Currently browsing posts filed under "Common Data Set"

Follow this category via RSS

New Common Data Set

hmm writes:

Williams just posted their CDS which could be worth looking through:

Massive surges in test scores this year than from any other year. Changing admissions strategy?

Another interesting points: Williams hit a 6:1 student to faculty ratio, which will likely be the lowest of any LAC. Good move. Most peers are 8:1.

Of Williams’s peer schools, only Pomona has posted their 2018 CDS as well:

The test score gap between the two is gigantic this year; in the past, Pomona has had equal or higher test scores than Williams. Pomona is more racially diverse and has a higher percent of students ranking in the top 10%, as well as a higher yield, so it seems they deliberately made test scores a weaker factor.

Would be interested in seeing what people think. Kudos to Pomona for attracting a super diverse student body (even Stanford doesn’t have the same %), but is that worth significant declines in testing? It’ll be interesting seeing the long-term implications of this for graduation rates.

hmm should join us as an author. As should others! Make EphBlog Great Again!


EphBlog Loves Provost Love

EphBlog loves new Williams Provost Dukes Love. Why? Recall our recurrent complaints about transparency with regard to already published College documents, like the Common Data Set reports. Formerly, Williams only provided the reports back to 2011. Now, it provides an archive back to 1998. Well done Provost Love!

But because this era of Perestroika might end, EphBlog has taken the precaution of saving permanent copies: cds_2010-11, cds_2009-10, cds_2008-09, cds_2007-08, cds_2005-06, cds_2006-07, cds_2004-05, cds_2003-04, cds_2002-03, cds_2001-02, cds_2000-01, cds_1999-00 and cds_1998-99.

It is especially nice to see a provost committed to transparency as Williams begins the re-accreditation process. Long time readers will recall that we devoted the month of January 2009 to going through the last re-accredidation report. Alas, we did not save a copy! Is one available somewhere?

UPDATE: A loyal reader points to this archive of material related to accreditation. Thanks! And kudos to Williams for making this material available even a decade later. Anyone interested in following this round of accreditation should study the last round closely.


SAT Score Changes

Interested in SAT score changes at Williams over the last 15 years? Me too! Alas, the College does not make it easy to study these things since they deleted the old Common Data Sets. Fortunately, I saved this link from 1998-1999 (although the link does not work):

C9. Percent
and number of first-time, first-year students enrolled in fall 1998 who
submitted national standardized (SAT/ACT) test scores. Does not include
partial test scores. SAt scores are recentered.

submitting SAT scores:

Number submitting
SAT scores:

submitting ACT scores:

Number submitting
ACT scores:

25th percentile
75th percentilee
SAT I Verbal



SAT I Math



ACT Composite



ACT English
ACT Math

of first-time, first-year students with scores in each range

SAT I Verbal
SAT I Math
















Here is the Fall 2014 data from IPEDS:


By the way, does anyone know how to get time series data out of IPEDS?

And here is a relevant table from the 2015-2016 Common Data Set (pdf):sat

1) I apologize that this is such a mish-mash.

2) It is not clear how comparable these numbers are over time. First, the rise of score choice and/or super scoring has made it easier (and more common) for students to take a test multiple times and only report the best results. Second, students are now more likely to take both the SAT and the ACT and either only report one. (Or, they report both and the College only uses the better in its own reporting.) But ignore those complications for now.

3) Scores have increased meaningfully over the last 15 years. But, given 2), I can’t say whether or not this is because the students have gotten smarter. Opinions from readers?


Common Data Sets


The College used to make its entire history of Common Data Sets available. Now it doesn’t. Am I the only person who finds this behavior pathetic?


Old Common Data Sets

Sadly, it is becoming more and more obvious that we can’t trust Williams to make publicly accessible documents that it has made public in the past. The most obvious example involve the Common Data Sets, which used to go back to 1998 but now only go back to 2011. Isn’t that pathetic? (And, yes, I have e-mailed to complain.) So:

1) Below are permanent copies of what I have now, less they disappear in the future.

2) Note how we have copies for 2009/2010 and 2010/2011, but these PDFs don’t exist on the Williams webpage anymore. Isn’t it sad that, if you want to look at these, you have to come to EphBlog?

A key part of transparency (and taking history seriously) is maintaining permanent copies of public Williams documents.

CDS 2009-2010
CDS 2010-2011
CDS 2011-2012
CDS 2012-2013
CDS 2013-2014
CDS 2014-2015

I think the below links used to work, but they don’t now. I hope to investigate this later.
Read more


Common Data Sets

The Common Data Set for 2015-2016 has been posted. Comments:

1) Williams has removed copies of the Common Data Set from before 2011 from its website. Pathetic! So, because we can’t trust Williams to be honest about its history, we have to start maintaining our own copies: 2011-2012, 2012-1013, 2013-2014, 2014-2015, 2015-2016.

2) Do readers notice any major changes over the last five years? Should we spend a week reviewing these documents? Let us know your preferences!

3) Diversity is always interesting. Compare the first year class 5 years ago:


With today:


The biggest change is probably the increase in Asian-Americans and the decrease in whites. (Also, I would not be surprised if many students the “Two or more races” category were at least part Asian. Indeed, I know many mixed-race white/Asian applicants who check the “white” box on the Common Ap because they worry (correctly!) that elite US college discriminate against Asian-Americans. A majority (?) of international students are also Asian by race.

Put it all together and I would wager that 20% of the class of 2018 at Williams is Asian. Hard to believe that that proportion is going anywhere but up over the next few decades . . .

Not that there is anything wrong with that!


Currently browsing posts filed under "Common Data Set"

Follow this category via RSS