The “Great Gatsby Curve” describes the positive correlation between “income inequality” and “intergenerational income persistence” (lack of income mobility). An academic Great Gatsby curve refers to a positive correlation between academic inequality and intergenerational persistence.

The existence and extent of an academic Great Gatsby curve is the subject of a new study by Ye Sun, Fabio Caccioli, Xiancheng Li, and Giacomo Livan published in the Journal of the Royal Society Interface.

The authors take “academic inequality” to refer to “the uneven distribution of opportunity and academic impact”, operationalized in terms of volume of citations, and “intergenerational persistence” as “the influence that a mentor’s status may have on their protégés’ academic success.” That latter quality is determined by measuring “the similarity between the positions of mentors and their mentees in the impact rankings of their discipline: the higher the rank–rank correlation, the more a mentee’s scientific impact is correlated to that of their mentor, the higher the intergenerational persistence.”

They analyzed data on over 300,000 academics across 22 disciplines who in total published close to 10 million papers between 2000 and 2013 and found that an academic Great Gatsby curve indeed exists. They also found that, by their measures, both academic inequality and academic intergenerational persistence were on the rise during the studied period.

A breakdown of the correlation by discipline finds it in philosophy, which, according to the study, has both the most academic inequality and the least academic mobility:

The academic Great Gatsby Curve: More inequality is associated with more impact persistence across academic generations. Citation inequality is measured by the Gini coefficient, using the cumulative number of citations authors (including both mentors and mentees) have received from the papers published within a 5-year time window before the final mentorship year, within 5 years after publication. Mentor–mentee pairs are pooled over time per cohort (with the same final mentorship year) to compute the impact persistence and inequality metrics. The error bars in vertical and horizontal directions refer to the standard deviation of the mean over cohorts. The point size of each discipline is proportional to the number of mentor–mentee pairs considered in our analysis. The solid line and the shaded area represent the regression line (with annotated Pearson’s 𝑟 and 𝑝-values) and the 95% confidence intervals, respectively. ***p < 0.01, **p < 0.05, *p < 0.1. Based on figure 3 from “The Academic Great Gatsby Curve” by Ye Sun, Fabio Caccioli, Xiancheng Li and Giacomo Livan.

One conclusion the researchers draw is that “academic impact—as quantified by citations—is to some extent inherited. As such, citation-based bibliometric indicators should be handled with care when used to assess the performance of academics.”

They also comment on the analogical basis of their study, and its limits:

We find that academia is not immune from the phenomenon of intergenerational persistence, which has been widely documented in the social sciences across dimensions such as income, wealth and occupation. We examined intergenerational academic persistence by analogizing academic mentors and mentees to parents and children, and academic impact (as measured with citations) to income. The persistence of income through genealogical generations and the persistence of impact through academic ones both reflect the transmission of resources and status, and they capture the extent to which the success of one generation may depend on that of the previous one. However, while there is a clear analogy between the mechanisms at play in these two contexts, there are also obvious differences. On the one hand, both mechanisms involve the inheritance of a network, be it social, professional or both. On the other hand, the transfer of economic status is—to a good extent—mechanistic, as it is grounded upon the inheritance of wealth. The transfer of academic status is instead grounded upon the inheritance of intangibles, such as knowledge and visibility. Additionally, an important distinction lies in the fact that mentor–mentee pairings are determined by choices made by both parties (as opposed to parents and children). More successful mentors may have the privilege of being more selective in their choice of mentees, and vice versa, leading to a positive correlation between their impact, even in the absence of causal effects.

You can check out the study here.