Peter Rousseau comments on our field experiment involving the Econ job market, in PNAS

 My post yesterday was about the experiment about social media and the job market for economists.  I only noticed later that the PNAS also posted a comment on our article, by Professor Peter Rousseau, the secretary of the American Economic Association, who has a long and intimate familiarity with that job market, which the AEA has played a giant role in organizing.

 Improving the job market in economics (and beyond…) by Peter L. Rousseau  PNAS   May 4, 2026  https://doi.org/10.1073/pnas.2609971123

Here is the part of his comment directly connected to our paper:

" the authors make a welcome and useful contribution to the market design literature with a fascinating experiment designed to substitute for and even improve upon the informal information channels lost to the economics job market in the new postpandemic normal. Given that some job candidates are less active self-promoters than others and that, conversely, excessive self-promotion can in some cases be viewed as a negative by prospective recruiters, the authors’ proposed mechanism offers serious promise for leveling the playing field, even if just modestly, for economics job candidates in terms of their visibilities, and perhaps even for expanding the number of jobs actually filled over the course of a recruiting season.
 

"In the experiment, an AI-based algorithm, supplemented with some human checking and reassignments, matched selected economists on social media (i.e., the “influencers”) with willing job candidates based on the closeness of their research. About 43 percent of willing candidates were selected for this treatment. The key to the experiment lies in the matches themselves, which were assigned in a manner that did not take the relative prominence or institutional ranking of an influencer directly into account. All candidate participants were invited to post a tweet about their job market papers on a social media site created for this purpose, and the influencers were asked to post neutral quote-tweets about the members of the treatment group to which they had been assigned. If executed according to design, recruiters viewing the quote-tweets receive information about the closeness of a given candidate’s research interests to those of the influencer. This may function as a partial substitute for the painstaking process of deducing such information across the hundreds of application packets that recruiters receive with only a brief period for making initial decisions. Knowing that a candidate’s research is close to that of Professor “X” is a tangible signal that could make that candidate more likely to be interviewed or receive a campus flyout or job offer from an institution seeking an entry-level economist like Professor X. The experiment indicates that individuals in the treatment group did indeed receive more campus visits and job offers than candidates assigned to the control group, and that the effect on job offers was especially strong for women. It also finds, however, that these effects were more pronounced for candidates matched to influencers with relatively higher citation counts than for those matched to influencers with relatively more followers, as these two measures of prominence in the profession are not that highly correlated. 

" The question of scalability then becomes paramount. Considering the experiment’s positive findings, it is natural to assume that, if universally available, all job candidates would choose to participate and receive the treatment. The process would otherwise go on as stated with perhaps additional influencers being selected by the organizers to serve the larger pool of candidates. Two observations seem reasonable at this point: first, in such a setup, better information about matches could lead to more open positions being filled, which would be a better aggregate outcome; and second, the treatment might in practice benefit candidates from outside the very top departments the most. This is because candidates from the highest ranked departments, who are often perceived by recruiters as having a higher probability of eventually becoming a star, will typically receive more interviews, campus visits, and offers, but in the end can still only accept one offer. With an enlarged set of viable matches, this means that some candidates who may have been otherwise overlooked will find jobs. Of course, the job market may take longer to clear under this mechanism as candidates will have more options to consider before departments go to second or third rounds of offers.
 

"Casual observations of the job market among economics departments and their chairs do suggest that a number of recruiters are unable to fill positions they have posted. The AEA does not currently collect information on just how many, but the very existence of the “AEA Job Market Scramble,” where recruiters and unmatched candidates can post their availabilities on an online message board each March, is indicative of the challenge (3). The design of a job signaling mechanism by the AEA and its implementation in December of each year (4), where job candidates can list two departments to which they would like to express interest in an interview, is another such intervention aimed at easing the congestion.
Another interesting result is that women appear to benefit most from the treatment, while this benefit does not extend to members of other groups traditionally underrepresented in economics. The authors point to existing evidence indicating that women on average tend to be less active promoters of their own research on social media than others and suggest that the additional visibility provided by the quote-tweets could be leading to more job offers. This potential channel, of course, could also be viable for any candidate with a tendency to self-promote less. To explain a special advantage for women, one could note the possibility of forces in the 2022–2023 job market where departments seeking to improve the gender balances of their faculties became aware of candidates through the mechanism who they may have otherwise overlooked. If this is the case, the next question to ask is why does the effect not carry over for members of other underrepresented groups? The answer, though no doubt a speculative one, may lie in the preexistence of other mechanisms and informal channels for promoting such candidates, rendering the marginal effects of the authors’ particular intervention not statistically significant.
Finally, while having the potential to increase the number of matches and raise their average quality, the effects of the authors’ intervention will be subject to some randomness based on the assignment of a given candidate’s influencer. For example, when any influencer posts a quote-tweet about a candidate who has been independently and objectively determined to have close research interests, that candidate’s post tends to receive more views and likes on X than those in the control group, and the extent of this visibility correlates with the size of the influencer’s following. Yet these effects do not seem to transfer downstream to job outcomes, where candidates receiving quote-tweets from highly cited influencers are the ones tending to see more offers. In a real sense, the adage “all publicity is good publicity,” often applied to economics research, may not be always true. The assignment of influencers to candidates, even if randomized, will matter for individual outcomes even though the aggregate effects of the intervention are positive. Given the potential individual benefits compared to nontreatment, however, job candidates would likely embrace the residual uncertainty and participate in the mechanism.
 

"The intervention designed by Qiu et al. may hold even greater promise outside of the economics discipline. In the natural sciences, for example, recruiting for scarce academic postdoctoral positions among new PhDs at a similar career stage, which are markets typically saturated with candidates, often moves directly to a very limited allocation of campus visits based in no small part on letters and other communications from mentors, some of whom could be less than ideally matched with their students or less well known than would-be assigned influencers. These cases are ones in which an enhanced visibility of candidates, when coupled with independent information about the closeness of their work to what senior researchers and their groups might be seeking, could lead to the greater advancement of science more generally.
 : 
"Competing interests P.L.R. has served since 2012 as Secretary-Treasurer of the American Economic Association, a 501(c)(3) non-profit deeply committed to improving the job market for new Ph.D. economists, and for which one of the companion article’s co-authors (A. E. Roth) served as President in 2017.

#######

 Peter's comment and our paper appeared online, but won't appear in print until next week in the May 12, 2026 | vol. 123 | no. 19 issue of PNAS.

 

Yesterday's post: 

Tuesday, May 5, 2026  Social media, job market outcomes, and ethics of field experiments, by Qiu, Chen, Cohn and Roth in PNAS

 

Social media, job market outcomes, and ethics of field experiments, by Qiu, Chen, Cohn and Roth in PNAS

 One of the fun things about our paper published in today's PNAS is that, as a working paper, it prompted a vigorous discussion of the ethics of doing field experiments in economics.  We discuss this more fully in the published version, below:

J. Qiu, Y. Chen, A. Cohn, & A.E. Roth, Social media promotion improves job market outcomes, Proc. Natl. Acad. Sci. U.S.A. 123 (19) e2528289123, https://doi.org/10.1073/pnas.2528289123 (2026). 

Abstract: Social media has transformed how academics disseminate research, but its effect on academic job outcomes remains unclear. Previous research has shown correlations between social media exposure and metrics like citation counts, but these relationships may be confounded by unobserved factors such as researcher quality or access to professional networks. We examine whether social media promotion causally affects job market outcomes in economics through a field experiment on Twitter (now X). We first collect tweets about job market papers from 519 candidates and post them from a dedicated account. We then randomize half of the posts to be quote-tweeted by established economists in the candidates’ fields, and measure the effects on both online visibility and hiring outcomes. We find that posts in the treatment group receive 441% more views and 303% more likes than those in the control group. Candidates whose posts were assigned to be quote-tweeted receive one additional flyout invitation compared to the control group average of 5.4 flyouts. Furthermore, women in the treatment group receive 0.9 more job offers than women in the control group, who receive 3 offers on average. Exploring mechanisms, we find that academic reputation drives these results, with stronger effects for quote-tweets from highly cited scholars and for candidates from top institutions. Our findings suggest social media promotion causally increases research visibility and improves academic job market outcomes.

 

  ...

"Ethical considerations.
"After the release of our working paper on the Social Science Research Network (SSRN) on May 20, 2024, a vigorous discussion arose on both social and mainstream media, particularly on Twitter, about the ethics of our experiment and of field experiments more generally (e.g., ref. 30). The main concern suggested that job markets are essentially constant sum, so that randomly promoting some candidates through having their JMPs quote-tweeted by influencers would necessarily (and unethically) disadvantage both those who were in the control condition of the experiment and those who did not participate in the experiment.
 

"We understand the importance of considering the ethical implications of any experiment and that ethicality is connected to the underlying economics of the job market. In this latter respect, given the information friction and congestion in the interview process, job markets are unlikely to be constant sum. Aside from the possibility of welfare gains from improved match quality, we note that, typically in matching markets, many employers fail to fill all their positions while at the same time qualified candidates fail to find one, so that welfare can also be improved by filling more positions. [In the 2022–2023 job market, the total number of jobs listed on JOE was 3,608, including 933 (1,083) full-time academic jobs in (outside) the United States and 718 full-time nonacademic jobs (any location). On the supply side, 1,386 Ph.D. students and postdocs applied to at least one job through JOE from August to December 2022 (31).] In economics, the job market often has unfilled positions by the end of February, leading to a scramble round each year starting in March. Similarly, the annual National Resident Matching Program (NRMP) for new physicians in the United States also leads to some positions being unfilled, despite having far more applicants than available positions. [For example, in 2024, 38,494 positions were offered to 44,853 active applicants and 2,510 positions were unfilled (6.5%), at the end of both the main match (a deferred acceptance algorithm, see ref. 32) and a centralized postmatch scramble called the Supplemental Offer and Acceptance Program (33).]
 

"The phenomenon of unfilled positions in a thick labor market may reflect congestion in the interview process. In such a market, since many positions receive more applications than the number of candidates who can feasibly be interviewed, the matching of interviews to jobs may be imperfect in the sense that an employer can find that none of the people interviewed can be successfully hired, but could have filled the position if more appropriate interviewees had been chosen. To mitigate this issue, signaling mechanisms have been introduced in both the economics and medical markets to facilitate a better matching of interviewees and employers (29, 34). In our context, the quote-tweeting of JMPs may similarly serve to help employers find better matches with their selection of interviewees who can be hired.
 

"We also propose that highlighting suitable candidates from underrepresented groups for a position could potentially expand the overall number of job openings. A notable example is the President’s Postdoctoral Fellowship Program, implemented across multiple institutions including the University of Michigan and the University of California system. This program seeks to recruit future faculty members “with the potential to bring to their research and undergraduate teaching the critical perspective that comes from their nontraditional educational background or understanding of the experiences of groups historically underrepresented in higher education.” (See, e.g., https://presidentspostdoc.umich.edu/, retrieved on August 29, 2025.)
 

"Finally, we consider trends in the broader context of job search in evaluating the ethical considerations related to our study. Social media has become a common channel for academics to advertise the JMPs of their students. Thus, we are not introducing a new channel for candidate promotion, nor are we excluding others outside of our experiment from availing themselves of this channel. Our goal is to understand the extent to which this channel may create visibility or improve outcomes for job candidates, especially since not all candidates may have equal access. Our paper belongs to the class of natural field experiment (35), a class that has seen a growing number of studies in which field experiments are used to assess the effects of market interventions. [A natural field experiment is one “where the subjects do not know that they are in an experiment” (35). In our context, participants were told only that we would arrange for their JMPs to be tweeted, but not that there would be a quote-tweet treatment.] One of the main benefits of conducting a natural field experiment is that it minimizes possible Hawthorne effects (36). These studies are widely accepted and even recognized, with the 2019 Nobel Prize for experiments in development economics. If it is ethical for economists to use experiments to evaluate interventions in other markets, it should also be ethical for economists to study the market for economists. And if it is ethical to promote students who are on the job market, then it should be ethical to study the effects of such promotion.
 

"In sum, from a normative perspective (should scholars promote candidates?), we argue that such promotion can reduce information friction and job market congestion, potentially leading to more efficient matching. From a positive perspective (does promotion matter?), we demonstrate in Results that it increases candidate visibility and improves job market outcomes, especially for women who are traditionally underrepresented in economics." 

#############

Earlier (a blog post about reference 30, above): 

Saturday, June 8, 2024  The ethics of field experiments in Economics, in the Financial Times

 

Market design search at WU Vienna's new Department of Business Analytics and Decision Science

 Ben Greiner writes to remind me that "Vienna is a beautiful place to live and work" and that there is a 27 July deadline for the first wave of recruiting for a new department, including two full professor slots for market designers.

" WU Vienna is establishing a new Department of Business Analytics and Decision Sciences, with a research focus on predictive and prescriptive analytics in support of data-informed strategic decision-making. This initiative reflects WU’s commitment to strengthening its academic profile at the intersection of analytics, artificial intelligence, and decision sciences.

As part of the department’s launch, approximately 24 new academic positions will be opened. These include 6 professorships, up to 6 tenure-track positions, 6 postdoctoral positions, and 6 pre-doctoral positions.

First call for professorships in June 2025

The first call launches on June 4, 2025 featuring 4 professorships.

These positions are distinguished by two different methodological orientations, with two different professorships per orientation:

• The first group focuses on candidates with a methodological focus on machine learning, symbolic or sub-symbolic AI (including deep learning, reinforcement learning, generative AI, and automated decision-making), or modern statistics and economicetrics.

• The second group focuses on candidates with a methodological focus in simulation, optimization, experimentation, algorithmic game theory, and/or market design.

To learn more about the call and application process, please visit the website for our job offerings.

 

Job search for Professor of Market Design: U. Mannheim and ZEW--Leibniz Centre for European Economic Research

Professor Achim Wambach writes with news of a job opening in market design:

Professor of Economics, Market Design (W3)
Department of Economics, University of Mannheim
 

"In a joint appointment process, the Department of Economics at the School of Law and Economics at the University of Mannheim and the ZEW – Leibniz Centre for European Economic Research in Mannheim are looking to fill the position of

Professor of Economics, Market Design (W3).

The job holder will be assigned to the Centre for European Economic Research immediately upon her/his appointment. She/he will remain a member of the department with a reduced teaching obligation of two credit hours per term (so called Jülicher Modell). The position is permanent.

We are looking for an individual who, thanks to her/his outstanding scientific qualifications, will strengthen both institutions' competencies in the analysis of markets and market design. She/he has proven her/his expertise in the game-theoretical analysis of market rules with a particular focus on auction or matching markets. The successful candidate has an affinity to field experiments and applied research.

The job holder will lead the Research Department 'Market Design' at ZEW, conduct research in market design and publish this research in internationally leading academic journals. She/he will also be responsible for third-party fundraising. The successful candidate should have experience in policy-advising, particularly in the practical application of market design. The position requires a distinguished academic record, demonstrated by high-level publications in international economic journals, and ideally experience by leading policy advisory projects. The candidate should also possess the ability to lead larger research teams and to effectively communicate research findings to a broader audience, including policymakers and the general public."

Harvard's lawsuit against the Trump administration

 Read it and weep for our country:

"7. Defendants’ actions are unlawful. The First Amendment does not permit the Government to “interfere with private actors’ speech to advance its own vision of ideological balance,” Moody v. NetChoice, 603 U.S. 707, 741 (2024), nor may the Government “rely[] on the ‘threat of invoking legal sanctions and other means of coercion . . . to achieve the suppression’ of disfavored speech,” Nat’l Rifle Ass’n v. Vullo, 602 U.S. 175, 189 (2024) (citation omitted). The Government’s attempt to coerce and control Harvard disregards these fundamental First Amendment principles, which safeguard Harvard’s “academic freedom.” Asociacion de Educacion Privada de P.R., Inc. v. Garcia-Padilla, 490 F.3d 1, 8 (1st Cir. 2007). A threat such as this to a university’s academic freedom strikes an equal blow to the research conducted and resulting advancements made on its campus.
8. The Government’s actions flout not just the First Amendment, but also federal laws
and regulations. The Government has expressly invoked the protections against discrimination
contained in Title VI of the Civil Rights Act of 1964 as a basis for its actions. Make no mistake: Harvard rejects antisemitism and discrimination in all of its forms and is actively making structural reforms to eradicate antisemitism on campus. But rather than engage with Harvard regarding those ongoing efforts, the Government announced a sweeping freeze of funding for medical, scientific, technological, and other research that has nothing at all to do with antisemitism and Title VI compliance. Moreover, Congress in Title VI set forth detailed procedures that the Government “shall” satisfy before revoking federal funding based on discrimination concerns. 42 U.S.C. § 2000d-1. Those procedures effectuate Congress’s desire that “termination of or refusal to grant or to continue” federal financial assistance be a remedy of last resort. Id. The Government made no effort to follow those procedures—nor the procedures provided for in Defendants’ own agency regulations—before freezing Harvard’s federal funding.
9. These fatal procedural shortcomings are compounded by the arbitrary and
capricious nature of Defendants’ abrupt and indiscriminate decision..."