Sidwell & Maret

[Anonymous]

@ SAM2 (8:06): Me again, yes I do have access to academic journals (I teach at a university). Your assessment of Franzen may be correct, but I can't confirm as the article he references is not available online and I would have to go to the library. However, from reading other articles that reference the study, it looks like they attempted to correlate beyond grade 3 but couldn't (weak correlation).

I did find some other studies:

1) Predicting children’s academic achievement from early assessment scores: a validity generalization study by Juhu Kim and Hoi K. Suen (2003)

They looked at many forms of assessment, including the WPPSI and found the correlation for long term academic success to be approximately 0.35. Therefore, that supports the PP's point that WPPSI scores at age 3-5 does not predict longer academic success.

2) However, I found another study (although dated --late 1980s) that supports the validity of the test in determining long-term IQ stability and that academic performance is moderately correlated (although not strong >0.8). Which is what I was attempting to point out that the test does predict IQ over time but does not strongly predict how a child will perform in school. A score of 125 should be thought of as a range, not fixed, i.e., within a range average to high/ high average to superior/ superior to genius--a score of 125 is really 120ish-130ish (I forgot the actual intervals) which can put someone within 2 ranges during their lifetime. That can also mean a child who scored in the 99 percentile can take the test again in 5 years and score in the 95th.

This is some text I was able to copy from the PDF I obtained through my online subscription:

LONG-TERM PREDICTIVE VALIDITY OF THE WPPSI: AN 11-YEAR FOLLOW-UP STUDY by
WILLIAM YULE, R. DAVID GOLD and CAROL BUSCH



". . . we can say that the correlation of 0.86 between the Full Scale WPPSI IQ and the Full Scale WISC-R obtained 11 years later indicates that these children’s relative intellectual status is remarkably stable throughout their school careers. Presenting our results in a parallel fashion to theirs we can say that there is a 9:1 chance that WISC-R IQ at 16 years will be within 13 points of WPPSI IQ at 54 years.

The present results confirm the widely held view that in the range of discrepancy scores which encompasses 80% of the population, Verbal-Performance discrepancy scores are not at all stable over an 11-year period. Discrepancies of less than 15 points in
either direction should be regarded as inherently unstable and therefore should never be interpreted. There is slight evidence that more extreme discrepancies are more stable, but there were too few cases to be other than tentative about such findings. Perhaps the most striking of the findings in the present paper are those relating WPPSI IQ scores to scholastic attainments at the end of compulsory schooling. The correlation between Full Scale IQ and Reading (0.61) was as high as found in contemporaneously administered tests, whilst that between IQ and Mathematics was a staggering 0.72."

In other words, IQ only accounts for 37 to 51% of academic achievement.

Hope this helps, I hope the PP posts some studies--I am also interested in this subject.

[Anonymous]

Anonymous wrote:@ SAM2 (8:06): Me again, yes I do have access to academic journals (I teach at a university). Your assessment of Franzen may be correct, but I can't confirm as the article he references is not available online and I would have to go to the library. However, from reading other articles that reference the study, it looks like they attempted to correlate beyond grade 3 but couldn't (weak correlation).

I did find some other studies:

1) Predicting children’s academic achievement from early assessment scores: a validity generalization study by Juhu Kim and Hoi K. Suen (2003)

They looked at many forms of assessment, including the WPPSI and found the correlation for long term academic success to be approximately 0.35. Therefore, that supports the PP's point that WPPSI scores at age 3-5 does not predict longer academic success.

2) However, I found another study (although dated --late 1980s) that supports the validity of the test in determining long-term IQ stability and that academic performance is moderately correlated (although not strong >0.8). Which is what I was attempting to point out that the test does predict IQ over time but does not strongly predict how a child will perform in school. A score of 125 should be thought of as a range, not fixed, i.e., within a range average to high/ high average to superior/ superior to genius--a score of 125 is really 120ish-130ish (I forgot the actual intervals) which can put someone within 2 ranges during their lifetime. That can also mean a child who scored in the 99 percentile can take the test again in 5 years and score in the 95th.

This is some text I was able to copy from the PDF I obtained through my online subscription:

LONG-TERM PREDICTIVE VALIDITY OF THE WPPSI: AN 11-YEAR FOLLOW-UP STUDY by
WILLIAM YULE, R. DAVID GOLD and CAROL BUSCH



". . . we can say that the correlation of 0.86 between the Full Scale WPPSI IQ and the Full Scale WISC-R obtained 11 years later indicates that these children’s relative intellectual status is remarkably stable throughout their school careers. Presenting our results in a parallel fashion to theirs we can say that there is a 9:1 chance that WISC-R IQ at 16 years will be within 13 points of WPPSI IQ at 54 years.

The present results confirm the widely held view that in the range of discrepancy scores which encompasses 80% of the population, Verbal-Performance discrepancy scores are not at all stable over an 11-year period. Discrepancies of less than 15 points in
either direction should be regarded as inherently unstable and therefore should never be interpreted. There is slight evidence that more extreme discrepancies are more stable, but there were too few cases to be other than tentative about such findings. Perhaps the most striking of the findings in the present paper are those relating WPPSI IQ scores to scholastic attainments at the end of compulsory schooling. The correlation between Full Scale IQ and Reading (0.61) was as high as found in contemporaneously administered tests, whilst that between IQ and Mathematics was a staggering 0.72."

In other words, IQ only accounts for 37 to 51% of academic achievement.

Hope this helps, I hope the PP posts some studies--I am also interested in this subject.

Sorry about the typos. Also, not sure why a happy face got inserted. Should read greater than 0.80 in part 2.

[Anonymous]

i'm not reading any of this because it's kind of boring. But i will make this point. We have three kids who are well on their way in their academic careers. One is about to graduate, another is in high school and the third in middle school. The WPPSI raised significant concerns about two of the kids which have proven to be accurate. The third kid scored extremely high on the test and does the best, by far in school, of the three.

[Anonymous]

Contrary to the mood of discouragement on this thread, I'd say apply and let the professions( AKA admissions team) decide. Please don't let others on the forum set the bar for your DC, just go for it. What does the app cost, after all : $75 ? Best of luck to you and chin up!!! I'm sure your DC is wonderful and will be just fine 20 years from now.

[Anonymous]

Regarding SAM2's detailed report, I'd like to ask how does one determine "later academic success" ? If that is based solely on GPA, I'd say that grades do not accurately reflect the future intellectual creativity or real life contribution to society of an individual. Striving to meet the objective "A" standard set by others doesn't often produce the next Nobel prize winner. As is more often the case, the very bright are bored in school and underperform and are contemptuous of being "graded", but will often do well in college , start a business on their own, etc..My point: all this analysis is a bit unhealthy and obsessive.

[Anonymous]

[quote=Anonymous]Regarding SAM2's detailed report, I'd like to ask how does one determine "later academic success" ? If that is based solely on GPA, I'd say that grades do not accurately reflect the future intellectual creativity or real life contribution to society of an individual. Striving to meet the objective "A" standard set by others doesn't often produce the next Nobel prize winner. As is more often the case, the very bright are bored in school and underperform and are contemptuous of being "graded", but will often do well in college , start a business on their own, etc..My point: all this analysis is a bit unhealthy and obsessive.[/quote]

I provided the detailed report, not SAM2. I was answering his/her valid request of a previous poster. And yes, academic success is measured by grades/graduating from school. We are making the same point as I was addressing PPs putting so much weight into WPPSI scores. The scores predict IQ over time, and in some cases if a child will do well in school (measured by grades). However, I was trying to make the point that kids who score in the 80 percentiles [b]should apply[/b] just like kids in the 90th+ percentiles as the score doesn't guarantee/predict better academic performance in the form of grades and also career success (your point). My intent was not to be obsessive or bore people, but to provide actual studies/research to make a point as people on this forum generally rely on opinion or something they heard without any real sources.

[Anonymous]

I found both the posts interesting, thank you (particularly given the non-substantive and silly nature of many posts on the Independent Schools board these days). Keep up the good work!

[SAM2]

@11:56 (university teacher): Many thanks for the articles. I will try to look those up. I am confused by some of the statements in the quotation you provided, and by your calculation that "IQ only accounts for 37 to 51% of academic achievement." How did you reach those figures? I will try to find that article so I can understand better. Any explanation much appreciated.

@13:39: All the credit for the detailed report belongs to 11:56, not me. I think that most research reflected in the articles we are citing defines "later academic success" as performance on other standardized tests. I've seen a few articles that try to correlate early test scores with less-objective measures like GPA or class rank. But I think most people studying this stuff focus more on the objective measures because they lend themselves to better measurement than softer standards like future intellectual creativity and contribution to society. I've seen some researchers that focus on "academic self concept," which seems to measure how a student thinks about herself and her abilities, rather than a pure objective measure. That's very interesting, but I haven't spent too much time reading those articles yet.

In the end, I agree with you that how a child measures on a couple standardized tests 15 years apart matters little in the grand scheme of life -- but what I'm focusing on is not evaluating the child's "life success" (which is a product of thousands of uncontrollable factors), but rather just whether a test given at a young age can predict (even with very middling accuracy) how a child will perform on another academic test 15 years later. Others may think that's boring, but I find the ability to make such predictions fascinating. I'm not trying to build any argument about my child or about some school; I'm just interested in the topic. (OK, admittedly, my child has taken the test, so part of my curiosity is based on learning what I can about my child, but I'm definitely not participating in any "smartest kid" contest.)

I also think 11:56 makes a great point about thinking of any score as a range, rather than a point. A particular score is just one snapshot of how that child did on one test. That one snapshot likely gives some sense about the child's abilities, but it's just one data point. If you add other data points, they might reinforce the first, or else reveal it as an outlier. Also, over time, thousands of confounding factors can intrude on a child's life, and dramatically change how that child develops. It would be wrong to think of any child as one single data point, when she's really a complex mix of experiences and attributes.

Thanks again for the feedback and info. I appreciate the good discussion and the chance to learn.

[Anonymous]

SAM2 wrote:@11:56 (university teacher): Many thanks for the articles. I will try to look those up. I am confused by some of the statements in the quotation you provided, and by your calculation that "IQ only accounts for 37 to 51% of academic achievement." How did you reach those figures? I will try to find that article so I can understand better. Any explanation much appreciated.

The correlation coefficient is referred to as "r". R^2 (r squared) is used to interpret the correlation coefficient. If you have a correlation of 0.6 then 0.6 squared equals 0.36. This translates to 36%. Therefore a correlation of .60 means that only 36% is equal to the percent of the variation of one variable (in this case WPPSI scores) that is related to the variation in the other (academic performance). So in my original calculation of 37 to 51% (0.61 and 0.72 from the Yule, Gold and Busch study) of the variation (correlation) is related.

Hope that makes sense.

[Anonymous]

PP and SAM2, I wouldn't argue that IQ tests don't have predictive value. I think they do. It's just that IQ is less stable over time than people tend to think. Also, testing in early childhood is less predictive than testing done after age 5.

I should have said that future giftedness cannot be predicted with a reasonable degree of certainty based on WPPSI scores at 4. There is, of course, some correlation between IQ at 4 and during adolescence, it's just remarkably low, too low to have significant predictive value.

I recommend The Development of Giftedness and Talent Across the Life Span edtd by Frances Degen Horowitz, et al for anybody who is really interested in this topic.

As for the rest of you parents with young kids, I encourage you to go with your gut; it's more reliable than a test.

[Anonymous]

SAM2 wrote:

Anonymous wrote:WPPSI scores through age 4 have NO value in terms of predicting future academic success. Anybody who tells you otherwise is not familiar with longitudinal studies ....

Please provide links to these longitudinal studies that show WPPSI scores have zero value in predicting future academic success. Thanks.

Read the book Nutureshock yet??? Just a place to start.

[Anonymous]

Anonymous wrote:According to www.petersons.com, 59 percent of applicants were accepted to Maret, whereas Sidwell's acceptance rate is only 15 percent. I think that proves that Sidwell is harder to get into, even considering its larger size. I suspect it has a larger applicant pool, too.

For comparison's sake, the rates listed for NCS, STA, and GDS are 30%, 49% and 35%, respectively.

Petersons uses a survey that is sent out to schools - they choose whether to answer it or not - and what to answer. Many schools don't want to give out the accurate numbers, as they don't want to discourage applications. The application due dates are not accurate for a number of these schools, so what makes you think the acceptance numbers are? If you don't work in admissions at one of the said schools, please don't act like you have any ideas of the facts and figures - its quite misleading.

[SAM2]

Anonymous wrote:

SAM2 wrote:

Anonymous wrote:WPPSI scores through age 4 have NO value in terms of predicting future academic success. Anybody who tells you otherwise is not familiar with longitudinal studies ....

Please provide links to these longitudinal studies that show WPPSI scores have zero value in predicting future academic success. Thanks.

Read the book Nutureshock yet??? Just a place to start.

I have read NurtureShock, and (embarrassingly) many of the articles cited in the IQ testing chapter. I think I even emailed Po Bronson a couple questions, but I can't recall if he wrote back. The bottom line is that those articles don't say "zero value." I think Bronson may have exaggerated that point to make the chapter more interesting. "All those IQ tests are worthless!" will sell a lot more books than "Those tests aren't quite as consistent over multi-year trials as most people think they are!"

[SAM2]

Anonymous wrote:

Anonymous wrote:According to www.petersons.com, 59 percent of applicants were accepted to Maret, whereas Sidwell's acceptance rate is only 15 percent. I think that proves that Sidwell is harder to get into, even considering its larger size. I suspect it has a larger applicant pool, too. For comparison's sake, the rates listed for NCS, STA, and GDS are 30%, 49% and 35%, respectively.

Petersons uses a survey that is sent out to schools - they choose whether to answer it or not - and what to answer. Many schools don't want to give out the accurate numbers, as they don't want to discourage applications. The application due dates are not accurate for a number of these schools, so what makes you think the acceptance numbers are?

I've spent a fair amount of time looking at Peterson's, and I think much of the data there is screwy. I tend not to trust it unless I can find a second confirming source. The numbers quoted above might be accurate, or maybe not. But I would not rely on them without more.

[Anonymous]

I'm a parent at Sidwell but there is absolutely no way that a majority of applicants who apply to Maret get in. That's just absurd on its face. Also, I can't believe that either Sidwell or Maret for that matter supplied that data.