"There's no difference between 2300 and 2400."

2CHiLLaXiN · January 30, 2011, 1:58pm

The difference between a 2300 and 2400 is a negligible advantage but an advantage nonetheless. However, the difference between a 2300 and a 2400 isn’t as profound as the difference between a 2100 and a 2200.

A 2300 is about a 770 on each section (or 770, 770, and 760, if you want to get all technical). A 2400 is an 800 on each section. On the SAT, a -1 in the raw score can be anywhere from a 770-800 (or 760-800 on the math section). So in theory, you could potentially get the same raw score on two tests but have wildly different scaled scores if the curve is different. I think that colleges recognize this and acknowledge that when you are in the 2300-2400 range, the variation in scores is due mostly to random variation and not so much due to aptitude.

vonlost · January 30, 2011, 2:10pm

Raw stats are sometimes a small part of admission criteria. At our D’s LAC GPA + test scores + rank together count for only 20%.

1029384 · January 30, 2011, 3:02pm

@jeandevaches
I didn’t discuss the causation of the socioeconomic correlation at all. Nor did I say anything about the SAT being a poor indicator of college GPA.

It’s dangerous to assume that rich parents = smarter children. It may mean children better equipped to deal with school, or extracurriculars, or standardized testing, but we can’t assert that the child of a factory worker is in general less intelligent than the child of a surgeon.

Expensive prep courses are not, in general, the reason for the socioeconomic correlation (although in the lower ranges of SAT scores, these courses will make a difference). A lot of the critical reading / writing questions are simply biased in favour of children who grew up in English-speaking families where reading is particularly emphasized. For example, a practice test I did before my SAT Reasoning contained questions with idiomatic expressions like “predilection for”, “predisposed against”; these phrases are considerably more challenging to teach to someone who never hears their family members use such language and does not have the time to read classic literature employing this vocabulary.

But back to the original question: combine a few minor errors that have no relationship to a student’s intelligence, like a Scantron bubbling mistake, a mis-understanding of a math word problem, and a simple calculator sequence-of-operations mistake… and a student’s score might vary 100 points when curved. I don’t think these reflect a significant difference between a 2300 or 2400.

shockazulu · January 30, 2011, 4:19pm

“It’s dangerous to assume that rich parents = smarter children.”

Actually it’s not, at least in terms of the average across these two populations ( rich and poor). This is because intelligence generally is considered to be derived from 50% genetics and 50% environment.

The relationship between intellectual environment and socioeconomic status, as discussed, is obviously quite high. Therefore one can assume that, on average, there is a positive relationship between intelligence and socioeconomic status (though only when one recognizes intelligence is not all genetics).

Indeed, statistics for this hold true. Minorities generally perform significantly worse on the SAT than their white or Asian counterparts. This is because while the quality of genetics is essentially the same, the intellectual environment in which children are raised is, on average, lower for minorities. This is because on, on average, they are of a lower socio economic status.

And yes, I have just generalized SAT results as evidence for an intelligence difference. Given that there is a strong relationship between SAT scores and IQ scores, this seems fair.

So as also previously discussed, if you wanted to truly asses the quality of SAT scores, you would have to account of socio economic level.

And about the point of this thread:
A 2300 is essentially a 2400. The point of the SAT is not to show admissions officers that you can master a standardized test. It’s supposed to measure the potential for learning in college.

Admissions officer are well aware that people take the SAT multiple times and that it is very much a skill, i.e, you can improve easily with practice. Basically, it you can get a 2300, with more practice, you are probably capable of a 2400, and admissions officers no that. This is because, as others have already noted, the difference between a 2300 and a 2400 can be as little as 3 questions. Who cares about that?

It’s declining marginal utility. Once you have shown you are capable of a 2300, the admissions officer has the information to determine that you have the ‘aptitude’ to be successful in college. Anything more than that really doesn’t matter.

Think of it as a curve. Explanatory variable is SAT scores, and response variable is acceptance rate. From scores <1900 or so, chances are very low. As in the difference between a 1200 and a 1500 (total score) to top colleges is essentially zero. You have shown, with such a low score, that you may not be successful at a top college. But teH difference is 300 points! Doesn’t really matter.

Around 2000 or 2100 or w.e you think it is, chances of acceptance to top colleges rise dramatically. The difference between 2100 and 1800 is HUGE, much larger than the difference between 1200 and 1500.

Once you get to about 2300 or so (point again debatable), you chances of acceptance won’t go up very much. Point of decreasing returns. If you score a 2300, you cannot significantly improve your application by bettering you score on the SAT because you have already proven your ‘aptitude’ (or w.e the SAT is considered to measure).

sMITten · January 30, 2011, 5:06pm

I have heard of schools bragging that they reject x number of 2400/4.0 kids. While there can be a number of reasons for this, I think that they may look at a kid with a 2300 differently than a kid at 2400 and other factors do begin to weigh in more heavily.

vonlost · January 30, 2011, 7:06pm

I have heard of kids claiming they have heard schools brag about this, but I’d guess it’s the students doing the bragging! ;)

2400Expert · January 30, 2011, 8:55pm

I got a 2400 on my SAT, and I believe that much of my success (especially obtainment of scholarships) was due to my SAT score. I don’t think I would have had the same success if I scored a 2300.

However, a 2300 is still an excellent score!

a94wang · January 30, 2011, 9:37pm

okay, so can you settle the debate in the critical reading 2011 january post? because you seem to be very knowledgable in this subject matter…

1029384 · January 31, 2011, 12:51am

The first sentence is true (other than the exclusion of Asian from minorities). I’m not sure about the rest of it.

I don’t think I would describe most Asian families as having “higher” socioeconomic status. Also, I don’t see any research in this field, but I would predict that for certain cultures, lower socioeconomics also produce nearly equal proportions of high scorers; there is a strong emphasis on education in my culture, and children from poorer families are likely to be encouraged (or forced) by their parents to work harder to break into a higher wealth bracket.

However, I fundamentally agree with you on this point:

I concur with sMITten on this:

Does a 2400 show an obsession with test preparation? Does the 2300 indicate a healthier balance of academics, test performance, and other (more productive) uses of the student’s time?

Everything, however, still indicates that a 2400 is not necessarily better than a 2300 – and having a 2400 may even have its drawbacks.

Metrical · January 31, 2011, 3:35pm

Many of you have the argument “2400 is 100 points more than 2300” which is really not necessary to point out. Obviously having a perfect score is better, but it is up to each college to decide how much better it actually is (in their admissions decision), which clearly goes back to all the other variables involved. I’d find it very hard to believe that of two students with identical applications the one with the lower score is chosen. I mean, this is like asking if a 100 is better than a 98 in an AP class. Just a waste of time to bother with silly things like that.
With that having been said, the whole “richer people, on average, are simply smarter than poor people, and that’s why they get higher test scores” argument is stupid. First, you must be confusing the SAT with an IQ test. Some people just do not do well on the SAT, but they all have the option of choosing to study for hours and raising their scores. Second, people with relative low IQs have had genius children, and vice versa. While it may be true that children of brilliant parents are more likely to be brilliant due to genetics, it is not always the case. Even so, most people who succeed in high school are not necessarily smart they often are very motivated and hard-working. Finally, I find certain professions pay less than others despite a clear difference in the amount of intelligence required. For instance, a physicist, in my opinion, would be much smarter than lets say a lawyer. Which one pays better? Socioeconomic status is not a good measure of intelligence. It simply happens that the amount of people in the lower-middle class and the lower class is far greater. It is conceited and ignorant to think that the children of richer parents are smarter.

shockazulu · January 31, 2011, 4:55pm

“I don’t think I would describe most Asian families as having “higher” socioeconomic status. Also, I don’t see any research in this field, but I would predict that for certain cultures, lower socioeconomics also produce nearly equal proportions of high scorers; there is a strong emphasis on education in my culture, and children from poorer families are likely to be encouraged (or forced) by their parents to work harder to break into a higher wealth bracket.”

To clarify, I was evaluating this solely within the context of the United States.

1029384 · January 31, 2011, 6:04pm

@shockazulu, #31

To clarify, my comment stands even in the context of the United States. Asians are a minority group, and it’s not like the “strong emphasis on education in my culture” is any less extant when they are immigrants. Just look at Amy Chua’s somewhat misunderstood experiences with childrearing.

@Metrical, I agree in general with your post. Of course, I can’t entirely say that there is no relationship between standardized test scores and IQ; you wouldn’t expect someone with an IQ of 75 to achieve a 2400 – I might even argue that the probability of this is close to 0. I DO agree, though, that when we get to the extreme ends [200, 400] or [2200, 2400] the SAT would not be a good measure of intelligence. Among my peers, people who are ‘smarter’ have scored lower than people who have crammed & attended prep courses for the SAT.

Metrical · January 31, 2011, 7:28pm

Yes, that is definitely true. One must have some sort of above average intellect to score a 2400. The point I was trying to make is that when you get to the higher scores, the argument that someone scoring a 2300 instead of a 2400 has a lower IQ, is flawed. Which makes the argument that the ‘rich kids getting 2350s are smarter than poor kids getting 2280s’ also irrational. The bottom line is, the SAT does have some ability to estimate intellect, but seeing as it is a test that one can study for and obsess over (i.e taking it 4 times because you have the time and money) its ability to measure intellect definitely breaks down at the higher end of the spectrum.

I find that to be true a lot too. I have met many people who are very smart, so they sat down down and took the SAT in one try with no preparation. Others obsessed over the exam since freshman year and scored a hundred points higher.

shockazulu · January 31, 2011, 8:44pm

@1029384

Yes, within the total population of the United States, Asians are a minority group. I misspoke (type?). Still, while the culture may not differ tremendously based on location (this is debatable also), the distribution of the socioeconomic status of these individuals is most certainly significantly different based on location, and that is an essential part of the discussion. People who can afford to move to the US are generally going to be wealthier than their non-immigrating counterparts.

And I haven’t it looked it up recently, but the correlation between IQ and SAT scores is something like .7-.8.

The predictor would be better if people only took the SAT once and didn’t study for it.

Metrical · January 31, 2011, 8:51pm

The predictor would ONLY be accurate if people took it once and didn’t study. I’d like to see that study, because as far as I am concerned many people taking the SAT do not take an IQ test. I don’t see how you can be believe such a strong correlation; it is not even a controlled experiment.

Actually, many of those who immigrate do so because of financial hardship, so that statement really cannot be based on facts. If you are wealthy, political reasons aside, you would be less likely to seek opportunities elsewhere (unless it is for the sole purpose of sending your children to a good university in the US, which even poor parents would wish to do.)

silverturtle · January 31, 2011, 9:40pm

I’m not so sure of that. The ideal scenario, in my speculation, would be a comparison after students have plateaued in their preparation.

shockazulu · January 31, 2011, 10:23pm

@Metrical

Here is a link:
[Formulas</a> converting SAT to IQ scores](<a href=“http://www.sq.4mg.com/IQformula.htm]Formulas”>http://www.sq.4mg.com/IQformula.htm)

It’s not a controlled experiment, but that’s irrelevant; causation is obviously the skills assessed by both of these tests (cognitive intelligence and test taking ability primarily). There’s a strong relationship between the two, that’s the relevant finding.

Also, about the immigrant thing, I understand your conjecture. I don’t have the data to back up what I said earlier. We don’t really have the information necessary to conclude the various levels of wealth of immigrants vs. those do not immigrate. I still think my hypothesis is the most likely, but again, no proof.

@silverturtle
why would you allow students the opportunity to practice the SAT but not an IQ test? That doesn’t make sense. You want to minimize the amount of test-taking skill that you are measuring as much as possible.

silverturtle · January 31, 2011, 10:29pm

The negative effect on validity with which we associate preparation is due merely to the practically inevitable non-standardization of that preparation. Because I expect that intelligence influences the fruitfulness of one’s preparation, adding that as a variable should bring the correlation closer to 1. Of course, a scenario in which everyone prepares to the point of plateau is unrealistic.

shockazulu · January 31, 2011, 11:18pm

@silverturtle

Obviously the optimal scenario, for the purposes of this discussion, would be if subjects would have the same amount of preparation in both the SAT and IQ test. Because this is real life, that obviously won’t happen. It follows then that the easiest way to get a similar level of familiarity with both the SAT and the IQ would to have test-takers equally unprepared (that is, not unprepared at all) for both.

As for intelligence influencing the fruitfulness of one’s preparation, yea sure, it does. But this also does not account for selection bias-only people who would prepare for the SAT have the work ethic to practice with the SAT. So again, if you would let people practice for the SAT in this scenario (and not the IQ test-that still doesn’t make sense), you would be increasing the importance of work ethic and test-taking ability. Again, we do not want to measure either of these things. Ideally, we would want these tests to perfectly assess cognitive/scholastic intelligence without measuring test-taking skill (/work ethic or w.e) at all. But that’s not possible, so you minimize the effect of that skill by having no one prepare for either test.

This assumes you can practice for a IQ test. Given that there is a IQ test improvement industry, just like there is a SAT test improvement industry, this seems likely.

1029384 · February 1, 2011, 12:33am

I have in front of me an article in Psychological Science, September 2005, titled “Unbelievable Results When Predicting IQ From SAT Scores” – unbelievable not as in “astonishing” but as in “absurd”. (Citation: doi: 10.1111/j.1467-9280.2005.01606.x)

The article attacks the conversion you have referenced, and opens with

and ends with

And don’t forget, even on the site mentioned above, “The formulas below do NOT work with “extreme scores” (perhaps over 1250 [out of 1600] SAT)” sounds pretty clearly like the correlation with IQ is considerably weaker at the high extremes.

"There's no difference between 2300 and 2400."

CONNECT WITH US