Open Syllabus Blog | Open Syllabus Blog

The OS Analytics Top 100 Books of the 21st Century

July 13, 2024, by Joe Karaganis

The New York Times recently published a list of the '100 Best' books of the 21st Century, drawing on recommendations from authors and critics -- and it is generally excellent. And also completely subjective in the way all best-of lists are, and that's part of the fun.

It got me wondering what the analogous '100 Most Frequently Taught' books of the 21st century would look like if one removed the primarily 'instructional' genres that crowd the top ranks in Open Syllabus Analytics -- the textbooks, reference guides, readers, introductions to X, and similar titles. So that's what I've tried to do here.

Let's start by noting that 'Most Frequently Taught' does not mean 'Best,' nor is it a stamp of approval of the author or the ideas. A title can serve multiple pedagogical purposes, including exposure to work that is, for better or for worse, 'influential.' But I think it's fairly easy to see when the selection of a contemporary title, especially, does appear to reflect a judgement about what's best to illustrate an argument, introduce a perspective, or illuminate a topic.

You have to go pretty far down the title ranks in Analytics to identify 100 titles that (1) aren't in those pedagogical categories above; and (2) were published after 1999. 2600 titles specifically. And you learn a few things when you do.

First, the corpus of teaching titles is old. There are very few widely-used titles published in this century. The approach for most instructional materials is to continuously update titles published -- often -- decades ago. And the most successful textbooks are no longer books so much as brands or product lines. Pedagogically, higher ed is very conservative.

Second, the results cluster around a few topics: globalization and economic development, leadership and management, cognitive science, new media, and science writing are the largest categories. I think that reflects three conditions for appearing in the top ranks: large fields do better than small fields -- e.g., business, economics, psychology, and political science over religion or anthropology or classics; fields that assign books (rather than textbooks) to undergrads do better than fields that don't, so very little STEM presence; and fields that track and develop new ideas around contemporary events have a stronger appetite for new titles. It would be possible to continue down the ranks and identify, for example, the top 100 assigned contemporary novels but in practice the literature side of English is a fairly small and mostly historical field and contemporary fiction is a niche within it. Only a few literary titles appear near the top.

So this is not a literary best list, but arguably best for learning something important about the topics that have clustered at the top of the ranks, and best for understanding books that have become influential in thinking about those topics. The overlap with the NYT list is small, but does feature a few rapidly canonized literary titles. I've indicated the overlapping titles with a smiley. And since I was enjoying this excercise I extended the ranking a bit to 116. So you get more best books from Open Syllabus than from the NYT.

The links go to title and author profile pages in Open Syllabus Analytics. You can sign up for free to explore the full data.

*** Updated to remove a couple mistakenly included older books and articles

Galaxy Brain

June 28, 2024, by Joe Karaganis

Joe Karaganis

The reception of the ‘Co-Assignment Galaxy’ has been a minor ongoing puzzle in our work. The Galaxy is generally recognized as the coolest thing we’ve built – a navigable, bird’s eye view of all of anglophone higher education and a very novel way of looking at fields and subfields. And it does attract a lot of oohs and aahs. What is doesn’t attract is substantive conversation, which has surprised me because it provides a way of visualizing and characterizing everyone’s field. And people argue about those constantly.

So I’ve taken a stab at the question of why the dog didn’t bark – why people are fascinated by the Galaxy but don’t know what to do with it, or how to argue about it. And the answer, I think, comes down mostly to the lack of habits of interpretation of large-scale plots, which are an increasingly common but also very recent way of representing data. New forms of representation need new forms of literacy, and then the specifics of this representation – of fields composed of teaching choices – bring additional requirements. I’ve explored both issues in a piece at Public Books, and in the process tried to convey what I think we can learn about fields and the structure of higher ed from the Galaxy.

Big, well-consolidated disciplines often resemble archipelagos that spread out along their major teaching subdivisions. This subfield geography in English is distinctive and reflects the continued periodization of teaching by century, country, and—to a lesser extent—genre.

The disciplinary gravity that holds these clusters together, though, can be fragile. In the real world, classes that tick more than one disciplinary box can be cross listed. In the Galaxy this tension has to be resolved through color and spatial location. Most fields have inflection points where disciplinary commonalities can be outcompeted by strong thematic or problem-centered curricula, which “pull” titles into their orbits. Such tensions between disciplinary and topical organizations of teaching and research have been the starting point for many new fields over the years, such as the area studies fields that emerged in the postwar decades in part through SSRC programs.

Area studies fields turn out to be strong attractors of disciplinary titles. There is, for example, a large semicircular cluster for Asian and South Asian studies, composed of an arc running from India in the southeast to China and Japan in the northwest. What’s the main thematic connector between South Asia and East Asia? The study of Hinduism and Buddhism, here in teal and gray. There is a corresponding regional Africa cluster with differentiated political science, history, and literature subclusters and a similar one for Latin America. And there is a more nebulous cluster covering the Middle East, North Africa, and the study of Islam—with the last contributing a strong interdisciplinary dimension to the field.

2.11 Update

May 3, 2024, by Joe Karaganis

Joe Karaganis

We launched a major update to Open Syllabus a few weeks ago that includes a new version of the dataset, new features, and some changes to how you can explore Open Syllabus data.

Data

First, the dataset (v 2.11). Our goal with every update is to increase both the size and quality of the dataset. With the new version, we’ve added about a million new syllabi from the US, 300,000 from the UK and Australia, and 150,000 from Canada. The total number jumps from 18.7 million to 20.9 million. The collection also has a bit more historical depth due to better digging through school archive sites — enough so that we’re comfortable pushing back the starting year for charts to 2008 (from 2009). We want to keep extending the historical window back, but this depends increasingly on faculty and school archival contributions.

Next, data quality. We have significantly expanded both the range of titles we can discover and the quality of our parsing of citations within syllabi. First, the number of unique discovered titles in the collection has jumped from 3.4 million to 3.8 million. This gain comes primarily from building a source catalog from ISBNs found within the collection, which has significantly expanded our view of non-US publishing in particular. We also updated the citation catalog that we build from linked content in the dataset, which means that we have a better view of periodicals, websites, videos, and other non-traditional sources. Here are the top-assigned titles, for example, from The Atlantic Monthly.

Combined with improvements in how we deal with the variety of citation practices used in syllabi, the number of citations discovered in the collection (across 21 million syllabi) has jumped from 56 million to 65 million. In short, we’re providing a more complete view of the same collection at the same time as we add to it.

Accounts

The biggest change affects how you can explore the data. We have shut down the old ‘free’ service, the Syllabus Explorer, in favor of creating an open tier for the new, better service, Open Syllabus Analytics. Analytics now has open, trial, and subscription access. The open tier is much better than the Syllabus Explorer, not least in the number of syllabi it includes, which jumps from 7.2 million to 11.9 million. It retains a 2019 date cap (up from 2018) and, compared to the trial and subscription versions, more limited functionality. You can sign up for a trial for free to explore these differences. Our goal is to get schools to subscribe.

Accounts have also changed. Now your account is just based on your email -- and if you sign up with a school email, it's immediately approved. Logging in with an email keeps you logged in on that decice and browser.

The Gender Dashboard

We’ve also improved the Gender Data dashboard, which is available to logged in users. Now you can search and filter across a wide range of dimensions, including country, field, and school. Want to know, for example, what gender ratios look like across academic fields in Australia or how those ratios have changed in law in the US? Now you can.

That’s it for 2.11! Check it out — and if you like it, talk to your data services librarian about subscribing.

OER Adoption Update: 2023

April 9, 2024, by Joe Karaganis

Joe Karaganis

With the release of version 2.11 of the OS dataset, we can update our reporting on adoption rates for open educational resources (OERs) -- specifically, open textbooks and 'open access' monographs.

The defining thing about OERs is that they are released under an open license and so can be freely circulated and used. Because they are free, OERs have become an important part of efforts to lower the cost of higher education, which often includes thousands of dollars per student per year for course materials. But free comes with a small catch: because there is no sale or single distribution point, it's difficult to track the use of OERs. And this difficulty makes it hard to know whether adoption efforts are working. And this lack of feedback complicates decisions about whether and where and how to invest in the open ecosystem. The OER ecosystem is weaker, we would argue, because it's bad at closing this information loop between supply and demand.

And this is where Open Syllabus can play a role. We track OER adoption via the appearance of titles on syllabi -- the point of use rather than the point of sale or distribution. With nearly 21 million syllabi in the collection, we can provide accounts of adoption across individual schools, states, and countries, in and outside the US -- with some caveats that we'll get to below.

The new dataset adds two million syllabi and extends the timeline to 2023. Broadly, it expands on the story we've been telling for several years: rapid growth in OER textbook adoption in the US and Canada, but from a very low baseline in the mid 2010s, and rapid OER monograph adoption outside the US and Canada from similarly low baselines. What does this mean concretely?

In 2013, OER textbooks were barely on the scene in the US and used in only around 1 in 400 classes. By 2023, it was 1 in 80. Over the past 10 years, the growth rate in OER textbook use has averaged around 17.5% per year in the US.

OER efforts have especially targeted two-year schools, where textbooks represent a larger portion of educational expenses and students are less able to afford them. This focus shows up clearly in the data. The adoption rate at two years is 1 in 40 classes -- with a growth rate of 26% per year over the past 5 years. Is this a tipping point? It's beginning to look like one.

Because US higher ed is very decentralized, OER adoption in the US is shaped by state, system, and school-level policies. California, which has the best-supported state-level program, remains the national leader by a large margin: Over 1 in 40 classes use OER titles and 1 in 25 at two-year schools.

Among the other states for which we have decent data: Texas and Michigan are close to average. Georgia, Florida, Colorado, and Illinois fall below it.

With respect to fields, math continues to be the leader, followed by computer science. 1 in 25 math classes use an OER textbook. Restrict to 2-years and it's 1 in 18.

Other fields have begun to catch up as OER authors begin to fill some of the common entry-level class niches. Almost all of the growth in English and Education, for example, has happened since 2019 -- and in both cases due primarily to the emergence of writing guides and student success manuals. Methods books have gained traction across a number of social scientific and adjacent fields, including business. And recently developed, topic-specific textbooks for core classes in business, psychology, and political science have begun to gain traction.

The open book ecosystem in other countries offers interesting points of comparison. Canadian OER adoption looks very similar to the US, both in terms of the rate of growth and the heavy focus on textbooks.

But adoption elsewhere is more heavily tilted toward monographs. Adoption in the UK is the extreme case, with rapid monograph growth and a very small role for textbooks.

That downslope after 2020 reflects some data access problems related to a change in reading list software platforms in the UK (and Australia), so we are inclined to discount it. The more general pattern shows steady and in many cases very rapid monograph growth elsewhere in Europe where we have no such data problems. For example, Italy:

And Portugal:

And the Czech Republic:

And to an extent in the Netherlands:

And also in Australia -- where overall adoption is lower but roughly evenly split between textbooks and monographs.

The data outside the US and Canada point to differences between OER textbook publishing and 'OA' or open access scholarly monograph publishing, which remains the province of university presses. For OER textbook authors and publishers, the goal is to provide substitutes for commercial titles used in popular classes. An OER textbook can have a steep adoption curve because it has a large potential pool of classes to convert. This conversion process, in turn, has been formalized into policy and advocacy approaches, and has become an official responsibility of the library at many schools.

OA monographs, in contrast, come out of a more-or-less parallel but distinct movement to ensure that published research is free -- beginning with journal articles but extending to books and other research outputs. As research, OA monographs are usually specialized titles and almost by definition not intended to be direct substitutes for existing (commercial) titles.

The incentives and growth potential are accordingly different. Publishers are publishing more OA titles (measured by growth of the Directory of Open Access Books ), but they tend to fill small course niches and show up in classes that already assign a lot of other (commercial) books. The US and Canadian advocacy model doesn't apply as easily in this context, 'free' has less of an impact, and the textbook-based curriculum of two-year schools is largely irrelevant. This is, we think, the context for what we see in the US and Canadian data, which shows modest growth at best in the adoption of OA titles over the past ten years.

But elsewhere, the monograph vs textbook data is different and I don't think we have a complete explanation of it. One important factor is that the market-size dynamic is reversed: it's entry-level textbooks that are fragmented by country and language -- and dominated by translated commercial textbook titles -- while the research culture is international and built around English. English-language research monographs 'travel' better in this context and cost/benefit decisions shift accordingly. There are also differences in the publishing and advocacy cultures that have grown up around Open Access research initiatives, with more active roles for libraries in funding new work. We'll just note this for now without venturing a full explanation.

We've alluded to a few caveats on the data already, and such a discussion could go on for some time. But let's address one that matters to the OER community.

There is reason to think that we undercount OER adoption -- though not dramatically with respect to anglophone countries. One source of undercount is our incomplete catalog of open titles. For the current dataset, we combine The Directory of Open Access books (for monographs), the Open Textbook Library (for textbooks), and parts of Academic Commons. (We also filter out public domain titles, which have begun to creep into OA/OER catalogs in new editions or translations). This the 'high end' of OER production in English -- most of it passing through peer review and/or press development processes. But these sources will miss a lot of non-English material relevant in other countries as well as less-formally-published work, such as material produced and circulated for departmental use or variations and remixes that are never recatalogued, and so on. There are archives for these materials but every new source that we integrate produces trade-offs between signal and noise in the data, so we move forward with caution.

A related source of undercount is the set of citation shortcuts that have grown up around some categories of OER title, including especially the replacement of authors by publishers in citations on syllabi. A typical example might list 'Chemistry by OpenStax' on a syllabus rather than Chemistry by some representation of the actual contributing editors and authors, which is our usual requirement for identifying titles. We can do a sanity check on this 'missing' material to some extent by searching our collection for the term OpenStax. It appears on 1 in 125 US syllabi and 1 in 65 at 2-year schools -- ballpark consistent, we would argue, with our updated approach to these citations. But there are certainly other scenarios and cases that we miss. These issues also affect all OA/OER cataloguing efforts -- and we inherit the problems of the catalogs we use.

Loosening concepts of authorship are a challenge for our work but also, arguably, for a community that treats attribution as a primary reward for OER publishing. This is an inherent tension in the OER community, which also values the right to remix open content. For our part, we think that we get the trend lines generally correct and undershoot a bit on the magnitudes.

The trend lines describe not just instances of adoption but also underlying processes of adoption that have institutional momentum -- and that momentum matters a great deal. Textbook choice, especially, is sticky -- costly to change -- and it can take a long time for new titles to take off. The major OER math titles are over a decade old. The computer science titles a bit older. Social science titles and business titles have begun to emerge in significant numbers only in the last five years. And this isn't long in curricular terms. Major commercial textbooks are often decades-old brands that become synonymous with their teaching subjects.

Last year, we asked whether the growth rate for OER titles will be sustained. This year's data provides a provisional yes. We'll continue to chart that progress and if we’re right, by charting it, accelerate it.

(April 18: updated with more thinking about the monograph data)

What Films Should We Teach

March 9, 2024, by Joe Karaganis

Joe Karaganis

Last year we published a roundtable discussion called "What Films Should We Teach: A Conversation About the Canon" in Public Books. For me, the goal was to spur some expert consideration of the Open Syllabus Movie Lab data, which I had assembled the year before, and which showed that the film curriculum in higher ed is still mostly organized around critical canons and principles developed in the 1960s and 70s. I wasn't equipped to unpack this data, but Jane Gaines, Jon Lewis, and Daisuke Miyao have built teaching and research careers around these questions and were generously willing to dig in.

Jane Gaines (JG): ... In the United States, film and media studies grew out of literary studies and, at many schools, remains either within it or closely tied to it. The Open Syllabus rankings strongly suggest the continued relevance of literary approaches to teaching film history, which tend to bring the criteria of personal authorship to bear on film.
Auteur theory received its major formulations by French New Wave directors and critics in the 1950s and 1960s, who argued that directors such as John Ford, Howard Hawks, and Alfred Hitchcock “transcended” Hollywood industrial production and impressed a personal “worldview” on their material. The persistence of Hitchcock and Welles at the top of the Open Syllabus director rankings suggests how strongly this tradition is still with us. Teaching these two directors carries on a tradition of exposing college students to “great authors” because of assumptions about the depth of their personal insight and vision. And much of the critical tradition around these directors is still fundamentally tied to frameworks and vocabularies used to evaluate literary works.

One casualty of this focus is the collective and industrial dimension of filmmaking.

Daisuke Miyao: Cinema is a collective industrial and cultural form, but there is very little work that considers films as the product of collaboration beyond the auteur-director’s authority. Much is lost through this narrow focus. The international impact of Rashomon, to take a major example, is inseparable from the contributions of its cinematographer, Kazuo Miyagawa, who experimented with film stock and lighting in ways that profoundly influenced a generation of artists in other countries, such as cinematographer Gregg Toland in Citizen Kane. What would film history look like if, as a thought experiment, we privileged cinematographers or editors or writers?

Predictably, one of the major factors that maintains the traditional canon is teaching bandwidth.

Jon Lewis (JL): When I arrived at Oregon State in 1983, there were only two movie classes: Film Comedy and Film Tragedy. Both were taught in the English Department. My colleagues were surprised to hear that I found the titles overbroad. But as a practical matter, they had to be. Only a fraction of schools in the United States have film programs: some sources put the number at over 300, depending on the definition; others say around 150 out of nearly 4,000 schools in the United States. This means that in most places the teaching canon is effectively whatever makes it into the Intro to Film class or its equivalent.
Moreover, teaching time is scarce and presents zero-sum choices. In a semester-long class, adding one title to the syllabus often means removing another. The relevant question for me in teaching is: What do students need to know if—as is often the case at Oregon State—their film education consists of one class from me? My answer is almost always “history.” Are there certain films students need to see in order to better understand contemporary cinema? Are there films that allow me to condense the discussion of a period or set of themes into one or two classes? This process of cropping is, for better and for worse, well suited to the auteur framework.
So even as over the years I have been researching and writing about films that challenged the canon or offered new ways of looking at it, the traditional approach has remained useful.

The whole conversation is worth a read.

LATEST POSTS

Data​

Accounts​

The Gender Dashboard​

Data

Accounts

The Gender Dashboard