Anonymous wrote:160 Nonverbal (>99), 132 verbal (98), and 132 quantitative (98). If she doesn't get in, then this is rigged.

Anonymous wrote:160 Nonverbal (>99), 132 verbal (98), and 132 quantitative (98). If she doesn't get in, then this is rigged.

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:My kid is at Navy. Lots of high test scores at Navy. Kid did not get universal screener

Navy is very tough. My kid had very high test scores and was in pool and still didn’t get in until 2 years later. Also it’s a center school, which means no principal placement. Center schools are disadvantage that way.

That’s incorrect. Center schools do have principal placements. We are in a center school and I know of a few kids who have bee placed in the AAP class by the principal.

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:My kid is at Navy. Lots of high test scores at Navy. Kid did not get universal screener

Navy is very tough. My kid had very high test scores and was in pool and still didn’t get in until 2 years later. Also it’s a center school, which means no principal placement. Center schools are disadvantage that way.

That’s incorrect. Center schools do have principal placements. We are in a center school and I know of a few kids who have bee placed in the AAP class by the principal.

Anonymous wrote:

Anonymous wrote:My kid is at Navy. Lots of high test scores at Navy. Kid did not get universal screener

Navy is very tough. My kid had very high test scores and was in pool and still didn’t get in until 2 years later. Also it’s a center school, which means no principal placement. Center schools are disadvantage that way.

Anonymous wrote:My kid is at Navy. Lots of high test scores at Navy. Kid did not get universal screener

Anonymous wrote:What is the total score?

Anonymous wrote:What is the cut off score for universal screener, are they taking avg. of nnat and ngat?

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:
Totally fair. Someone did post a url that had a PowerPoint that made it seem like the distribution was 100 with a std dev of 15 or 16, but it was not a bastion of clarity.

Immaterial now, but in case anyone is keeping track for future years, 139 on NGAT was not a universal screener score for my kid.

I'm the PP who posted the link. Assuming the sample report in the slide describes the same set of 3 tests that FCPS 2nd graders took, it seems clear to me that total standard score as well as the 3 individual standard scores are based on the same normal distribution with mean 100 and sd 15. Note that the percentile is given for total score in that sample report. Disclaimer is that I have no way of being certain that it is indeed the same scoring and sample that generated FCPS results.

So total score 130+ likely still means the student's total of 3 tests is better than 2sd above national average. A plausible guess is that the individual total score just sums the 3 raw scores (none of which are known to us) and then was mapped to the national sample distribution of such raw score totals.

Regardless, it is surprising that 139 total score wasn't enough to be in top 10% within that school. In statistical sense, a small (=high sampling error) AND very selective (=high sampling bias) sample can still make that happen.

My understanding is that that slide was for a version where the total score is also maxed at 160. Apparently there are two ways to combine, one maxed at 160 and the other at 175.

Regardless, 139 not making the cut seems statistically unlikely, but likely better explained by a model with a higher standard deviation?

160 max for individual and 175 max for total standard score is only mentioned in FCPS report and don't conflict with the information in that linked sample. Notice that it is just 4 standard deviations vs 5 standard deviations. If you take the total of 3 component scores, you have finer score grid over which you can map distributional norm. In other words, meaningful distinction between 4 vs 5 sd may not be feasible with a single component score. i.e. with more than 0.003% of the population getting perfect score. Still, it would be even rarer to get perfect scores on all 3 test components, thus it becomes a bit easier to separate 5 sigma from 4.

On this board and elsewhere, I haven't seen a single composite of 170+

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:Do you know how many numerical points each percentage point adds?

There is no point system. It’s holistic once they are in the review pool.

I meant on the exam score. For instance, a score of 128 is 93%. Would a score of 148 be, for instance, in 99%?

The distribution is bell shaped, so each percentage point corresponds with a different number of numerical points. Also, at the top, anything over a certain number is 99th percentile.

I think it is very unfair to only consider the NGAT. I know my child scored 98th in NNAT and 93 in NGAT. So we just discard the score of kids who did well in that? Very unfair. Don’t give a test you are going to toss.

98th percentile in the NGAT might not get you in-pool at most schools. It used to be the county wide cut off was at the 99th percentile.

This area has a lot of educated parents with graduate degrees. That tends to translate to kids who are advanced in school because their parents have been reading to them and exposing them to academic material from a young age. Genetically, it is likely that their parents are above average intelligence to smart, which means the kids are likely above average intelligent to smart. This means that the advanced programs in this area are serving the top 1-2% in scores and not the top 10% that much of the rest of the country serves.

There are schools where the top 10% of kids are scoring in the 99.5% range.

Mathematically, there aren't enough top 1-2% kids to fill up two AAP classrooms at most centers.

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:

Anonymous wrote:
Totally fair. Someone did post a url that had a PowerPoint that made it seem like the distribution was 100 with a std dev of 15 or 16, but it was not a bastion of clarity.

Immaterial now, but in case anyone is keeping track for future years, 139 on NGAT was not a universal screener score for my kid.

I'm the PP who posted the link. Assuming the sample report in the slide describes the same set of 3 tests that FCPS 2nd graders took, it seems clear to me that total standard score as well as the 3 individual standard scores are based on the same normal distribution with mean 100 and sd 15. Note that the percentile is given for total score in that sample report. Disclaimer is that I have no way of being certain that it is indeed the same scoring and sample that generated FCPS results.

So total score 130+ likely still means the student's total of 3 tests is better than 2sd above national average. A plausible guess is that the individual total score just sums the 3 raw scores (none of which are known to us) and then was mapped to the national sample distribution of such raw score totals.

Regardless, it is surprising that 139 total score wasn't enough to be in top 10% within that school. In statistical sense, a small (=high sampling error) AND very selective (=high sampling bias) sample can still make that happen.

My understanding is that that slide was for a version where the total score is also maxed at 160. Apparently there are two ways to combine, one maxed at 160 and the other at 175.

Regardless, 139 not making the cut seems statistically unlikely, but likely better explained by a model with a higher standard deviation?

160 max for individual and 175 max for total standard score is only mentioned in FCPS report and don't conflict with the information in that linked sample. Notice that it is just 4 standard deviations vs 5 standard deviations. If you take the total of 3 component scores, you have finer score grid over which you can map distributional norm. In other words, meaningful distinction between 4 vs 5 sd may not be feasible with a single component score. i.e. with more than 0.003% of the population getting perfect score. Still, it would be even rarer to get perfect scores on all 3 test components, thus it becomes a bit easier to separate 5 sigma from 4.