Nonsuch 2010 top 80 and in catchment cut-off scores

[runningmum]

Thanks stressedparent,
That explains it well.

[WP]

My query relates to the fact that the example given by NFER shows a range from roughly 70 - 140 ie. the lower 50 % of the population squashed within 30 marks whilst the upper 50 % have a spread of 40 marks.

As stressedparent says, they normalize to a symmetrical bell curve with an average of 100 and standard deviation of 15. Then any scores that would have been over 140 are reported as 140 (or 141 in some places), probably because the distinctions are unreliable that far out (and they don't need them). The result is that the number of people getting each score tends to decrease as you move from 100 to 139, following the normal distribution, but then you get a spike at 140, because the tail of the bell curve has been squashed up at that point. They similarly truncate the curve at the bottom end at 70.

[WP]

Maybe the normalisation is not done just vs the population that sat the exam but against a national population, or maybe the difficulty of the test has an effect on the scores (but not on the ranking).

Different school, but here's an incidental quote from a recent adjudication that went against Tiffin on a couple of minor matters:

Standardisation of test scores (in this case, a verbal and a non-verbal reasoning test provided by the NFER) is the process by which a raw score in the tests is adjusted to take account of the age of the individual candidate. This is done by obtaining the scores of a sufficient number of candidates (the standardisation sample), who are representative of the group taking the test (the test sample), to establish what the performance of this group generally will be. So the effect of the age of candidates generally can be seen and then “removed” from the scores of individuals in the test sample. This means that, in order for the standardisation to be accurate, the ages of pupils in the standardisation sample must be the same as the ages of pupils in the test sample. However, my understanding of this process is that by definition the two are distinct groups, and so I would expect that allowing pupils from outside the age range to take the test would have no effect whatsoever on the standardisation of scores of those within it.

[Giulio]

Thanks WP
when my DS took the Wallington test, they added 1 mark for age (born in December) and 3 marks for early test (it was before the CAF deadline).
I suppose the latter was given to everybody that sat the same test.
For the age adjustment, do you have any more details of how much that is? For example if the child was born in February or in July.
Maybe other parents have examples to share?
G

[stressedparent]

Hi Giulio,

When it comes to age standardisation, statistically no body is awarded or deducted any marks for been younger or older. The way it is done is very fair to all the children sitting the test. They split the cohort into 4 different groups SEP-NOV, DEC-FEB, MAR-MAY and JUN-AUG. every group will have it own average mean score, this is used to compile the standardised score. On average older children will have slightly higher average mean score? This varies from year to year. Lets assume the younger cohort this year were very bright which will produce a higher mean average then age standardisation becomes very negligible or nearly irrelevant. 11 years old will be compared with an 11 year old and vice versa. This keeps it fair I guess.

stressedparent

[la boume]

What i heard recently, is that if the child born for example in December and the child born in July (again, for example) scored the same amount of marks, the younger child would be placed higher on the waiting list. Don't know is it true or not.

[KeepCool]

What i heard recently, is that if the child born for example in December and the child born in July (again, for example) scored the same amount of marks, the younger child would be placed higher on the waiting list. Don't know is it true or not.

No, that's not true - it goes on distance to the school.

[la boume]

I thought so before I spoke with one mum whose daughther narrowingly missed Tiffin Girls. She claims that the distance rarely applies as all the places from the waiting list went to summer-born girls. But officially it is the distance criteria. It will be interesting to hear from the other parents whose daughters made it to Tiffin through the waiting list.

[mitasol]

The waiting list should be run strictly according to the original over-subscription criteria.

[WP]

What i heard recently, is that if the child born for example in December and the child born in July (again, for example) scored the same amount of marks, the younger child would be placed higher on the waiting list. Don't know is it true or not.

Younger children tend to score lower than older children, so if they had the same raw marks the July-born girl would have scored more highly in comparison with others of her age than the December-born girl, and thus would get a higher standardized score. But there's no reason to expect that a greater proportion of July-born girls would get in than December-born one.