Paul Nation argues that deliberately teaching vocabulary is one of the least efficient ways of developing learners’ vocabulary knowledge because only a few words and a small part of what is required to know a word can be dealt with at any one time. If deliberate teaching of vocabulary is such an issue, what about deliberate vocabulary testing?
First of all, I believe that if a test should be 100% fair, you only need to test what has been covered in some way. So while some of the popular techniques of dealing with new words are highly effective and quite natural, such as guessing words from context or using a dictionary, they are not suitable for testing.
What also bothers me is that it’s probably in the teacher’s power to make a test 100% fair, but can we make it 100% reliable?
I’ve recently encouraged myself to reconsider one of my ways of testing vocabulary. The main reason for this wake-up call is that after a break time of several years, I got a group of 11-year-old students again. L2 instruction is not totally new to them – they already learned English for a few years in primary school – but they are new to our institution and they are the youngest, so I feel the need to treat them with great care, at least for starters before they get used to all the hustle and bustle of a grammar school.
Normally, I test my students’ vocabulary the most boring way you can imagine, i.e. via the L1>L2 translation of individual words. They have alphabetical lists of words at the back of their workbooks, so at home, they are required to memorize a section of words we previously covered in class and then I pick 12 words for them to translate into English. It’s quick, easy to mark and students mostly like it because they know what to expect.
I needn’t say out loud that this is not one of the most fantastic ways of testing vocabulary. Ironically, stronger students sometimes get bad marks because they simply skip revision (secretly hoping they can remember something on the spot). Clutching at a straw, they usually do come up with something but it’s often some sort of circumvention, which, to be honest, I consider a handy learning strategy, but I can’t always let them get away with it. On the other hand, students who struggle in most areas of language learning pass the tests with flying colours. Why? Because it’s not that difficult to memorize individual words. One may do so quite successfully without even knowing how to use the words.
It’s apparent that passing or failing such a type of test doesn’t really say anything about the scope of a student’s knowledge; it only proves that more diligent students remember to do their homework. That being said, here in the state sector of education we try to educate students, not just teach them, so diligence is one of the character traits we value and support. Also, such a type of test can be a lifesaver for a student who normally struggles and needs to improve their general score. That’s why I think I still use them.
Over the years, however, I’ve tried various types of vocabulary tests. Unfortunately, I believe all of them are flawed in some way, though maybe some of them are a bit less flawed than the one I described above.
Gapfill. In advance, I create a gap fill where I leave out words I would normally ask students to translate from L1 to L2. Wait. Looking at one of my recent gap fills, I’m not so sure I would pick the same words. The words in a gap fill are usually chosen with greater care and the choice appears to be less random. After all, students need to understand the whole test to be able to come up with the words, which requires more of their mental energy. The problem is that if a weaker student fails to understand something in the surrounding context, they may fail to complete the missing word too, even if they know the word. Conclusion: it takes a while to create a good gap fill test. You need to carefully consider the words you are testing as well as the context. Also, some students are slow readers so even when they understand all the words, it takes them more time to process the text and then there is little time left for the actual ‘test’.
Gapfill with the first letter provided: This is something I’ve been doing recently to help students a bit (and sometimes to prevent stronger students from cunningly avoiding the word I want them to come up with). However, I do realize that this type of assistance may make things even harder for weaker students. I recently introduced this type of vocabulary test in one of my classes where I had only tested the infamous way mentioned above (L1>L2). To be able to pass a test, students had to revise a longer text in their coursebook. To my surprise, some students struggled a lot. This time, the class was not divided as usual – into the stronger and weaker ones. The results were more varied. I think it’s because to pass the test, a combination of diligence AND skills was required.
Gapfill with the words provided: This seems to be the best option because apart from being a test, it’s also a nice revision exercise. The grammar of the surrounding context may help a lot. This type of test certainly encourages students to make a wild guess if they are not sure but this too is an effective learning strategy. After all, Paul Nation maintains that we need to see learning any particular word as being a cumulative process where knowledge is built up over a series of varied meetings with the word. And we primarily want our students to learn vocabulary, right? So the test is just another means to an end.
Matching the word with its definition: A bit more time consuming than L1>L2 translation. But, once again, is it a reliable test? What if the student doesn’t understand the definition? Does it mean they don’t know the word itself?
Lexical sets: Write five things you can find in the kitchen. Well, a student may know that there is something called an oven in the kitchen but do they know what exactly an oven is? A more meaningful exercise of this sort could be: Name five things (or choose five from a given list) you’d take with you on holiday and say why (I’d take a large towel because ….).
Now that I think about it, what is really the problem? Is it the fact that you simply can’t design an ideal vocabulary test or because testing itself is a terribly flawed concept? 🙂