CiC3-TB

How can I mark speaking tests? There are numerous ways to go about it. Different types of tests emphasize different aspects of conversation. How you go about marking your speaking tests will depend on what you prioritize. Whatever way you chose, make sure that your students are crystal clear about what is expected of them. Typically, speaking tests are scored on a set of Likert style scales. Each aspect of conversation is given a score of 1 to 5, then the points are added up. For example:

To counteract this situation, you can try an occasional test marked in an open-ended manner. This means there is no top score, only a limited amount of time to collect as many points as possible. The onus is therefore on the student to come up with as much grammatically correct sentences within their allotted time. One idea is to give two points per correct turn, one point for a turn containing a small error, and zero points for a turn with a major mistake. Alternatively, you could just count turns that are mistake free. One advantage of open-ended marking is that it solves the problem of clustered test marks, which are not very motivating for students in the long run. What to measure? Here is a non-exhaustive list of commonly assessed aspects of conversation and some ideas for how they can be reliably measured: Fluency can be measured in several ways. First, you can look at the total number of turns or words, or you can measure noticeable gaps of silence. These are two sides of the same coin; students can get credit for speaking more or marked down for long gaps. Turns can be kept track of on a tally sheet, as can noticeable silences. If students record and transcribe a short conversation, they can calculate average words per turn, which can be tracked over time to measure progress. Another way of looking at fluency is in the quickness of student responses to your questions. If they respond quickly, then points up. If they lag, then points down. Accuracy can be measured by counting the number of grammar errors. If using a Likert-type scale marking sheet, then a high score would mean the absence of any mistakes. Again, you’ll need to clarify how strict to mark this point. Too strict and you’ll run the risk of demotivating students; too lose, and this criteria won’t have any meaning. Intonation is all about how students speak, the quality of their voice. This is a good thing to measure because of problems Japanese learners typically have with speaking in a clear, positive voice. Most of the time they are either too flat or over-the-top genki. Making intonation part of a test would push students to work on this important part of speaking because how words are said affect their meaning. Before the test, it would be a good idea to model what each score sounds like so that your students are clear what you are looking for.

Fluency Accuracy Intonation

1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5

Content

As Nation mentions, it’s not a good idea to look at too many aspects at once because it becomes too difficult to make reliable judgements. If you use such an approach, you will need to clarify what each number means and communicate this to your students. For example, what constitutes a 5 for accuracy? Is it the total absence of mistakes, or will you allow small ones to slide by? A good idea would be to write all this out in your notes and create a bi lingual handout for your students so that they will know exactly what they need to do to get a high score. Open-ended marking

One challenge with marking speaking tests is balancing the strictly evaluative aspect with the motivational one. You want your students to do well and feel positive after a test so that they continue

putting effort into getting better. When the class has taken on a good dynamic, you may find that you end up giving scores within a very narrow bracket, say between 16 and 19. The occasional outstanding student can get a 20, while a poorer one a 12, but the bulk of the class will often get similar marks. Since there is not much variation between test scores, this could be potentially demotivating.

How do I test my students? 122