How Many Students Go to BC?

Googling “Bellevue College Enrollment” produces the following result:

This is a screenshot from the results of a google search engine. It indicates that Bellevue College had a Total enrollment of 14,675 in 2010.

WIKIPEDIA, on the other hand tells us that:

With an annual enrollment of 37,000 students, Bellevue College (BC) is the largest of the 34 institutions that make up the Washington Community and Technical Colleges system.

A bit more searching yields a bounty of enrollment numbers summarized below

This is a bar graph. It shows enrollment numbers for Bellevue College from a number of different sources. The point of the chart is to show the wide variation numbers from different sources. The numbers are: 5,178, 5,684, 5,787, 8,376, 12,305, 13,398, 13,469, 14,675, 14,675, 18,200, 32,725, 37,000.

What’s going on? What is our enrollment? Can’t we even answer a simple question like how many students go to BC?

It turns out that the question, ”How many students go to BC?” isn’t as simple as it seems. The answer requires assumptions, clarifications and definitions that drive the differences in the numbers reported above.  In particular we have to be clear about:

  • The time period we are interested in or, in the case of these data the time period they are reporting on. A typical answer is “right now” but that isn’t always the best answer.  There are times when enrollment is fluctuating, or school is out of session, or we are only part of the way through the year.[1]
  • The time span we are interested in. If I tell you that I have earned $100,000 on my job it’s more impressive if I’ve had that job for one year than if I’ve had it for 20 years.  In educational contexts, enrollment can be reported by quarter, semester, and year or, in some cases across some other time span.
  • What kinds of students are we interested in. A typical answer is “all of them” but people don’t often think about all of the many kinds of students an institution serves.  This is particularly true of community colleges which serve a wide range of student needs and desires.  Some of our students need a specific skill like instruction on a new piece of software, others are trying to graduate from high school while others want to gain new job skills.  Our enrollment includes all of these learners but in many cases people want to know about a segment of these students, like those enrolled in credit bearing courses for example.

So how do these factors contribute to the variation in the data above?

Let’s start with time period.   What academic year or years are being reported?  Some of the variation in the table above results from data being reported from different years (often without any notation).  Let’s look at the subset of data in the middle range of numbers

This is a barchart showing the middle section of the barchart above.






It turns out that all of these numbers measure the same thing in the same way. In fact they are all using the same source data[2] but for different years. We can plot the data and see that the variation makes a rough sort of sense.

This is a scatterplot showing that the varying enrollment numbers relate to different years. When the numbers are plotted against time on the x-axis the variation looks more sensible.











We can go a bit further and gather all of the data for all of the years from the original source and we see what we would expect from enrollment data; gradual change over time. We also see an increase in enrollments following the financial crisis which is also expected. The key point is that these varying reports of enrollment are not contradictory, just poorly documented. This is a scatterplot just like the one above but with additional data points added. The variation across time is now clearly much smoother than in the previous plot.

OK, that takes care of some of the data but what about the numbers in the 5,000 range?

This is another subsection of the first barchart showing the three smallest enrollment numbers.

The issue here is the kind of student included in the count.  Two of these numbers are also from IPEDS but instead of looking at all total enrollment they look at something called “Degree Seeking Undergraduates.”   The differences are, again mostly due to differences in the year reported.  In this graph the numbers reported online are in red while the rest of the data series from IPEDS is in black.

 This is another scatterplot similar to the two above. Again, variation in estimates makes more sense when time is included on the x-axis. The chart also shows a green X in the upper right hand corner.

So far so good but the BC at a Glance number (represented by a green X) is from Fall 2015 and doesn’t match the IPEDS number in the graph above even though it reports degree seeking undergraduates just like the IPEDS data.

What Gives?

What gives, is the definition of degree seeking. The U.S. Department of Education does not count dually enrolled high school students as degree seeking. Therefore, none of our many Running Start students are reported as undergraduates through the IPEDS system. However, Running Start students earn a lot of degrees and almost always intend to use their credits for degrees. For this reason we count them as degree seeking when reporting our own data.[3] This is why our BC at a Glance number is higher than the corresponding IPEDS number and part of the reason why IPEDS data don’t always give a very accurate representation of our institution.

This, leaves us with just a few numbers to explain

This is another subsection of the first barchart. This one shows the four largest numbers.

The time span also affects these numbers. The 8,376 reported in BC College at a Glance is an annual number of unduplicated degree seeking undergraduates. Some students are here for all quarters in an academic year while others are here for only one or two quarters. BC at a Glance reports 5,787 as the average number of degree seeking undergraduates in a single quarter of the regular 2015-16 academic year[4]. The 8,376 represents annual enrollments meaning the number of different degree seeking individuals who were enrolled for any number of quarters.

Finally, what about the largest numbers in the table? These numbers result from a combination of type of kind of student and time span selections. As a community college, Bellevue serves a wide range of learners in a wide range of contexts. These include a large number of learners taking non-credit continuing education courses who are not reported to the U.S. Department of Education. 32,725 is the number of unique number of individuals who are connected to the college as learners in any way in the 2015-16 academic year. The 37,000 number in Wikipedia is that same number from a few years ago. 18,200 is the quarterly version of that number.

Here are a few takeaways from this discussion.

  • What is BC’s enrollment? Is a question that should only be answered when the time period, time span and kind of student are specified.
  • Most external sources that report data on lots and lots of colleges (Cappex, Peterson’s, the College Board) rely on data collected from or prepared for IPEDS reporting.
  • IPEDS reporting has some deficiencies particularly with regard to dually enrolled students and can be misleading.
  • Your best source for accurate data about BC college enrollments is the Office of Effectiveness and Strategic Planning which produces quarterly enrollment reports that can be found here.

[1] In the age of big data we have fetishized so called “real time data” but real time data isn’t always accurate for a number of reasons.  Lots of really important data is quite old. For example monthly unemployment data is reported about a week after the month ends because it takes time to collect, clean and report the data.  Data on economic growth is reported in three waves one two and three months after the end of each quarter with each wave thought to be more accurate than the previous.  While there is a lot of interest in leading indicators of growth, everyone recognizes that they are subject to error and no one seriously wants “real time” GDP data because it would be so inaccurate as to be worthless.

[2] The Integrated Postsecondary Education Data System (IPEDS) is a federally mandated reporting system for all institutions of higher education. Most sources that report on large numbers of colleges (Peterson’s, College Board, google) rely on IPEDS data.  The problems with IPEDS data are discussed below.

[3] On the other hand we generally exclude Tech Prep and College in the High School students because, while they earn college credit through BC they are not taught by our faculty.

[4] Here, regular academic year means fall, winter and spring quarters.  BC enrollments are substantially lower in the summer quarter.

Last Updated October 25, 2016