The number of Marine Protected Areas (MPAs) is continuously increasing worldwide because of the increasing recognition of the profound effects that humans can have on marine systems. A large body of literature deals with marine reserves and their potential as conservation and management tools. In several cases, empirical evidence demonstrated that reserves can harbour greater diversity, higher abundance, and larger organisms than unprotected areas. In most cases, however, reserve design and site selection involved little scientific justification, with no direct test of most of the mechanisms assumed to work in a marine reserve. Field investigations of subtidal marine reserves are generally confounded by intrinsic ecological differences between sites investigated inside and outside reserves, by a lack of site and reserve replication, or by the absence of information about the biota before reserve establishment. This is particularly true in the Mediterranean Sea. The aim of this paper is to show that, at least in the Mediterranean basin, the effectiveness of MPAs has been rarely demonstrated because of lack of appropriate sampling designs. An MPA can be considered as a zone subjected to human impact, presumably a positive one. As a consequence, we propose the use of experimental procedures generally utilised for detecting environmental impacts.