Dear Ms Morgan: in grammar there isn’t always one right answer

Writing his latest open letter to Nicky Morgan in the Guardian, Michael Rosen says he believes the main purpose of the primary Spag test is to grade children rather than examine their knowledge of language. 

I can understand that you may not have had a chance to look at the 2016 sample test [pdf] for key stage 2 English grammar, punctuation and spelling. It suffers from a severe case of terminology-itis. The symptoms are: a) an assumption that there is universal agreement on all the names, structures and functions of bits of language in this test – there isn’t; b) the best way to achieve coherence and effectiveness in children’s writing comes from getting them to learn these names – there is no evidence for this; c) that the hours of teaching-time required to teach these names could not be better spent helping children to do detailed comparative work on different kinds of texts, investigating, interpreting and experimenting, while keeping in mind the objective of enabling all children to write coherently and interestingly.

Though this test’s apparent purpose is to examine children’s knowledge of language, I think its main purpose is to grade children. That’s why some of the questions are based on stacked-up levels of abstraction and some are trick questions.

As some of the test involves four-way multiple-choice questions, will your department put out the directive: “If you are in doubt about an answer, tick any box – you have a one in four chance of being right. Remember that a part of doing tests well has nothing to do with what you know about the subject, and a lot to do with what you know about passing exams”?

The Spag test was brought in on the evidence-free assumption that spelling, punctuation and grammar questions have “right and wrong answers” (see Bew Report 2011 [pdf]). This statement is factually wrong as shown by this test: question 2 asks children to choose verb forms to put into a sentence, for which there is only one “right” answer. I can think of two correct answers. Question 3 asks children to match prefixes to words. There are two common right answers for one of them and a rare one for another…

But are the people who devised this test really interested in writing? I doubt it…

More at: Dear Ms Morgan: in grammar there isn’t always one right answer


In his full letter, Michael Rosen points to a series of specific areas where he would challenge assumptions made in the 2016 sample test and others where he points out aspect likely to lead to confusion. 

Is he making a fair point about the test fundamentally missing the point about being able to use language, or write, well?

Please give us your reactions or feedback in the comments or via Twitter…


Don’t forget you can sign up to receive our daily email bulletin every morning (around 7 am) with all the latest schools news stories. Your details will never be given to anyone else and you can unsubscribe at any stage. Just follow this link

Rise in number of GCSE and A-level candidates given extra marks for extenuating circumstances
Encourage boys into cheerleading to 'challenge gender stereotypes'
Categories: Primary.


  1. Doing these type of tests which consider words or sentences in isolation kill the love of English stone dead.  While I admit that know terms such as adjective, adverb etc are useful to discuss how authors use language, I’m not convinced that being able to spot a ‘subordinate clause’ or the ‘past progressive’ adds anything to appreciation of the beauty of the English language.
    As Rosen perceptively remarks, these tests seem to have been devised by those who have no real interest in how language is used to delight, amuse, move or persuade.  Their fussy focus on elements blinds them to the qualities of the whole.

Let us know what you think...