About four years ago, I looked into data on how US college students’ choice of majors had changed since 2011 and whether the University of Oregon (UO) differs notably from other R1 (Carnegie “Very High Research Activity”) universities. The old post is here, but there’s no need to look at it. I’m now revisiting this question with new data and better analysis and plots. I was inspired to do so in part by Benjamin Schmidt’s recent posts on degree trends. Do UO students choose unusual majors?

The takeaways of this post: across R1 universities, there’s a massive increase in computer science majors; across R1 universities, the number of humanities majors has fallen; at UO, there’s a striking and unusual increase in journalism and communication degrees — **112% more** in 2021 than in 2011, compared to +21% at R1s on average.

### Trends in majors since 2011

The data come from the Integrated Postsecondary Education Data System (IPEDS), from which one can download CSV files of degrees awarded in a given year at a set of universities. I modified my 2018 program to trawl through these datasets, tabulating the degrees awarded [1] and lumping the National Center for Education Statistics classifications into my own categories. The lumping is in many cases obvious, for example “40.0801 Physics, General” maps onto my category of “Physics,” and “Economics” encompasses “45.0601 Economics, General, 45.0602 Applied Economics, 45.0603 Econometrics and Quantitative Economics, 45.0604 Development Economics and International Development, 45.0605 International Economics, and 45.0699 Economics, Other,” though only the first of these is likely to be relevant for Bachelor’s degrees. Some lumpings are somewhat subjective — I wanted to separate mathematics and statistics, and there are classifications that could fit either. There’s room for improvement, but I think my categories are a good match for what degrees we usually think as corresponding to various fields.

Let’s start with the punchline plots, showing the rate of increase of various majors from 2011 to 2021. “Rate of increase” here is the **average annual percentage increase [2].** A value of 2%, for example, would correspond to a 1.2-fold increase when compounded over 10 years, and 1.5-fold over 20 years. On the vertical axis is the rate of increase at UO (bar = fit uncertainty). On the horizontal axis is the average of all the R1 universities’ rates of increase (N = 99 universities; bar = standard deviation [3]). The size of the points relates to the number of degrees at UO in 2021. If UO were perfectly average, all points would lie on the dashed diagonal. I’ll plot two versions: the first simply shows the percentage rate increase in each degree category; in the second, degrees are normalized by the total number of degrees awarded to account for overall growth in a university’s population [4]. (The total is of “first” degrees, and so shouldn’t double count double majors.)

Growth rate of various degrees:

Growth rate of various degrees, normalized by total number of degrees [revised/fixed Feb. 7, 2023]:

In case you’re wondering what happened to statistics and all engineering degrees, these don’t exist at UO. (Yes, that’s unusual.) The relative annual percentage increases at all R1s are 15.9 +/- 9.3 % (statistics) and 0.7 +/- 4.1 % (all engineering). Yes, that’s +16% **per year** for statistics.

Comparing the un-normalized and normalized graphs, there’s an overall shift down/left, but nothing radically changes position. Several majors that are growing in absolute terms (e.g. Physics, Biology) are declining in relative terms.

Returning to Oregon: We see general agreement overall with other R1s, but with some clear differences.

Computer Science has become more popular at UO (+11% *per year* and a 172% increase between 2011 and 2021), but its popularity has grown even more elsewhere (average +17.8% per year, +423% from 2011 to 2021).

My own subject, Physics, is below-the-average in both absolute and relative terms. (Yes, we know that we should do something about this…) Humanities are similarly stunted. This reminds me that I was chatting last week with a History colleague who first pointed me to the amazing Vaclav Smil. We brought up the idea of making a course about the history and physics of energy use, essentially mirroring Smil’s stunning “Energy and Civilization,” but in addition to being a lot of work to create, the course might have an audience of zero!

The most striking major category, I think, is journalism / communication: Even normalized by the size of the student body, it shows a 5.4 +/- 1.0 yearly percentage increase here, compared to -1.0 +/- 4.1 % at R1s in general [5]. Why is there such a proliferation of journalism and communication majors at UO? I don’t know. I can’t even construct a sensible hypothesis! Here are some thoughts, which add up to nothing:

- Journalism tops polls of the “most regretted” college major in the US, with 87% of degree holders wishing they had chosen something else (source). Communication is #4.
- In my “physics of energy and the environment” for non-science-majors course, I assign a lot of newspaper/media articles about energy and the environment. It is, of course, an important topic. It is amazing how low the quality of most articles is, and even more amazing how devoid they are of basic scientific understanding or comfort with numbers.
*The Economist*is the only reliably good source. I’m teaching the course this term; of 95 students, 1 is a journalism / communication major. - Coincidentally, chatting w/ three other physics faculty under 50 a few days ago, we all realized we subscribe to the physical, delivered, on-paper local newspaper. “I’m not the only one!,” we exclaimed.
- Many of the degrees in this category are not journalism per se, but advertising and other fields. In 2021, for example, UO awarded 126 Journalism degrees, 224 degrees in Public Relations/Image Management, and 317 in Advertising. Maybe this just highlights generational differences, but when I was in college (1990s), telling someone you wanted to go into Public Relations or Advertising would be like saying you wanted to go into arms dealing, or maybe worse. Kids these days…

Since 317 advertising majors can’t be wrong, I’ll insert my usual links to my pop-science biophysics book, now one year old! My description, Publisher, Amazon.)

I’ve spent many hours on the analysis code for this post, and there are a lot of stories it enables, such as differences among various universities and methods for data visualization. I’ll likely milk this for more posts; if there’s something you’d like to see, please **email me or comment below!** Perhaps I should make a quiz: By what factor did the number of computer science degrees at UC Berkeley increase between 2011 and 2021? (I’ve already made the plot…)

## Today’s illustration

A bottle with red liquid. Despite a smudge, I like how it turned out.

— * Raghuveer Parthasarathy.* February 4, 2023

### Notes

I added these explanatory notes on Feb. 7, 2023.

[1] “Degrees” is the sum of “First major” and “Second major” numbers in the IPEDS database.

[2] In other words, let’s fit the number of degrees *D* over time *t* as D = D_0 exp(rt). The growth rate is *r*. We can express *r* as an average annual percentage increase (i.e. like compound interest) via *annual pct. rate* = 100 * (exp(*r*)-1.0). Here’s an example of degrees vs. time for Math, where each line corresponds to a different university; UO is magenta, and its fit is the magenta curve:

[3] The error bars are the standard deviations of the average annual percentage increases (see [2]) for all the R1s. In other words, each university, for each major, has some growth rate. I calculate the mean and the standard deviation; that’s what’s plotted on the “x” axis.

[4] The normalized growth rate is calculated by first dividing number of degrees by the total number of “First major” degrees, in all subjects, for that university in that year. The “first major” number should match the overall graduating population. Total degrees will not, because of some number of double majors.

[5] I corrected the normalization, Feb. 7, 2023. Previously, it was slightly wrong (3.9 for Journalism at UO, rather than 5.4, for example.)

Very nice graphs. Your choice of colors and shapes for symbols made them easy to differentiate. It would be fun to animate the graphs to see rates of change. Thanks for your hard work.

Thanks! Animation is a great idea. I’ll put it on the list…

A few other thoughts on Journalism:

1) UO’s J School is in the top 10% nationally. Maybe that ranking has changed over time, and coupled with UO being a state school (and inexpensive for some by comparison to other options?), this draws a crowd?

2) The J School has a new (well, newly rebranded) center for Science Communication Research, and a new minor in science communication. They are looking for collaborators across campus from all disciplines (https://scr.uoregon.edu/collaborate-with-scr/). You could do something about the dismal state of science/math literacy in journalism! I took a science communication course with them in grad school, and was glad for the cross-disciplinary perspective on what constitutes science literacy.

3) I myself wanted to be a journalist in middle school, and I consider it a noble profession (advertising, not so much). It’s also been widely portrayed as an industry in decline / under attack, and in need of saving by more paid subscribers. If you calculate journalism separately from Advertising / PR, does it have such a striking increase? Or is Advertising / PR driving the growth?

All great points! (1) I wonder how much students are influenced by measures like this — I really don’t know. (2) Yes, I’m a member of the Center for Science Communication Research, and I hope that this leads to some interesting activity! (3) Journalism alone at UO has a +3.2 % increase per year compared to -3.2 +/- 7.5 for all R1s — about half the value than if we include advertising, etc., but still surprisingly large. (And the opposite sign of most universities!)