Bias in the Academy: Counting Co-Authors

In light of the frequent campus climate issues of recent years, many of us in higher education have been thinking about inherent biases in our institutions’ appointment, promotion and tenure systems. How might faculty of color and women be systematically thwarted when they try to move up the academic labor market? One fundamental way such biases manifest themselves is how academe gives credit for single-author and multiple-author journal article publications.

In my field of economics, the number of authors per paper has increased monotonically over time. David Card and Stefano DellaVigna have observed that 75 percent of journal articles were single authored in the early 1970s. In contrast, less than 50 percent of journal articles were single authored in the early 1990s, and less than 25 percent of journal articles were single authored in the early 2010s. When I look around casually today, about 80 percent of economics journal articles have at least two authors. And Card and DellaVigna report that, in 2018, the mean number of authors per journal article was almost 2.5.

The reasons for the growing number of co-authored journal articles are not clear. One possible explanation is that the field of economics has become increasingly complex. Consequently, it is less likely that any individual economist has all the talent and skills necessary to complete a journal article. Another possible explanation is that academics are under greater pressure to publish. And if departments do not distinguish between single-authored and co-authored journal articles, it is easier to increase the number of publications with co-authors.

When I talk with faculty members of color, they express a concern about this practice of co-authoring papers. They tell me that it is harder for faculty of color to find co-authors. In many ways, finding a co-author is like finding a spouse. We tend to marry people who look like ourselves. Tall people tend to marry other tall people. Educated people tend to marry other educated people. White people tend to marry other white people. There are similar patterns with co-authors. They tend to have ties to the same graduate schools. They have interests in the same subfields. And faculty members of color tend to write with other faculty of color. But with fewer faculty of color in academe, it is harder for those scholars to find appropriate co-authors.

Unfortunately, while the practice of co-authoring articles creates a bias against faculty of color, we can do little to change the situation immediately. If we can increase the number of faculty members of color in higher education, that will help, but it will take some time.

For today, we must focus on being careful about properly crediting the work in co-authored journal articles when we evaluate faculty members. While single-author papers send a clear signal about skills and abilities of the author, co-authored papers do not provide specific information about each author’s skills and abilities. That ambiguity can result in systematic biases. We must make sure that we recognize the work of co-authors in a fair and consistent way.

Measuring Productivity

Let’s begin with an illustration. Suppose you have three faculty members in your department named Peter, Paul and Mary. When they come up for review, Mary has published one journal article all by herself, and Peter and Paul have published one journal article together. So Mary was able to publish as much as Peter and Paul put together. Everything else being the same, Mary gets credit for being more productive than either Peter or Paul. On average, Mary is twice as productive as Peter or Paul.

The extensive literature on scholarly productivity is uniform in measuring productivity this way. It basically divides each publication by the number of co-authors and adds up the total. It gives twice as much weight to an article with a single author over an article with two co-authors. It gives three times as much weight to an article with a single author over an article with three co-authors. And it gives equal credit to all the co-authors.

In addition to the number of authors, scholarly productivity literature tries to adjust for everything else as well. It adjusts articles for their word length (measured in standard pages) and for quality by looking at things like the number of citations, journal impact factor and so forth. The literature on how to measure productivity, in principle, is relatively clear. The typical equation to measure productivity looks something like this:

But on review committees and in academe in general, it is not clear that productivity is actually being measured this way. Some departments and review committees seem to give equivalent weight to co-authored journal articles and single-author journal articles. They often behave as though Peter, Paul and Mary are equally productive, according to previous research on the topic.

When review committees look at the CVs of faculty members, perhaps they are merely impressed by the length of the publication list. Suppose, for example, that Karen and Richard are two assistant professors in your department. (While representative, these individuals and articles are fictitious. No identification with actual journal articles is intended or should be inferred.) When they come up for review, Karen has published the following paper:

• Karen C., “The geopolitics of Argentinian soybeans,” Thought & Expression, Vol. 43, Issue 2, 2016, pp. 348-371.

But Richard has published the following three papers with some co-authors:

• Richard C., Leanna D. & Yufeng S., “Understanding the New Alliance for Food Security and Nutrition: Eat Drink Man Woman in Taiwan,” Thought & Expression, Vol. 43, Issue 2, 2016, pp. 22-45.
• John A., Richard C., & Suyapa L., “Acquiring Land for Public Purposes in Papua New Guinea: The Land Act of 1996 and the Lands Acquisition Act of 1974,” Thought & Expression, Vol. 43, Issue 2, 2016, pp. 521-544.
• Gustavo A., Sarah B. & Richard C., “Regenerative Organic Agriculture in Patagonia: Soil Health, Climate Change and Carbon Absorption,” Thought & Expression, Vol. 43, Issue 2, 2014, pp. 251-274.

When you see a longer list of publications like this, it gives you the impression that Richard has been more productive than Karen. The longer article titles also make the list of publications seem more substantial. But Karen and Richard have been equally productive here. (We are holding things like journal quality and the length of the paper constant.) But it is not obvious that review committees looking at these lists will recognize the equivalent productivity levels.

Who’s on First

With co-authored papers, another issue that review committees and others consider is author order. Sometimes the order in which the authors are listed signals the relative contribution of each author to the paper. Sometimes the authors are merely listed in alphabetical order. Sometimes regular co-authors merely alternate the order in which they list their names. This all makes the issue of author order difficult to interpret properly.

In the field of economics, the usual practice has been to list co-authors alphabetically. However, in the world of academe, the first author seems to get more credit than the other authors. For example, Eugene Fama and Kenneth French have written a number of important finance papers together. People always talk about the Fama-French model. They do not talk about the French-Fama model. Is it a coincidence that Fama has won a Nobel Prize and French has not? N. Gregory Mankiw, David Romer and David N. Weil wrote an important paper on economic growth, but it is often cited as Mankiw, et al. Romer and Weil don’t even get mentioned.

This practice of listing authors alphabetically has resulted in a very perverse outcome for economics professors. Professors with last names that start with a letter near the beginning of the alphabet are more likely to receive tenure than professors with last names that start with a letter near the end of the alphabet, as Liran Einav and Leeat Yariv found when examining assistant professors in top-ranked economics departments in the United States. The key to success seems to be to find a smart co-author with a last name that falls later in the alphabet.

Here is a list of some fairly famous works in economics. While not a scientifically selected sample, the first author of these works below seems to have received more fame for their joint work, even though the authors are listed in alphabetical order. The asterisk here means the author has won a Nobel Prize in economics.

• Stiglitz*, Joseph E. and Andrew Weiss, "Credit Rationing in Markets with Imperfect Information," American Economic Review, June 1981, vol. 71(3), pages 393-410.
• Engle*, Robert and Clive Granger*, "Co-integration and Error Correction: Representation, Estimation, and Testing," Econometrica, March 1987, vol. 33(2), pages 251-76.
• Fama*, Eugene F. and Kenneth French, "Common Risk Factors in the Returns on Stocks and Bonds," Journal of Financial Economics, February 1993, vol. 33(1), pages 3-56.
• Akerlof*, George and Janet Yellen, "Can Small Deviations from Rationality Make Significant Differences to Economic Equilibria?," American Economic Review, September 1985, vol. 75(4), pp. 708-20.
• Friedman*, Milton and Anna Schwartz, A Monetary History of the United States, 1867-1960, Princeton University Press, 1963.

This practice of listing co-authors alphabetically, along with the privileging of the first author, may disadvantage minority group members, whose names often appear later in the alphabet than white scholars.

The last two works listed above raise an additional issue. As Heather Sarsons has noted, “Women … suffer a significant penalty when they coauthor. The results hold after controlling for the total number of papers published, quality of papers, field of study, tenure institution, tenure year and the number of years it took an individual to go up for tenure. The result is most pronounced for women co-authoring with only men and is less pronounced the more women co-authoring a paper, suggesting that some gender bias is at play.”

So if you are a woman and your last name starts with a letter near the end of the alphabet, you are not going to get adequate credit for your co-authored work in economics. Anna J. Schwartz is the poster child for this problem. She gets almost no credit for co-authoring one of the most important economics books of the 20th century. Thus, it is vital to account carefully for co-authored publications.

Counting Consistently and Appropriately

The practice of author order varies by field. In the field of psychology, for example, the norm appears to be to list authors in order of their contribution. Consequently Einav and Yariv have found that, for psychologists, the probability of getting tenure does not depend on the where your name falls in the alphabet. Your last name can be Kahneman or Tversky, and you are equally likely to receive tenure, everything else being the same.

In the review process for psychology professors, the practice of privileging the first author is certainly appropriate. How much privileging does the first author deserve? Consider the three papers written by Richard C. above in the Thought & Expression. Let’s assume that the first author in this journal does 30 percent more work than the second author, and that the second author does 30 percent more work than the third author. Then Richard C. would receive 42.4 percent of the credit for the paper on food security, 32.6 percent of the credit for the paper on Papua New Guinea and 25.1 percent of the credit for the paper on Patagonia. That’s because Richard is the first author on the food security paper, second on the Papua New Guinea paper and third on the Patagonia paper.

In the field of economics today, the vast majority of journal articles are co-authored. Since peer-reviewed journal articles are the gold standard in the field, it’s crucial that we evaluate single-author and multiple-author journal articles properly in tenure, promotion and review processes. All these publications should obviously count in the review process. But we need to count them appropriately and consistently. On average, single-author publications should be weighted twice as much as two-author publications, and three times as much as three-author publications. And we should not privilege the first author unless it is appropriate to do so. This holds for men and for women, for white faculty and for minority faculty. Otherwise we will end up further stacking the academic deck unfairly.