Measuring Nepotism Through Shared Last Names: Response to Ferlazzo and Sdoia

TL;DR

This paper investigates the presence of nepotism in Italian academia by analyzing last name scarcity, addressing previous critiques, and demonstrating that geographic, demographic, and gender factors do not fully explain the findings, thus supporting nepotism evidence.

Contribution

The study refines the analysis of last name scarcity as an indicator of nepotism, accounting for various confounding factors and providing stronger evidence of nepotism in Italian academic disciplines.

Findings

Last name scarcity persists after controlling for geographic and demographic factors.

Higher nepotism likelihood in southern Italy and Sicily.

Inheritance patterns of academic positions from father to son support nepotism evidence.

Abstract

In a recent article, I showed that in several academic disciplines in Italy, professors display a paucity of last names that cannot be explained by unbiased, random, hiring processes. I suggested that this scarcity of last names could be related to the prevalence of nepotistic hires, i.e., professors engaging in illegal practices to have their relatives hired as academics. My findings have recently been questioned through repeat analysis to the United Kingdom university system. Ferlazzo & Sdoia found that several disciplines in this system also display a scarcity of last names, and that a similar scarcity is found when analyzing the first (given) names of Italian professors. Here I show that the scarcity of first names in Italian disciplines is completely explained by uneven male/female representation, while the scarcity of last names in United Kingdom academia is due to discipline-specific immigration. However, these factors cannot explain the scarcity of last names in Italian disciplines. Geographic and demographic considerations -- proposed as a possible explanation of my findings -- appear to have no significant effect: after correcting for these factors, the scarcity of last names remains highly significant in several disciplines, and there is a marked trend from north to south, with a higher likelihood of nepotism in the south and in Sicily. Moreover, I show that in several Italian disciplines positions tend to be inherited as with last names (i.e., from father to son, but not from mother to daughter). Taken together, these results strenghten the case for nepotism, highlighting that statistical tests cannot be applied to a dataset without carefully considering the characteristics of the data and critically interpreting the results.