Standardised scores - Kent
Moderators: Section Moderators, Forum Moderators
Re: Standardised scores - Kent
I have just been reading up on the Bell curve, standard deviation Iq tests etc and unless I have made a mistake there is no reason to think a score in Kent Test of 390 would be top 2%.
Because if you look at the graph and info here: http://www.nfer.ac.uk/nfer/research/ass ... scores.cfm" onclick="window.open(this.href);return false;
when a score of 130 (390) =2%
then score 120 (360) is in top 10%
But we know those pass Kent test at 120 (360) are top 20-25%
Because if you look at the graph and info here: http://www.nfer.ac.uk/nfer/research/ass ... scores.cfm" onclick="window.open(this.href);return false;
when a score of 130 (390) =2%
then score 120 (360) is in top 10%
But we know those pass Kent test at 120 (360) are top 20-25%
-
- Posts: 3
- Joined: Mon Nov 05, 2012 10:29 pm
Re: Standardised scores - Kent
Assuming a normal (i.e. bell) distribution all that is required to define the shape is the mean and standard deviation of the aggregate scores. We know 11,451 children took the exam, and the numbers at any point can be calculated (or looked up in stats tables) as area under the curve (which equates to probability).DebP wrote: We would need more data to check the shape of this curve is as discussed in previous posts. I am considering about the best wording for a FoI request...?
Thinking about this, the real curve (as opposed to the model predicted) will be distorted because the standardisation procedure imposes artificial minimums and maximums of (I think) 69*3 and 141*3 respectively.
This is correct.DebP wrote:I have just been reading up on the Bell curve, standard deviation Iq tests etc and unless I have made a mistake there is no reason to think a score in Kent Test of 390 would be top 2%.
Because if you look at the graph and info here: http://www.nfer.ac.uk/nfer/research/ass ... scores.cfm" onclick="window.open(this.href);return false;
when a score of 130 (390) =2%
then score 120 (360) is in top 10%
But we know those pass Kent test at 120 (360) are top 20-25%
The NFER figures are based on mean 100 and standard deviation 15. A score of 130 is 2 standard deviations above the mean. In stats terms the z-value is 2, giving a p-value of 0.02275, so you would expect 2.275% to exceed 130.
In Kent, substantially greater than 2.275% score higher than 390, I suspect due to self-selection and tutoring.