Extended Data Fig. 4: Effects of prestige.
From: Quantifying hierarchy and dynamics in US faculty hiring and retention
Logistic regression coefficients, expressed as a change in log-odds of faculty being a self-hire (a), being a non-U.S. faculty (b), or a woman (c) for a one-decile increase in prestige, stratified by domain (colours) and academia (blue), for newly hired faculty (filled symbols) and for existing faculty (hollow symbols) and connected by a line. Circles, significant (two-sided t-test, Benjamini-Hochberg corrected p > 0.05); crosses, not significant. (a) Existing faculty are more likely to be self-hires at more prestigious universities, but this effect attenuates or disappears for new hires, indicating that the positive relationship between self-hiring and prestige is likely driven by attrition. (b) Newly hired faculty are more likely to hold a non-U.S. doctorate than existing faculty. This likely results from higher rates of attrition among faculty with a non-U.S. doctorate (Fig. 1c). (c) We observe no universal relationship across domains between prestige and gender, but both new and existing faculty are somewhat more likely to be men as prestige increases for academia as a whole
