Newby question about standardisation

Eleven Plus (11+) in South West Hertfordshire

Moderators: Section Moderators, Forum Moderators

11 Plus Platform - Online Practice Makes Perfect - Try Now
Cutting42
Posts: 186
Joined: Thu Nov 17, 2011 10:06 pm

Re: Newby question about standardisation

Post by Cutting42 »

I started writing out a more comprehensive explanation and found the below in this very forum. This is a pretty good explanation:

http://www.elevenplusexams.co.uk/advice ... xplanation#" onclick="window.open(this.href);return false;
thisisnuts
Posts: 797
Joined: Thu Mar 05, 2015 11:23 am

Re: Newby question about standardisation

Post by thisisnuts »

Thanks Cutting

That was the first explanation I came across when I started browsing this forum. The NFER website says much the same. But neither of them give a sample table going up to 100%.

There are still conflicting opinions on whether a 100% raw score will always give a 141 standardised score, regardless of age.

As DG advises, our target is QEB followed by schools further away that have a reasonable out-of-catchment allocation.

Actually discussed the issue with DS. Thankfully, he's amused by the Parmiters situation rather than upset.
madsystem
Posts: 64
Joined: Tue Mar 13, 2012 8:44 pm

Re: Newby question about standardisation

Post by madsystem »

Hi i.ve pmed u
ToadMum
Posts: 11946
Joined: Wed Jan 18, 2012 12:41 pm
Location: Essex

Re: Newby question about standardisation

Post by ToadMum »

I may be wrong, but isn't age standardisation done by comparing within age group divisions -children aged 11 years being compared with other children aged 11 years, 10 years 6 months with 10 years 6 months etc?

In that way, the highest score (or scores above a certain raw score) are allocated a standardised score of 141, the lowest raw score(s) 69 or whatever. That way a standardised score of e.g. 121 is 'the same' across age groups regardless of the underlying raw score.

(Of course it could work the other way if you have a particularly high scoring group of younger kids / lower scoring autumn borns).
Outside of a dog, a book is a man's best friend. Inside of a dog it's too dark to read.Groucho Marx
thisisnuts
Posts: 797
Joined: Thu Mar 05, 2015 11:23 am

Re: Newby question about standardisation

Post by thisisnuts »

ToadMum

What you're saying makes sense - that there are separate distribution curves for 11 year olds, 10.5 year olds and 10 year olds (and all ages in between?) - so an average older child will get the same standardised score as an average younger child, even if the younger child's raw score is lower. This also suggests that the very brightest children in each age group should have a standardised score of 141.

The following extract is from the detailed explanation of standardised scores that Cutting42 posted a link to earlier:

"If both children in this example – the oldest one and the youngest one – achieve the identical “raw score”, the youngest child’s final 11+ score will be higher than that of the oldest child. The standardisation process has “awarded” extra marks to the younger child to compensate for their younger age."

This mostly supports your explanation, as a younger child scoring (for example) 75% will presumably be further along the distribution curve than an older child, so will have a higher standardised score.

But if a young and old child both have raw scores of 100%, and the younger one WILL have a higher final score, then how can the older child get 141?

I can't even bother to be confused any more. :?
scary mum
Posts: 8841
Joined: Mon Mar 15, 2010 3:45 pm

Re: Newby question about standardisation

Post by scary mum »

My understanding is that you are correct - children are out into a group with others birn in the same month & their scores plotted on a standard curve. Therefore if, for example, you are a particularly bright set of July-born children their raw scores could in theory match those of the September-borns. The slight oddity about the curve is that there will be a kick up at the right hand end of the curve, so more children will get the top score than the other high scores leading up to it. So you could have (plucking figures out of the air). 5 children with scores of 135,136,137,138,139 & 140 but 20 with scores of 141 because you can't score more than that, so whatever month the child is born, if they score 100% they will get 141. I think this is correct, but someone like Sally-Anne should be able to answer your question.
scary mum
Hera
Posts: 856
Joined: Mon Feb 22, 2010 3:50 pm

Re: Newby question about standardisation

Post by Hera »

Kids will be compared with the cohort in the month they are born and each month will have a different average as it will be taken from the kids in that month. If the average score for a child born in December is 70%, 70% will be 100 if the average score for an August born child is 69%, 69% will be 100. There will then be standard deviations of 15 so the sores will go up, up to 100%. Therefore although the average starting point may be different for children from different months they will all get to 100% which is 100%, it is the marks inbetween which can vary. In reality it is not that simple but that is how it works.


Cross posted with Scary Mum
Turboless
Posts: 39
Joined: Wed Oct 08, 2014 12:05 am

Re: Newby question about standardisation

Post by Turboless »

My understanding is that you don't actually need full raw marks to get top marks after the standardisation is applied. If you look at the standardised example (it's just an example though) on the page with all the admission stats page you'll see that the top 0.31% of the population would get full marks of 282.

This should mean that both the oldest and the youngest children should be able to achieve full marks after the standardisation has been applied (I'm pretty sure they would have have to enforce this otherwise it would be unfair). However the younger children would need a lower raw score to achieve full marks.

Where these exams are starting to fall down is they were designed to be used for the 11+ which was to try and differentiate between the top 20-25% of the country but now they're being used to separate the top 2-5% of the country which in a lot of cases simply comes down to how many silly mistakes a child makes on the day under pressure.
loobylou
Posts: 2032
Joined: Thu Nov 27, 2014 5:04 pm

Re: Newby question about standardisation

Post by loobylou »

I think that, if you apply your logic fully, then it would mean that no child who wasn't in the August cohort would be able to get full marks. That definitely isn't true because I know March born children who did achieve the full marks. However I don't really understand how it actually works in practice!
Turboless
Posts: 39
Joined: Wed Oct 08, 2014 12:05 am

Re: Newby question about standardisation

Post by Turboless »

No it would be possible for all ages to get 100% standardised results.

Here's a simple example the actual numbers are made up but should illustrate the point:

e.g.
If when they worked out the standardisation (which would have been done years ago getting many children to take this exam and the other exams which they standardise against) they find the following:

0.31% of the oldest children in the year get 90% or over (raw score)
0.31% of the youngest children in the year get 85% or over (raw score)

Then any child taking the exam getting over 90% (raw) would be guaranteed to get full standardised marks however the youngest of the year could get 85% (raw) and get full marks but the oldest would need to get over 90% (raw).
Post Reply