The president and vice-president of the Italian National Agency for the Evaluation of Universities and Research Institutes (ANVUR) claim that Italy’s rise in international research-impact rankings is a real effect (P. Miccoli and R. I. Rumiati Nature 574, 486; 2019), and not (as we have argued) the result of Italian scholars citing one another’s articles more heavily (see Nature http://doi.org/dcgj; 2019). We question their evidence for this claim.
First, they say that scientific productivity in Italy has risen in the past decade, possibly stimulated by the introduction of performance-related university funding. More articles are indeed being published, but the yearly growth rate of Italy’s scientific production has in fact slowed down since the introduction of performance-related targets in 2012, according to ANVUR’s own statistics (see go.nature.com/34ms9n; in Italian).
Second, they state that ANVUR recognizes the importance of correcting gaming behaviours, including self-citation. They point out that, in an evaluation of 2011–14 work, the agency established a criterion for ‘downgrading’ papers in which self-citation exceeded a given threshold. ANVUR’s own reports, however, show that this downgrading was never applied (see go.nature.com/2jn2si; in Italian).
In our view, ANVUR’s claim needs to be grounded more in fact and less in aspiration.
Nature 576, 213 (2019)