2014 Warks data - an interesting read.
Moderators: Section Moderators, Forum Moderators
2014 Warks data - an interesting read.
Warks County Council offered 6 places in error. 3 at Rugby High and 3 at SGGS
On 7th Sep, 2013 (first sitting) the mean score was 199 with 30% (515) being offered on 3rd March. 1700 sat the test.
On 14th Sep, 2013 (second sitting) the mean score was 217 with 22% (58) being offered on 3rd March. 260 sat the test.
What does this show if anything?
Can any of the following hold:
a. Children taking the test on the second sitting are on average cleverer.
Or is it second sitters knew content whilst first sitters did not?
b. A sub-group of children sitting on second sitting scored massively higher marks.
(The lower offers was because they didn't apply for places or massively higher marks for a few).
Did the sub-group scoring massively higher marks know content of the exam?
Is the 11+ compromised?
Food for thought.
On 7th Sep, 2013 (first sitting) the mean score was 199 with 30% (515) being offered on 3rd March. 1700 sat the test.
On 14th Sep, 2013 (second sitting) the mean score was 217 with 22% (58) being offered on 3rd March. 260 sat the test.
What does this show if anything?
Can any of the following hold:
a. Children taking the test on the second sitting are on average cleverer.
Or is it second sitters knew content whilst first sitters did not?
b. A sub-group of children sitting on second sitting scored massively higher marks.
(The lower offers was because they didn't apply for places or massively higher marks for a few).
Did the sub-group scoring massively higher marks know content of the exam?
Is the 11+ compromised?
Food for thought.
Last edited by gideon on Sat Jun 21, 2014 12:45 pm, edited 1 time in total.
Re: 2014 Warks data - an interesting read.
How many children sat the test on each day? - Thank you for adding the figures Gideon
I suspect a far greater number of children, with a huge range of abilities sat on the first. (The ability range was certainly large in DS's school.) The second, smaller cohort, the majority having sat the B'ham exam the previous weekend, would probably push up the average on that date ...
Just my thoughts ...
JD
I suspect a far greater number of children, with a huge range of abilities sat on the first. (The ability range was certainly large in DS's school.) The second, smaller cohort, the majority having sat the B'ham exam the previous weekend, would probably push up the average on that date ...
Just my thoughts ...
JD
2014 Warks data - an interesting read.
I updated it.. 1700 first sitting and 260 second.
So, is it the case 1700 is a wide range of ability and 260 are a small focused group and not of a wide ability range? Is a 9% difference significant?
8 childen sitting the test on 17th Sept received offers (mean 192 from 65 candidates).
Did these 8 know content? Low mean, so diverse scores.
On 10th Oct there were less than 10 sitters (actual number not disclosed) 2 offers and a mean of just 191.
So, is it the case 1700 is a wide range of ability and 260 are a small focused group and not of a wide ability range? Is a 9% difference significant?
8 childen sitting the test on 17th Sept received offers (mean 192 from 65 candidates).
Did these 8 know content? Low mean, so diverse scores.
On 10th Oct there were less than 10 sitters (actual number not disclosed) 2 offers and a mean of just 191.
Re: 2014 Warks data - an interesting read.
Well the extra months practice certainly didn't help the 10th Oct sitters!
JD
JD
-
Sally-Anne
- Posts: 9235
- Joined: Wed Jan 11, 2006 8:10 pm
- Location: Buckinghamshire
Re: 2014 Warks data - an interesting read.
If anything, the data proves the complete reverse.gideon wrote:Can any of the following hold:
a. Children taking the test on the second sitting are on average cleverer.
Or is it second sitters knew content whilst first sitters did not?
b. A sub-group of children sitting on second sitting scored massively higher marks.
(The lower offers was because they didn't apply for places or massively higher marks for a few).
Did the sub-group scoring massively higher marks know content of the exam?
Is the 11+ compromised?
Food for thought.
As JD says, the 260 who took the second sitting almost certainly took the Birmingham test the previous weekend. They had therefore already experienced test conditions, giving them a slight advantage. To add to that, children taking more than one 11+ test are more likely to have been tutored/be of selective ability or both. That creates the higher mean score for the second sitting.
However, the lower figure of 22% does indeed suggest that they did not apply for places, having been successful in the Birmingham test.
If the 11+ had been compromised, and the children taking the second sitting were in possession of test content, you would expect to see a higher mean score and a higher percentage of offers.
The lower mean score for the subsequent sittings suggests that those children were a more “all ability” group who were taking the test late due to illness or for family reasons.
2014 Warks data - an interesting read.
The higher mean score is already there - 9% higher.
In 1 week children may know a few syonyms, antonyms, general subject of comprehension, type of maths and NVR questions. They would not have access to the entire paper, but enough to score a few more marks.
Many may have applied for the Birmigham schools, but could the 58 offered be aware of content? Surely 1700 children and 3000+ adults cannot keep quiet.
One cannot say the second cohort were tutored any more than the first or they genuinely wanted a Warks grammar school - perhaps a backup. A few of the first cohort could have been from other areas eg Bucks using it as a free mock. Does tuition make a difference if the claim it is not required? (Joke!) Are you suggesting a take more than one 11+ is a good idea - i.e. taking free mock in Bucks to experience test conditons
The third sitting is primarily for religious groups who refuse to be tested on Saturdays, but seems a high number. School had started and people knew years in advance when the testing dates were. I see no reason for 65 children sitting the test so late. I wonder if all were for religious reasons.
I would have thought, if all things are equal and we believe what WCC and Durham claim (which I do not) then one would expect a similar mean on the second sitting. What is clear is the second sitting scored on average 9% higher and not 3-5%. I think this is a significant difference.
One would need the actual scores to do more analysis. I doubt Warks would release this.
In 1 week children may know a few syonyms, antonyms, general subject of comprehension, type of maths and NVR questions. They would not have access to the entire paper, but enough to score a few more marks.
Many may have applied for the Birmigham schools, but could the 58 offered be aware of content? Surely 1700 children and 3000+ adults cannot keep quiet.
One cannot say the second cohort were tutored any more than the first or they genuinely wanted a Warks grammar school - perhaps a backup. A few of the first cohort could have been from other areas eg Bucks using it as a free mock. Does tuition make a difference if the claim it is not required? (Joke!) Are you suggesting a take more than one 11+ is a good idea - i.e. taking free mock in Bucks to experience test conditons
The third sitting is primarily for religious groups who refuse to be tested on Saturdays, but seems a high number. School had started and people knew years in advance when the testing dates were. I see no reason for 65 children sitting the test so late. I wonder if all were for religious reasons.
I would have thought, if all things are equal and we believe what WCC and Durham claim (which I do not) then one would expect a similar mean on the second sitting. What is clear is the second sitting scored on average 9% higher and not 3-5%. I think this is a significant difference.
One would need the actual scores to do more analysis. I doubt Warks would release this.
Re: 2014 Warks data - an interesting read.
So are you accusing a majority of that second sitting as having cheated Gideon? Maybe I am naive but I would be shocked and astounded if this were the case. Most parents do their best but wouldn't dream of cheating, or even to put it more nicely, going to all the trouble to pre see the content in order to give their child an advantage.
I think the numbers mean it is not a statistically significant group in terms of comparing with the other, it's just an anomaly.
I think the numbers mean it is not a statistically significant group in terms of comparing with the other, it's just an anomaly.
Re: 2014 Warks data - an interesting read.
There is a simple explanation for this, the 1st exam includes lots of "no hopers" who were mandated to enter the exam by their Warks primary school(and my not even want to go to a grammar), whereas the exam 2 cohort will include a much higher proportion of children who are consciously targeting a grammar.
As has been stated, the 2nd exam cohort will include those targeting the B/Ham KE grammars as well
As has been stated, the 2nd exam cohort will include those targeting the B/Ham KE grammars as well
2014 Warks data - an interesting read.
No, I am saying a number of child will know things about the paper. They may know some synonyms, they may know if NVR was of a certain type and practice. They may know a certain type of maths question was in the paper and practice. Children talk, parents talk. Is this cheating? I say technically "No". Is it unfair? "Yes".Yamin151 wrote:So are you accusing a majority of that second sitting as having cheated Gideon?
I am sure you know tuition centres ask for feedback and use this to re-create questions. They also pass it on to late sitters. Many have seen mock papers from centres in the Midlands and have seen past questions as children had described (and even repeated in other areas).
If certain religions refuse to be tested on Saturday, why is the test delayed for 10 days for this group and not held the very next day as in Birmingham or the day before the main day? I cannot understand the logic. Why isn't the second sitting on the next day?
Clearly children remember content as content about the Birmingham 11+ appears on this site, I assume from children. I know many children who know they will score high marks pass on content to their "weaker" friends. Many children do not care and want their friends at the same school. Many parents would do anything to get grammar school places! Why isn't the test made more secure my holding the second date closer to the first?
Re: 2014 Warks data - an interesting read.
Your argument is condemned by your own statistics. Lower percentage offered places, full stop.
