use pauses, repetition and noises appropriately to support communication?
Speakers need to be able to understand and use the social meaning of speech and many aspects of this can be evaluated formally. We can for example assess the ability to use formal and informal language, and the degree of directness, by using suitable tasks which recreate social factors such as status and age. The speaker's understanding of conversational principals and rules can be observed in interaction with others, e.g. in a group task. Connotations of language can be included as part of evaluation of use of vocabulary.
How do we evaluate?
As we can see most aspects of spoken language can be evaluated formally. The challenge is to find a form for testing which enables us to do so. An effective format for evaluation should enable us to isolate and analyse various elements, possibly under a series of general headings, but also use tasks which allow us to measure the speaker's communicative competence in general. Clearly if we want to measure a speaker's language we want them to perform to their best ability, so we also need to consider the best ways to reduce the impact of emotional factors such as stress and nerves. Finally, we need to think about practical concerns around available resources, such as time, examiners, equipment if we are going to record the speaker, and space.
There are a wide range of test types and elicitation available. In this article we are limiting discussion to formal evaluation by an examiner, as opposed for example to self-assessment or informal evaluation during classes. Below then we review some of the most common ways for a teacher or examiner to evaluate speaking formally; many speaking tests use a mixture of different types, as will be seen in part 3 (Ed. to be published later) of this article.
The candidate has a conversation with the examiner, or with another candidate with the examiner observing. Natural conversation gives the examiner an opportunity to evaluate a wide range of areas (and helps the candidate produce a relaxed and so realistic performance) but this is very difficult to achieve and requires skilful handling by the examiner.
A discussion activity can be given more focus and drive by asking the candidates to complete a task which requires them to talk together and then make a decision. This is an effective way to evaluate functions such as agreeing and disagreeing, and making suggestions, as well as conventions of conversation such as turn-taking.
Presentations and descriptions
The candidate has to give a short presentation on a topic, or describe or explain something. The examiner just listens. Topics can include personal experiences and current issues. Candidates can be asked to describe a process or a machine, give advice and provide instructions for how to do something. The amount of time candidates are given to prepare this presentation can vary from one minute to days before, depending on the language focus and resources.
Roleplays/making appropriate responses
The candidate is given a role or a situation and has to complete a task in an appropriate way. This can be carried out with the examiner, or with other candidates while the examiner listens. The advantage of this kind of activity is that certain candidates will feel more comfortable in a role and so perform better; the reverse is of course also true.
Interviews/Questions and answers
The examiner asks the candidate a series of questions. In an interview these may be related, and changeable, depending on what the candidate says. In questions and answers these are usually unrelated, although usually increasingly complex, and fixed. Interviews have a similar potential to produce useful samples as conversations, but questions and answers enable a focus on specific aspects of speaking, require less training for the examiner, and are easier to evaluate against a marking framework.
Using visual prompts
The candidate is required to describe a visual prompt such as a photograph or a diagram. This can be developed by asking the candidate to compare, order or link a sequence of pictures. This kind of test is suitable for all levels of candidate and enables the examiner to focus on a wide variety of language across a range of levels.
Re-telling a story
The candidate is required to re-tell a story which they have read or listened to before the test, or based on notes which the examiner gives them. They can also be asked to comment on an extract from a set text that they have read before the test. This kind of evaluation not only tests spoken language but also the ability to retain, organise and recall information; how much this is emphasised depends on the time between receiving the information and having to reproduce it, and on the marking scheme used.
The candidate is given a text to prepare and then read aloud to the examiner. The advantages of this kind of evaluation are that it can be controlled very easily, so the examiner can focus on specific items of language, such as minimal pairs or sentence stress, and that it is highly consistent, as all candidates work with the same or similar tasks. The disadvantage is that reading aloud is not a realistic task to evaluate communicative competence and can be challenging even for native speakers.
In this part of the article we have discussed which elements of spoken language can be evaluated and what kinds of questions we can ask when we consider each. These questions measure communicative competence, in other words with the underlying message of ‘How good is the speaker at communicating?'. We have also considered some of the many types of tests and elicitation techniques available. These need to be matched to testing aims, types of candidates and the resources available. In the third part of this article we will look at three formal speaking tests, identify how they work, and consider their effectiveness.
Written by Paul Kaye, British Council, Syria