The scholarly ‘Great Awokening’ seems to be winding down
After 2011, there was a rapid change in discourse and norms around social justice issues, particularly among knowledge economy professionals (i.e., people who work in fields like journalism, the arts, entertainment, law, tech, finance, consulting, education, and research).
As I detail in my forthcoming book, this “awokening” manifested in everything from poll and survey responses, to media outputs, to changes in political alignments, and beyond. Within academia, there was a sharp increase in student protest activity beginning in 2011, accompanied by growing tensions around “cancel culture” and self-censorship. There were ballooning investments in (demonstrably ineffective) mandated diversity-related training and rapid expansions of campus “sex bureaucracies.”
Changes were also apparent in research outputs.
In a recent paper for the National Association of Scholars, computer scientist David Rozado analyzed 175 million scholarly abstracts from articles published from 1970 to 2020. He found that, after 2011, there was a sharp increase in the use of prejudice-denoting terms. This held for virtually all forms of bias and discrimination (racism, sexism, transphobia, Islamophobia, ableism, ageism, fatphobia, and derivatives of the same). Statistical analyses suggested that a single underlying shift, likely among the people who produce academic research, could explain most of the observed change across all of these terms since 2011.
In Compact Magazine I recently argued that, by several measures, the “Great Awokening” seems to be winding down. Starting in late 2021, and continuing throughout 2022, there appeared to be a moderation trend across many social indicators. I was curious whether this pattern could be observed in academic research as well. I was also eager to replicate Rozado’s general findings in alternative data sets.
Analyzing trends in different academic research databases (described below) over the last 23 years, I found roughly the same patterns of behavior that Rozado observed. There was a significant uptick in research focused on various forms of bias and discrimination starting in 2011 and persisting through 2020. Rozado’s findings were therefore not an artifact of the specific data set he used but replicated across a range of scholarly databases.
However, the additional two years of data I was able to analyze were also quite revelatory. After 2020, there were declines across the board in published research focused on identity-based bias and discrimination. Academic scholarship seems to have passed peak “woke.”
A Quick Methodological Note
Several different databases collect scholarly papers and/or research abstracts.
Rozado relied on the Semantic Scholar Open Research Corpus and observed patterns from 1970 to 2022. For my replication and extension, I analyzed articles published between 2000 and 2022 that appeared in Google Scholar, Web of Science, EBSCO Essentials, JSTOR and Scopus.
Although there is a strong core of work present across virtually all these databases, myriad papers also appear in some databases but not others. Each draws from different combinations of indexes and has slightly different criteria for inclusion in the database and slightly different means for pulling up works in response to various queries.
Google Scholar, for instance, contains a number of works that are preprints, dissertations, white papers from think tanks, academic books, or even non-scholarly articles cited in scholarly research, alongside published peer-reviewed papers. Other databases are less “noisy” in that they are more restricted to published peer-reviewed journal articles, but they also tend to be less inclusive even with respect to published journal articles per se.
To account for these differences, I started by looking at the trends for each of the five chosen databases separately, analyzing the number of articles per year referring to racism, sexism, transphobia, homophobia, and variations of those terms. Each told the same general story. However, the overall quantities of articles and the slopes of the lines between data points were generally nonidentical from one year to the next.
To get a sense of the overall trends in the data, I averaged the counts across the databases for each genre of prejudice from year to year. The specific terms analyzed, the number of mentions for each genre of terms per year in each data source, and the annual averages across the data sets for each genre of prejudice are all available here. The averaged results are visualized below. But first, let me offer a brief guide for interpreting the chart.
Expected and Observed Patterns in Academic Research: 2000–2022
More journals are being created all the time. The number of people taking part in academic research has consistently and dramatically grown as well. Moreover, contemporary scholars are significantly more productive in terms of publication output than previous cohorts of scholars (albeit, largely through co-authorship). The total amount of research produced each year in most fields has continued to climb in accordance with these trends.
As a function of these realities, when we look at total mentions of key terms over time, for virtually any significant term, the expected pattern of mentions would be a secular increase. That is, a steady upward slope in raw mentions should be the normal expected pattern for most terms, irrespective of any unique events happening in the world. Periods of gradual and steady increase are therefore not analytically interesting. They are the norm.
What is interesting and informative are periods when the slopes of the lines change dramatically in either direction. With respect to racism, sexism, homophobia, and transphobia, we see significant and roughly simultaneous increases in scholarly discussion after 2011 and persisting for most of the decade that follows. The graph below visualizes the trend for each genre of bias and discrimination from 2000 to 2022, averaged across all five databases:
However, the chart above does not just illustrate a significant increase in scholarly discussion of identity-based bias and discrimination after 2011. We can also see that there has been a significant decline in scholarly discussion of these issues in recent years across the board.
The timelines run a bit differently for different types of prejudice or discrimination. Work on sexism and misogyny plateaued first, in 2018. Work exploring prejudice against racial and ethnic minorities, and lesbian, gay, or bisexual people, reached its zenith in 2020. Work discussing bias and discrimination against trans folks peaked a bit later, in 2021. By the end of 2022, however, all four had retreated a bit from their high-water marks.
The declines over this particular period are particularly striking because they diverge sharply from the broader trends in scholarship overall during these years. Many scholars, their usual schedules and workflows interrupted by COVID-19, used the pandemic period to convert completed research into publications. As a recent study in Nature observes, “2020 saw a sharp increase in articles on all subjects being submitted to scientific journals — perhaps because many researchers had to stay at home and focus on writing up papers rather than conducting science.”
That is, what we would expect to see on most topics in most years is a secular rise in publications. What we would expect to see for research on most topics during the COVID years is an unusually large rise in publications. Instead, for work on identity-based bias and discrimination, the opposite pattern is apparent. Commensurate with trends explored in my recent essay looking at other social indicators, it seems as though the “awokening” may be winding down in academic research.
It would be interesting to see if similar patterns were revealed for books. Unfortunately Google Ngram Viewer, the best book database available at present, covers only work published before 2019. Consequently, it is easy to visualize the post-2011 rise of terms like racist, sexist, homophobic, and transphobic (or more niche and fashionable terms like settler colonialism, racial capitalism, or toxic masculinity) in published books. However, any declines would be largely outside of the available data.
Gold Rushes
In a 2012 paper, sociologist Gil Eyal explored how social and historical developments can often give rise to academic “gold rushes” — periods of intense interdisciplinary scholarship on particular topics wherein large numbers of scholars who are not specialists on a given subject matter, and have not previously published much (if at all) on a particular topic, suddenly shift their scholarly attention to capitalize on dramatically increased public attention to a particular issue and the opportunities entailed thereby.
As I illustrated in a 2018 paper on the then-burgeoning field of “Trump studies,” the research produced during these gold rushes is often subpar, because the incentive structures are even more messed-up than usual.
The goal in a gold rush, for scholars and journal editors alike, is to deliver big, splashy results that affirm the narratives key stakeholders are eager to advance. This is how one can secure tons of press and lots of citations, and capture the attention of deep-pocketed foundations and donors (with all that entails for one’s disciplinary standing and career prospects downstream).
Scientific research already faces significant problems related to fraud, bias, negligence, and hype. These problems tend to be radically exacerbated during gold rushes.
And what a rush it was! Looking at averages across each of the four genres analyzed, more than half a million papers produced between 2012 and 2022 focused on identity-based bias and discrimination. Few of these are likely to be read or cited five years from now (especially if self-citations are excluded from analysis). Fewer still will have any significant and demonstrable positive impact on policymaking or any other form of practical social change. But that’s almost beside the point. CVs have been padded with new publications. Grants have been awarded. Careers have been enhanced. The gold rush mission has been accomplished.
Secular Increases
The post-2011 academic gold rush on identitarian bias, discrimination, and inequality seems to have run its course. Production on these topics has plateaued or declined as a result — at least, relative to their hitherto unprecedented peaks.
However, although the rate of increase in academic research produced on these topics may normalize into something that approximates pre-2011 levels, the overall volume of research on identity-based bias and discrimination would consequently increase each year (as was also the pattern prior to 2011) and eventually exceed the current high-water marks.
Again, the typical trend for literature produced on most topics will be an increase in total papers published each year relative to the previous year (as a function of the ever-increasing numbers of journals and researchers, as well as ever-intensifying publication pressures and expectations). Research on bias and discrimination is not unusual in these respects. A secular annual increase in research produced was the norm before the “Great Awokening,” and it will likely be the norm again … probably sooner rather than later.
The intensity of focus on identity-based bias and discrimination that defined the last decade of scholarship could not go on forever. However, absolute declines in production on these topics will likely be even more short-lived.
Yet, as the awokening recedes, it seems likely its main legacy will not be the huge volume of fairly questionable yet rarely read or cited academic research it produced, but rather the equally widespread, dramatic, and contemporaneous changes in policies, procedures, practices, and administrative structures — many of which will likely persist for the foreseeable future (even if some of the most extreme and pernicious changes over the last 10 years do eventually get rolled back).
Put another way, we may be returning to “normal science” in most fields; however, this neither entails nor implies that higher ed institutions will simply revert back to their pre-2011 status quo.
The awokening may be ending, but that means it’s time to sleep in the bed we’ve made — for better and for worse.