SAT: Part hoax and part fraud?

<p>^^ Agreed. It is no more than that. The kid was not allowed to get away with non sequiturs at home and thus recognized those on the SAT. This is just one example, but the point is made.</p>

<p>

</p>

<p>Many of the high schools at the top of the list are in high income areas, or are academic magnet schools. The ones in the Silicon Valley area are probably getting extra boosts due to the fact that many students’ parents have been selected by immigration system (graduate students on F-1 visas, later to become skilled workers on H-1B visas, etc.).</p>

<p>^^the City of Irvine has all four public high schools on that list. The median household income is $87k in Irvine.</p>

<p>$87,000 (although probably considered “low income” on these forums) is considerably higher than the California median household income of about $61,000. Also, Irvine probably has some boost due to the university.</p>

<p>Yes, Irvine is higher than the state average. But no where close to the mythical $200k on this thread. And to the extent that they live in the City, many/most of the faculty brats would attend the adjacent HS, University. Don’t forget, in addition to highly educated faculty, a UC campus also has a thousands of non-academic staff, such as cafeteria workers, custodians, gardeners, bus drivers, etc. </p>

<p>The City has~230k residents, so it hardly qualifies as a “college” town.</p>

<p>“Many of the high schools at the top of the list are in high income areas” I’m sure they are. If the scores keep going up linearly with income 21 points per 40K income, you would need an average family income of $880-920K for town residents for an average student score of 2003. I know salaries in California tend to be high but that seems a little much. (And is it the salary that matters, or the disposable income?)</p>

<p>

</a>
While it is true that there is a correlation, like nearly every other discussed measure, when adding sufficient controls, the contribution from SAT scores becomes quite small in comparison to other factors. Table 17 of the study you linked indicates that with sufficient controls, SAT score had a smaller contribution to dropping out of demanding majors than application LORs, application essays, and just about every other measured criteria except for NOT being rated as having good personal qualities (good personal qualities led to a negligible increased chance of dropping out of demanding major).</p>

<p>I don’t think any one is disputing that test scores have a non-zero benefit in predicting chance of dropping out of major, grad, rate, GPA, or similar criteria. The question is the size of this benefit is and how if it should influence policies. All referenced studies in this thread with controls for HS GPA and a measure HS quality or course rigor suggest the benefit beyond the other available information about the applicants is quite small. I’d expect it to be even smaller, if unique cases for which the grades are not as meaningful are separated, such as home school students and students attending HSs with severe grade inflation.</p>

<p>

</p>

<p>It seems to me that the SAT does a pretty good job of controlling for HS quality, course rigor, and some other things. And its pretty easy to run. It has the effect of sidestepping all of the machinations and politics that would go into attempting to “equalize” high school GPAs. Essentially, that is one of the reasons why some people don’t like it. </p>

<p>

</p>

<p>My point was not that SAT scores explain change of major. In fact, I noted that Arcidiacono et al. argue that differences in academic background can explain everything. Instead, my point was that one can run a regression in which SAT appears to be negatively related to something we care about (i.e., improvement in GPA or class rank), yet that negative relationship disappears when one looks a little deeper. </p>

<p>

No, but all the ones I am aware of do post percentage of kids who graduated in the top 10% of the class. Which is probably a reasonable proxy for top grades, no?</p>

<p>The top ten % of the class, unweighted, still reflects the basic inequality of grades of high-level or lower-level courses. Across the board, if you take AP English or regular-level English, the A is the same in terms of class rank, and that must make class rank as a differential less meaningful than it appears to be. Which is why the SAT and other tests remain relevant: GPA and class rank are not immediately comparative, but must be considered in context, while SAT has at least some claim to objectivity. I would not want the SAT to be considered on its own, but I recognize its use in conjunction with other indicators.</p>

<p>

</p>

<p>As has been reported on these forums, some high schools use weighted GPAs to determine class rank, and some students complain about facing “dilemmas” about whether to take an unweighted elective that they are interested in (commonly art or music) which can cause them to lose weighted GPA points for class rank purposes.</p>

<p>Our school doesn’t release unweighted class ranks. At the high end, class rank is largely determined by how many arts classes you take. Actually, it’s worse than that. The best way to rank high is to take study halls.</p>

<p>Our school ranks with weighted grades. Theoretically only academic courses are used, but for some reason orchestra was considered an academic course. It actually helped my son since even though orchestra carried no extra weighting, his grades were higher in orchestra than any of his other courses, and he took two orchestra classes every single year of high school. I’ve come to the conclusion there is no fair way to weight, but even in the worst systems the top 10% will generally be good students.</p>

<p>The concept that grades are that ties to mental ability over a certain basic level is the real issue. The SAT measures IQ. Grades often measure other things. </p>

<p>

</p>

<p>Perhaps that is what some of the users originally wanted (as perhaps implied by an earlier meaning of the A in SAT), but there is no real way to separate innate IQ from the effects of educational experience (e.g. how well was the person taught reading, math, and writing in school?).</p>

<p>The SAT is derived from IQ tests, but the many changes over time have diminished its connection to that history. </p>

<p>I think the tests still have an important role to play in distinguishing among high schools, and in comparing students within demographic bands. While it isn’t always talked about, that’s the way that holistic schools often do use them: how well did this student do compared to others with similar backgrounds? This is how homeless/rural/public housing kids with 1200 CR+M can end up in Ivies.</p>

<p>ucbalumnus wrote: “there is no real way to separate innate IQ from the effects of educational experience”</p>

<p>Here comes the dreaded anecdotal experience comment - a family with parents taking the SAT in the mid 1970s with no prep at all, coming from lackluster schools, and their 3 kids taking the test in the last six years, with no prep (other than vigorously ignoring the practice book bought by the parents), coming from strong schools: All of them National Merit. Not a one of them a hard working, “good” student, each preferring to read and learn for pleasure, not much motivated by grades. In the humble opinion of this family, scoring well on the SAT is kind of a useless party trick, like being double jointed, which doesn’t at all indicate likely success or failure in school. The SAT is fun and leads to unearned fleeting glory. :slight_smile: </p>

<p>

</p>

<p>That is probably an extremely powerful educational experience.</p>

<p>According to Wikipedia:
"Some measures of educational aptitude are essentially IQ tests – for instance, Frey and Detterman (2004) reported a correlation of 0.82 between g (general intelligence factor) and SAT scores;[79] "</p>

<p><a href=“Intelligence quotient - Wikipedia”>http://en.wikipedia.org/wiki/Intelligence_quotient&lt;/a&gt;&lt;/p&gt;

<p>In the US News data for top 50 national universities, the correlation between SAT 75th percentile and percent of class in top 10% is +.67. Perhaps top 10% can be considered a proxy for average HS GPA?</p>