The following section repeats and expand the headline
The pass mark for 2014 entry was a minimum of 118 for each of the three papers, verbal reasoning (VR), non-verbal reasoning (NVR) and mathematics, together with an aggregate score of 360. This standard is determined to allow 21% of Kent children to pass (excluding a minority from what are called the outer areas). The other 4% of successful candidates taking it up to 25% comes from a separate process, called Headteacher Assessment (HTA). This latter process is explained in my information pages. For the national sample, a score of 113 is sufficient to become part of the top 21% of children.
As part of the standardisation process, children are only compared with children from the national sample born in the same month, so that age is not a factor. Last year I carried out an analysis on age related results and confirmed this for the Kent Test, although in Medway younger children perform significantly worse. The following analysis does not apply to Medway which has “local” standardisation, a very different process from Kent, explained here.
Nearly every child taking the Kent Test will have been prepared in some way, if only by being exposed to the practice papers that are part of the testing process or by their parents buying a few example papers and working through them for familiarisation. Others more fortunate will attend an after school booster club. In a state school, this will not explicitly be called an 11 plus preparation session, as 11+ coaching is banned from the curriculum. But, as a result of this preparation, the national standard for selecting 21% of children rises to some degree from the 113 of non-prepared children. You will find a good basic article on the process of standardisation of 11+ tests here.
However, I do not accept this explains the majority of the rise, with the coaching industry in full flow, not just individual tutors but increasingly commercial organisations chasing business opportunities. Many private schools owe their existence to a coaching culture for the 11 plus, including curriculum time.
The numerical result of my analysis for Kent candidates can be summarised as follows.
Kent Test Scores September 2013
Number of Candidates
|Scoring 118 or above||5944||5209||3660|
|Scoring 138 or above||1995||1454||355|
|Scoring 141 by paper||1555||1120||225|
3660 children by paper
Conclusion One: Mathematics is by far the likeliest determinant of whether a child passes or fails as children in general perform far better on their two reasoning tests, especially verbal reasoning, than on mathematics.
A total of 3167 Kent children passed the Kent Test before HTAs were taken into account, all of whom are amongst the 3660 maths passes above. Another 3941 Kent children, who were not given an automatic pass, failed on their mathematics (and possibly one or both of the other two tests). Just 493 passed the mathematics test and failed on one or both of the other two. 2003 children passed their verbal reasoning but failed on one or both of the others.
The child who just gains a pass in their mathematics at 118 will on average score 7 points more in NVR and an astonishing 11 points in VR. Independent Appeal Panels considering appeals for grammar school will be aware of this expected differential as it is reflected in most of the cases they see. Parents preparing appeals should be aware of Panel expectations.
Conclusion Two: The only logical explanation of why the pass mark has risen by five marks over that expected of a child who has carried out no preparation, is because of coaching, for I cannot see that practice tests (set as part of the test process) will have made that much difference. What is more stark is the gap that opens between scores in maths and the two reasoning tests which comes into far sharper focus looking at the high scoring children. At the top of the range, just 161 Kent children scored the maximum aggregate score of 423. This excludes the 1394 children who gained the maximum 141 in verbal reasoning but not in one or both of the other two papers. BY contrast, just 64 children with 141 in maths did not gain maximum marks in all three papers.
In total therefore, more than four times as many children scored 138 or more on their NVR than in their maths, and more than six times as many with 138 or more in their VR. The pattern at 141 is even more extreme.
So either it is much easier to increase marks by coaching verbal and non-verbal reasoning tests than maths, or else tutors are failing to give a high enough priority to maths preparation.
Conclusion Three: The Judd School is absolutely right to shift away from reasoning tests and focus on curriculum achievement (not a conclusion I was expecting or have assumed in the past) where there is less evidence that coaching can make such large differences. Other grammar schools should take note although some remain in denial about the effects of coaching.
Conclusion Four: The new Kent Test to be taken in September may dilute the problem somewhat with the introduction of an element of multiple-choice English, but does not address the central issue identified here, which is the limited value of VR and NVR Tests. I can’t therefore see it lasting too long.
Conclusion Five: Medway’s system of “Local Standardisation” eliminates the problem of Conclusion One. I am highly critical of the Medway pattern of tests, which gives undue weight to a single piece of extended writing (40% of the total marks). However, by reducing the influence of the VR and NVR, Local Standardisation has much to commend it, even if it is not easy to understand (a KCC Report on possible changes to the Kent Test did not!).
Conclusion Six: If you can afford it, get some good coaching or other preparation. I hate this recommendation, but children who are not coached are at a considerable disadvantage.
Note: I have heard it said that the maths test was harder than last year, because fewer children scored maximum marks. This conclusion is false. Standardisation is designed to remove any differing level of difficulty from the mark allocation. The pass mark of 118 is the same as last year, so there is no difference in the standardised marks.