Statistics can be manipulated to say what you like?

Eleven Plus (11+) in Birmingham, Walsall, Wolverhampton and Wrekin

Moderators: Section Moderators, Forum Moderators

11 Plus Platform - Online Practice Makes Perfect - Try Now
Dopey2012
Posts: 24
Joined: Sat Oct 13, 2012 10:46 pm

Re: Statistics can be manipulated to say what you like?

Post by Dopey2012 »

Isn't it swings and roundabouts? How about the very bright child who even on "top table" is not working to best of his/her ability? That child also misses out on extension activities at the requisite level? Looking at the table, is the mean "pass" mark on an upward curve? There do not appear to be huge surprises so can we deduce that one year is moderated in relation to all past years to truly standardise results ?
Okanagan
Posts: 1706
Joined: Mon Aug 22, 2011 9:20 pm
Location: Warwickshire

Re: Statistics can be manipulated to say what you like?

Post by Okanagan »

Dopey2012 wrote:Isn't it swings and roundabouts? How about the very bright child who even on "top table" is not working to best of his/her ability? That child also misses out on extension activities at the requisite level?
Been there done that. And the "we have to show two sublevels of progress a year, so we can't rate them now at above level x or we wouldn't be able to show progress excuses. It shouldn't be about ticking boxes to show progress, it should be about actually making progress!
Dopey2012 wrote:Looking at the table, is the mean "pass" mark on an upward curve? There do not appear to be huge surprises so can we deduce that one year is moderated in relation to all past years to truly standardise results ?
You would expect if the proportion taking it is roughly equal, the absolute standard of those ranking high enough to get a place would be roughly equal too, year on year, once it has been standardised. Passmarks for GCSE's (supposedly, but that's a whole different debate!), SATs, degrees etc are set to ensure a consistent standard over time, but as the only purpose of 11+ is to determine entry, on a local level, then it's only relative to the cohort.

Variation in the numbers required for entry, doesn't necessarily mean variation in the absolute standard though. If you start with a scenario where the top 32% of the whole Y6 population take it. But the following year the top 40% take it and all the additional entrants score lower than the original group. In the first year a candidate X who scored better than 50% of those who took it would get a standardised score of 100, but in year two X would be ranked ahead of 60% of the new cohort, with an expected score of 103.8. Similarly a candiate Y ranked in the top 25% of the original group would get a score of 110.1 - but would be in the top 20% and score 112.6 in the new group. The score has changed, but the absolute standard hasn't - e.g. if the number of places was fixed at 8% of the whole population it would be the same candidates who would form that 8%, albeit armed with slightly higher scores.

It will be interesting to see what, if any, effect moving the exam forward, and in front of the Warwickshire one has. The figures from Warwickshire last year suggest that there were quite high numbers who took both, but then didn't apply to Warwickshire, giving themselves a free CEM mock, and extra preparation time to work on any areas of difficulty between the 2 exams. That opportunity wasn't there this year.
Last edited by Okanagan on Sun Oct 21, 2012 12:55 pm, edited 1 time in total.
Dopey2012
Posts: 24
Joined: Sat Oct 13, 2012 10:46 pm

Re: Statistics can be manipulated to say what you like?

Post by Dopey2012 »

Thanks for the detailed reply. The dice have been rolled its just that they don't land until March!
Post Reply
11 Plus Mocks - Practise the real exam experience - Book Now