"Figures 1 and 2 show the distribution of impact factors for two subject categories of the ISI: “mathematics” (N = 191 journals) and “genetics” (N = 133). As noted, the journal selection and the ISI-categories are themselves debatable, but for the purpose of this argument the comparison between these two sets is revealing. The values of the x-axis differ not slightly, but by an order of magnitude!" (...) "The problem is well-known among scientometricians (e.g., Martin & Irvine, 1983). However, it has an important policy implication: one is not allowed to compare evaluations based on citation rates across disciplines. Consequently, it is not legitimate to allocate funds across fields of science on the basis of comparisons among citation rates for different disciplines (National Research Council, 2007). In other words, one has to disaggregate. The rules for the disaggregation, however, are far from obvious." (min fremhævelse, CE).
Eller, som Loet L. selv bøjer det i sarkastisk neon på en nyhedsliste:
"For example, administrators are well advised to close down the mathematics department for upgrading their university. Mathematicians do not cite and are not cited as much as geneticists. Thus, the average citation rate is comparatively lower with an order of magnitude."
Men lukker man helt ned for matematik, kan det jo være, at samfundet om 30 år mangler kompetencer nok til at indikere og evaluere noget som helst. Det er ikke så sært, at Videnskabsministeriet her i sommer så sig nødsaget til at udsætte offentliggørelsen af deres indikatormodel til at ranke de danske universiteter. Måske skulle ministeriets nye Kvalitets- og forskningsevalueringsenhed ansætte nogle matematiker og humanister...