Many teachers, researchers, parents and children have said that the Grammar, punctuation and spelling test (formerly 'SPaG' now 'GPS') for Key stage 1 and 2 :
involves too much preparation and squeezes out of the curriculum time that could be spent on exploring language through experimenting and practising with many different kinds of reading and writing;
is skewing what it means to 'write well' so that it now means 'include specific grammatical features' like 'expanded noun phrases' - it is quite possible to write an absurd, pointless 'expanded noun phrase' or a non-expanded noun phrase that is delightful, eye-catching, surprising, interesting or exciting;
some (not all) of the grammatical terms are too difficult and/or unreliable and/or inconsistent - e.g. we say when talking about functions that a 'noun' can be a 'subject' or 'object' but by the same system, a 'verb' is a 'verb'! There is no function name for it in GPS that is equivalent to subject and object;
it skews a view of language towards a view that language is about abstract rules and away from looking at language-in-use, used by real speakers, writers in relation to real listeners and readers;
it skews a view of reading and writing away from saying that 'meaning' is important towards saying that reading and writing are about obeying these grammatical instructions;
the kind of grammatical knowledge required may be appropriate for older pupils but for many of primary age it is not only too much but also too difficult conceptually e.g. defining and non-defining relative clauses; or the fact that the terms are overlapping in confusing ways e.g. a relative clause is a type of subordinate clause but there don't seem to be any other kinds of subordinate clause (according to GPS) (there are, but not in this test!);
the GPS test is flawed because some of the questions have more than one answer; some questions are seemingly about right/wrong matters when in actual fact, language-in-use shows us that it is not a matter of right/wrong but of style (e.g. the Oxford or serial comma), or of very varied use (e.g. in the punctuation of defining and non-defining relative clauses);
grammarians themselves are in genuine disagreement over what the test tries to show is a 'fact' (e.g. preposition or subordinate conjunction heading a subordinate clause using 'after', 'before', 'until');
some of the questions are not grammar, punctuation or spelling but arbitrary and invalid tests (e.g. synonyms and antonyms) leading to absurd questions as in 'what is the antonym of 'fierce'?' (nb the marking criteria is reported to have said that 'timid' is not an antonym of 'fierce'!). Even so, the idea that a word, taken out of context, has a specific meaning for which there is an 'opposite' or 'antonym' is an absurd and misleading idea about how language works;
some of the questions have different parts leading to the fact that you can be mostly right yet receive no marks - this means that the test is at times 'not valid' in measuring whether children know things or not;
the test as a whole was not devised to teach children grammar but to test whether teachers could teach this particular skewed and narrow form of grammar - it's to measure 'accountability' (see Bew Report 2011);
the grammatical knowledge required by these tests is spawning hundreds of pre-test booklets some of which have errors or have used slightly different terminology;
the terminology itself is at times confusing and much disputed by grammarians themselves - words like 'command' or 'exclamation' are used in a highly specific way, specific only to this system and yet in the questions on the test, candidates are asked to exclude the usual ways in which such a word is used: e.g. excluding 'you must go out' as a 'command' ie the child is required to leave their usual sense of the word 'command' outside. Again, something that as adults we can do, when given the instruction, something not very easy when you're a child, particularly when 'distracted' by a construction using 'must';
absurd and incorrect 'rules' are being invented by the people devising the tests in order that the questions produce right and wrong answers only e.g. when a GPS test determined that the word 'bright' could not be used as an adverb. Professor David Crystal pointed out quite clearly that it could and such constructions have been common in English since at least Shakespeare's time;
many people from the world of editing, translation, writing, journalism are reporting that they do not understand why their 6 or 10 year olds are having to learn descriptions of language which they, as practitioners haven't needed to know e.g. 'fronted adverbial';
because there is no full rationale for this scheme of work, a term like 'fronted adverbial' is supposed to be just accepted by teachers and pupils as a 'fact'. In fact (!) it's problematical. At the very least, there is a problem in that some 'fronted' phrases are clearly 'adjectival'. Children learn 'adverbs' and 'adjectives' so why are they learning 'fronted adverbials' and not 'fronted adjectivals'? What so special about 'fronted adverbials' that they need to be singled out? 'Fronting' may well be an interesting aspect of language to explore (there's a lot of it in poetry) but it's not just a matter of being 'adverbial';
each one of these issues would be slightly problematic but not disastrous. It's the accumulation of all them (and others, no doubt) that is the real problem.