To those who think A 145 cogat or nnat score is not rare, you are wrong.

[Anonymous]

OP here...Thanks to 16:35 for reviewing what I wrote instead of silly insults that others wrote.
So moving along. I actually agree it does feel like everyone and their "Dog" makes the 130 cut to make the pool for GT but there is a huge difference scoring 130 than a 145. I Think this is where the confusion is setting in.
http://www.iqcomparisonsite.com/IQtable.aspx
A score of 130 means a rarity of 1 out of 33 based on the chart above. That means on each test there are at least 394 kids who score 130 or above. With 4 tests that means there should be a minimum of 1576 kids who make the pool. With Fairfax county students being well above the national average you can at least multiply 1576 *1.5 which equals 2364 students in the pool initially. Again this is a forecast. 2364/13000 = 18% overall so I could see how it can feel like "everyone and their dog" qualified to make the pool initially. Of course not every kid who scores 130 makes it and of course some kids make it who do not score above 130.
I think the system can be gamed a little but the difference between scoring a 130 and a 145 is huge on the nnat, cogat, or an Iq test.

On a side note I think this is interesting.
http://www.iqcomparisonsite.com/GREIQ.aspx
The average sat score at langley is 1210 or about.
Based on the chart you have average kids at langley who have iqs around 128-129.6. Yes this is an estimation as well as it says on the chart, but I think it says a lot about the type of kids we have going to school in the area.

[Anonymous]

NNAT and CogAT are achievement tests, not IQ tests. Why are you taking those scores and attempting to treat them as IQ scores for comparison and "measurement of rarity?"

[Anonymous]

I answered your question in the OP. Because Whether its the NNAT or an IQ test, they are scored the same. The rarities are the same nationally and I used those figures to forecast numbers for fairfax county. If someone scores a 145 on the nnat its just as rare (or very close to it) as someone scoring a 145 on an iq test. Yes it is well understood that someone scoring a 145 on NNAT does not mean they will score a 145 on IQ or vice versa. The premise of the OP was that many on this site seem to think thousands of second graders are scoring over 145 on one of the four AAP tests and thats just not true. I am not treating the "achievement scores" as IQ scores. Again the point is the rarities are the same for both.

[Anonymous]

I think the CogAT author makes a good point in his presentation:

http://faculty.education.uiowa.edu/dlohman/pdf/Best%20Practices%20in%20Using%20Standardized%20Tests%20for%20Talent%20Indetification.pdf

where it may be better to compare students to local norms vs. national norms, as it compares classmates.

[Anonymous]

Anonymous wrote:I answered your question in the OP. Because Whether its the NNAT or an IQ test, they are scored the same. The rarities are the same nationally and I used those figures to forecast numbers for fairfax county. If someone scores a 145 on the nnat its just as rare (or very close to it) as someone scoring a 145 on an iq test. Yes it is well understood that someone scoring a 145 on NNAT does not mean they will score a 145 on IQ or vice versa. The premise of the OP was that many on this site seem to think thousands of second graders are scoring over 145 on one of the four AAP tests and thats just not true. I am not treating the "achievement scores" as IQ scores. Again the point is the rarities are the same for both.

Thank you, OP!

I am the poster of one of the other threads. THIS is what I was trying to ask.

Not AAP worthiness, not prepping vs without prep, but statistical rarity of the higher score.

I appreciate your explaination. Thanks!

[Anonymous]

Am I the only one that suspects that the OP keeps replying to his/herself on this thread?

[Anonymous]

As the OP this would be post 5 of 21. My time is too valuable to sock puppet. I cant think of one good reason why I would even want to sock puppet on this thread or any other for that matter.

[Anonymous]

My DC got 160 on NNAT. I suspect DC's IQ is mid 140.

[Anonymous]

Anonymous wrote:When you consider the fact that there is an endless supply of prep materials for the cogAT and virtually none for the WISC, this thread loses all meaning.

nut what if i was in second grade and never studied and got a 145

[Anonymous]

TL;DR

[Anonymous]

Most tests that are supposed to follow a normal distribution don't quite do it at the higher end due to ceiling effects and the Flynn effect (accelerated by increased preparation effects) over time--so there's often a bump in scores at the highest level.

[Anonymous]

Anonymous wrote:My DC got 160 on NNAT. I suspect DC's IQ is mid 140.

My friend's daughter got a 160 on the NNAT. She's your average smart/hard working kid that's a dime a dozen in this area. IQ probably in the 120s, same as my daughter.

[Anonymous]

Anonymous wrote:NNAT and CogAT are achievement tests, not IQ tests. Why are you taking those scores and attempting to treat them as IQ scores for comparison and "measurement of rarity?"

Yes. Her snowflake is extra b/c of their group test score. A real IQ score doesn't matter to her b/c she doesn't have one. Her kid will get into AAP but it will be too easy for them b/c they are so extra with their 145.

OP, 145 IQ is different, like it or not. But go prep and get thee to George Mason. They do iq testing for a few hundred. I'm sure you'll get the score you want and if you don't, you'll still have your group tests.

[Anonymous]

There are two huge problems with your premise, OP.

The first is that the NNAT and CogAT are normed on a group of kids who are taking the test blind, without any prep or preview of the types of questions being asked. FCPS has a significant kids who prep, and thus will get inflated scores relative to their actual ability.

The second issue is that CogAT and NNAT are multiple choice tests with a very low ceiling. They are not designed to differentiate between a kid with a 140, a 145, a 150, etc, and it isn't even possible to norm the tests correctly at the high end. All of these tests have disproportionately many kids scoring at the top ends.

[Anonymous]

Anonymous wrote:Many post on here its not rare to see nnat or cogat of 145 but say it is rare to see 145 iq scores. The reason this is wrong is simple. In general they are scored the same way. Granted nnat/cogat are supposed to be scored at a standard deviation of 15 instead of 16 they both show significant rarity. Google iq rarity to view to rarity chart. For example an iq of 145 with standard deviation of 15 means the rarity is 1 out of 741 nationally. With an iq of 145 and standard deviation of 16 which is what I believe nnat cogat is using the rarity goes to 1 out of 407 nationally. So with fairfax county I think you can expext the students to be well ahead of the national average. As a guess I would say that rarity goes to 1 out of 200 for students scoring over 145 on one of the tests. With 13000 2nd grade students that means I would forecast 65 to have scored 145or better on any one of the tests. Lets say we go back to a 16 standard deviation for iq tests for fairfax county I would forecast 35 students have above a145 iq. 13000/741 for national rarity then a fairfax county adjusted forecast times two the national average to come up with 35. In general whether its a 145nnat /cogat or iq test the rarities should be close since they are scored basically the same way. To say a 145 nnat or cogat is common in relation to iq is wrong again because the methodology for scoring has to be very close. Only question is standard deviation used but even with that the rarities still cant be that far off unless you think 35 of 13000 is rare but 65 of 13000 is not.

You're naive. For all of these models (cogat, nnat, nwea, wisc, iq tests) the model fit basically breaks down at the 97.5 percentile.
So whatever "rarity" you compute by converting the z-Score to an area under a Gaussian has such a high error rate that it's essentially meaningless.
This is why nobody reports how well the models fit in this range, and nobody reports the ground truth in the sampled population (what percentile actually scored in this range).