Talk at Conference on Admission to Grammar School and the Kent Test

Thursday, 22 October 2015 20:09

Talk at Conference on Admission to Grammar School and the Kent Test

Be the first to comment!

Kent Test Presentation

21 October 2015

NOTE: This presentation was made the day before KCC announced, on the initiative of the Leader, Paul Carter, that a commission was to be set up to explore wider access to grammar schools. My presentation closes with a section below that addresses precisely this issue.

I have amplified the presentation in places to pick up several discussion points.

You will find my article about the Conference here.

Kent Test Conference

Photo of Matthew Bartlett at Kent Test Conference

Courtesy of Kent on Sunday

This presentation explores two main themes: admission to Kent grammar schools and the Kent Test. Some may think these are the same, but the reality is that the Kent Test provides only part of the opportunity for children to gain entry to grammar school, with alternative routes being such that there are no two grammar schools out of the 32 in Kent (nor indeed of the six in Medway), that have the same expectation of potential applicants.

Of course the hot news of the moment is the new annex in Sevenoaks and how it will affect school admissions, and I will also say a few words about that......

On a personal note, some of you will know that I am stepping back on my work on admissions and appeals, and have chosen to take a complete break this month apart from this evening. KCC used to offer a free independent advisory service at this time, but sadly government cuts have seen this removed to the specific disadvantage of those who least understand processes. Please make no mistake. Whilst in many parts of Kent and for most people, decisions on which schools to apply for are fairly straightforward, but in some situations, those least able to understand the process make serious mistakes, and of particular relevance this evening, are put off applying for grammar schools being bewildered by the appeals process, when children have the potential to succeed at grammar school.

32 Grammar Schools, 32 Admission profiles

Of Kent’s 32 grammar schools, three are what is called super selective, expecting entrants to achieve high scores in the Kent Test, although all three use different cut-offs and different rules for residence. Another six are partially super-selective, the Dartfords, Wilmingtons, Maidstone Grammar, and Simon Langton Boys, which each give priority to a differing proportion of high scoring pupils again operating on individual rules. Then there are the five that operate their own tests as an additional route to admission alongside the Kent Test, the Dover Grammars, the Folkestone Grammars and Mayfield in Gravesend. Several of these admit more than half their pupils via the local test, including Dover Girls Grammar which demonstrates the validity of the process by producing some of the best GCSE and A Level results in the county. Between them, these add in nearly 2% more of Kent children to the total passes.

That leaves 18 that admit on the straight pass score. The Kent Test, pass mark is set to admit 21% of the children of the county, but only passed 20% in 2014 the latest date for which I have this information at present, showing more success for girls, an issue to which I shall return later.

2014 Kent Test	Boys	Girls	Total	% Boys	% Girls	% Total
Number of Kent Pupils	7986	7608	15594
Took Kent Test	4883	5004	9887	61%	66%	63%
Automatic Pass	1555	1557	3112	19.5%	20.5%	20.0%

An interesting phenomenon is that the percentage of girls taking the test is significantly higher than that of boys, with a similar pattern in Medway. Why?

However, this is only part of the story and the other two routes break the myth that the Kent Test is a one off pass or fail to enter grammar.

Headteacher Assessment

A second tranche of children are found of grammar school ability, by what is known as the “Headteacher Assessment”. This rightly comes at the question of ability from a different direction, looking at children’s work, past test grades and headteachers recommendation that can cover any issues thought relevant and Emma will go into more detail on this.

Headteacher Assessment is supposed to find another 4% of children selective across the county, but the reality can be higher, and in recent years has been as high as 6%, although falling over the past two years, to around 4.6 % for the most recent assessment. To complicate it further, the assessment of work traditionally favours girls so this splits, giving passes to 7.5% of girls and 5.5% of boys for the 2014 test (the 2015 figures being my best estimate from the data currently available).

Kent Test Headteacher Assessment Outcomes 2014 & 2015
	HTA	Boys	Girls	Total	% Boys	% Girls	% Total
2014	Entered	990	1082	2072	13%	14%	14%
2014	Passed	443	567	1010	5.5%	7.5%	6.5%
2015 (estimated)					5.2%	7.4%	6.3%

An even greater issue is that with pressure of places in the west of the county, children are nearly twice as likely to get through in the east, with mid-Kent somewhere between.

The target overall pass rate is 25%, but for 2015, it is 25.6%, for the second time again favouring girls, by 26.6%, with the boys 24.7%. Under the old Kent Test, with no literacy component, boys led on the test itself, with girls coming out top on HTA, producing an even split on the final outcome. Now, we have seen in both years of the new Test, a considerable gap opening, favouring the girls, with consequent changes in the pressure on grammar school places for the two genders.

Overall pass Rates for Kent Children 2014 & 2015
	Boys	Girls	Total	% Boys	% Girls	% Total
2014	2104	2176	4280	24.7%	26.6%	25.6%
2015	1998	2124	4122	25.0%	27.9%	26.4%

Grammar School Appeals

This of course is before grammar school appeals, which come from yet another direction, allowing parents to produce whatever evidence they wish to show their child is of grammar school ability. Nearly 700 of these were successful this summer, over a hundred more than in 2014, but some can be discounted for this purpose for various reasons, such as children already passed seeking places in oversubscribed grammar schools and discounting out of county children, allowing at least 500 more Kent children to be found suitable for grammar school. The variation, school by school, in success rates at appeal is staggering, ranging from the astonishing figure of 146 from Chatham and Clarendon Grammar, in Ramsgate, with 33 successful, to 70 at Dartford Boys, with just 6 successful, five of who had already been found of grammar school ability by the Kent Test. That excludes Cranbrook School, the only grammar in the county with entry at age 13 plus, which had just one appeal which was unsuccessful.

Another specific example from me: Weald of Kent Grammar in 2014, had 48 successful appeals out of 69 as it sought to expand numbers, possibly in preparation for the Sevenoaks Annex. Last summer, when the annex looked in doubt, the school “discovered” it had no more room and the success rate fell sharply to 6 out of 70. For 2016 admission the school has amazingly found this no longer to be the case and is admitting an additional 60 girls, so I guess that successful appeals will once again rise in number, building towards the annex opening in 2017, unless the annex falls to a challenge when the expansion could suddenly vanish. Such variations although not usually as extreme as this are not uncommon as schools situations change. For, increasingly, appeal panel decisions tend to reflect the meaning of “selective ability” according to demand for places at the school and the schools wishes.

Overall the appeal numbers amount to a further 3% of Kent children being offered places meaning that in total over 30% of Kent children will have been awarded grammar school places in the County for 2015 admissions and, for the above reasons, with a profile that can vary widely from school to school. As you can see, there is no way one can estimate the chance of success at appeal any more without understanding something about the school concerned and its priorities, although too many ill-informed prophets try. I cross checked my figure with the 2014 school census for comparison when 29% of Kent children were in Year Seven in Kent grammar schools, the discrepancy probably explained by the rise in successful grammar school appeals this year, and by children dropping out of the system to go to private schools.

The New Kent Test

So let us turn to the Test itself and the reason for the rise in appeals.

Part of the answer for the rise in appeals lies in the nature of the new Kent Test itself, now in its second year of operation, and introduced for two main reasons: to reduce the effect of coaching and also to introduce an element of literacy into the test.

Let us first look at the makeup of the test, for those unfamiliar with it.

Structure of the Kent Test

There are two separate tests, each lasting about an hour, including practice sessions

The two papers are both multiple choice

Verbal and Non-Verbal Reasoning (two sections)

English and Mathematics

There is a third paper, a writing exercise, only used if there is a HTA

Marking the Kent Test

The two tests are both Multiple Choice, marked by Optical Mark Readers.

For 2015, results were divided into three equally weighted sections:

1) Reasoning Paper

2) English Section

3) Mathematics Section

Kent Test Marks

The Kent Test pass score is selected annually so that 21% of the Kent cohort of children will be found selective.

The marking for the Kent Test is nationally standardised, so that 21% of a sample of the national population achieves a score of 113 or better on each test.

National standardisation ensures that children born in each calendar month are measured against children of the same age. Research I have carried shows no significant advantage to being born in any month (as distinct from Medway)!

I must admit that after seeing the outcomes of last year’s test, requiring candidates to achieve a standardised score of just 106 in each of three multiple choice assessments in English, maths and reasoning, together with an aggregate score of 320, I forecast that the pass standard would rise this year, but I have been proved wrong. This should be a matter of concern.

A nationally standardised score of 106 places the child in the 34^th percentile, far below the 21^st percentile of the grammar school standard. This would be 113 in each test. I analysed individual scores last year (2015 figures still to come) in an attempt to find out the reason for this low mark and came to the following conclusions. Coaching still has a significant effect, with boys and girls achieving above national averages in each section but less than in the old Kent Test. However, it is impossible to be precise as at KS2 Kent children were performing above national averages at both Level 4 and also Level 5, the target for grammar school performance which may account for much or all of the three points average raise above the national average. There is however, an effect clearly seen in the reasoning tests, which averaged 5 standardised points higher than both English and maths. The main reason that the cut-off of 106 is so low, is that children (mainly boys) scored higher in mathematics or (mainly girls) English but too many were not up to standard in both.

Scores in Individual sections of the Kent Test 2014
2014	Maximum Mark	Pass Mark	21% Score Boys	21% Score Girls
Reasoning	141	106	121	121
English	140	106	114	116
Mathematics	141	106	118	116

What I do believe this also shows is that the reasoning test is more susceptible to coaching in any case, and as it is different from normal classwork, that is where the emphasis has again been placed too highly by those preparing children for the test.

Literacy and numeracy should be central to any programme of preparation for the eleven plus and good coaching here can only be of benefit to children’s achievement.

There is now a further worry about the new Test, and that is that there will be children who are passing automatically with scores that can be as low as 106,106, 108. These could be children who are not up to standard and I have already heard reports from several grammar schools of such children who are already struggling. Of course, some may just be the product of poor teaching and will flourish in their new environment. Many of those they have displaced will have evidence of grammar school ability, and so hopefully will win through on headteacher assessment or appeal.

So where now? I am confident the new Test is a more effective model than the “Old Kent Test” in that it both reduces the effect of coaching and also places more reliance on curriculum assessments. However, the low pass standard is certainly letting too many children into grammar school who may not be up to standard, and I once again expect it to rise for next year, when the message gets home that English and maths are the key to success. It does beg the question that, if with most children who fail, the deciding factor is the English or maths, what then is the purpose of the reasoning, the most coachable element of the test?

Wider social Access to Grammar Schools

Then there is the recent Report of the Sutton Trust, committed to promoting equality of opportunity, into the performance of grammar schools.

It is highly critical of too many grammar schools for shutting the doors of opportunity and rewarding the effects of coaching. In typical Trust fashion it does come up with some possible ways forward, several of which are already developing in Kent.

Sutton Trust Recommendations

1) Ensure the testing system does not disadvantage pupils from low and middle income backgrounds.

2) Provide a minimum ten hours test preparation for all pupils to provide a more level playing field.

3) Improve outreach work significantly, actively encouraging high achieving students from low and middle income backgrounds to apply.

4) Schools should consider the merits of powers available in the admissions code to attract high achieving students who are entitled to the Pupil Premium.

5) Primary schools could do more to encourage their high achieving children to apply to grammar schools in selective areas, and develop partnerships with grammar schools.

6) Build new partnerships with non-selective schools to support their high achieving students

Comments:

1) Any Test will place such children at a disadvantage. The task is to minimise this, as can be seen in several of the new local alternative tests, such as the ones in Dover and Folkestone.

2) An excellent proposal, especially as the only research (now quite old) suggests this is the optimum amount of time for preparing for reasoning tests with a specific pass mark (the grammar school standard).

3) This is already happening with at least two grammar schools, those in Sittingbourne, who run “Aim High” classes for primary school children.

4)Queen Elizabeth’s Grammar School in Faversham offers this as a priority in its oversubscription criteria.

5) I agree, but there is much good practice around.

6) There are already examples of this in existence, such as the one between Canterbury Academy and Simon Langton Boys’ Grammar School.

Conclusion

As the Sutton Trust and others remind us, admission to grammar school can be fiercely competitive. Outside the state schools, many private schools exist primarily to secure places at grammar school. The large and rapidly expanding tutoring industry can be very effective in the same aim, where it is good, but it is as yet unregulated – so beware. This is an understandable outcome of parental aspirations to get their children into grammar school or perhaps the right grammar school, some going so far as to use both private and coaching routes to get there, in extreme cases, from the age of five.

Kent has a selective system, comprising both grammar and non-selective schools, which overall performs above its expectation at both GCSE and A Level. There is no appetite to change this, so the task is to make the selection process as fair as possible and I believe the new test and the admission flexibilities are a step in the right direction; implementation of the recommendations of the Sutton Trust on a wider scale would take it further.

Last modified on Wednesday, 30 December 2015 20:12