Tuesday, March 13, 2018

Early Decision and Early Action Advantage

There is a lot of talk about admission rates, especially at the most competitive colleges and universities, and even more talk, it seems, about how much of an advantage students get by applying early, via Early Decision (ED, which is binding) or Early Action (EA, which is restrictive, but non-binding).

I license the Peterson's data set, and they break out admissions data by total, ED, and EA, and I did some calculations to create the visuals below.

Two important caveats: Some colleges clearly have people inputting the data who do not understand our terminology, who don't run data correctly, or who make a lot of typos (a -500% admission rate is probably desirable, but not possible, for instance).  Second, not every university with an EA or ED option (or any combination of them, including the different ED flavors), breaks out their data.

Start with the overall admit rate.  That's the one that gets published, and the one people think about. It's the fatter, light gray bar.  Then, the purple bar is the regular admit rate, that is, the calculated estimate of the admit rate for non-early applications (this is all applications minus all early types).  The light teal bar is the early admit rate: ED plans on the top chart, and EA plans on the bottom.  Some colleges have both, of course, but most show up only once.

You can use the filter at right to include colleges by their self-described level of admissions difficulty.

Working on another view to show the number of admits scooped up early vs. regular.  Stay tuned.  Until then, what do you notice here?  Leave a comment below.

Thursday, March 1, 2018

Tuition at State Flagships

The College Board publishes good and interesting data about college tuition, including a great table of tuition at state flagship universities. (I realized while writing this that I don't know how a university is designated a state flagship.  Maybe someone knows.)

There is some interesting stuff here, but I'll leave it for you to decide what jumps out at you: If you live in North Dakota, you might wonder why South Dakota has such low tuition for non-residents.  If you live just outside Virginia or Michigan, you might wonder why it costs so much to cross the border.

Anyway, using the tabs across the top, there are five views here:


Four maps, showing (clockwise from upper left) in-state tuition, out-of-state tuition, non-resident premium index (that is, how much extra a non-resident pays, normalized to that state's in-state tuition), and the non-resident premium in dollars.  Hover over a state for details.  You can change the year, and see the values in 2017 inflation-adjusted dollars, or nominal (non-adjusted) dollars.

States in Context

This arrays the states by tuition over time.  Use the highlight functions (go ahead, type in the box; you won't break anything) to focus on a region or a specific state. You can view resident or non-resident tuition, adjusted or non-adjusted.

Single Institution

Just what it says.  The view starts with The University of Michigan, but you can change it to any state flagship using the control at top right. Percentage increase is best viewed in 2017 adjusted dollars, of course.

Percentage Change

Shows change of in-state tuition by institution over time.  The ending value is calculated as a percentage change between the first and last years selected, so use the controls to limit the years.  Again, highlight functions put your institution in context

Non-resident Premium 

This shows how much extra non-residents pay, and trends over time.  Again, highlighter is your best friend.

Feel free to share this, of course, especially with people who are running for office in your state.

And, as always, let me know what you think.

Monday, February 26, 2018

College Board AP Data

The College Board recently released data on its AP Exams.  I've downloaded several workbooks already, and of the one I've dug into, I've only been able to get through two worksheets.  The data presentation is clunky (please, agencies, provide un-pivoted data without merged cells and totals and all that stuff, if not by itself, then as a companion), but it reveals some interesting patterns.

Well, I think so.

I've visualized it in five views: The source of the data is here, in case you want to download it yourself.

View 1, Totals (using the tabs across the top) is just totals: Use the controls to show males or females, or certain scores, or certain exams.  I think it's very compelling, especially if you look at the high scores the College Board claim about AP opening access to selective institutions.

View 2, Scores by Ethnicity and Exam, shows score distributions of the four largest ethnic groups.  Filter by a single exam if you'd like.

View 3, 100% Stacked Bars, shows the same data, presented by ethnicity.  Again, filter to a test if you'd like.

View 4, Mean Scores by Ethnicity and Exam, arrays all tests, and breaks out mean scores (yes, I know you shouldn't take averages of string variables.  So sue me).  Use the highlighter if you'd like to make any of the groups stand out visually, and filter by gender if you'd like.

View 5, Mean Scores by Gender and Exam, shows the differences between males and females. Filter to a single ethnicity if you'd like.

Tell me what you see.  Does this change your perspective on the College Board claims, or does it strengthen them?  Does it help you make up your mind?

I'd love to hear.

Wednesday, January 31, 2018

How is College Enrollment in the US Changing?

College enrollment is down.  Or maybe it's up.  Or maybe it's both.

When you read headlines, you don't get a lot of nuance. And in a country as big as ours, with such an incredible diversity of programs and widely divergent institutions, nuance is important.  So this may help do the trick.

This is enrollment data from about 6,600 post-secondary institutions in the US, and goes back as far as 1980.  It includes every institution, including those that grant degrees, and those that don't; four-year private, not-for-profits, for-profits, and publics; liberal arts colleges, research universities, and technical institutes.  All here.

It's on two dashboards.  The first shows all undergraduate and graduate enrollment at all these institutions, since 1980.  (Note: The data skips from 1980 to 1984, and I took out two years of data--1998 and 1999--because they looked a little funky.)

On the first dashboard, there are several controls to filter the data.  So for instance, if you want to look at just doctoral institutions, you can do that.  Just colleges in New England? Yes.  Only care about full-time enrollment? Just use the filter to select it.  If graduate enrollment is your interest, it's easy to get rid of the undergraduate data.  Just use the controls.  The top chart shows raw numbers, and the bottom chart shows percent change over time.  If you want a longer or shorter window, there's a control to limit the number of years.  This is especially helpful to show percent change.

Then, you can break out what ever enrollment you've selected.  Use the control titled "Color Lines By" and you can split the data shown into groups. 

Try it.  You won't break anything.  You can always reset using the little reset button at the bottom.

The second dashboard (using tabs across the top) shows similar data, but you can choose an individual college.  Once you've done so, you can limit the data shown, and you can also split it out according to your interest.

Have fun.  I've found some interesting little ditties I'll be tweeting out, and I encourage you to do the same.

Thursday, January 25, 2018

A Quick Look at the NACUBO Endowment Data

Each year NACUBO releases its study of endowment changes at about 800 colleges and universities in the US and Canada.  For this post, I'm including only those institutions in the US, and only those who reported two years of data to the survey, or about 787 institutions.

Higher Education in the US, of course, is a classic story of the haves and have nots; a few institutions near the top of the endowment food chain have amassed enormous endowments, allowing them great freedom in the programs they offer and the students they enroll. In fact, the 21 most well endowed institutions control over half, or about $280B of the $560B held overall, leaving the other 766 to divvy up the remaining $280B among them; the top 93 own 75%.

What's more interesting, I think, is the astonishing endowment growth: Stanford added $2.4B to its endowment in one year.  That amount is bigger than all but 38 of these institutions' total 2017 value.  In other words, if the gain on Stanford's endowment was an endowment, it would be the 39th largest endowment in the nation.  And in total value, it still trails Harvard by about $12B.

A couple of notes: Endowment growth is not the same as investment performance.  Some of the growth or loss can be accounted for by additions and withdrawals as well.  Second, endowments are not a big pot of money the college can spend as it wishes.  Some percentage of the income from endowments is restricted to certain programs, and often carry additional expenses the college has to come up with on its own.

Still, I think this is interesting and compelling.  Let me know what you think.

Monday, January 15, 2018

National Trends in Applicants, Admits, and Enrolls, with Draw Rates

If you read this blog regularly, you'll know I'm interested in the concept of the Draw Rate, a figure seldom used in college admissions.  Many people, when thinking about market position in higher education use selectivity or admit rate (the percentage of applicants admitted), or yield rate (the percentage of students offered admission who enroll) by themselves.

But in the market of higher education, these two variables often fight against each other. (BTW, if you object to the use of the word "market" in higher education because you think it debases our profession, see what Zemsky, Wegner, and Massy have to say about that here.)

Colleges, driven by market expectations, have for a long time tried to increase applications, believing that what the market wants is greater selectivity in the institution they choose, based on the Groucho Marx effect. Except that in order to enroll the class you want, you have to take more students when apps go up (at least in the case of the bottom 90% of colleges).  That's because your incremental applications almost certainly have a lower propensity to enroll.

So, Draw Rate (yield rate/admit rate) helps account for that.  Higher Draw Rates are generally a sign of higher market position.  Think about it mathematically: A very high numerator (high yield) coupled with a very low denominator (low admit rate) is the thing many colleges pursue.  If you pursue greater selectivity and don't account for the lower yield, you won't be in enrollment management too long.

The problem, of course, is that, in general, people who were not born 18 years ago don't apply to college.  And the number of people who will turn 18 in any given year continues to drop going forward.  So no matter how many applications each student makes, they can only go to one college next fall.

Over the past several years, the "Winner Take All" mentality has driven demand at the most selective institutions.  The need to keep up trickles down to each tier below, and the annual "We received a record number of applications for this freshman class" shtick gets old fast, even if colleges have not gotten that message yet.

The take away: Colleges have been spinning their wheels, working harder and harder to generate more applications just to stay even.  The national psychosis weighs heavily on the minds of parents and students, and they respond by hedging their bets, applying to--guess what--more colleges.  And the spiral spirals out of control.

Here are five views (using tabs across the top) to show the data.

Dashboard 1 is a high level overview of applications, admits, and enrolls at four-year public, and four-year, not-for-profit institutions (open admission institutions do not report application activity to IPEDS).  You can use the control at top to show all institutions, or just public or private.  Top view is raw numbers; bottom is percent change.

Dashboard 3, the next tab, shows the same data on bar charts, with the draw rate as a brown line hovering over the bars.  Note how it's dropped over time: This is the effect of soft applications.  You can look at any region, or any single institution if you want, but the really interesting filter is at top right: Compare colleges by their 2016 selectivity.  You see that the only institutions who have collectively increased their draw rates are exactly the ones who had the strongest market position already: The most selective colleges.  Step down from Most to Highly to Very, etc, and watch the trend on the brown line.

Next comes Dashboard 2, showing Applications per Seat in the Freshman Class, and draw rate by region.  This might explain why we in the Midwest are fascinated with the obsession with college admissions by East and West Coast media.  Y'all are welcome to come to the Midwest and chill, if you'd like.  You can use the filter to select groups of colleges by Carnegie type.

Dashboard 4 shows four key metrics to reinforce the relationship between and among them.  Again, select by 2016 Selectivity to see how they make a difference.

Finally, Dashboard 5 allows you to compare individual institutions.  I've put Harvard, Stanford, and MIT on to start, but you can choose any colleges you wish.  (I recommend no more than three or four at a time.)  To remove a college, hover over its name in the filter and X it out.  To add, type any part of the name and hit "Enter" on your keyboard.  You'll be presented with all possible matches, and just choose the ones you want.  I recommend choosing similar institutions for scaling/charting purposes.

I hope this is interesting to you; let me know what you see, and if you spot any problems.

Wednesday, January 3, 2018

Freshman Migration, 2010-2016

This is perhaps the most popular, as well as my personal favorite, post, and I'm sad that I can only do it once every two years (as the IPEDS reporting cycle only requires this data be reported bi-annually.)
This shows patterns of freshman migration within and outside of state boundaries. It's valuable to people because you can see the composition of freshman classes at colleges: Where do the students come from? You can also see patterns of state exports: Which states keep students at home, and which send them out-of-state (of course, the size and educational offerings of the various states means it's often unfair to compare, but it's still interesting.)

For this, I've limited the universe to four-year, public and private, not-for-profit institutions. Community colleges and for-profit colleges tend to have very local enrollment patterns, and high numbers of part-time students. I've also taken out institutions whose primary focus is religious training, as well as those from a few obscure Carnegie categories.

The freshmen in this analysis are only those who graduated within twelve months of enrollment in college. A word of caution: If you are afraid to click buttons and interact, stop now. This won't be of any help to you. You can't break these, and you can always reset using the controls at lower right. So click around and explore the data.

Finally, this shows the data I downloaded. Some of it is pretty clearly wrong, but that's not my problem. Contact the IR office at the offending institution and ask them what they were thinking.

So, first up: If you want to compare any four colleges on the geographic composition of their freshman classes, start here. I've added four colleges that start with "D" but you can use the controls to look at any four you want. Note: Students labeled as "in-region" are from the region, but not the state. Therefore someone "in-region" in a New Hampshire college would be from one of the five other New England states. Got it? Good. Play away on this one:

Next up: Looking at the bar charts: It's a little more complex, but you can do it.  If you want to see which colleges enroll the most (top chart) or highest percentage (bottom chart) of students from in-state, in-region, or out-of region, this is your visualization. Choose a year (it defaults to 2016), and if you wish, limit it to colleges in a region (The Southeast, for instance).  You can limit to public or private as well.  Then choose which group of students you want to explore: In-state, in-region, or out-of-region.  Again, comparing Texas to Rhode Island should only be done for the "interestingness factor," not to draw conclusions.

Here is the same data, represented on a scatter plot, in case you want to step back, and see the data all at once.  The two scales are the number of freshmen, and the percent from the region selected.

Which states export the most students, and when they export them, where do those students end up?  If you've wondered that--or if you're from Illinois or New Jersey and lament our students' mobility--this is the visualization for you.

Choose a year, and see (on the top bars, in purpley-mauve) which states exported the most students.  Then, click on a bar representing a state to see where students from that state enrolled, in the bottom chart.  If you want the college destinations to be limited to public or private, or a certain region, you can use those controls to do so.

And finally, if you're interested in which states keep students at home, you can see that, too, on this visualization. The top view looks at colleges in a state, and where their students come from; the bottom looks at students from that state, and whether they go out-of-state or stay in-state.  Again, choose a year or institutional type, if you want to look at colleges or students going to those types of colleges.

I hope you have enjoyed looking at this data as much as I have enjoyed playing with it. If you spot any errors that I've made (Tableau still has no spell check....) let me know, and I'll get to fixing them right away. Otherwise, leave a comment below with questions or observations.

Wednesday, December 13, 2017

How Many Colleges are There in America?

Seems like an easy question: There are 7,284 post-secondary options in the US.

But everyone has a different definition of what they want when they ask for a count of colleges.  This should give you some clearer sense of the right answer for you.

At top left is "The Answer," and that will not change as you navigate through this.  But you can use the controls here to change the number of colleges and universities you're looking at, and to change how they're broken out.

Those controls change the number (in orange, at top) and the splits.

For instance, at the far right, on the control labeled "Region, choose "Great Lakes," and you'll see that there are 1,079.  On the gray box at top right, choose "State" and you'll see 354 in Ohio.  Under "Control of Institution" choose "Public" and you'll get 266.  And so on.  Now break out by "Campus Location" and see most are located in cities.

The reset button is at lower right.

I hope this is helpful to you as you wonder about the shape and size of American higher education.

Monday, December 11, 2017

What's All The Fuss About, Redux

My tireless crusade continues.

Everywhere you look, it seems most of the discussion and ink spent on higher education focuses on the most selective institutions in America.  In addition, if you listen to parents and students and counselors talk, you'll learn that there is a perception that college is increasingly hard to get into.

So, I broke the whole world of 1.403 four-year private, not-for-profit and public colleges and universities into bands, based on the absurd input measure of their freshman selectivity.  On the visualization below, they range from red (less than 15% admitted) to purple (over 60%) admitted.

Each institution falls into one of these boxes.

The four charts, clockwise from top left: The number of colleges in those categories, the number of freshmen they enroll, the total number of freshmen with a Pell grant, and the total undergraduate enrollment.

If you think you see a lot of purple, you do.  And this is before anyone enforces any sort of standard definition of what an "applicant" is.  Sometimes, it's just a person who accidentally clicks on an email link.

Of course, sometimes the scarcity of a good is exactly why people freak out about it. And of course, this doesn't even consider open admissions colleges (nine percent of all college enrollment in the US is in California's Community College System). So, this won't change the world, but I feel better for sharing.  Now you can't say you weren't told.

Friday, December 1, 2017

2016 IPEDS Admissions Data

Fresh from IPEDS, just months after the wrap up of the 2017 admissions cycle, comes the 2016 admissions data.

I've done something a little different this year to focus your attention, using five views of data, navigable via the tabs across the top of the visualization:

Admissions data (first tab) is pretty clear.  Colleges display admit rates (overall, in red) and then admit rates by gender (men are in blue; women are in orange).  If the blue bar extends beyond the orange, you can see that the admit rate for men is higher, and vice versa.

On the right are standardized test scores, showing calculated means.  In other words, since no one publishes averages and everyone wants them, I took the mid-point of the 25th and 75th percentiles to approximate the 50th percentile.  Note that IPEDS does not allow colleges that are test-optional to report test score information.  Also note that I've taken out a lot of colleges with extremely limited or suspect data.

As always, you can play with the filters (if there are any) to limit the colleges displayed, and you can sort columns by hovering until you see this little icon, and then click on it.

You can reset the view by clicking this little icon at lower right.

The four other views show a limited scope of colleges: Selective, wealthy, mostly men, and Land Grant institutions, and plotted them using some variables that should both answer and generate questions.

You be the judge. 

Tuesday, August 15, 2017

Chasing the Endowment Unicorn

Higher education is struggling these days, and there are a lot of solutions from a lot of pundits, all of which tend to be macro in nature: Delivery, cost structures, optimization, curricular adaptations, and many other ideas abound.

On the micro level., however, the vast majority of the 1,700 or so private, four-year colleges and universities will point to "increasing our endowment" as one of the most crucial solutions to our internal institutional challenges.

This is, in all probability, because the wealthiest institutions in the nation (in terms of endowment resources) are also the best known, and much of the brand of any institution is driven by wealth and reputation and prestige.  And even in this decade and these trying times, some of these institutions have parlayed considerable investment income into one-year operating surpluses of over a billion dollars. No, that's not a typo; it's a problem every university president would love to have. (Reminder to self: Update this chart.)

I once had a finance professor suggest that every institution should multiply the amount of money spent on Advancement each year by 20, then consider these options:

Let's say your Advancement Office budget is $8 million per year.  It would take an endowment increase of about $160 million to throw off that $8 million in cash each year forever (at 5%). Thus, shutting down the Advancement function completely would be the equivalent of raising $160 million in unrestricted endowment overnight. Unrestricted dollars are the hardest to raise, of course, because people don't tend to say, "Here's five million dollars; do with it whatever you want."

(It's also a good time to remind people that much endowment money is restricted; the $20 million gift from a big donor doesn't usually provide general operating relief but instead is used to fund some center or institute or faculty chair the donor thought was a good idea.  So in some sense, total value of the endowment can be occasionally misleading. It's still generally better to be bigger, though.)

Due to head starts and compounding, the wealthiest institutions are so far ahead of the rest of us that even trying to catch up seems futile.  Of course, that stops no one from relying on the old "tried and true."  In reality, our only hope of catching up with them would be a catastrophic market crash with no rebound; even then, we'd all be poor.  No solace there.

Take a look at the interactive visualization below.  Each bubble is an institution.  Hover over a bubble for details.
  • The SIZE of the bubble indicates endowment value at the end of FY 15 (probably June 30, 2015)
  • The COLOR of the bubble indicates tuition dependency (in IPEDS, "Percent of core revenues from tuition and fees.) Orange is low; blue is high.
  • The relative position on the y-axis (up and down) indicates one-year endowment value change (note: This is just subtraction, so it is not endowment performance).
  • The relative position on the x-axis (left and right) shows the one-year percentage change.  I cut it at 50% each way for clarity as there were a few extreme outliers.
If you'd like, you can use the filters at the top right to limit the types of institutions shown, or the range of endowment values.  Use the highlighter at the top left to highlight a specific institution.  Just start typing any part of the name to do so.

How do you feel now?

Saturday, July 8, 2017

Changes in College Attendance by State and Ethnicity, 2005-2015

Note: If you haven't read my post about the 2016 election results and educational attainment, it might be of interest to read that first.  Or later.  Or not at all. Your choice.

This started simply enough: A couple of tables from the Digest of Education Statistics, (tables 302.65 and 302.70) showing the percentage of adults aged 18-24 who were attending a degree-granting college by state and ethnicity in 2005 and 2015.  If you've read this blog enough, you know I have a love/hate relationship with the digest: Great data, but horrible formatting.  The tables are made to be printed on a single 8" x 11" sheet and handed out.  The crucial distinction between data and insight is lost.

Regardless, I reformatted the sheets into something workable for Tableau, and started to look at them. I wasn't having much luck: Some of the states didn't have data on African-American students, for instance, in 2005.  The variable for "Asian/Pacific Islander" was relatively new then, and only a few states had that data available.  Beyond that, I was looking to add some color-coding into the visualization to help make a point, and it wasn't going well.

But I've been fascinated since the election by some of the tweets and writing of Chris Arnade and Sarah Kendzior, who are thinking about what the election results mean in "flyover land."  And my blog post about the election results and attainment has stuck with me, mostly because of the reaction people had to it.

So I colored the states by the 2016 election results, and it got more interesting, as you can perhaps see below.

It's easy for us to look at things like this and chalk it up to "uneducated people voted for Trump." While that may technically be true, leaving it at that makes it too convenient for us in higher education to forget that educational attainment is only partially something you earn; it's also something you're born into.  Some of the ten charts on this post might make that clearer.

This can also, of course, be a post about urban and rural, divides. The division in our country might be as much about opportunity as it is about attainment.  If history tells us anything, it's that people start to rebel when they feel they don't have a chance via any other path.

So as we look at the current reality, the question, as always, remains: What are we doing to change the future?

Thursday, June 29, 2017

What Happens if Federal Money Goes Away?

Strategic planning at universities is always an important process, but it's even more crucial to do correctly these days.  And lots of institutions might be missing a really critical element in scanning the external environment: The extent to which federal financial aid programs contribute to the essential revenue streams that run the enterprise.

This is a fairly simple, if crowded, visualization, showing about 900 private colleges and universities who have good data in IPEDS.  Each dot is a bubble, colored by region, representing a single institution.  Its position along the horizontal axis shows student loans as percentage of core revenues, from left (low) to right (high).  I've included subsidized undergraduate direct loans, unsubsidized undergraduate direct loans, Parent PLUS loans, graduate subsidized, and Graduate PLUS loans in the calculation.  I did not include private loans.

Some of these numbers may seem high, but understand what this says and what it doesn't say: Loans go to pay other things (computers, gasoline, rent, books, food, etc.) so the colleges don't actually see all this money.  But presumably, the funding does make attendance and the paying of tuition possible.

And the IPEDS definition of Core Revenues can be confusing, too, as there are many revenue sources you might not consider.  This is what IPEDS puts into the category of Core Revenues:

  • Tuition and fees revenues (F2D01) 
  • Federal appropriations (F2D02) 
  • State appropriations (F2D03) 
  • Local appropriations (F2D04) 
  • Federal grants and contracts (F2D05) 
  • State grants and contracts (F2D06) 
  • Local grants and contracts (F2D07) 
  • Private gifts, grants, and contracts (F2D08) 
  • Contributions from affiliated entities (F2D09) 
  • Investment return (F2D10) 
  • Sales and services of educational activities (F2D11) 
  • Other revenues (F2D15) 

And if you have investment losses, your core revenues drop.  In other words, it can be misleading. And even if it doesn't, most places don't spend all of their investment returns, so while it shows up as a revenue, it is usually never touched.

Got it?

Second, on the y-axis, is Pell Grant revenue as a function of your Core Revenues.  Same idea as above, but using Pell as the numerator over Core Revenues.

Add these two together, and you'll see what happens to your revenue stream if federal aid goes away.

The bubbles are sized by tuition dependence; the calculation is not standardized, so for the sake of simplicity, I looked just at tuition revenue as a percentage of tuition plus investment income.

If you want to show a single or a group of institutions in context, use the filter.  Just type part of the name and select it.  If you want to look at fewer institutions, choose a region, a state, or reduce the range of core revenues (for instance, type $100,000,000 in the left hand box of the filter, or use the slider, to eliminate very small institutions.)

As always, hover over a bubble for details.

You'll notice some interesting things, I hope.  Mostly, I hope this doesn't frighten you.  Depending on where you work, it can be a bit daunting.

Friday, June 16, 2017

The Discount Dilemma

"You should write something about discount rates."

I hear that a lot these days.  Even though NACUBO does its annual discount study, people still want and crave more.  There is no topic, it seems, as much on the minds of people in universities as discount rate.

But despite my desire to make you, the loyal readers of this blog, happy, there are a lot of reasons I haven't written about discount rates:

  • First, data are old.  It's a long story, but financial reporting (where you learn about financial aid) is reported about a year after the freshman class enrolls.  So the ability to calculate discount is always behind the most current admissions data.  The viz below is for 2014 freshmen for instance, and it's the most recent publicly available.  It's hard to describe to people how much things have changed between 2014 and 2017. (And even harder to figure out why 2016 admissions data are not out yet.)
  • Second, discount is not as important as accountants think it is.  "WHAT?!?" say the accountants! I politely suggest that what you really care about if you're running an institution is net revenue per student, and total net revenue, the cash you use to run the university.  
  • Put it this way: If your tuition is $50,000 and you have a 40% discount rate, you net $30,000 cash per student.  But you're generating less net revenue than your competitor who charges $55,000 in tuition with a 43% discount rate, who nets $31,350 per student.  Still, people insist on comparing disparate institutions on this single accounting measure. Lower discount is not always better.
  • In that same vein, average and total revenue are both important.  If you took only full-pay students in the scenario above, you would average $50,000 with a zero percent discount.  But your numbers of enrolling students would go way down, as would your total; you wouldn't have enough to cover your overhead.  Inversely, you can generate more total revenue by discounting more and enrolling more,  but your costs go up faster than revenue, which is of course not good.
  • Third, most people and stories focus on freshman discount.  If you're a small, tuition-driven, liberal-arts college, that might be meaningful, as freshman may make up 25% or 35% of your revenue. But it's less important at big, complex universities, where we have freshmen, transfer, law, medical, graduate, and other types of enrollment too.  Overall discount is far more important, but usually less discussed.  At some institutions, where educating undergraduates can almost be said to be a side business, the revenue from student tuition is tiny, dwarfed by things like research dollars and endowment return.  Large discounts on tiny fractions don't add up to much.
  • Fourth, net revenue is not the cost to the student.  Pell and state grants come to the university as cash, and look the same as money from a student's pocket, even though they are very different to the student.  You can't tell how much a college costs a student by looking at this data.
  • Fifth, not all discount rates are the same, even at similar institutions with similar tuition rates. You can get to a 50% discount rate by having the class half full-pay and half-full need that you meet; you can also get there by having everyone at 50%.  These are impossible examples, of course, but you get the picture.
  • Finally, discount rate--an accounting measure--is something we used to look at only after the fact, where three decimal points are very satisfying.  It's essentially impossible to manage to that rate in a fast-changing, dynamic environment unless you constrain other outcomes.  As I've said before, when we send out aid awards, we're not planting saplings; we're casting seeds.  You can predict with some precision what percentage of which seeds will take root and grow, but you can't control the wind, the rain, or the temperature, which are all critical to success.
As you look at this, there are a couple of things to consider in addition to the usual caveat about the accuracy of IPEDS.

I calculated discount by taking the Financial Aid Cohort of freshmen (first-time, full-time, degree-seeking) students and multiplying that number by the tuition, then using the total institutional aid.  At most places, institutional aid is not funded, it's just a contra revenue.  But at others, income from restricted endowment funds may actually fund the aid.  Here's a long boring argument about whether this matters or not. For this purpose, I'm taking all aid as unfunded discount.  Thus, discount = Institutional Aid/Gross Tuition Revenue.

Some numbers seem a little crazy.  While a lot of institutions have freshman cohort numbers lower than total freshman enrollment (which makes sense) some have freshman total numbers lower than the aid cohort, which suggests someone in the IR office is counting wrong.  Not my problem.

This does not include all 2,200 four-year, degree-granting private colleges and universities.

Discount only makes sense for private universities, for one thing, which takes us down to about 1,600.  Several hundred of them don't accept freshmen, others have incomplete data, some are rabbinical institutes or schools of theology or massage therapy that aren't of great interest to many. Others have tiny freshman classes.  Thus, this view starts with 675 institutions, and shows five things. The four columns, left-to-right show Calculated freshman discount, Average Net Revenue per Freshman, Endowment Per Freshman, and Percentage of Freshmen with Pell.  The color shows mean SAT CR+M scores, approximated using the 25th and 75th percentiles; gold is low, and purple is high.

Use the filters on the right to limit the view.  For instance, you might only want to look at colleges in your state, or with similar sized freshman classes, or test score averages.  Or perhaps you only care about Doctoral institutions.  Filter to your heart's content.  If you want to sort the columns, just hover over the bottom of the column until you see the small icon appear, like this:

Click through the cycles to sort descending, ascending, or alphabetically.  Click on undo, redo, and reset to, well, you get the picture.

Let me know what you think and what you see here.  That is, if you've gotten this far and I haven't whomped the enthusiasm right out of you.

Tuesday, April 4, 2017

Undergraduate enrollments by ethnicity, 2015

I was doing some research for our own internal discussions, and decided to take it a few more steps to look at enrollment of undergraduate students by ethnicity at about 2,000 four-year, public and private institutions in the US.  (And when you look at the data and wonder why, rest assured, I checked: Miami Dade does offer Bachelor's degrees via online programs.)

It's here, and the first two views are pretty easy to navigate:  Each chart shows a separate ethnicity and lists each institution in descending order.  The first view is by counts, and the second by percentages.  Thus you can see the institution that enrolls either the most Hispanic students, for instance, or the institution with the greatest percentage of Hispanic students, depending on your preference.

If you'd like to focus on a single state, just public or private, or colleges of a certain size range, use the filters at the top.  You can always reset the views using the control at the bottom.

The third view allows more customization.  Each point represents an institution, arrayed on the x- and y-axis.  But you can control what values the axes show: For instance, percentage White on the x-axis, and percentage Asian/Pacific Islander on the y-axis.  The points are colored by control: Orange is for private institutions, purple for public. Again, you can limit by undergraduate enrollment or by state, if you'd like.  But this view has the advantage of choosing a highlight institution: Use the highlight box to put a university of interest in context.  Type part of the name, and select it, and it will show up all by itself.

I hope this is helpful for use with students who are interested in thinking about and comparing colleges and universities by enrollment profile.  And if you're interested in seeing how an ecologist might look at enrollment diversity, check out this piece I wrote for Academic Impressions last fall.

Friday, March 10, 2017

Is This Why Democrats Support Education Funding?

This post started off simply enough: I found some cool data on changes in educational attainment over time. I was going to take a look at how far we'd come as a nation in the last 40 years (even though I had already published this), and show where the biggest gains were.

It wasn't very compelling, at least at first.

Then, I decided to get ambitious (my wife was in her evening class, so I had the night free), and wondered if there were any interesting connections between changes in educational attainment and voting patterns in 2016.  I found a data set with election results by county, (a handful of counties are missing) and merged it in. And thus, this.

Before I start, there are a few points to make about the data.  First, the definitions changed slightly over time.  For instance, in 1970, the field is labeled "College degree," while in 2010, it's labeled "Four or more years of college."  Not the same thing, but we'll have to go with it for now.  Also, 2010 is not really 2010; it's the data from the five year American Community Survey of the Census Bureau, but there's no reason to believe it's not as accurate as the census itself.  In fact, the ACS is used to test the accuracy of the census.

So, onward.

Using the tabs across the top of the visualization:

First, a scattergram, plotting attainment in 1970 and 2010.  The regression line suggests that attainment has essentially doubled in 40 years; those bubbles (counties) above it have done better; those below worse.  Bubbles are sized by votes in the county in the 2016 presidential election.  If you want to look at just one candidate, use the highlighter function.

Note that the counties that went for Clinton tend to be larger (more urban), with higher levels of attainment (moving toward the top right) and more above the line.  Counties that went for Trump tend to be the opposite.  And of course, there are many exceptions (Johnson County, Kansas; Apache County, Arizona), and clearly the binary blue/red can be misleading; some victories are by a point, some by 15 points or more.  Finally, it's almost certain that much of the change in attainment is due to people moving in and out; not everyone lives where they were born. But it's still interesting.

Second, the bar chart (I know from experience that many people won't click on the second tab. Please. You will be glad you did).  The x-axis is broken into college degree attainment. For instance, the long bars in the center show counties where 30--34.99% of adults have a college degree.  You can see how many votes these counties cast for Trump, and how many for Clinton.  I double checked this; it is perhaps the best story I've ever told with one chart.  And although the left end changes as you select single states (using the filter at the top), the right end is fairly stable.

Finally, the last chart just shows three variables: 1970 and 2010 college-degree attainment, and the change over time.  See the box to the right of the chart if you want to sort the data.

Admittedly, this election was a strange one, so perhaps there are no lessons to be learned.  But over the past few decades, Republicans have been fairly staunch opponents of increased educational funding, and you have to wonder if this doesn't explain why; people who lived in areas with higher levels of education voted for Democrats in the last election.

Fifty years ago, the Republicans were the party of the college-educated, white collar classes; the Democrats the blue collar, working-class, high school educated citizens.  That's all changed, if 2016 is any indication.

Agree? Disagree? Let me know in the comments below.

Friday, March 3, 2017

2016 Freshman Admissions Data

Note: I've just discovered that although this data set is labeled 2016, it is for the 2015-2016 data year; thus, this is Fall, 2015 admissions data, not Fall 2016 as I had thought.)

This always proves to be a popular post: The 2016 Admissions Data summary.

Here you'll find ten views, showing test scores, admit rates, need data, and international student information (which should only be used as a guide, as you'll see.)

Use the gray boxes and/or arrows across the top to navigate this information, and the filters to limit the views.

Note: This data is "as reported" to Peterson's and is presented as is.  If it's wrong or your college is missing, it's almost certainly a reporting error; most institutions left at least some fields blank.

It comes from the Peterson's Undergraduate database and the Peterson's Undergraduate Financial Aid database, both copyright 2016 by Peterson's-Nelnet.  The data here are used with permission of Peterson's.

Wednesday, February 1, 2017

Welcome to the Hen House, Mr. Fox

Jerry Falwell, Jr., President of Liberty University, has just indicated that he will head a new task force to examine the Department of Education's policies on colleges and universities: Things like, "overreaching regulation” and micromanagement by the department in areas like accreditation and policies that affect colleges’ student-recruiting behavior, like the new “borrower defense to repayment” regulations," according to this report in the Chronicle of Higher Education (subscription required, but if you're in higher ed and you're not subscribing, you should. I've not been paid to endorse the Chronicle. Or anything, for that matter.)

Thousands of people on Twitter and in other social media have already pointed out why this is a bad idea, especially coming on what appears to be the approval of Besty De Vos as Secretary of Education.  But if you want to see how far Liberty has taken recruitment, even under the current regulations that attempt to make admissions and recruiting more ethical, I challenge you to fill out the inquiry form on their website.  Then count to ten, and I'm guessing you'll soon be connected to one of their friendly, helpful, sales agents who will tell you how you, too, can become a champion for Christ, and how you can use federal dollars to do so.

Here is how much federal aid Liberty students (and those from every other university receiving your tax dollars) received in 2015-16 (a few of the programs are 2014-2015, but they don't change much.) Interact to your heart's content.

And feel free to share this with your elected officials.

Tuesday, January 31, 2017

What kind of jobs can English Majors Get?

What kind of jobs can I get if I major in English? (Lots) Do I have to major in science to go to medical school? (No) Do actors have to go to a Theater program? (No).

All these sound like conventional wisdom, but now, thanks to my friends at Human Capital Research Corporation, we have some better answers.  The data set they put together is based on The American Community Survey (ACS) of the Census Bureau, a small but statistically significant sample of the US Population.  It asks questions that include occupation and college major (for those who are working, and for those who have a bachelor's degree).  The data below contains over 3 million individual responses to these questions (for people in the labor force between the ages of 25 and 60 with a bachelor's degree).

One the first dashboard (using the tabs across the top), you see two views.  On the blue chart on the left, choose a major (cluster) at the top.  The chart below will show you the professions (also clustered) of people with a bachelor's degree in that area.  Hover over a square for details, including the number and the percentage of the total.  Multiply by about 20 to convert the sample to the total.

One the red chart, choose the profession, and see the majors of the people working in that area.

Most engineers majored in engineering; most nurses in nursing, most teachers in education, and most accountants in business.  But beyond that, you get a rich sense of the wide range of careers open to people with almost any degree.

On the second tab, look at the majors on the left, and see how people are distributed by going across the row. Look for larger, blue bubbles to see clusters: 37% of people with a degree in library science, for instance, work as a librarian; 29% of architecture majors are architects.  The rows total 100%. Unfortunately, the number of professions makes labeling the professions impossible, except in the box that pops up when you hover.

Then, on the third tab, the view is the same, but the columns total 100%.  So you see the majors of people in professions.

On the last two views, the story is not the large bubbles, I think, although the add to understanding; the story is the small bubbles: People from all majors doing all jobs.

And a word of caution, of course: I defaulted the first two views to biology and medicine, and the tendency will be to conclude that you must be a science major to go to medical school.  In fact, this is likely driven by the fact that the vast majority of applicants to medical school major in the sciences.

What else do you see here? What surprised you?  Let me know in the comments below.

Wednesday, January 11, 2017

NY City Public Schools, and what they might tell us about the SAT

Recently, I received a message from Akil Bello who pointed out a data visualization he had seen.  It was originally posted to Reddit, but later was edited to eliminate the red-green barrier that people with color-blindness face.  The story was here, using a more suitable blue-red scheme.

There's nothing really wrong with visualizing test scores, of course.  I do it all the time.  But many of the comments on Reddit suggest that somehow the tests have real meaning, as a single variable devoid of any context.  I don't think that's a good way to analyze data.

So I went to the NY City Department of Education to see what I can find.  There is a lot of good stuff there, so I pulled some of it down and began taking a look at it.  Here's what I found.

On the first chart, I wanted to see if the SAT could be described as an outcome of other variables, so I put the average SAT score on the y-axis, and began with a simple measure: Eighth grade math and English scores on the x-axis. Hover over the regression line, and you'll see an r-squared of about .90.

Scientists would use the term "winner, winner, chicken dinner" when getting results like this.  It means, for all intents and purposes, that if you know a high school's mean 8th grade achievement scores, you can predict their SAT scores four years later with amazing accuracy.  And--here's the interesting thing--the equation holds for virtually every single school.  There are few outliers.

Ponder that.

But critics of the SAT also say that the scores are reflective of other things, too; an accumulation of social capital, for instance.  So use the control at the bottom to change the value on the x-axis.  Try economic need index, or percentage of students in temporary housing, or percentage of the student body that are White or Asian. The line may go up (positive correlation) or down (negative) but you'll always see the schools with the highest scores tend to have the characteristics you'd expect.

Jump to the second tab.  This is more a response to the Reddit post: The top map shows the ZIP codes and a bubble, indicating the number of schools in that ZIP.  The bottom map shows every school arrayed on two poverty scales: Economic Index and Percent in Temporary Housing.  The color shows the mean SAT score in the school (Critical Reading plus Math, on a 1600-point scale.)  Purple dots represent higher scores.

Use the ZIP highlighter, and you'll see the top map show only that bubble, and the bottom will show the schools in it.

Got the lesson?  Good.  Now, think about why the colleges with high median test scores a) have them, and b) tend to produce students with high GRE and MCAT and LSAT scores,  and c) point to excellent outcomes for their students.

And let me know what you think.