Aggregate impact factor of scientific fields

Science journal impact indicators are not comparable because of inherent differences in publication and citation behaviors from field to field. A breakdown of the field aggregate impact factor of databases shows that for the 22 fields and four areas considered by Thomson Reuters, the leading provider of science indicators, five variables largely explain variance in impact factor of a given field. Therefore, it is necessary to consider all these sources of variance in the standardization process of the impact indicators. A Principal Component Analysis is employed to find the sources of the variance and a Cluster Analysis is used to detect similarities.