PDA

View Full Version : SAT results


miyavvii
06-05-2010, 12:27 PM
Hi there,
Recently I have undertaken 2 SAT II subjects - the Math Level II and Physics.
I have done well for the math scoring 800 while I have done really bad in Physics , scoring 640 ( did not managed to complete ).

I was wondering if when I retake, would the better of the scores for the Physics subject replace the old one ? Other there would be a history of the scores for the subject taken.

Thanks

Butterbeer
06-05-2010, 01:51 PM
most schools take into account your highest score. even though some schools may get to view a record of all your scores, normally they won't penalise you for improving on your score in subsequent tries.

but there is a caveat to this: too many tries (e.g. more than 3 tries) but only marginal improvement in scores (e.g. 640 to 660 to 680) won't reflect well on you, as it seems that you are unable to improve your proficiency of the subject significantly despite repeated attempts and preparation.

miyavvii
06-05-2010, 03:42 PM
oh icic.
Thanks for the verification. Btw, is there a validity period for the results ?
I've ask the invigilator previously but was not able to get a clear answer.
Thanks

Butterbeer
06-05-2010, 04:59 PM
5 years i think. but please check the college board website.

windowsmaclinux
06-08-2010, 05:25 PM
I personally doubt the merit of advising students not to take the SAT more than 3 times. I doubt whether schools really have time to dig up your testtaking record and see how many times you have tried. If a school explicit says it only looks at your highest score/combination (please check this), they couldn't care less about your other scores. Even if a school needs all your scores and doesn't say the above, it does not really mean they will look at all your past scores. Nowadays in most admission offices, test scores are entered into a computer or transferred electronically from testing agencies (eg collegeboard), updated and then printed out on a small slip of paper attached to your file. This is done by a different person (euphemistically called "data entry specialist") from the "readers" (who will decide your fate). So don't fret over this too much.
Anyway I think taking the same test more than 3 times won't probably increase your scores by a lot and is probably just a waste of time which you could have spent on other things.
If a school allows ScoreChoice, use it to give you some peace of mind. And yes, feel free to take physics again.

LockT31W
06-08-2010, 06:17 PM
Colleges also have an incentive to consider only your best scores regardless of how many attempts you've taken, because it would boost their admitted students' score profile. This is also why many of them combine your best score for each section of the SAT I across multiple attempts, and not your best total score from a single attempt.

Butterbeer
06-08-2010, 06:57 PM
Anyway I think taking the same test more than 3 times won't probably increase your scores by a lot and is probably just a waste of time which you could have spent on other things.

probably the most important point in this entire thread. which leads to the merits of advising one not to take the SAT more than three times.

Colleges also have an incentive to consider only your best scores regardless of how many attempts you've taken, because it would boost their admitted students' score profile.

a somewhat irrelevant argument. High scores boost the admitted students' score profile, but only because the admitted students probably have (generally) the highest scores amongst all applicants in the first place. in this case, the technique of considering best scores for applicants with both low or high scores gives neither an advantage.

LockT31W
06-08-2010, 07:54 PM
a somewhat irrelevant argument. High scores boost the admitted students' score profile, but only because the admitted students probably have (generally) the highest scores amongst all applicants in the first place. in this case, the technique of considering best scores for applicants with both low or high scores gives neither an advantage.

If colleges look unfavourably on retaking the SAT many times, they would only hurt their admitted students' score profile, since they would then exclude some people with high scores from many attempts. By disregarding the number of attempts, they can admit those high-scoring students, which boosts their admitted students' score profile. Therefore,

1) You won't be disadvantaged no matter how many times you attempt the SAT, because colleges have an incentive to treat you the same way they would treat someone with the same score from fewer attempts.

2) Because colleges don't care how many times you take the SAT, if you have the money, you can keep on retaking the SAT until you get a superscore you like. The superscore can only increase, whether by luck or through actual improvement.

Butterbeer
06-08-2010, 08:54 PM
If colleges look unfavourably on retaking the SAT many times, they would only hurt their admitted students' score profile, since they would then exclude some people with high scores from many attempts.

If college looks unfavourably on retaking SAT many times: People who have high scores from many attempts would be excluded because there are people who have the same scores (or better) from fewer attempts.

If college has no preference for either type: People who have high scores from many attempts would be included, but so will those who have the same scores from fewer attempts.

Thus in either case, ceteris paribus, the incentive does not exist since the admitted student profile will not be worse with either outlook.

Following the line of argument, it will then not be safe to assume that colleges will assume the "no preference" outlook if they are rational, since there are no actual incentives to do so.

Hence, you MAY be disadvantaged depending on how many times you take the test and how you fare for each of your attempts.. The discounting effect is not linear and is of course dependent on the quality of the other parts of your application.

Of course, this is under the assumption that colleges rank and choose students according to SAT score only, which is not true in reality.

haha sorry for the seeming excessiveness of this post, its just a rather intellectually interesting issue to stimulate my mind in the midst of NS. the real practical advice is that: don't spend too much unneeded time on the SAT, especially when you know that your test scores aren't going to improve despite repeated attempts, as seen probably from your practice test attempts. . Save the money and the energy for building up a stronger application in other areas. SAT scores ultimately add little marginal benefit after a certain threshold (usually 2200).

LockT31W
06-08-2010, 10:01 PM
If college looks unfavourably on retaking SAT many times: People who have high scores from many attempts would be excluded because there are people who have the same scores (or better) from fewer attempts.

If college has no preference for either type: People who have high scores from many attempts would be included, but so will those who have the same scores from fewer attempts.

This is exactly what I mean. In the first case, you are excluding some people with high scores from the admitted pool. The college's admitted pool declines in SAT score profile. It's as simple as that. Say you want to admit 2000 students, and it happens that the top 2000 of your applicants have scores >2200, and half of them took the SAT more than 3 times. If you don't exclude them, your score profile is exactly >2200. If you exclude them, you need to make up the difference of 1000 using applicants with lower scores, which hurts your score profile. It's simple arithmetic here.

Butterbeer
06-08-2010, 10:42 PM
If you exclude them, you need to make up the difference of 1000 using applicants with lower scores, which hurts your score profile.

As mentioned earlier, my point was (and always has been) that when i say a particular college is biased against those who take SAT too many times, it means that in the case of two applicants having the same set of final scores, the applicant who takes too many attempts may be at a disadvantage.

Hence, in the case of your example, these 1000 applicants with repeated scores will only be excluded if and only if there are applicants with equivalent scores but who achieved them in fewer attempts. Hence the use of the ceteris paribus assumption.

I don't think any college (or even myself for the matter, if i were an admissions committee member) would put such a heavy penalty on people who did SAT many times, as to totally discount their scores from consideration altogether. The concept of exclusion is not based on principle, but relative advantage here.

LockT31W
06-08-2010, 10:54 PM
But there will always be applicants with equivalent scores from fewer attempts. I recognize the assumptions behind the model, of course, and the exclusion scenario is hypothetical. But even in the relative advantage model, the college continues to disadvantage itself by imposing any penalty at all on people who do repeated attempts. The score profile will still be hurt. A primary function of the admissions office is also to increase the prestige of the college on paper, and SAT score profiles of enrolled students are heavily used by potential applicants as a measure of the college's academic strength.

Anyway, most colleges have an explicit policy of considering only the highest scores attained by the applicant. I have not yet seen any reason or evidence to doubt their claims.

Butterbeer
06-08-2010, 11:20 PM
Let's try an illustrative model. Supposing a college selected the students by first ranking them according to SAT score and attempts. Let's say it turns out something like this (starting from supposed

2400 1 attempt
2400 2 attempts
2400 3 attempts
2400 more than 3 attempts
2200 1 attempt
2200 2 attempts
2200 3 attempts
2200 more than 3 attempts
2000 1 attempt
2000 2 attempts
2000 3 attempts
2000 more than 3 attempts

The college then decides on a certain % to accept.

If the college was unbiased, it would first arrive at a cut-off score based on the %, regardless of no. of attempts, since if its unbiased, the score itself is the only way to set a parameter to the shortlist. Say this cut off score is 2200 this time. So everyone from "2200 more than 3 attempts" onwards is considered qualified. So even though the actual % accepted is more than the initial one decided, the college cannot do otherwise since it purports not to take into account number of attempts.

If the same college was biased, since given the extra parameter of the number of attempts, it is now able to differentiate between candidates with the same raw score. BUT the number of attempts only matters if the two sets of scores (one with many attempts, one without) are very similar or are the same to each other. since it makes no sense to condemn a "2400 with more than 3 attempts" in favour of a "2000 with 1 attempt". Thus, this college will go down the list,accepting students until the quota % is reached. In the case, perhaps it is "2200 with 3 attempts" or maybe even "2200 with more than 3 attempts" (assuming the college's candidate score distribution is such that the % coincides with the threshold of the score exactly). It is not possible to go any lower than that since the first approach only serves to increase the % while the second approach preserves the same % so the second approach will always be more selective (if not equally selective) as the first.

As such, using the second approach which is biased, we can see that the cut-off scores for both cases are actually the same. The distribution of scores for the second approach might even be more favourable than in the first approach. Hence, there would be no incentive to want to be unbiased so as to improve the score profile. There may be other incentives such as convenience but the argument of improving score profile by being unbiased will not apply in this case.

This is strictly an illustrative model btw :)

with regards to practical advice, i agree that colleges do probably use only the highest scores for any computation purposes when deciding an initial cut of applicants in the admissions process. But at more highly competitive colleges, impression starts to play a slightly larger role, and having too many sets of SAT scores with little improvement from one to the next is not exactly the preferred one.

LockT31W
06-09-2010, 09:36 AM
It becomes clear that we were talking about two different things. My hypothesis was that anyone with more than 3 attempts are excluded. But this is a minor issue.

This is the major flaw in your illustration:


2400 1 attempt
2400 2 attempts
2400 3 attempts
2400 more than 3 attempts
2200 1 attempt
2200 2 attempts
2200 3 attempts
2200 more than 3 attempts
2000 1 attempt
2000 2 attempts
2000 3 attempts
2000 more than 3 attempts

...

BUT the number of attempts only matters if the two sets of scores (one with many attempts, one without) are very similar or are the same to each other. since it makes no sense to condemn a "2400 with more than 3 attempts" in favour of a "2000 with 1 attempt".

This model breaks down once you use a more graduated distribution.

2400
2390
2380
2370
2360
2350
.
.
.
2210
2200 with 1 attempt
2200 with 2 attempts, 3 attempts, 4, 5... etc
2190 with 1 attempt
.
.


for example, which is also what happens in real life. In your model, the arbitrary cut-off happened to be at 2200, so we shall use that again. Once again, in the unbiased model, it happens that all students from 2200 and upwards perfectly fit the college's quota.

Then we go to the biased model. According to your model of decision-making, where admission officers make quality judgments based on number of attempts AND the score, a 2400 in 3 attempts is considered better than a 2000 in 1 attempt, despite the number of attempts. That is reasonable. However, what happens when you have a 2400 in 3 attempts and a 2390 in 1 attempt? According to you, there is a small penalty for taking the test too many times. Therefore, there's a chance that the slightly lower-score, lower-attempt applicant "replaces" the higher-score, higher-attempt applicant. This "replacement" takes place throughout the >2200 range without changing the score profile, since "replacement" can only occur when scores are close - at scores higher than the cut-off, a "replaced" 2400 individual still falls within the acceptable range of >2200. To understand it better, we could assume that the college deducts 10 points from your score for every attempt beyond the 3rd, etc.

However, at the margin of the >2200 score range, something happens. Going by the "replacement" decision making model, a 2200 with say 4 attempts will be replaced by a 2190 with 1 attempt. The score profile drops to >2190.

with regards to practical advice, i agree that colleges do probably use only the highest scores for any computation purposes when deciding an initial cut of applicants in the admissions process. But at more highly competitive colleges, impression starts to play a slightly larger role, and having too many sets of SAT scores with little improvement from one to the next is not exactly the preferred one.

This is also major speculation. We have never worked inside an admissions office and cannot claim knowledge of admission officers having subjective feelings about things. On the other hand, it is equally valid for me to say that top colleges are in even fiercer competition for the best students and have bigger reputations on the line. I can say that they therefore would disregard any supposed bad "impression" for the sake of their score profiles. We are better off taking the colleges at their word, instead of doubting them unncessarily.

Also, could you explain what "bad impression" could possibly come out of taking the SAT many times? You have said earlier that other things on the application also matter, which is true - so why do you now claim that colleges penalize multiple attempts? Multiple attempts don't say much from the academic strength of a student. It all seems petty and trivial in this context, especially when compared to the larger objectives of having an impressive SAT score profile and well-rounded students.

A disclaimer: both our models, where students are ranked according to scores, is definitely not used as a final cut because, as we both said, colleges consider other parts of application. But we can safely assume that such a "weeding-out" process happens somewhere upstream, because the admitted students' scores are clearly clustered near the top every time - this is no coincidence. Such an initial weeding-out is then followed by consideration of the other parts of the application.

Butterbeer
06-09-2010, 07:15 PM
First, technical matters:

My principle behind the model of a biased college has always been:

score x with y attempts < score x with z attempts < score a

where x < a, y < z.

However, what happens when you have a 2400 in 3 attempts and a 2390 in 1 attempt? According to you, there is a small penalty for taking the test too many times. Therefore, there's a chance that the slightly lower-score, lower-attempt applicant "replaces" the higher-score, higher-attempt applicant.

Hence, you are committing the fallacy of imposing the "penalty" for multiple attempts a second time, when this disadvantage has ALREADY been taken into account by virtue of the initial ranking which goes according to the principle as listed above.

To understand it better, we could assume that the college deducts 10 points from your score for every attempt beyond the 3rd, etc.

As such, this doesn't apply.

What I have been trying to say is that, the incentive for colleges to treat multiple attempts the same as fewer attempts is not present. This in no way leads to the implication that colleges are hence biased towards multiple attempts. It merely says that there is no reason for colleges to be opposed to treating multiple and fewer attempts equally, since no compelling incentive pushes them to want to treat both equally.

Of course, in the real world, my model would not hold, and a 2200 with 4 attempts may get trumped by a 2190 with 1 attempt. However, the incentive at best remains negligible or insignificant because:

(1) such action at the margin is rather uncommon and
(2) colleges do not use the '> ' notation to denote their admitted student score profile. instead, they use percentile and box plots. hence, given the large population of data, and use of this presentation mode, the admitted student score profile cannot be significantly affected even if colleges are biased.

Now on to the other part:

This is also major speculation. We have never worked inside an admissions office and cannot claim knowledge of admission officers having subjective feelings about things.

This, as I'm sure you're aware of, is a forum. I'm expressing an opinion, and I'm not a fact-checker who merely aims to regurgitate admission policies on college websites. I base this opinion on what I have gleaned from college counsellors, admission books and general sentiment.

For example, top colleges often purport to not have any lower limit for their SAT score for admission. Yes, this indeed seems to be true, as people with sub-2000 scores do get into these schools. But would this in any way lull you (likely to be an unhooked applicant) into not working your best for your SAT score, aiming to surpass the informal 2200 mark?

Similarly, top colleges purport to only take the highest SAT superscore into account, and this indeed may be true. But would this in any way lull you into adopting the strategy of taking attempts, each with marginal increments in preparation and effort, all with the viewpoint that ultimately, you can inch yourself up to a high score?

On the other hand, it is equally valid for me to say that top colleges are in even fiercer competition for the best students and have bigger reputations on the line.

Yes, it is valid. But now I don't see how this point runs into conflict with the topic of debate.


We are better off taking the colleges at their word, instead of doubting them unncessarily.

While there is no explicit proof of this, its a tenable opinion that I wouldn't fault another for holding.

Also, could you explain what "bad impression" could possibly come out of taking the SAT many times?

I've explained that in my earlier post.

You have said earlier that other things on the application also matter, which is true - so why do you now claim that colleges penalize multiple attempts? Multiple attempts don't say much from the academic strength of a student. It all seems petty and trivial in this context, especially when compared to the larger objectives of having an impressive SAT score profile and well-rounded students.


Please don't hijack my arguments by changing their context. The entire topic of this thread was on how an independent variable (SAT scores) affected a
dependent variable (one's application to colleges) and I am merely attempting to draw some links between the two. I have never insinuated that since colleges [I]may penalise excessive attempts, the SAT factor hence occurs at the expense of trivializing the other parts of the application. Indeed, you can infer from my first response post that this generally isn't a make-or-break issue. I have gone in-depth into this issue for the prime purpose of disproving your "incentive" hypothesis, NOT to cast any opinions on relative importance of various parts of the application.

Conclusion: The majority of my later posts has largely been an intellectual exercise for me (i.e. to disprove the 'incentive hypothesis"). I have in no way any intention for anyone to draw implications of my opinions on the attitudes of the admissions committee or any sort of dishonesty or conspiracy on the part of the admission office.

So let's close this thread soon, if not now.

Run
06-10-2010, 11:31 PM
Basically, just do well in as few tries as possible. I'd recommend that if you still can't do well after a few tries in the same year, don't spend any more time and money on it, and spend more time on your other components like your essays, or 'A' level grades or poly GPA if you are still in school.