In the world of education research, approaches often work better in laboratory conditions than they do in the classroom.
A good example of this is the testing effect: a well-established theory suggesting that repeated testing (or quizzing) before and after learning can improve further test performance.
It’s often assumed in schools that this effect has been proven beyond all doubt. Yet this is a good example of the gulf that can exist between cognitive science research and classroom practice.
There is a huge difference between a carefully controlled experiment - in which participants learn a small number of words or word pairs, before being retested shortly afterwards - and the weekly spelling test, which involves handing out 10 individual words to learn to spell on a Monday and testing them on a Friday.
Even if the child gets all the words correct on Friday, it is often the case that they do not use these spellings accurately when writing. Some children seem to catch spelling; some do not. And the weekly spelling test, used in this way, doesn’t seem to make much difference - although, on the surface, the testing effect suggests it should.
Testing effect: when does testing help?
So, when does testing help? A recent paper by an international collaboration of researchers attempts to identify when this approach is most helpful in the classroom. Their meta-analysis of 221 studies only included studies completed in the classroom, allowing a more accurate picture to take shape of how this effect works outside of the lab.
The research points to three possible theories as to why testing works. Firstly, a test gives additional exposure to the material to be learned - quite simply, the pupils have another opportunity to learn it.
Secondly, testing in the format that the final test or exam is delivered in may help pupils learn the material in a way that is compatible with that format - it helps them shape the process of understanding in the context of the questions asked.
Finally, knowing that you will be tested can be motivating and encourage pupils to do the work. This does not seem to be affected by whether the test is high or low stakes for the pupil.
However, there are quite a few additional things to consider. For one, the age of the children matters. This approach seems to be most effective in secondary classes.
Testing is also more effective when it is conducted in school, after learning, and the more frequently the material is tested, the better.
How the test is structured makes a difference, too. Matching answers and “fill in the blanks” cloze formats are more effective than free recall of the answers and tests that include a variety of test structures. But, at the same time, there is a greater positive effect if the test is structured in the same way as the exam - hence why practice papers can be so helpful.
In addition, providing corrective feedback after a test makes a big difference in how successful the learning is.
This gives us some idea of when the testing effect might work. But it’s important to remember that recalling knowledge to pass an exam that is formatted in a particular way is just one part of learning how to think about a subject and continue to learn about it.
When we consider whole subject disciplines, recalling answers to test questions is only one piece of the jigsaw.
Take writing, for instance. The spelling tests are one aspect of the learning process. They help children learn how to write the letters in sequence, in response to the teacher’s prompt, and provide them with some handwriting practice. But they do not help the child learn the meaning of the word, how it is used in writing, other words that are generally used at the same time or when it might be good to use these words.
So, does testing work? Yes. But the testing effect is not as straightforward as it sounds.
Megan Dixon is a doctoral student and associate lecturer at Sheffield Hallam University
For the latest research, pedagogy and practical classroom advice delivered directly to your inbox every week, sign up to our Teaching Essentials newsletter
