Standardisation
Moderators: Section Moderators, Forum Moderators
Having read the standardisation table on this forum, I was under the impression that it's not that large a % point allocation as we might hope / be led to understand, but then I may have misinterpreted it...my DS is September b'day but year ahead, and took 11+ on 10th birthday. He scored max marks, so I checked his raw scores: 85% maths, 88% NVR and 93% VR...he reads voraciously, but don't ask him to write a story - the style of the test suited him, I think. I'm sure there are lots of components to the individual scoring, not least having a good or bad day. Our problem has been finding schools to accept him 'early'...Good luck to all those embarking on appeals - not an easy time.
Thanks, Tipsy. DS has been ahead a year since year 2, but even with that there aren't many options (Only Sevenoaks of the local indes. was interested, and he has a place there - but the others might become a bit more keen in future if fewer can afford the fees) It happened to me, 40+y ago, kicked my heels for a year and then found it hard to 're-boot'...funny how some things don't seem to change....'rules is rules'
-
- Posts: 459
- Joined: Tue Jan 22, 2008 3:36 pm
- Location: Rugby
Standardisation
Please browse the Warwickshire section for a readable explanation of how the process works, it is a standard statistical approach applied to all kinds of data so that like can be compared with like. The fact that this technique is so misunderstood and mistrusted is a terrible indictment of mathematics teaching in this country, especially how it is regarded in the maintained sector where most teachers are not sufficiently qualified to teach it as the indispensible tool (among many others) that it is!
.
.
-
- Posts: 893
- Joined: Thu Oct 11, 2007 8:59 am
- Location: Cloud 9
Re: Standardisation
Sorry Sassie'sDad, I feel like I'm stalking you today (and I'm going to turn the PC off now), but this is unkind and untrue! On what evidence do you base this statement? If we're talking about unqualified teachers we need to look further than the maintained sector...Sassie'sDad wrote:The fact that this technique is so misunderstood and mistrusted is a terrible indictment of mathematics teaching in this country, especially how it is regarded in the maintained sector where most teachers are not sufficiently qualified to teach it as the indispensible tool (among many others) that it is!
.
This is a very interesting topic. I am quite familiar with statistics and using distrubutions.
As I understand, this means that you cannot compare marks from one year to another, unless the mean scores of all the candidates remains the same. For example if the mean mark on each test was 65/100 then the score would be standardised around a mean of 65%, if the following year the mean mark was 70/100 then the score would be standardised around a mean of 70%
So if a child gets 80/100, this would only translate to a standardised score of 100 ( ie 71% 100/140) if the mean score for the test was 80%.
Therefore if a child regularly gets 90% in NFER practice papers, that DOES NOT mean that the child would expect to get a standardised score of 126 (90% of 140) because you cannot know beforehand what the mean will be that the score is standardised against.
This would certainly explain why many people express surprise by the differences that their child achieves in practice papers and the actual standardised score
Before looking into this subject, I was concerned that my son would be marked down as he is November born, however, I am now much happier because I know that his score will be standardised against other kids born in the same month - and given the number of children who take the 11+ I would have no concern that there were enough kids to provide a statistically valid sample to standardise against.
Out of interest, does anyone know how many children take the same NFER tests?
As I understand, this means that you cannot compare marks from one year to another, unless the mean scores of all the candidates remains the same. For example if the mean mark on each test was 65/100 then the score would be standardised around a mean of 65%, if the following year the mean mark was 70/100 then the score would be standardised around a mean of 70%
So if a child gets 80/100, this would only translate to a standardised score of 100 ( ie 71% 100/140) if the mean score for the test was 80%.
Therefore if a child regularly gets 90% in NFER practice papers, that DOES NOT mean that the child would expect to get a standardised score of 126 (90% of 140) because you cannot know beforehand what the mean will be that the score is standardised against.
This would certainly explain why many people express surprise by the differences that their child achieves in practice papers and the actual standardised score
Before looking into this subject, I was concerned that my son would be marked down as he is November born, however, I am now much happier because I know that his score will be standardised against other kids born in the same month - and given the number of children who take the 11+ I would have no concern that there were enough kids to provide a statistically valid sample to standardise against.
Out of interest, does anyone know how many children take the same NFER tests?
-
- Posts: 1245
- Joined: Fri Jul 06, 2007 9:31 pm
Exactly! If the exam room is full of candidates that all were scoring 90% in practice tests who then score 90% in the real test they would still only get a standardised score of 100. If the top child got a raw score of 95 and the bottom one got a raw score of 85 then there would only be 10 raw points between the standardised 70 and 140 scores; given the gradient of the normal curve to which these are then fitted it is clear that a couple of raw points could make all the difference if a cut off is somewhere up in the 120s.mattsurf wrote:Therefore if a child regularly gets 90% in NFER practice papers, that DOES NOT mean that the child would expect to get a standardised score of 126 (90% of 140) because you cannot know beforehand what the mean will be that the score is standardised against.
It would be unfair otherwise as it would not compare children on a like-for-like basis.mattsurf wrote:I am now much happier because I know that his score will be standardised against other kids born in the same month
Regards
SVE
Animis opibusque parati
-
- Posts: 1806
- Joined: Mon Feb 20, 2006 2:29 pm
- Location: Berkshire
Mattsurf, this is rather interesting and I was with you, up until this point....
Could you try to explain this part again, please? I think I got the bit after, though.mattsurf wrote:
So if a child gets 80/100, this would only translate to a standardised score of 100 ( ie 71% 100/140) if the mean score for the test was 80%.