<p>Yeah, I’d like to see those rankigns with WASP in there.</p>
<p>“You also point out that a school with an exceptionally large economics department would make an Engish department proportionally smaller…(snipped)… I think that is negligible.”</p>
<p>I was just illustrating the general principle, not mandating that there is only one other good department at the other school… It doesn’t have to be only an economics department. It could be 10 departments that are equally strong, and the other school sucks except for English. So indeed students overwhelmingly and disproportionately flock to that school’s English department, since that’s all it’s got going for it. I can create a hypothetical where the numbers are not trivial.</p>
<p>My earlier example, e.g. Clarkson vs. Stanford for engineering is a case in point for that discipline. You may have a case of it, to a degree, on your English rankings as well.</p>
<p>The fact that income and SAT score are associated does not prove causation in any way. As C said, it might just be that industrious smart parents pass these traits on to their children. The fact that such traits also sometimes lead to high incomes could be incidental. Also in making high SAT scores a mere derivative of exogenous factors you cheapen the achievements of the less fortunate students who still manage to do well on the SAT.</p>
<p>Umm… what would you say of rich families who spend money to have their children do SAT prep-courses? Also, a lot of private schools offer SAT preps. In contrast, students at public schools where their peers are going to state or junior colleges may not even realize it is a good idea to study for the SAT. Certainly, students at private schools on average receive better hs educations -you can see the difference in students’ initial performances in college. I think it’s remarkable how the best colleges accept 1/2 of their students from private colleges -obviously this shows that lower-income students have some sort of a disadvantage in life before they even get to college. I think there is definitely an advantage in having money when it comes to standardized testing. If you don’t have money and weren’t highly prepped for college, you are just that more brilliant for being able to do well on SATs.</p>
<p>If you select your pool carefully enough it will work within those limits, because, for schools of similar size and makeup, the % ranking will be very similar to the ranking based on absolute # faculty & course offerings, which I contend is actually the more relevant data.</p>
<p>I would expect this approach to break down more when you start applying it to compare schools of different sizes. You will start rating the larger, more diverse schools lower using your methodology, even though they actually have more faculty in the given department and offer more courses to undergraduates. </p>
<p>For example, you may conclude, using your methodology, that Whitman is superior to Stanford in Biology. Or if not Stanford, then some other school with biology capabilties dwarfing Whitman’s.</p>
<p>Moreover, few people trying to apply this will be aware of all the adjustments to the denominators (eg strip out conservatory, other specialized colleges) that are necessary to make these results even do what you want them to do.
And to to it right, when you strip out the irrelevant specialized colleges you need to adjust the SAT scores for that college as well.</p>
<p>For example, The last report I saw showed the average SAT scores for Oberlin’s College of Arts & Sciences was 70 points higher than for its Conservatory of Music. Not surprising, given the differing aims and selection criteria. Don’t know what you’ve got in there. But even if you, this time, got Oberlin right, there will probably be a number of schools for which the sub-data on the scores for the individual colleges within a university are not readily available.</p>
<p>I have long thought the ACT conversion chart is BS. Most schools using both tests have a lower relative average for the ACT if you use the conversion chart. That is and average ACT of 28 and SAT of 1300 or the like.</p>
<p>xiggi-
I excluded some top-tier schools because I wanted to challenge this technique and see if it distinguished between schools that were similar in SAT. I don’t know if I can find the data I need to do economics departments in LACs but I will try.</p>
<p>ecape-
I agree that poverty places kids at a disadvantage and that they deserve a lot of credit when they succeed. I happen to be one of those (relatively) poor people.</p>
<p>monydad-
I think I understand what you are saying about colleges with uneven or disproportionate departmental enrollments. My brain is still chugging away trying to grasp the implications. Your point is hypothetical but with some specific examples. Maybe the phenomenon is not common or not pronounced. I wonder if using the overall SAT for the college in the formula helps adjust for the unevenness. Schools with one or two “star” departments might also have lower overall SAT. I am still thinking that through.</p>
<p>Your point about Stanford versus Whitman biology departments intrigued me. Does Stanford have an outstanding bio department? I was able to find the undergraduate enrollment in Stanford’s School of Humanities and Sciences, which is where biology is housed. The UG enrollment is 2624, about the same size as many LACs. I think this provides the proper denominator. Accoring to the IPEDS COOL website
Stanford had 131 biology bachelors grads. Stanford’s ratio is therefore .0499 compared with .0324 for Whitman. Multiply that by SAT 75th percentile 1550 and Stanford gets a score of 77, compared with 47 for Whitman and 56 for Haverford. Does this seem valid?</p>
<p>In post #45 I said:
“…Or if not Stanford, then some other school with biology capabilities dwarfing Whitman’s.”</p>
<p>It doesn’t have to be Stanford vs. Whitman. It is inevitable that distortions will exist using your method. Someplace along the line you will (falsely IMO) conclude that a department of 10 (or so) professors is “better” than a department of 30 that has much more extensive resources in biology, simply because the school with the department of 30 has many other strong, and large, departments as well. </p>
<p>And no, I don’t mean exactly 10 or exactly 30.</p>
<p>“I wonder if using the overall SAT for the college in the formula helps adjust for the unevenness.”</p>
<p>I’ll help you with that. No it doesn’t, It just creates more distortion.</p>
<p>What you’d really want is the SAT for the students taking courses in that department. Which is rarely if ever available.</p>
<p>IF the overall SAT you’d propose to use is not just for the college, but includes SATS for students in a conservatory of music, a teacher’s college, a college of agriculture, or whatever other specialized, and separate, college there is there that has completely different admissions practices, and you’ve lumped them all together, then you really have something useless.</p>
<p>monydad-
For undergrad biology, I tried lumping together SAT scores and enrollments from widely diverse units just to see what would happen, this time with some large publics that have lots of diversity in the curriculum. My data again came from IPEDS COOL website. I then compared the ranking with the US News ranking for grad programs in biological sciences. (I know this is a very crude validation.) The US News rank is in parenthesis (the rank within this list alone, not the published rank.)</p>
<p>I used total undergraduate enrollment and the overall average SAT 75th percentile for the entire university. The department index was number of biology bachelors degrees awarded.</p>
<h2>BIOLOGY</h2>
<p>myrank usnewsgradrank college qualityscore
- (4) UCLA 46
- (1) Berkeley 45
- (10) UC Irvine 39
- (2) Wisconsin 32
- (4) U Washington 28
- (16) UC Santa Barbara 25
- (6) U Illinois 22
- (8) UT Austin 22
- (3) U Michigan 18
- (7) U Colorado 17
- (10) UVA 15
- (15) Ohio State 15
- (19) U Missouri 15
- (13) Indiana U 14
- (8) U Minnesota 13
- (14) Penn State 13
- (18) USC 12
- (16) U of Utah 9
- (10) Purdue 6</p>
<p>There are a few big differences between my rank and US News, but only 6 of the 19 ranks differed by more than 3. It certainly isn’t perfect agreement but it is pretty good confirmation that this method has some merit even when numbers are used loosely.</p>
<ol>
<li>What is the correlation coefficient? Is it significant?</li>
<li>What would the correlation be if you simply ranked the colleges by SAT score, and ignored all information about the departments? </li>
</ol>
<p>Is this “departmental ranking” a global ranking of reputation of colleges based on the prominence of the college -is reflected both in the SAT scores of the students while the USNEWS grad school rank is contaminated by halo effects, also from things like SAT score?</p>
<p>“… it is pretty good confirmation that this method has some merit even when numbers are used loosely.”</p>
<p>I think you could have arranged this group of schools randomly and you still would have found confirmation that your method has some merit.</p>
<p>afan-
The correlation between the biology rank based on my “quality score” and the biology rank based on US News graduate edition is .65 (moderately high) and it is statistically significant at the point zero one level with 19 pairs of scores. The US News rank is not the published rank but the rank from one to nineteen within this list based on the US News published rank.</p>
<p>The correlation between my “quality score” and overall SAT 75th percentile was -.35, not statistically significant.</p>
<p>The correlation between the US News rank (1-19) and total undergrad enrollment was -.19. Between US News rank (1-19) and the number of bachelors grads was -.68. So, my quality score did not do any better at predicting the US News rank than did the number of bachelors degrees awarded, at least within this pool of very large public universities. However, I think the “quality score” would excel when the enrollment sizes varied more widely.</p>
<p>monydad-
The LAC “quality scores” in post #39 are on a somewhat different scale than post #51 because, for the large publics, the total undergrad enrollment denominator in the ratio for the large publics included their schools of engineering, communications, and so on. It was not just the undergrad enrollment in arts and sciences. This would have made the ratios smaller for the large publics and their quality scores lower relative to the LACs.</p>
<p>afan-
The correlation is -.46 (moderate) between the Gourman Report score and my “quality score” for biology departments at the large publics, which was statistically significant at the point zero five level.</p>
<p>Even though the calculation was rough, not using quite the right numbers, there is some cross-validation of the idea that the undergrad size of a department, relative to total undergrad enrollment and adjusted for selectivity, does provide some indication of the department’s academic quality.</p>
<p>I tried this ranking method with a department from the area of social sciences: psychology. Here are the psychology rankings for some LACs using degrees awarded divided by total enrollment multiplied by SAT 75th percentile.</p>
<h2>PSYCHOLOGY</h2>
<p>college qualscore</p>
<p>Barnard 45
Davidson 42
Haverford 39
Vassar 37
Claremont McKenna 36
Macalester 36
Middlebury 33 (tie)
Wellesley 33 (tie)
Kenyon 31
Whitman 24
Colby 20
Washington & Lee 19
Colgate 16 (tie)
Bowdoin 16 (tie)</p>
<p>The rankings of this same group of LACs have been quite different for Engish, Biology, and Psychology even though the SAT 75th and the total enrollment have remained constant. The difference has been the size of the department within the college, as measured by the number of bachelors awarded.</p>
<p>For the English departments in the LAC group listed earlier, I looked up the rank of PhDs in English awarded to bachelors graduates in English at these colleges. My source was “Postbaccalaureate Origins of Doctoral Recipients 1986-1995”. This publication does not attempt to adjust for the size of the college when ranking PhD production. </p>
<p>Again, I calculated relative size “quality scores” by dividing the number of English bachelors degrees awarded by the total enrollment and then multiplying by the SAT 75th percentile.</p>
<p>The rank in English PhD production was closely related to the “quality score” (correlation = .68, statistically significant at the .01 level). This further validates the idea that the size of a department, relative to the total enrollment, adjusted for selectivity, is a good indication of departmental quality.</p>
<p>What’s more, within this group of LACs that were similar in selectivity, the best predictor of PhD production in English was simply the number of students who graduated with bachelors degrees in English. (correlation = -.86, statistically significant at the .001 level) In other words, when you control for SAT scores, English PhD production seems to be very closely related to the number of bachelors degrees awarded in English. So, if PhD production is an indicator of quality, then so is departmental size, because PhD production mirrors size. </p>
<p>qualityrank collegename qualityscore enrollment englishbachelors SAT75th EnglishPhDsrank EnglishPhDsnmbr</p>
<p>1 Kenyon 66 1634 76 1420 17 21
2 Davidson 55 1714 66 1440 20 18
3 Haverford 51 1172 41 1460 22 16
4 Colby 47 1821 60 1430 24 14
5 Middlebury 42 2357 66 1500 12 23
6 Barnard 41 2287 66 1430 15 22
7 Oberlin 40 2357 80 1440 1 52
8 Vassar 36 2475 61 1460 5 28
9 Colgate 34 2831 68 1430 18 20
10 Wellesley 33 2289 52 1460 2 45
11 Bowdoin 33 1677 38 1460 47 8
12 Whitman 24 1481 25 1440 47 8
13 WashingtonandLee 19 2174 29 1450 72 6
14 Macalaster 15 1900 20 1450 47 8
15 ClaremontMcKenna 13 1124 10 1460 83 5</p>
<p>i have a question that relates to this. how important is is really to go to a school with your intended major as it’s strength? There are a few schools i’m looking at where I like the location and other factors, but some of the programs i’m interested in are not listed as being strong. For these types of programs, what does it mean if they aren’t strong? Will it really be a lot worse than a strong program? What sort of situations exist?</p>
<p>In most situations it’s not very important, but in some cases it can be extremely important. </p>
<p>For example, if you’re interested in engineering, you need to find a school strong in engineering. However, if you’re interested in fields like history, biology, etc. it’s not as essential you find a strong program, although that’s a plus. You’re more trying to find schools with an overall fit.</p>
<p>Say you really like Brown. It’s physics department isn’t the greatest, but that’s ok. You need a solid foundation, and then you’re off to graduate school. However, the location, student body, etc. are more likely to make you happy than a school with a stronger physics program (like Caltech).</p>
<p>I would go so far as to say that it’s more important that the school you choose be an “overall good fit” than being very strong in your intended major. Students change majors very often, for very good reasons; they grow and change, they didn’t know what the major entailed until they get there, or they discover a new, more exciting passion. As long as the major is “strong enough” at the school - has good facilities, good teaching and a range of opportunities within the major- I would be more concerned about other factors.</p>