SSAT-to-SAT

Prep School Admissions
1

<p>According to SSATB, one's SSAT scores are a good predictor of his/her future SAT scores. This is probably based on its own survey and statistical analysis. So I did a quickie of my own to see if such a prediction model is any good and, more importantly, to find those schools that were able to help their students more than predicted, and more than peer schools.</p>

<p>To this end, I turned to boardingschoolreview.com, where you can find average SSAT and SAT scores for individual schools. Out of 321 schools, 57 schools showed both SSAT and SAT scores. Indeed, there was a correlation, a strong one (r=0.8) , between SSAT and SAT scores. (Pardon my language. Here, 'r' denotes correlation coefficient based on Spearman analysis. r=0 means no correlations, and r=1 means perfect correlations. There was less than 1 in > 1 trillion of chance that this happened by accident.) Then, I did what else might be correlated with SAT scores, and found none came close. Not class size, not % of teachers with advanced degrees, not # of AP classes, not teacher:students ratio,... and you can guess the rest. One thing that came close (r=0.57) was endowment size.</p>

<p>From this analysis, I derived an equation to quantify and compare how much improvement (or reverse) schools helped their students with SAT scores beyond what's predicted. And I grouped them by their average SSAT percentiles before comparison, because given everything being equal, it is a lot more difficult to improve 5 percentile from 90 percentile than from 50 percentile. Now, here are the results:</p>

<p>85-94 percentile group
Boarding School | Avg. SSAT percentile | Avg. SAT score | Expected SAT | Observed/Expected SAT
* Middlesex School | 88 | 2080 | 2000 | 1.04
* Groton School | 90 | 2080 | 2016 | 1.03
* Phillips Exeter | 94 | 2097 | 2046 | 1.02
* Milton Academy | 90 | 2065 | 2016 | 1.02
* Choate Rosemary Hall | 85 | 2018 | 1977 | 1.02
* Phillips Academy | 94 | 2076 | 2046 | 1.01
* The Hotchkiss School | 87 | 2013 | 1992 | 1.01
* Deerfield Academy | 87 | 2000 | 1992 | 1
* St. Paul's School | 89 | 2014 | 2008 | 1
* Indian Springs | 85 | 1970 | 1977 | 1
* Georgetown Preparatory | 90 | 1975 | 2016 | 0.98
* St. Andrew's School, | 85 | 1928 | 1977 | 0.98
* The Thacher School | 90 | 1962 | 2016 | 0.97
* Woodside Priory School | 85 | 1844 | 1977 | 0.93</p>

<p>75-84 percentile group
Boarding School | Avg. SSAT percentile | Avg. SAT score | Expected SAT | Observed/Expected SAT
* Linden Hall | 78 | 2095 | 1923 | 1.09
* Lawrenceville School | 84 | 2095 | 1969 | 1.06
* Concord Academy | 84 | 2050 | 1969 | 1.04
* Cate School | 83 | 2027 | 1962 | 1.03
* The Loomis Chaffee | 79 | 1935 | 1931 | 1
* St. George's School | 81 | 1920 | 1946 | 0.99
* St. Mark's School | 84 | 1940 | 1969 | 0.99
* The Taft School | 84 | 1928 | 1969 | 0.98
* Blair Academy | 75 | 1840 | 1900 | 0.97
* Hill School | 79 | 1855 | 1931 | 0.96
* Verde Valley School | 80 | 1700 | 1938 | 0.88</p>

<p>65-74 percentile group
Boarding School | Avg. SSAT percentile | Avg. SAT score | Expected SAT | Observed/Expected SAT
* Emma Willard School | 66 | 1950 | 1830 | 1.07
* Cranbrook Schools | 68 | 1939 | 1846 | 1.05
* Westminster School | 70 | 1926 | 1861 | 1.03
* Miss Porter's School | 70 | 1920 | 1861 | 1.03
* Stevenson School | 67 | 1866 | 1838 | 1.02
* Holderness School | 65 | 1800 | 1823 | 0.99
* Portsmouth Abbey | 72 | 1807 | 1877 | 0.96
* Lawrence Academy | 65 | 1732 | 1823 | 0.95
* The Ethel Walker | 65 | 1675 | 1823 | 0.92
* Colorado Rocky Mountain | 73 | 1701 | 1884 | 0.9
* Hawai'i Preparatory | 65 | 1628 | 1823 | 0.89</p>

<p>55-64 percentile group
Boarding School | Avg. SSAT percentile | Avg. SAT score | Expected SAT | Observed/Expected SAT
* Suffield Academy | 55 | 1910 | 1746 | 1.09
* Dana Hall School | 58 | 1885 | 1769 | 1.07
* Canterbury School | 55 | 1820 | 1746 | 1.04
* Asheville School | 60 | 1860 | 1784 | 1.04
* The Bolles School | 56 | 1800 | 1753 | 1.03
* Garrison Forest | 55 | 1790 | 1746 | 1.03
* Saint Andrew's | 60 | 1820 | 1784 | 1.02
* Westover School | 61 | 1827 | 1792 | 1.02
* The Williston Northampton | 57 | 1790 | 1761 | 1.02
* Pomfret School | 62 | 1786 | 1800 | 0.99
* Avon Old Farms | 60 | 1705 | 1784 | 0.96</p>

<p>45-54 percentile group
Boarding School | Avg. SSAT percentile | Avg. SAT score | Expected SAT | Observed/Expected SAT
* St. Anne's-Belfield | 50 | 1827 | 1707 | 1.07
* Salisbury School | 54 | 1810 | 1738 | 1.04
* Ridley College | 50 | 1750 | 1707 | 1.03
* Rabun Gap-Nacoochee | 45 | 1691 | 1669 | 1.01
* Trinity-Pawling | 50 | 1710 | 1707 | 1
* The Gunnery | 50 | 1680 | 1707 | 0.98
* Church Farm School | 50 | 1670 | 1707 | 0.98
* Proctor Academy | 54 | 1660 | 1738 | 0.96
* Vermont Academy | 50 | 1550 | 1707 | 0.91
* Brewster Academy | 54 | 1548 | 1738 | 0.89</p>

<p>Schools are sorted based on the ratio of observed or actual SAT scores over expected scores, and the higher the ratio, the more improvement in SAT scores. Now, just like SSAT, we all understand a higher SAT score does not guarantee the "better" colleges. (You define "better" here.) And for many boarding schools, higher SAT scores might not be one of their goals. But lower SAT scores can limit one's college choices. So no lecturing is necessary how significant or insignificant SAT scores in college application. </p>

<p>Having done this analysis, beside those top ranked elite schools with super statistics, two schools stand out from the rest. One is Linden Hall and another is Suffield Academy. I wish I could show the chart that I'm seeing, but they just stand out by merely looking at the chart. Both shows 9% improvement in SAT scores than predicted, which is 164 and 172 and which I believe is very significant improvement. (Disclaimer: I hadn't known anything about these schools until I finished this analysis.)</p>

2

<p>Thanks for sharing. I had just been wondering if anyone had done that analysis.</p>

3

<p>What a really fantastic analysis. Really interesting stuff. For me, the bottom line is: quality in, quality out. The school may keep them occupied for 4 years, but these kids – on average – would probably have done about the same on SATs no matter what school they went to (that is, an SSAT 94th percentile kid would score very high on SATs regardless of their high school). This same phenomenon happens in sports. Holderness has a great varsity ice hockey team – even just had someone go to the NHL – but is that because of the coaching or is it that Holderness is able to recruit superior players. I wonder… is boarding school really just about the quality of talent that can bee attracted to attend the school or is there actually some value-add that happens at the school. Aside from the couple of schools noted – where there is a boost seen – it looks like schools really don’t contribute meaningfully to the student’s academic development. If true, then really all a diploma from an “elite” boarding school is is a signaling device – hey college AO, this kid was smart enough to get into Hotchkiss and smart enough to graduate. Along the way he got a 2,000 on his SATs. Admit him with no risk because he’s certified quality. $250K for a signal.</p>

4

<p>Not useful. The SAT should be the scores of the graduating seniors, while the SSAT would be either admitted students, or the entering freshmen. You’re comparing apples and oranges. If you follow the schools for some time, you’ll notice both SAT and SSAT scores vary.</p>

5

<p>Test scores are not as strong indicators of collegiate and professional success as transcripts. The value of the BS experience is not only in superlative academic culture, but also in time management, independent living skills, self advocacy, tenacity & resilience, etc…</p>

6

<p>Regardless of whether SAT scores are a predictor of life success in anything, they are a very important factor for college admissions for the applicant. And for both prep schools & colleges, SAT scores are a very important part of their marketing-- schools can’t brag how high their own avg GPA is because that only suggests grade inflation. </p>

<p>As Sharing gift pointed out, a high SAT score will not guarantee u admission to selective colleges, but a low score may significantly hinder u. Regardless of how much colleges profess SAT scores aren’t that important, you’ll observe that the avg SAT scores for selective colleges don’t decrease because there will be hell to pay in the USNWR college ranking. </p>

<p>The higher a college’s avg SAT score, the more important it is for an applicant to have a high SAT score. For prep schools, what can be better for their college matriculation record than to pre-select the kids who will statistically have the high SAT scores, i.e. the ones w the high SSAT scores. </p>

<p>The kids that are accepted into prep schools w significantly lower than avg SSAT scores will have a strong hook to get them accepted to a selective college regardless of their SAT scores. For the rest of us, test scores matter a lot.</p>

7

<p>Great job with analyzing these data! </p>

<p>As you point out, ceiling effects on SSAT may limit the correlation (since admitted students score a mean of 93% at some schools, and the distribution is probably not normal near the upper limit). </p>

<p>But schools purposely do not admit based on SSATs alone. A reasonably high score is expected, but many kids with very high SSAT scores are not admitted to top boarding schools every year. A passion, contribution to the community, athletic interests, etc., are valued just as much as high test scores. The school would be pretty boring if it were composed of excellent test takers without other interests.</p>

8

<p>When I shared this analysis, my intention was just that–to share my analysis, not to elicit opinions on how important or trivial SAT scores are to college application. I guess we all know that SAT is just one of the many factors affecting college admission, especially for those nationally competitive colleges. It is up to individuals–colleges as well as applicants–how much value they place on SAT scores… and accept consequences for their value judgment. </p>

<p>To elaborate, what I intended was not to deliver the information that was not readily available to public (i.e., SSAT and SAT scores, which took less than 30 min to download, process and reformat), but to share an insight that was not immediately apparent (i.e., whether a school outperform or underperform against expectation). My understanding is no one has ever tried (or at least published) to compute actual SAT scores vs. predicted scores and compared schools based on this metric.</p>

<p>As for the apples-oranges analogy, I believe Periwinkle had a point. It would have been more accurate if we started a cohort of entering freshman class with SSAT scores and followed them through their senior year when they had gotten SAT scores (a longitudinal study). I would readily accept the notion of year-to-year variations in SSAT and SAT scores, but the degree of variation must be rather small, if not insignificant, considering the high correlation coefficient value (0.8). That means, say, incoming and outgoing classes at Phillips Academy share a remarkably similar academic profiles (or at least SSAT and SAT score distribution) over several years. This can’t be the result of random events. So my conclusion is despite this caveat, the analysis is still valid and useful… that is if you appreciate its utility.</p>

9

<p>One of the most excellently researched and informative posts I’ve seen here at CC. SSAT and SAT may not be a panacea for admission, but they are the only independent gauges publicly available I am aware of that allow apples-to-apples comparisons. I am heartened to see that two of our family’s favorites, Groton and Middlesex, are the top-tier’s biggest “value adders”. That confirms our experience with these schools. They are both absolutely amazing places.</p>

<p>We have added Linden Hall and Lawrenceville to our pool of prospective schools for son #2.</p>

<p>Thanks again for sharing this, excellent analysis and info.</p>

10

<p>

</p>

<p>Sorry to disappoint you but Linden Hall is an all-girls school.</p>

11

<p>should do more research of my own before posting :-)</p>

12

<p>Thank you SharingGift for indeed sharing your gift with this excellent analysis and insight.</p>

<p>I am not a statistician which may well be obvious in a moment, but I wondered whether the option many students have to take the ACT in lieu of the SAT could skew overall data. For instance, if more students at a particular school choose to take the ACT, the SAT results would reflect a smaller sample that may no longer be representative of the class/school as a whole.</p>

<p>Also, because the ACT and SAT often appeal to different types of students/learners, I wonder if some schools, culturally and pedagogically, tend to turn out test takers more inclined/suited toward one test vs another, obviously not in any hard line kind of way.</p>

<p>That said, my oldest is just 3 weeks into his freshman year at BS, making us total neophytes at this, so I could be missing a whole bunch of factors obvious to those with more experience, but these were some early morning musings while slurping my Saturday coffee.</p>

13

<p>Very interesting. I’m wondering if there is any correlation to the international population. Don’t know much about Suffield, but Linden Hall has a huge Chinese population and also has required SAT prep. Many of the students have very limited English when they get there so the ESL and SAT prep probably help a lot.</p>

14

<p>I wonder if the schools showing greatest improvement from ssat to sat, tend to have more drop outs from the lower end of the ssat distribution. That is, a gain from ssat to sat could for some schools represent the culling of weaker students. If so, the gain in test scores could reflect that the school provides less support to struggling students.</p>

15

<p>

</p>

<p>Not sure many schools use their academic support systems for students who don’t struggle. If there are such schools, obviously you’d want to stay away from them. I doubt that is the case with any of the schools discussed here. The support systems are there specifically for the students who need them most. Even after taking advantage of the help that is there for them, some students will leave, but the schools are absolutely not looking to winnow them out.</p>

16

<p>Oh, no, I didn’t mean that some schools might provide less support to kids that struggle than to kids that don’t struggle. I meant some schools might provide less support in general than other schools, so they are more likely to lose struggling students. I also didn’t mean to imply that some schools are “looking to winnow them out”. I meant that perhaps some schools just don’t have great support services.</p>

17

<p>^^Ah, that makes sense.</p>

18

<p>I went the Linden Hall so I can tell you why there is that difference. SAT Prep is all that school cares about. SAT Prep is constant. SAT II exams are finals for honors classes. My freshman year we took 14 (I counted) SAT II practice exams throughout the year. SAT vocab practice is part of every english class. SAT prep math warm-ups. SAT prep books are supplementary to every class. SAT prep classes are mandatory junior year, and when you get to your senior year most people take the SAT test at every test date available. A certain number of them (3 or 4) are mandatory, so everyone takes the actual test a few times. Also, there were some rumors that the previous headmaster of 8 years used fraudulent test scores. I don’t know if it is true, but I do know that the English SAT scores do not match the english fluency in the hallways at LH. That headmaster was also fired at the end of last year.</p>

<p>The SAT scores may look good, but the constant test prep is actually one of the downfalls of the school. Real learning and deep thinking was replaced by multiple choice and test prep every day.</p>

19

<p>were replaced** sorry I was pretty tired when I wrote that</p>

20

<p>

</p>

<p>That’s an interesting idea. Since SSAT and SAT are similar in a sense that they tend to reflect innate intelligence than achievement/knowledge (ACT?), you could imagine better correlation between them than with ACT. I think though there is a strong correlation between SAT and ACT for the majority of students. Individual outliers won’t affect statistics and overall conclusions.</p>