Relying on efficiency analysis, we evaluate to what extent policy makers have been able to promote the establishment of consolidated and comprehensive research groups to contribute to the implementation of a successful innovation system for the Spanish food technology sector, oriented to the production of knowledge based on an application model. Using data envelopment analysis techniques that allow calculation of a generalized version of the traditional distance function model for productive efficiency, we find pervasive levels of inefficiency and a typology of different research strategies. Among these, in contrast to what has been assumed, established groups do not play the pre-eminent benchmarking role; rather, partially oriented, specialized and ?shooting star? groups are the most common patterns. These results correspond with an infant innovation system, where the fostering of higher levels of efficiency

and the promotion of the desired research patterns are ongoing.