Monday, July 28, 2014

Where Students in the US were enrolled, 2012

A while ago, I wrote about the wide range of institutions in the US post-secondary education system.  The point I was trying to make was that the prestigious, extraordinarily selective institutions you hear about all the time represent just the top of iceberg when it comes to colleges and universities in the US.

That visualization counted institutions, for the most part.

But another way to look at this is to show where students enroll, and I think this can be equally surprising, and, I hope a bit enlightening.  So take a look at this, again using Tableau Story Boards.  Each gray tab across the top shows the data presented in a different way; keep clicking from left to right to see the interesting tidbits this data reveals to us.

Did you know, for instance, that one out of every eleven undergraduate college students in the US (excluding for-profits) enrolls at a California Community College?  Or that there are no private, not-for-profit colleges in Wyoming? Or that the state with the highest percentage of students enrolled in four-year public universities is...well, you'll have to find that for yourself.

Interact:





Tuesday, July 22, 2014

Changes in Faculty Salaries Over Time

Note: The first version of this had a bug; I forgot to add the "Gender" filter to the final view, so everything was showing up at about 3 times the actual value.  It's fixed now, with a thanks to Tableau Zen Master Allan Walker of Utah State University for catching it. While I was at it, I added a second view to show differences by gender.

The last time I wrote about the salaries of educators, I said I'd never do it again: There were too many people who didn't know the difference between nominal and constant dollars ("No one made $75,000 in 1980!") and those who didn't understand averages ("How could the average in that state be $60,000? I only make $53,000!").

But this is interesting, I think in light of discussions about the rapid increase in tuition over time (in case you've been under a rock recently.)  It shows the changes in average (there's that word again) faculty salaries by rank since 1975, in either nominal or constant (inflation adjusted) dollars.

Compare these changes to this and see what you think:



Tuesday, July 15, 2014

A DataViz Reboot: WICHE Projections of High School Graduates

A while ago, I used Tableau Software to visualize trends in High School Graduates provided by WICHE.  I think it was good, but with Tableau's new Story Points feature (in which you create pages of the story you want to tell) I think it's an even better story.  If you scroll through these points, you can get a sense of how America has become more diverse, and how those changes vary pretty dramatically by region.  That last point, especially, is often lost on people who talk about changing national demographics.

Just like all politics is local, almost all enrollment is too.

So, first, if you want, look at the old version, then take a look at the new visualization below.  What do you think?



Saturday, July 12, 2014

Where do International Students Enroll in the US?

In recent years, Colleges and Universities have turned their focus to International Enrollment as a source of new students. But is that a good idea?  It depends on what type of institution you are, apparently.

This data comes from the Institute of International Education's Open Doors project, and while it's valuable, it still points out the problems with pre-aggregated data.  On the site, you'll find good stuff about students by enrollment level (graduate and undergraduate); you'll find good information about enrollment by institution; and you'll find information like this about enrollment by Carnegie Classification.  But you can only ask one question of each data set.  The result is that this information is intriguing, but not granular enough: For instance, what if I presumed that graduate students would naturally flock to doctoral and research institutions and a) I wanted to test that theory or b)I  just wanted to look at undergraduates to see where they went?

Still. have a look, and especially use the year filters to see percent change over long and short periods.



Wednesday, June 25, 2014

US Post-secondary offerings

If you only read the papers, you'd think US Higher Education consisted of a dozen or so high profile institutions.  But fortunately, there are "more things in heaven and in earth than are dream't of in their philosophy," with all appropriate apologies to Shakespeare.

When I started this blog, it was in response to a new Tableau Software feature I had seen pre-viewed last September, called "Story Points."  In fact, the very title of the blog has a lot to do with that: Believing that data can and should be used to tell narratives that provide people with memorable insight.

This is my first attempt to use Story Points to tell a story; one I hope sticks with people as we think about a pretty amazing selection of post-secondary options for students.  To navigate the story points, just use the grey boxes along the top, and a new chart or dashboard should point the way to insight.


Monday, June 23, 2014

Predicting EFC with One Question

Most everyone who knows anything about our Financial Aid system thinks it needs some improvement.  And almost everyone who actually goes through it, it seems, is astonished by the outcome: They expect me to pay how much for college? And that's just for one year?

For those of you who don't know, all federal financial aid begins with the FAFSA, or Free Application for Federal Student Aid.  It's a form that collects information about income, assets, and family size in an attempt to estimate how much a family should be able to contribute to the cost of higher education.  Should being the operative word.  The figure it calculates--EFC, or Expected Family Contribution--is really a misnomer, sort of like the Peacekeeper Missile.  It's really just an index number designed to estimate federal expenditures on financial aid programs.  Many colleges find it so unreliable that they use another form, such as the College Board's Profile, or their own proprietary form to collect more information.

Many people believe the complexity of these forms causes many low-income students, especially those from immigrant families, to opt out of college.  And a new bill has been introduced in Congress to simply the process, by possibly asking only two questions.

Debate is underway, and you can search it to read about it, but I began to wonder: What if we asked just one question: Parental Adjusted Gross Income.  What is the current relationship between AGI and EFC?

I took about 60,000 records from seven years of freshmen applicant to DePaul data, and plotted them (after I made them anonymous and created a fake ID number).  Then I created a polynomial trend line to fit the data, and added confidence intervals to the lines (well, the software did, anyway.)  Note that only about half of the students who apply for admission fill out a FAFSA.

This is messy, of course, so do this:


  • Try choosing a single year.  Or, if you want to show all years, click on a year in the color legend to highlight that year and gray out the others.
  • Use the filter to limit the range of incomes (it's currently set between $0 and $250,000, since negative incomes are associated with business owners, and almost everyone over $250,000 doesn't qualify for any need-based aid), or include only Pell Grant Recipients. 
  • I don't recommend going below zero.


What do you think?  Could we simplify the process?



Wednesday, June 18, 2014

Bachelor's Degrees by Program and Ethnicity, 2010 and 2011

The previous post, about Doctoral Degrees by Program and Ethnicity, generated a followup question from Jennielle Strother at Seminary of the Southwest about similar data for undergraduate enrollment.  While I couldn't find that exact data, I did find this from the Digest of Education Statistics, showing degrees awarded by race and program, so I spent a half hour to pull it into a visualization.

Some data visualization experts don't like tree maps because it's hard to make precise comparisons of area across distance, but I do like it for this purpose: You can pretty easily see the data in one view with minimal effort, and since precise comparisons aren't vital, you can get a good sense of the lay of the land.

It's also very easy to ask your questions of this chart.  For instance, if you want to see how degrees shook out within a program (like engineering, or English)  you can quickly make those selections and see the results by ethnicity.  If you want to exclude non-resident students, for your analysis, you can. If you want to see what Hispanic students majored in, you can look at all programs but select just"Hispanic" on the ethnicity filter.  Just make your selections in any combination and click "apply" on the filters.  And choose which year you want to look at.





Tuesday, June 17, 2014

Doctorates by Discipline and Ethnicity

A recent article in Inside Higher Education touched on a subject I've written a lot about on my other blog (the one with more words than pictures), specifically the role of standardized tests, in this case the GRE in selection of students for graduate programs.  The article cites another article in Nature blaming the dearth of minority and women doctoral graduates in science and engineering, at least in part, on the GRE.

For anyone who is at least knee-deep in the debate about the value of standardized tests, the arguments are familiar ones: Too much emphasis on the tests means that too many candidates with strong potential are being overlooked, especially when you consider the predictive validity of the tests.  The authors are pretty blunt: " The GRE is a better indicator of sex and skin colour than of ability and ultimate success."

So, in light of that, take a look at this data on 2012 Ph.D. recipients, which was downloaded from the NSF Survey of Earned Doctorates.  The patterns are obvious: The two groups who score the highest on the ACT, the SAT, and the GRE--Asians and Caucasians--dominate the newly minted Ph.D. classes.

And if you assume that there is a need to get minority canididates into faculty roles at colleges and universities to effect change in this area, you see even more cause for concern.  For instance, take a look at the production of African-American Ph.D.s in areas like math or computer science.  When you consider that there are about 2500 public and private degree-granting institutions in this country, you see how unlikely the chances are that things will turn around quickly.


Sunday, June 8, 2014

Yes Education Pays. But maybe not how you think.

You have probably seen the headlines: College graduates make $800,000 more in the course of a lifetime than high school graduates do.  It's statistically correct.  And the conclusion it probably leads many people to is completely wrong.

It's true, of course, as you'll see below, that income increases with every increment of education: A high school graduate earns more than someone who didn't graduate from high school; a person with a bachelor's degree earns more than someone with a high school diploma; and someone with a master's degree earns more than someone with just a bachelor's.  (This is not true for every person, of course, just for groups on average; Bill Gates, whom I'm pretty sure earns more than yours truly with a Master's Degree, never finished college.)

But it's wrong to say that graduating from college is the cause of the income difference. It's true that earning a degree opens new doors to you, and new opportunities for income, but some of that can be explained by the fact that the best students in high school--the ones who are probably likely to earn the most later in life--are also the same ones who go to college.  In other words, the same factors that get you into college are the ones that increase your chances for success.

Still, this visualization tells an interesting story, in four views, via the tabs across the top: On the first view, you can choose your own comparison: Use the filters at the top to select the values for the blue bars and the orange dots, to see the gap between any two education levels.  You can choose men or women, and you can pick inflation-adjusted dollars or nominal dollars.

The second view, Income by Attainment, you see the whole world laid our for you.  Pretty simple, and again, you have some choices to make.

The third view is where it gets really interesting, and where you see the disparity between men and women. This (and all the views) show workers over 25, so some older women may be skewing this, but it's still telling.  And shameful.

Finally, the last view shows how much each step up gains you in income over the last one.  A couple points are annotated for demonstration purposes.

As always, don't just look at these: Interact and explore the data.  And let me know what you find.



Tuesday, June 3, 2014

MOOCs: What Harvard and MIT Data Reveal

If you know much about higher education, you know that Massive Open Online Courses (known as MOOCs) are all the rage.  These courses are open to anyone, anywhere, for free, and promise opportunity for students who wish to learn on their own.  They are exciting in concept, and threaten to turn higher education on its head.

So the recent release of HarvardX and MITX data on MOOCs is exciting.  The data is scrubbed to protect the privacy of the students who took the courses, but still yields a wealth of interesting stuff.  But you must interact.

This dashboard starts with the intro, but has five views you can see by clicking across the tabs at the top. Once on a view, you can limit the data shown by (depending on the dashboard), by gender, education levels of students, home country, or institution, and whether the student registered or completed the course for a grade.  Data are shown by course, age, home country, and institution.

Have fun, and let me know what jumps out at you.



Friday, May 30, 2014

The Diversity of American Higher Education

This is not a post about the diversity of American college students, but rather about the diversity of institutions.

Recently, on a list discussion I subscribe to, someone made a point about non-selective or open institutions in the US: That there were many more of them than you might think.  So that got me thinking about our collective fascination with the most selective colleges and universities.  But even that visualization only looked at four-year, degree granting, freshman-enrolling institutions.  And of course, despite what popular culture might suggest, there are "many things in heaven and on earth that are not dreamt of in your philosophy."

So, this: Just about all of the over 7,000 post-secondary educational institutions in the US, shown a couple of ways: Control (Public, Private not-for-profit, and Private, for-profit: Open- and non-open admissions policies; and bachelor's-awarding or not.

Use the filters at right to select a subset, and you'll see how many institutions and how many freshmen we're talking about; the bottom chart shows by state, in case you're interested in that.

Surprised? You're almost certainly not alone.



Thursday, May 29, 2014

Miles to go before we sleep...

President Obama has framed a very specific goal for the United States: To increase the proportion of 25-to-34-year-olds who hold an associate degree or higher to 55 percent by the year 2025 in order to make America the leader in educational attainment in the world.

It's also very ambitious: Just see for yourself.  This is data from the US Education Dashboard showing where we are as of a few years ago.  It uses PUMS Data of the American Community Survey, and averages the results over three years to reduce the standard error of measure.

You can see how far we've come in three years by using the filter at the top right to toggle between 2005-2007 and 2007-2009.  These were, I'd remind you, very good years for higher education in the US.

More important, you can see how far we have to go; the reference line is set to 55%.  And as a bonus, of course, you can compare states and different ethnic groups.  US Totals are indicated by the blue bar; states are in grey.

Where would you put the odds of accomplishing this task?


Tuesday, May 20, 2014

Some More on Tuition and Fees

There is a lot of data out there on college costs, but the story is not so easily seen.

That's why I like this data set from the 2013 Digest of Education Statistics, Table 330.30. It breaks tuition charges into percentiles by college or university control and further into two-year and four-year. Use the filters (in the blue boxes, at the top) to select Tuition and Fees only, or Tuition, Fees, Room, and Board (on the left); and institutional control and type (on the right). The top chart shows the charges, in nominal (not-adjusted for inflation) dollars. For your reference, inflation over the time period of 2000 to 2012 was 33 percent.

You can see the labels at the end of the lines: Purple is 90th percentile, for instance, and the orange line shows the 25th.

But toggling through the different types of colleges and universities can be less than insightful, other than seeing they've all increased pretty dramatically over time.  Look instead at the bottom chart, which shows percent change over the same time period.  What does this suggest to you?


Monday, May 19, 2014

We're Number...Well, actually, we're pretty average

In some sense, using weighted averages to compare the US to the average value of any set of countries (especially when you don't include Russia, India, or China) is tautological; we're so big, we pull the average toward us by our mass.  Still, there is some interesting stuff here.

This visualization shows the number per 100 of students who are at the typical bachelor's degree-earning age who actually received a degree in the year shown.  In other words, if the typical age to receive a BA in the tiny country of Karpathia is 23, and there were 100 students aged 23 in 2009, how many of them earned a degree?

The visualization is sorted by 2011, but you can look at any year by hovering just to the right of the year label at the top of any column, and clicking the little icon that appears.  Clicking again will sort in reverse order, and clicking again will sort alphabetically.

It also starts off showing the "Total" value, but you can choose just women or just men; I encourage you to do so, as even more interesting details emerge.

Note: Finland has some funky numbers due to a government incentive that caused many students to accelerate a year, so they spike way up in 2008 and way down (for them, at least) in 2009.

So next time you want to say the US is the best, please be sure to make sure you're not talking about degree attainment.


Friday, May 16, 2014

Student Loans in Detail

Note: Functionality restored.

A few posts back, I wrote about Enrollment by Institutional Type.  I've also taken a stab at Student Loans before, but find the federal data very hard to work with, as the types of loans rolled up into different categories is not consistent over time, thus leading to just flat-out-wrong conclusions about what's happening over time (note: In case it's not clear, that link leads to a visualization that proves how important it is to know your data; I didn't think those numbers could be right, and it turns out they weren't, but it was only apparent when I took a look at graduate loans, which were rolled up with undergraduate loans one year, but not the other.)

Anyway, it's interesting to take a look at the world of federal student loans: Who gets them, the balance between and among the different programs, and how different institutions benefit from them.

This visualization shows both macro- and micro- student loan data.  On the top two charts, you can see the whole universe: Each circle represents an institution, sized by the total volume of loans selected.  On the left, they're colored by School Type; on the right, by Loan Type.  At the bottom, you'll find a bar chart, showing the volume by school, and the bars are colored by institutional type.

Use the filters in the middle to limit your selection, to see only the type(s) of schools or the type(s) of loans you're interested in, or to view institutions from a state or group of states.  All three views update.

What you'll see here is interesting, to say the least.  The first thing to jump out at you is The University of Phoenix, but with about 400,000 students, that's to be expected.  Even if you control for size, you'll see Phoenix has a high percentage of its budget generated from student loans, and although it's easy to compare them to Princeton, for instance, on a per capita basis, you also have to remember that Phoenix enrolls a lot more low-income students (in both number and percentage) than their distant cousin in New Jersey.

What do you see?  What would you like to see?



Wednesday, May 7, 2014

Pricing Public Education in the States

Everyone is--or perhaps should be--interested in how the state you live in thinks about public higher education pricing.  Even if you don't have a student in your family enrolled in a state institution, you probably pay taxes to support higher education; even if you don't pay taxes, you're probably interested in how your state develops educated citizenry who someday will.

So it's very interesting to look at this data from the Digest of Education Statistics several different ways.  In order to get the most out of this visualization, you must interact, however, and there are only two ways to do so.  Both are very easy.

First, use this control in the middle of the page:


I promise you, you won't break anything.  With this control, you decide what value to show on the bars and to use to color the map (orange is low; purple is high).  It starts off with 2012 Resident Tuition and Fees for Public, four-year universities.  But choose anything: Non-resident tuition, for instance; Resident tuition at two-year colleges; average private tuition in the state; the non-resident premium (that is, how much more does a non-resident pay); the non-resident premium percent (how many times base tuition is the "up-charge" for non-residents; the 2-to-4 year premium (the difference between average two year publics and four-year publics in the state for residents): the 2-to-4 year premium percent (that value again, expressed as an up-charge to move up from a community college to a state university); the in-state private premium (if a student stays in state, how much more does he or she pay for a private university, on average); or out-of-state premium, for students who want to leave the home state but would consider either a public or private university in that destination state.)

Second, you can limit the view to just certain regions, which makes the New England states, for instance, easier to see.  Use this control


to make those selections to your heart's content.  Both controls work on the map and the bar chart underneath it.  The bottom has a scroll bar on it, and if you show the entire US, you can see the US Average, in blue.

This is at once astonishingly simple and very rich in details.

Some caveats: These are averages, so Michigan does not show Michigan State or the University of Michigan tuition: It shows the average of publics in Michigan.  Things are a little less clear when you compare average publics to average privates, of course, but it's still interesting. Second, these data show sticker price, not the net price.  Even flagship public universities are heavily discounting non-resident tuition, so your results, as they say, may vary.

Start interacting.




Tuesday, April 29, 2014

The Growth of the For-profit Institution in Higher Education

There is a story in this data, but you can't see it yet.  You'll have to click your mouse to see it.

But let's start with what is there: Total enrollment in degree-granting, public and private not-for-profit colleges and universities in the US from 1995 to 2012.  What you see is mostly stability, if you assume the whole world of higher education is made up of these two sectors only: Some growth in each in the top chart (the population has grown, of course, over that time); a relatively stable distribution of private and public enrollments (where private colleges have always enrolled about 20% of the students); and percentage changes since 1995 that are virtually identical. (Hover over the line at any point to get the details).

Now, for the work: In the filter in the right, check the box labeled Private Not-for-profit, to add that population to the mix.  The blue line appears, and you begin to see its effect on the other two sectors: By 2012, the for-profit sector enrolled over 11% of all students.  Although the percent of total had fallen a bit by 2012, it still represented over 10% of all students enrolled.

What's more astonishing is the rate of growth: From 240,000 in 1995 to 1.8 million students in 2012, a 650% increase.

Before I publish the next chapter, what percentage of federal grant and loan aid do you think goes to these three types?



Tuesday, April 15, 2014

Freshman Wanderlust

When freshmen students go to college outside their home state, where do they go? It's a question with lots of answers, and the insight is not always easy to figure out, let alone communicate. But I took a stab at it anyway.

There are three views here, using the tabs across the top: If you want to know where students from a particular state enroll out-of-state, you should use the default view: When Freshmen Cross State Lines, Where Do They Go?  Pick any freshman home state (the view shows Michigan to start). You can also limit the colleges displayed by filtering on college region or Carnegie Classification. You can see that 372 freshmen left Michigan to go to The University of Toledo in 2012; 117 went to my institution, DePaul. Note that IPEDS data sometimes has mistakes (choose Arkansas, and you'll see one jump right out at you.** See note below for an update.) But overall, this data looks pretty clean.

If you want to see which colleges enrolled students from specific regions, use the second tab. Again, limit your selections as you wish by using the filters.

Finally, the third tab shows colleges by in-state freshmen, out-of-state freshmen, and percent out-of-state, all colored by the percent out of state. You can sort the institutions by the values in any column, by hovering over the x-axis label and clicking on the little bar icon that pops up. Subsequent clicks resort descending, ascending, and alphabetical. You won't break anything. Click away.

If you're a counselor looking for geographic diversity, this can be helpful. I found lots of interesting stuff that I can use tomorrow as we think about recruiting. What did you see?

**Note: I looked at the data and figured out that Harvard having 226 students from Arkansas was pretty unlikely, as was Harvard having zero students from California, so I took a leap and figured someone typed the data in the wrong box.  It's fixed now.  I also added a map where you can see the number of imports to any college from out of state by choosing the dropdown box on the fourth vis, the map.

** Note Two: I've added two more views, to the right: A bubble chart and a "Percent from out-of-state" chart.

Thursday, April 10, 2014

Is Admissions Fair to Women?

We go through cycles in college admissions, it seems, and the topic of interest in recent days (at least based on my limited view on things) has to do with gender discrimination in college admissions.

Most men readily admit that women are smarter, especially when it comes to high school performance. Others point out that men score higher on standardized tests, which predict far less with regard to college performance than grades do, and probably shouldn't carry as much weight as they do.

The focus on the treatment of young men and women in college admission goes back at least as far as this article in the New York Times in 2006. And the topic has been popping up a lot lately, most recently when Patrick O'Connor sent me this article, and asked for my opinion.  I thought it would be an interesting idea to look at the data.  So I did, using IPEDS data from the Fall of 2012.

The story here is interesting: The thing that jumps out at you, or at least might jump out at you, is that women file far more applications than men, which drives the fact that 55% of college students are now women. (Note: The number of applications from women does not necessarily mean that more women are applying to college, although in this case it does.)

The second thing that might jump out at you is that this is not true at the most selective institutions, where men file about the same number of applications.  Why is this?  Lots of reasons you could speculate about, and almost none of them reflect well on our society.  I'll leave the answers to researchers.

Anyway, have at it.  Use the filters liberally here (you won't break anything); the top three charts show summaries, and the bottom one shows individual institutions.  You can choose by selectivity, state, Carnegie Classification, public or private, in any combination.

And the bottom filter allows you to see the places where men have the greatest advantage in terms of admit rates.  Enjoy.


Wednesday, April 9, 2014

Staffing in Public Schools in the US, 2011

A while ago, I found some interesting NCES data on teacher salaries over time Teacher Salaries; it quickly became the most popular (currently 34,000 views) and the most contentious piece I'd ever posted, and months later, it's still getting hundreds of hits a week.  I simply found the data interesting, and wasn't making any sort of political statement.  I vowed at that time to never post about teachers again, so of course I'm breaking that rule today.

This is from the 2011 Common Core of data, showing all personnel in each state who work in public primary and secondary education.  Click here to see the data table and read all the footnotes, especially if you want to argue.

There are many things that can explain this data: States that have mostly urban areas (like California and New York) are likely to have bigger schools and classrooms, and thus fewer teachers per student.  Different states with similar populations may have more district administrators if they manage schools locally, or fewer if they do it in bigger geographies (by county, for instance.)

I added a couple data pieces to the table: First, the 2010 census data of the population under 18, which is not a perfect proxy for public school enrollment, but it's probably close enough.  I used Data Ferret to extract the data, and used those figures to compute the ratios (students per teacher, for instance.)  Again, a state with a younger population (more in elementary schools) might have fewer students per teacher than one with an older population (more in secondary schools).

Second, I downloaded 2012 election results, to see if data might sort out by Republican and Democrat states.  I'm pleased to say they generally don't, as I suspect that might avoid more colorful "discussion" on the topic.  The colors are applied to all but the top tab; Republicans in red and Democrats in blue.

There are four views, with the tabs across the top.  If you want to sort any of this data by any column, just hover over the x-axis labels at the bottom and click on the little bars icon to the right of the text.