Formative Assessment in University English Conversation Classes

This is an account of one teacher’s use of formative assessment in Japanese university EFL conversation classes. Formative assessment was used in these classes in the ways advocated by Clarke (2013) for use in UK primary schools; that is, through the use of decontextualised learning objectives, success criteria for meeting the objectives, student examples, talk partners, and selfand peer-assessment. The ways in which these tools of formative assessment were used and the benefits they brought to the classes are discussed.

does not have a universally accepted definition, but it can be thought of as any assessment which is used to change the way that teaching or learning occurs.There is no particular method of assessment that is formative.A test used to assess achievement for the purposes of grading, reporting results to parents, or comparing schools would be considered summative assessment.
However, if used to find out what students know and the areas where more help is needed, thereby informing future lessons, the same test would be an example of formative assessment.Moreover, formative assessment does not have to be in the form of a test.Teachers constantly assess their students in various ways such as through pieces of written work, participation in regular class activities, and homework.All of these methods can be used as ways of informing what teaching and learning needs to take place.
It has been argued that assessment can only be really classed as formative when it is used by students themselves to inform learning (Black & Wiliam, 1998;Hall & Burke, 2003).Similarly Clarke (2008) identifies self-assessment as crucial for formative assessment to be effective.When formative assessment is defined in this narrower way, where assessment is something done and used by students themselves, it can be viewed as a classroom application of self-regulated learning (SRL).Paris and Paris (2001) identify self-assessment as one of three main areas of direct classroom application of SRL, which is "the self-directive process by which learners transform their mental abilities into academic skills" (Zimmerman, 2002, p. 65).While SRL is a construct that may encompass a person's whole learning behaviour, formative assessment is a strategy that can be used in the classroom for a particular learning objective that will help learners develop as self-regulated learners.Formative assessment involves using self-assessment to help learners see gaps between their current level and desired level, and provides ways of bridging these gaps.Clarke (2008) identifies five tools as necessary for effective formative assessment: pure learning objectives, success criteria, student examples, talk partners, and self-and peer-assessment.Her work is based on the findings of several action research teams made up of groups of teachers in UK and US primary schools.My attempt to use formative assessment in university conversation classes in Japan is based upon Clarke's five tools.The forthcoming sections will introduce each tool and describe how it has been operationalized in my classes.

Pure Learning Objectives
A pure learning objective is one that has been separated from the context of the learning goal.
For example, "using topic sentences correctly" is a pure learning objective while "writing a topic sentence for a paragraph that compares two cities" has the context embedded in it.In conversation classes and textbooks designed for speaking, there are different types of learning objectives that a lesson might have.For example, we might have a topic-based learning objective such as "to be able to talk about daily routines", a grammar-based learning objective such as "to be able to use the simple present tense correctly", or a function-based learning objective such as "to be able to place an order in a restaurant".A de-emphasis on grammar in many EFL classrooms in recent years has made topic-based or function-based learning objectives more likely.
Topic-based objectives such as "to be able to talk about daily routines" have the context as part of the objective.When the class moves on to another topic, such as family or shopping, although the teacher may see connections between the last learning objective and the next, it is less likely that the students will.They feel as if they are starting from scratch.This is likely to be the case with function-based objectives as well.If we use pure learning objectives, however, the students can see how these are transferable to different contexts ( i.e. different topics or different situations.The skills necessary for holding conversations in English can be decontextualised relatively easily.
Learning objectives related to conversation skills, such as providing extra information in answers, asking follow-up questions, and being able to continue a conversation even when we are not asked a question are pure learning objectives that can be transferred to conversations on any topic.If teachers set these kinds of skills as learning objectives for a course, and make sure students are aware of this, students can see how what they are learning is easily transferable to any conversational context.In one conversation course I teach, each class has a specific skill learning objective and a topic.As an example, the first class has a learning objective of "giving extra information in answers" and the topic is "hometown or neighborhood".In subsequent lessons with different topics, the "giving extra information in answers" learning objective is referred back to and used with the new topic.In this way students see that each skill is relevant to all topics and they begin to use the skills with increasing frequency.I typically choose learning objectives based on the conversation skills that I feel many students are lacking.
When students understand very clearly what the learning objective is, and what is necessary to meet this objective, they are more able to take control of their own learning.It can be extremely demotivating for students to not know what is expected of them.In Japan, going from high school English classes to a university conversation class with a native English speaker can be quite a change, and many students may feel at a loss as to what is expected of them.Topic-based learning objectives such as "to be able to talk about my family" may not really help them in this respect.
Decontextualised skill-based learning objectives can help a lot more, especially once broken down into success criteria.

Success Criteria
Success criteria are the details of the learning objectives.They break the learning objective down into smaller parts, telling students exactly what they need to do to meet the objective, and helping students see where they need to improve.For example, for the learning objective "giving extra information in answers" the success criteria might be: 1. Give a basic answer plus two extra pieces of information.
2. Don't repeat the words from the question in your answer.
Success criteria can be given by the teacher at the beginning of the lesson or generated by students by having them look at examples of answers that meet or do not meet the learning objective and thinking about why they do or do not meet the objective.Understanding the success criteria seems to really help students understand how they can improve their conversation skills.
Very quickly students go from answers reminiscent of junior high school English textbooks, which often have grammatically correct but simple and repetitive exchanges, to answers that seem more natural and are more likely to keep a conversation going.For example, prior to defining success criteria, a typical answer to the question, "Have you ever been overseas?"may be: "No I haven't.I have never been overseas."However, after considering the success criteria, students become able to make responses such as: "No, but I'd like to go to Italy and France.I love pasta and I want to see the Eiffel Tower." Ideally, success criteria should be generated by students themselves.This gives students more ownership of the criteria rather than the criteria being something imposed by the teacher.In an English Communication class I teach, students were instructed to generate success criteria for the very general learning objective of having a conversation in English.I asked them to consider in pairs or small groups what is important in conversation.They came up with criteria such as: "talk a lot", "smile", "listen carefully", "use only English", and "have eye contact".The activity of making a student-generated list of success criteria also enabled me to deal with some misconceptions among students about conversation skills, such as needing continuous eye contact.Once the class list of success criteria had been generated, students could then use it to evaluate themselves and their peers after each conversation they had.From this self-and peer-assessment, students can see where their strengths and weaknesses lie and then work on improving their weak areas.

Using Student Examples
Being able to see examples of good conversation skills can help students understand what is required of them, and help them to generate their own success criteria.An actual student conversation, rather than one from a textbook, can be effective as students see it is a realistic goal to which they can aspire.Transcribed, anonymous conversations can be used if these are available from another class of a similar level.It is effective to have both a high quality example and a less high quality example which the students can compare in order to see why one is better.The difference in quality needs to be clear and should be related to the learning objective.For example, for the "give extra information in answers" objective, one example should have lots of answers with extra information while the other lots of short, basic answers.If the examples are different in terms of grammatical accuracy, students may focus on this rather than the skill the teacher wants them to practise.Although I haven't yet tried this myself, video or audio recordings of conversations would probably be even more effective but cannot be made anonymous.A way around this would be to film two willing student volunteers at the beginning of the course and again at the end.This video could be shown to future classes at the beginning of their course.Seeing examples of good skills and poor skills can help students really understand what is expected of them, and show them where they need to improve.Through seeing that other students have managed it, it also helps them see that what is expected is an achievable goal.

Talk Partners
Most teachers probably already use a lot of pair work in conversation classes for maximising English speaking practice.Another way of using talk partners, however, is perhaps less common.
Talk partners can be used very effectively as a replacement for a hands up approach (where students raise their hand to answer a question).The problem with asking students to raise their hands is that it tends to be the same few students who do not mind being the centre of attention and are relatively confident in their English ability.This means that the teacher has no idea whether the rest of the class understand, and therefore does not know whether further explanation or practice is needed.It also means that some students may not really think about the question, instead passively waiting until the teacher or another student tells them the answer.Calling on students by name is one way to encourage quieter students to become more involved in class.However, students who have had bad experiences of being asked to answer questions in front of the whole class may feel uncomfortable and it could take time for them to overcome this.The use of talk partners provides a relatively non-threatening environment that allows all students to get involved.
Asking a question and having students answer or discuss it with their partner (using the L1 where necessary) means that all students have to think about the answer rather than wait to be told it by the teacher or another student.Also, the teacher and the students themselves can get a better idea of what all students in the class know.After a given amount of time, if appropriate, the teacher can call on two or three pairs to share their ideas with the class (again using the L1 if necessary).Even if the teacher cannot speak the L1, they can get a good idea of how much the students know by the level of noise and confidence in their discussion.When sharing ideas, pairs can say their ideas in the L1 and the class can help translate into the L2 for the teacher.I have found using talk partners in this way is good for various situations including: 1. Brainstorming vocabulary or questions for a given topic; 2. Discussing (in the L1) whether certain sentences are correct or appropriate; 3. Discussing (in the L1) aspects of learning such as self-evaluation or success criteria.

Self-and Peer-Assessment
The hallmark of formative assessment, as mentioned above, is that it is used to inform subsequent teaching and learning, in particular by students themselves.The most effective formative assessment makes use of the students' own assessment during the actual writing of an assignment or the practising of a skill, rather than afterwards (Clarke, 2008).Students can evaluate themselves and their peers using the success criteria.I have found that many students are uncomfortable pointing out another student's weak points even after seeming to understand they are helping their classmates by doing so.After trying out a 1 to 5 score system, a ranking system, and a circle-triangle-cross system (where a circle means something has been done well, a triangle means there is room for improvement, and a cross means something has not been done well) I have settled on a ranking system.Students rank the success criteria according to how well they have been achieved, both for themselves and for their partner.This avoids the problem of students not having to give a low score or a cross to their partner.Once students have identified their weak areas through ranking how well the success criteria have been achieved, they can try to improve these areas in subsequent conversations.
Students may need strategies to help them improve a given area.The teacher's role here is important as students may not know how to go about improving their weak points and may use inappropriate strategies.For example, several of my students felt they were using too much Japanese during conversation time.A common suggestion for a way to improve this area was to memorise lists of English words so that their vocabulary increases and they have less need for Japanese.Clearly this is, at best, a very long-term strategy and is likely to become an extremely inefficient use of their time if indeed they continue to do it at all.I suggested, instead, that they keep on hand a pen and a piece of paper during conversation time and keep track of how many times they use Japanese.Documenting their use of Japanese in numerical terms helps them to focus on reducing the tendency little by little, which is a more achievable goal.They can also keep track of what Japanese they used, and then find out how to say those words or expressions in English later.
Helping students find strategies to address their weak areas, and ensuring they use these strategies in subsequent lessons, is perhaps the most important job of the teacher in classes using formative assessment in this way.It has also been the biggest challenge for me personally.Students have needed a lot of reminding and encouraging, which isn't surprising, perhaps, as they have probably never had to do anything like this in their educational experiences so far.
One tool that has been something of a breakthrough in my classes with self-assessment is using a voice recording application that I had students download onto their smartphones.They can record a conversation with their partner, and then transcribe and analyse it.When they see their conversation written down, it seems much easier for them to analyse it for weak areas.For example they can count how many answers they gave with and without extra information.As they get used to listening to their recorded conversations, or when they are practising a skill that is not new, they can listen without transcribing.Recording makes it much easier for students to notice the things they had trouble saying in English, which makes it easier for them to look up these language points later.

Conclusion
In summary, using decontextualised learning objectives helps students see that what they have learned is transferable to other contexts and understand what is expected of them in a conversation class.Using success criteria helps students to meet the learning objectives and identify their weak SiSAL Journal Vol. 5, No. 4, December 2014, 423-429 429 areas.Using student examples of high quality conversations helps students understand what the teacher expects and helps them generate or understand the necessary success criteria.Talk partners force students to engage, help the teacher to identify students' weak areas, and provide students with self-and peer-evaluation.They can also assist students in generating success criteria for themselves.Self-and peer-assessment allow students to take responsibility for their own learning.
With plenty of support from the teacher, as and when necessary, it will also hopefully help students to take their first steps on the path to becoming active and self-regulated learners.

Notes on the contributor
Carla Wilson has been teaching EFL in Japan since 1998.She currently teaches speaking and writing classes at four universities in Hiroshima.