When comparing universities there is a statistic that some people like to use called “selectivity”.  This is simply the ratio of (# of offers)/(# of applicants).  So, a school that has a lot of applicants and only sends out a few offers is considered to be “highly selective”, because their ratio will be very low.  Selective schools are sometimes viewed as more “prestigious” by administrators, parents and applicants.

This is not the whole picture however.  There is another interesting statistic called “yield”, which is the ratio of (# registered)/(# offers).  If this number is low, it means that a school sends out many offers that are not accepted by the applicants.  There are various reasons why this might be high or low, but to illustrate let’s look at some real data for 2011.  The following table was constructed using data from the American Society for Engineering Education (ASEE) college profile database, which you can access at this link:  http://profiles.asee.org/   From this database, we can look up the number of applicants, offers, and registrants (i.e. those who accepted the offer and attended that school) for Canadian and select US engineering programs.  Then we can calculate the selectivity and yield for each, as shown below.  I also added the QS Engineering and Technology University Ranking, where available.

First, you might notice that a few Canadian schools are missing from the table.  There was no information in the ASEE database, so I presume they didn’t participate in the survey, and Western was missing one piece of data.  The McGill data was only available up to 2010, so I listed that instead.  For the U.S. schools, I just randomly selected a few well-known ones from the top end of the QS rankings.

Comments:

Looking at selectivity, we see the hyper-selective U.S. schools such as MIT, Stanford, Columbia, where the ratio is 0.1 or less.  There is nothing in Canada that is quite that low.  It looks like UBC is the most “selective” in Canada, followed closely by Toronto, McGill and Waterloo.  These selectivities are not much different from many of the well-known and highly ranked U.S. schools like Michigan.

Looking at yields, Calgary comes out on top at an amazing 82%, with Alberta not far behind.  The only U.S. school coming close is Stanford.  Why are these Canadian yields so high?  I suspect is has to do with the amount of local choice.  In Calgary, you either accept the offer to your local engineering school, or move away some distance, so the tendency is for higher yields.  In southern Ontario where there are many choices within a reasonable driving distance, the yields are lower.  Likewise in the U.S., the highly ranked UC Berkeley is quite selective, but has a relatively low yield because there are plenty of top engineering schools to choose from in California.

Admission offices in universities are usually quite aware of their yield values.  They usually send out more offers than they actually have space for, knowing that only a certain fraction will likely be accepted.  It is sort of like airlines over-booking their seats, knowing that a few passengers won’t show up or will re-schedule.

Consider also that there is a connection between selectivity and yield.  If a school tends to have a high yield, they will send out fewer offers and that will drive down their selectivity ratio.  And vice versa.  So, although people tend to compare just selectivities, they can’t really be viewed in isolation.  I’ve run across rumours that some universities try to drive up the number of applications they receive, so that their selectivity ratio will be driven down and they will appear to be more “selective”.  Certainly the hyper-selective U.S. schools don’t need to encourage more applicants.  Getting admission there is already more like a lottery than a selection process.

You might think that highly selective schools would be ranked higher, so I plotted the selectivity versus QS Rank.  It looked like a random scatter graph, so there is no obvious correlation between selectivity and QS Rank (which is all reputation based).

Finally, looking under the column labelled “Registered”, we can compare the relative sizes of Engineering programs in a few schools.  Waterloo appears to be the largest in Canada, but is still quite a bit smaller than some U.S. schools like Georgia Tech.  Is a smaller school better than a larger one?  I think that’s more of a personal preference, and there’s no clear correlation between QS ranking (reputation) and the size of the school.  My view is that a larger school offers more opportunities to interact with different faculty and other students.