SAT- Blunt tool

[mifune]

<p>Although score discrepancies may very well be attributed to a basic measurement error, that cannot always be the general assumption. Discriminating among separate performances must exist across the entire scoring scale.</p>

[silverturtle]

<p>

The idea that the gap between 2350 and 2380 or 2380 and 2400 has any sort of importance in and of itself is absurd.

</p>

<p>

Obviously the correlation exists.

</p>

<p>These are inconsistent claims.</p>

<p>

Although score discrepancies may very well be attributed to a basic measurement error, that cannot always be the general assumption.

</p>

<p>Absolutely.</p>

[nemom]

<p>Correlation does not imply causation.</p>

[silverturtle]

<p>^ That reminds me of my recent post with that exact statement in the “Be as unhelpful as possible” game on the High School Life forum. Seriously, though, whom are you directing that at? I did not notice anyone committing that fallacy.</p>

[nemom]

<p>Interestingly, in the various stats cited by Mifune, we find that at Stanford, transfer students with an 800 in critical reading were admitted at a lower than than transfer students with a 700-799. (799 actually being rather silly as you cannot score this.)</p>

[ThisCouldBeHeavn]

<p>

These are inconsistent claims.

</p>

<p>No they aren’t. </p>

<p>Have any of you guys ever heard of “significant figures?” It’s an interesting concept; maybe you should look it up.</p>

[silverturtle]

<p>

Have any of you guys ever heard of “significant figures?” It’s an interesting concept; maybe you should look it up.

</p>

<p>That comes across as rather condescending.</p>

[nemom]

<p>Re Posting 182
No. These are not inconsistent. A correlation that is meaningless is quite possible. Common causal correlation is often meaningless.</p>

[silverturtle]

<p>

Interestingly, in the various stats cited by Mifune, we find that at Stanford, transfer students with an 800 in critical reading were admitted at a lower than than transfer students with a 700-799. (799 actually being rather silly as you cannot score this.)

</p>

<p>Interesting until one realizes the numbers of people that the data correspond to–only one or two people were accepted with a CR score of 800.</p>

[mifune]

<p>Non-correlation claims between the SAT’s and its effect on admission probability are not quite correct.</p>

<p>From a portion of a previous post:</p>

<p>

Of the other separate contending factors in college admissions, below is a realistic account of what exactly the SAT may or may not be “correlating” with:</p>

<p>[ul][li]GPA: Somewhat; there is a mild positive correlation between GPA and aptitude demonstrated on standard instruments. However, the evaluation of GPA is often contingent on one’s comparative performance relative to an applicant’s classmates – since there are obviously varying degrees of difficulty among schools and specialized grading scales. In the absence of such information, GPA becomes an even greater subjective criterion (although I, personally, am not ideologically supportive of the idea of precise ranking systems). </p>[/li]
<p>[li]SAT II: Similar to GPA, achievement test scores are somewhat, but far from exclusively, correlated with aptitude performances.</p>[/li]
<p>[li]Awards: Perhaps if the awards involve recognition for accomplishments on SAT-like evaluations. </p>[/li]
<p>[li]Extracurriculars: Although co-curricular commitment may be readily distinguishable among students (which is fairly irrelevant in regards to SAT performance), this is a highly skewed category and often appeals to institutional needs. </p>[/li]
<p>[li]Job/Work Experience: No.</p>[/li]
<p>[li]Volunteer Experience: No.</p>[/li]
<p>[li]Summer Activities: No.</p>[/li]
<p>[li]Essays: Once again, although perceptible levels of demonstrated skill exist between essays, this part of the application is not quite the best measurement for obtaining a grasp of one’s verbal aptitude, given the revisions and near-absolute certainty of outside assistance.</p>[/li]
<p>[li]Teacher Recommendations: Maybe; but the quality of letters of recommendation is often contingent on a subjective interpersonal assessment and that individual’s ability to effectively transmit positive regards.</p>[/li]
<p>[li]Interview: No.[/ul]</p>[/li]
<p>Essentially, it is the same scale and elite institutions are securing an applicant’s competitiveness through multiple measurements, as is necessary due to the slight statistical measuring error of any one. Given the more substantial difficulty of evaluating other criteria on non-standardized continuums, it is fully reasonable that there is an associative desensitization with respect to the aforesaid criteria. </p>

<p>Thus, based on an intuitive appraisal of the competing dynamics, it is practical, logical, and discernibly accurate to believe that the highly conspicuous correlations between SAT scores and acceptance rates are due, in large part, to the profound causality of the SAT itself. Given the SAT’s palpable influence on bettering the selectivity ranking of universities, the greater preference in favor of higher-scoring students is certainly explicable.

</p>

[silverturtle]

<p>

A correlation that is meaningless is quite possible.

</p>

<p>Actually, it’s impossible in our context.</p>

[ThisCouldBeHeavn]

<p>

That comes across as rather condescending.

</p>

<p>Frankly, I felt that your consistent refusal to acknowledge the point deserved a hint of condescension.</p>

<p>The idea that a weak correlation justifies minute differentiation is, again, absurd.</p>

[mifune]

<p>

Interestingly, in the various stats cited by Mifune, we find that at Stanford, transfer students with an 800 in critical reading were admitted at a lower than than transfer students with a 700-799. (799 actually being rather silly as you cannot score this.)

</p>

<p>As silverturtle correctly points out, transfer statistics at the upper score levels do not quite reach a level of statistical significance.</p>

[silverturtle]

<p>

The idea that a weak correlation justifies minute differentiation is, again, absurd.

</p>

<p>The unavoidable implication of your assertion is absurd: at what point should a college start discriminating among different scores if not at the difference threshold facilitated by the metric?</p>

[nemom]

<p>Re: posting 177 (Mensa/SAT/IQ)
You have it right, bovertine. I considered fooling with Mensa back in my mispent youth. My old SATs were good enough but I couldn’t dig up the official paperwork so I think I took their test. (or maybe I took the test then eventually found the paperwork). I never ended up going to any meetings because the stuff I got was just unappealing. </p>

<p>I think we can all agree that the SATs are no longer linked strongly to IQ (and may never have been anyway). The way in which SAT scores and IQ scores are calculated is similar - using test data and statistical analysis to produce a numeric scale. But I think trying to talk about the math/stats of the SATs by using IQ analogs just muddies the waters.</p>

[calmom]

<p>

I don’t think there is any reasonable non-literal interpretation of the phrase.

</p>

<p>Silverturtle, as nemom and momofthreeboys correctly pointed out, my post re IQ was using an example or an analogy, and was never intended to be “literal.” One of the difficulties I have had with the train of discussion online is that you seem to have an over-literal response to my posts – I do not know why this is. Maybe its a limit of online posting – maybe its a reflection of the fact that you are a teenager trying to keep up the flow of conversation on the parents thread, and people with middle aged brains tend to engage in discussion on a different level. (“We are better at getting the gist of arguments,” she says. “We are better at recognizing categories. And we’re much better at sizing up situations” See [Fresh</a> Air Interviews: Writer Barbara Strauch : NPR](<a href=“http://www.npr.org/templates/story/story.php?storyId=125902095]Fresh”>Fresh Air Interviews: Writer Barbara Strauch : NPR) - quoting Barbara Strauch, author of The Secret Life of the Grown-up Brain: The Surprising Talents of the Middle-Aged Mind ) </p>

<p>However, one irony of this situation is that I feel that the SAT test format tends to reward students who are very literal and prosaic in their thinking process. It’s all about getting the “right” answer in a time-limited fashion – kind of like a game of Jeopardy – so it tends to trip up highly imaginative or nuanced thinkers, who are also likely to be bored by any sort of rote learning or study, and thus averse to extensive test-prep. So there is a rather significant subset of very bright students who flub the test because they were overthinking the questions or ran out of time… and then don’t bother with a retake because they don’t have patience for the process. </p>

<p>And that’s another aspect of the “blunt tool” – not only does the test over-predict the potential of student whose high-end scores are obtained through manipulation (extensive prep and repeat testing), but it under-predicts the potential of students who are averse to the standardized, multiple-choice test format. (How many fit this group is anyone’s guess – but that’s why colleges stick with a holistic admission process. Essays and recs can tell an entirely different story than the test scores, and its a story the ad coms want to know).</p>

[nemom]

<p>Perhaps worth considering:
[Study</a> on Accuracy of SAT Prompts Schools To Accept Other Tests - The Tech](<a href=“http://tech.mit.edu/V128/N44/sat.html]Study”>http://tech.mit.edu/V128/N44/sat.html)</p>

[nemom]

<p>RePost 198
calmom has some good points. I know my child had to work on the writing section because he often found all the choices for sentence improvement to be bad.</p>

[ThisCouldBeHeavn]

<p>

The unavoidable implication of your assertion is absurd: at what point should a college start discriminating among different scores if not at the difference threshold facilitated by the metric?

</p>

<p>The point at which the difference in scores is larger than the natural variance between different sittings. At the top end of the scale, a question or two per section is a normal degree of difference. That can result in a difference on the order of 100 points overall. I don’t think that’s an absurd conclusion at all. And again, with the flawed structure of the test, it is far easier to reach this tail than it should be.</p>

[silverturtle]

<p>

I think we can all agree that the SATs are no longer linked strongly to IQ (and may never have been anyway).

</p>

<p>I am sure that there is a strong link, though probably not as strong as it had once been (prior to 2005). The study I alluded to earlier demonstrated quite clearly that there was a strong link prior to 2005; because the test is not fundamentally radically different now, I assume the link remains similar.</p>