Supporting Families
  • banner2
  • banner4
  • banner12
  • banner6
  • banner7
  • banner11
  • banner13
  • banner10
  • banner9
  • banner8
Thursday, 13 November 2014 00:00

The Conundrum of Kent Test scores solved

Like many others, I have puzzled over the low pass mark of 106 required in each of the three assessments of English, Maths and Reasoning to produce 21% of children taking the Kent Test assessed of grammar school standard.

The Tests are nationally standardised so one would expect an untutored child on each test to score 113 to come in the top 21%. An initial look at these figures might suggest that Kent children are less bright than average, but a closer investigation of scores for the individual subjects, shows a very different picture and provides a full explanation of the conundrum.

Quite simply, whilst the majority of children have scored considerably more highly in the reasoning test than in the mathematics or English, a large number have failed to reach the standard in one of maths or English, dragging down the pass mark to provide the numbers.

As a result 4446 Kent children reached the pass level of 106 in English, and 4884 in maths, out of a total of 9902 taking the test, but less than half this figure will have passed in both!

In summary, Kent children have outperformed the national standard in all three assessments, whether through natural ability or the effect of tutoring on maths and English being open to question. However, the tutoring effect is still seen to the full in the Reasoning assessment, although this now counts for just one third of the assessment compared with the two thirds of previous years.

In my view, this data shows the new Kent Test has been highly successful if its aim was to select children with ability in both maths and English, and reduce the effect of tutoring, although the days of the bright male mathematician whose literacy skills are poor are over, if this pattern is repeated in future years.

The Judd School, which has been influential in the design of the new test, with its call to reduce the effect of coaching and improve standards of literacy in its intake, should be well pleased with this outcome and is surely likely to back off from its plan to introduce its own test for the 2016 intake.

I have already published an article on the Kent Test outcomes, and another on my reflections of the admissions season this year, both of which now need to be read in the context of the above. As soon as I receive the necessary data from KCC, I shall also publish a full analysis of Kent (and Medway) test outcomes. .....

Nationally, on the standardised tests, children should score 113 to finish in the top 21%. In Kent, girls needed to score 116 in each of English and maths, comfortably outperforming the national norm. For boys it is 114 for English and 118 for maths. However, for the reasoning tests, the 21st percentile rises to 121 for both boys and girls showing the powerful influence of tutoring in this subject.

There will still be children who have passed on scores of 106 or thereabouts in both English and maths, who appear to be about the 40th percentile in both subjects, and may well struggle at grammar school. KCC could have cut these children out by asking for a higher score in just one of English or maths to retain the 21% level, but this of course would have defeated its original aim. A number of these will in any case have lost out on a grammar school place through failure to reach the 320 aggregate required of all successful candidates.

Appeal Issues

For other families, what does this mean for chances at appeal? There will clearly be a very high number of children appealing who have passed in two out of three subjects, but one can hardly see appeal panels upholding more cases as a result even when, as I predict, the number of appeal cases is likely to soar.

Whilst I can think of a couple of boys’ grammar schools that would still like to see priority given to the high scoring mathematician, I don't envy the appeal panels trying to choose between some 1500 Kent children spread amongst the 33 Kent grammar schools who have missed an automatic pass by three or fewer standardised marks in one test, a very high proportion of whom will appeal for a grammar school place.

I am sure that many appeal panels will be looking for a good score in either maths or English to balance a miss in the other subject, rather than just a bare pass level of 106.

What is sadly clear is that a standardised score of 100 in a single test remains the sign of a child in the 50th percentile, or who is completely average, and appeal panels are still going to require very strong evidence to up hold an appeal with a score of 100 or lower against the many others who are much closer in terms of marks.

Exclusions

None of the above makes reference to the additional four percent of Kent children found selective by Headteacher Assessment, Panels having taken into account not only the test results, but Headteacher recommendation, the pupil’s classwork seen, the written English Test and grades achieved at school.

Nor does it look at outcomes for children from outside Kent who tend to fall into higher ability levels, hence the decision to look at Kent grammar schools, and so cannot be analysed in the same way.

Last modified on Sunday, 25 January 2015 17:02

1 comment

  • Comment Link Sunday, 15 May 2016 11:23 posted by James Coombs

    Hi, I'm very interested to hear that Kent tests are nationally standardised. I've been trying to get raw test marks from CEM because these are 'locally standardised' (sic) in that individuals scores are calculated against the mean and standard deviation of just those who've applied preventing any comparison between different schools.

    Admissions law says you need to be of 'grammar school standard' to be admitted but whilst in Kent 28.4% of places are selective in Devon it's only 1.7% http://trak.org.uk/proposed-annexation-of-devon
    Either Devon children are *really* falling behind Kent ones, or alternatively CEM's describing these results as 'standardised' is an abuse of the word amounting to wholesale public deception.

    CEM won't release the raw test marks because they say that would make it possible for tutors to teach to the test. Unfortunately the Information Tribunal has fallen for this fairy tale so the fact that GL are providing properly standardised scores could come in handy if I'm allowed to take this to the Upper Tribunal.

    Whilst anyone who tries to deny that tutoring skews the results is clearly living on a different planet I'd split hairs slightly over the inclusion of the word "untutored" in your second paragraph. If the score is standardised then tutored or untutored alike, a score of 113 reflects a position in the top 21% (... erm actually 19.3062% according to Excel but point taken it's not 106). PETER: The issue is surely what is a grammar school standard? In Kent it is defined by the 21%. In Devon it can be completely different. Medway, as with CEM uses local standardisation to meet its definition of a grammar school child as in the 23%. I have never found raw scores helpful, although in Kent they are made available on request. Just lead to lots of pointless analysis and conclusions that go nowhere. I advise parents accordingly. I have re-read my second paragraph and am happy that the meaning is clear. Having looked at your website, as is well known in this part of the country, I also am a strong supporter of grammar schools being for local children.

Leave a comment

Make sure you enter the (*) required information where indicated.
Basic HTML code is allowed.