Friday, 23 September 2016

Evidence on last year's SATs for whoever wants it

Below is the evidence I sent into the National Association for the Teaching of English when they called for comments on last year's SATs. I notice that the Select Committee on Education is now calling for evidence too.

I offer my testimony below to anyone who wants to use it in any way they like.

(One area I have missed out in the testimony below, is the fact that children who used a comma or semi-colon in the right place and broadly of the right shape, could receive no mark, if the comma was the wrong angle, the semi-colon was too big. I have recorded this in detail in an earlier blog. The effect of these right punctuation but supposedly wrongly positioned or angled punctuation was that borderline passes became fails. This could have the effect of a school being not 'outstanding' or not 'satisfactory' and therefore suitable for forced conversion.

Let us picture the idiocy of that: a school has to be turned into an academy (which will not necessarily make it better) because a child correctly put a comma where it should be but drew it 'wrongly' angled. This is education 2016!)

Below is the evidence I gave to NATE:

I observed at close quarters the way the tests worked as one of our children is 11 and was doing the tests.

1. The booklets that schools use to train the children to do SATs have mistakes in. Some of these arise out of the fact that the terminology used in the GPS test keeps changing and the people who produce the booklets or the schools can't keep up with the speed of change. It is pointless for the authorities to say that they have produced a glossary to clarify things when schools are squeezed for budgets and need the booklets quickly. 

The booklets contain mistakes in how questions are worded. So, for example, I noticed that one of the questions had two possible answers. Close observation of previous years' tests showed that this was the case in the tests themselves. 

2. One of the systems the tests use is multiple choice. The people designing the tests must be aware that this poses the so-called 'plausibility of the distractor' problem. This means that one multiple choice question does not necessarily have the same weight as another. One multiple choice question can have very plausible distractors ie three alternatives all of which could be true, while another mcq could have implausible alternatives. The marking of an exam can't distinguish between these. This applies to the GPS tests and makes them less valid.

3. There are non-grammatical elements in the GPS test, most notably questions about antonyms and synonyms. These have no grammatical content and are simply in the test for historical reasons. 

4. The antonym question this year was for the word 'fierce'. This emphasised the incorrect and pointless use of words ripped from context. There is no antonym for 'fierce' when it is taken from context as exemplified by such variant usages as e.g. a fierce argument, a fierce storm, a fierce speech and - more recently a fierce singer like BeyoncĂ©. 

5. The terminology for types of sentences is one example of how GPS terminology is problematical. So, this kind of grammar classifies sentences into four types. (This in itself is highly questionable). One of these types is called by GPS a 'command'. However, behind this definition is the need for this sentence to include an imperative form of the verb. In the questions on this, the child has to choose which of four is a 'command'. However, at least one of the sentences in last year's test included the modal 'must' which in common usage is of course a 'command'. In other words, 10 and 11 year olds are asked here to reject their common sense of the word 'command' - which is, after all, knowledge about language - in order to select one specific usage as a means of spotting the imperative.

6. Several of the terms that are compulsory for GPS are highly specific terms and it is not clear why they have been included e.g. 'fronted adverbial', 'determiner'. The term 'fronted adverbial' is clearly a 'fuzzy' term as it seems to apply to phrases that are essentially 'adjectival'. It also complicates matters concerning differences between single words (ie adverbs), phrases (ie collections of words that have no verb) and clauses (ie collections of words that include some verb form or another not necessarily finite). 'Determiners' has created a new layer of classification over and above 'articles' which now includes such words as 'every' and 'each' but no one can decide if it includes numbers. In other words a category has been introduced which has no clear boundaries and it is by no means clear how it is useful.

7. Some of the punctuation that is considered right or wrong is clearly not right or wrong. The Oxford or 'serial' comma is outlawed when it is in fact common and correct usage. This was a question on this year's paper. The commas re distinguishing between defining or non-defining relative clauses have become redundant in many accepted places. 

8. The requirement that 'exclamations' must begin with 'How' or 'what' and include a finite verb is clearly an absurd use of the word 'exclamation'. 

9. The GPS test continues to use language ripped out of context, artiificially constructed sentences, which creates a curriculum spent looking at language that children do not read, write, hear or speak. This is a major problem. Language belongs to all of us, not to the people who devise such sentences and such usages. The sub-text to this seems to be concerned in prescribing such sentences as 'ideal'. However, written language is much more diverse than this, as expressed through poetry, adverts, brochures, notices, bulletins, emails, texting, song lyrics and much more besides. If the purpose of these tests is to reveal the grammar of written English, it fails on that count alone as it does not reveal the variation in grammar of the varieties of written English. There are text books on this. I referred to them in my talk at NATE two years ago. 

10. All language is in context. There is always a context for all language. The context for the language of the GPS is the fact that it was not introduced because of the intrinsic properties of grammar, or grammar-teaching. It was introduced for the sole reason that a tests in grammar supposedly produces right or wrong answers. This is stated quite clearly in the Bew Report (2011) which is the sole justification for introducing this test. This point was an add-on to the Bew Report after the April interim report and was added on without a single academic reference piece of evidence. The test was solely designed to test teachers not children as it came under a brief to report on 'accountability'. The idea is that teachers' are tested on their ability (or not) to teach GPS. This is the true context for this particular use of language. In other words, the grammar is twisted into absolute alternatives in order to fit the requirement to be right or wrong. This is precisely where Nick Gibb fell down in the famous Martha Kearney interview. To all intents and purposes he wasn't wrong, it was the test that was wrong. Gibb 'wobbled' over whether 'after' was a preposition of subordinate conjunction when this is a matter of dispute between linguists themselves, as evidenced by Geoff Pullum's comments on the matter. There is no reason other than the test's requirement to produce right/wrong answers for this to be an absolute matter of being a subordinate conjunction or preposition. 

11. There is a further problem concerning the sole use of this kind of grammar as a way of describing language. Essentially, it's a system to describe language that is derived from meaning and structure. However, at least as important a determinant is what might be called 'social function' or 'use in context'.  The reason why the terminology changes or varies is because ultimately this kind of grammar is self-referential. It can only keep referring to itself in out of context situations. This is why ultimately there is no resolution to the subordinate conjunction OR preposition argument. Again, when terms like 'command' or 'exclamation' are used, these are words that refer to social function and yet the 'grammar' requires specific grammatical structures for these terms to be valid. This is a confusion and trivialisation of language use. There is a serious and useful discussion to be had with children about 'ways of commanding' or 'ways of exclaiming' and examining the differences in tone and meaning in these. Reducing it to tick-boxing particular forms of the verb or sentence is virtually useless in terms of language-use. 

12. Talking of language-use, the most pernicious aspect of all this is the effect it is having on children's writing. In the exemplification required for e.g. 'working at the required level' many of these GPS features have to be obeyed. The consequence for all this is that we are arriving at 'writing by numbers' - teachers teaching writing according to the number of these grammatical features being included. Meaning, purpose and function have become sidelined. This is writing in order to fit an idealised picture of correctness rather than concentrating on how to interest, excite, persuade and convince. The core basis of what in fact was the classical education (!) ie 'rhetoric' has been dropped. Ironically, in the rush for 'core knowledge' (in this case, 'grammar') another aspect of traditional core knowledge (rhetoric) has been dropped. 

13. Historically speaking, it is quite clear that other ways of examining language have been trialled and these tests represent a rejection of these. So in the early 1970s, the Schools Council, representing hundreds, if not thousands of teachers, teacher-researchers and academics, working with M.A.K. Halliday, produced 'Language in Use' an extensive programme of 110 units involving school students of all ages investigating language in use. A teacher's booklet produced by Peter Doughty accompanied the units. This was and still is a model for how such work can be produced and then modified through practice. It is important to remember that GPS came from a completely different process: it was imposed through one stroke of the pen - a comment made in the Bew Report of 2011 and 'accepted' by Michael Gove. There is no evidence of teacher input and no evidence of knowledge of pedagogy involved. The knowledge of language that is embodied in the test is of one specific kind only. There is no evidence offered as to whether most children of this age really understand the terms being used, no evidence that it helps children write for a purpose. 

14. We urgently need to open up the discussion again as to what kinds of 'knowledge about language' are age-appropriate and useful in primary schools.