Reader Comments

Post a new comment on this article

Response to Ferlazzo & Sdoia

Posted by stefanoallesina on 30 Aug 2012 at 22:15 GMT

I have written a short response to the findings of Ferlazzo & Sdoia.

My manuscript (submitted to PLoS One) can be accessed here:
http://arxiv.org/abs/1208...

I show that the scarcity of last names in the UK disciplines is largely due to field-specific immigration. This is why fields of research in which the presence of immigrant researchers is relevant (such as Computer Science, Pure Mathematics and Applied Mathematics) display more names than expected, while for example low-immigration disciplines such as English Literature or Celtic Studies show a significant scarcity of last names.

I also show that the scarcity of first names in Italian disciplines is due to male/female representation.
The disciplines with lower p-values have few women professors:
Electronic Eng. 13% women
Industrial Eng. 14% women
Physics 18% women
Earth Sciences 26% women
Civil Eng. 27% women
while the disciplines with the highest p-values have more women than men:
Linguistics 66% women
Psychology 56% women
Education 54% women
Anthropology 53% women

Clearly, having very few female (or male) professors means having fewer first names.

I also perform analysis that strengthens the case for nepotism:
- Analyzing professors at the regional or macro-regional level yields comparable results.
- The scarcity of last names vanishes when considering only women (women maintain their maiden name, daughters take the last name of their father) while is relevant analyzing only men (where father-son pairs share the last name): academic positions tend to be "inherited" as last names are.

Stefano Allesina

No competing interests declared.

RE: Response to Ferlazzo & Sdoia

ferlazzo replied to stefanoallesina on 03 Sep 2012 at 18:32 GMT

A number of factors are called to account for accounting our results on the UK dataset (frequency distribution of names is different, common British names have higher frequencies than their Italian counterpart, immigrant researchers, gender composition). All these considerations show how demographic factors can determine the results of the method. Exactly what we wrote in our manuscript.
It is worth noting that the Allesina argues that a pattern of results in Italy is due to nepotism, while the same pattern in the UK is due to something else. However, it should be specified when a given result can be taken as a sign of nepotism and when the same result should instead be considered as due to demographic factors. The question is: what is the proportion of immigrant researchers, or of women, or of rare names that makes it possible to interpret the results of the analysis as due to nepotism or not? In other words, what are the boundary conditions?
Anyhow, we totally agree with the final remark Allesina proposes in his reply that the results of an analysis must be critically interpreted.
We are preparing a more detailed reply that will be submitted as soon as possible.

No competing interests declared.