Writing tests are not the answer you are looking for.
We’ve written this post because there has been a lot of discussion about the Writing Framework recently and this has caused some to romanticise the days of writing tests.
How the DfE/STA decides to assess writing tells you a lot about its philosophy and epistemology towards the craft of writing and its feelings towards apprentice writers. It also has profound effects on the ontology, methodology and writing pedagogy of the teaching profession. It influences the way things are taught. Therefore, what is deemed important in a test will inevitably lead the way teachers teach. So, you have to ask yourself, is it likely that a high-stakes test will test what should be taught? I ask this question because we as teachers know full well we will be asked to do what needs to be done in terms of writing instruction and activity to produce good scores. I have no problem with this in principle, and indeed it can be the strength of any assessment system, but will a writing test encourage good writing instruction and activity? I have my doubts and I explain why below.
Writing tests are an attempt to determine whether a student can do something. The first question needs to be: why is this information needed and who is it going to be useful to? Test writers need to describe what writing behaviours are most important and are therefore the ones they want to test. As teachers, we would then need to say: now we’ve given these tests, what do the results mean, and what do we do with them? How am I to put this information to work? Are the consequences of test scores legitimate? Are these writing tests valid? Do they tell us effectively what needs to be known? Are they reliable? Would a child get the same score if they did the same test on a different occasion? Was the test too narrow and insensitive to measure all the things our school is trying to achieve? Does the test focus on too slender a slice of what is important in writer development? What would your answers be if there was a return to writing tests as we knew them a few years ago?
Next, you probably should, maybe, possibly, potentially consider the children in all this. You know, the ones that have to take the test. Any assessment should have a clear and realistic purpose for the person who has to take it. Writing is a social act. If children are faced with a set of questions to answer with no purpose or authentic context in which to tackle the writing task – then the writing won’t even represent their normal writerly behaviour, nor will the test represent what research says is good classroom practice which insists on writers focusing on the potential audience and purpose for their writing. How a child interprets and engages with a test task significantly affects their response to it and therefore the quality of their writing product. Beyond this, you also need to consider whether a writing test favours children who are better able to maintain focus for long periods, able to write using a certain type of writing process, have a higher threshold for stress and a have greater level of social maturity. Additionally, children have little experience in taking writing tests – how do you suppose they will gain this experience? What will the consequences of this be?
Assessment at its best has what is called ‘consequential validity’. This means the assessment gathers a variety of information, at diverse times, and under differing circumstances. It establishes connections between assessment, policy and teaching practice. Assessment such as this, throughout primary school, is better than a writing test because it gives information about a student’s development as a writer and importantly gives you plenty of opportunity to act on what you find. Despite educational research for a long time stating that focus should be on children’s processes and not be overly orientated on their writing products, it’s laughable that writing tests historically focus only on product and have little or no regard for a child’s writing processes or behaviours. This means, through testing, you’re not only teaching children a misconception about writing but you also won’t be able to infer from their test how well a child might perform under normal, everyday writing conditions. Conditions which enable children to use all the writing processes, in their preferred way, and at their own pace.
A ‘best test,’ a test that can provide exactly what you need to know, that can guide future teaching and be easy to administer and interpret, simply doesn’t exist. The ultimate goal of assessment should always be to improve teaching. The idea of a writing test is therefore wholly unsuitable, would be inaccurate, would encourage misconceptions about writing to be taught and therefore would not serve the desired purpose. If reinstated, tests would cause damage to children’s ongoing writing development. Writing tests are not the answer you are looking for.
This article is based on the following papers:
- Evaluating Language Development by Farr, R., & Beck, M., (2003) In J. Flood, D. Lapp, J. R. Squire, & J. M. Jensen (Eds.) Handbook of research on teaching the English language arts (2nd ed., pp. 590–599). Mahwah, NJ: Lawrence Erlbaum Publishers.
- High-stakes assessment in the language arts: the piper plays, the players dance, but who pays the price? by Hoffman, J., Paris., S., Salas, R., Patterson, E., Assaf, L., (2003) In J. Flood, D. Lapp, J. R. Squire, & J. M. Jensen (Eds.) Handbook of research on teaching the English language arts (2nd ed., pp. 590–599). Mahwah, NJ: Lawrence Erlbaum Publishers.