Marine protected areas (MPAs) are a well-established conservation practice worldwide, but their effectiveness in protecting or replenishing fish biodiversity remains uneven. Understanding the patterns of this heterogeneity is central to general guidelines for MPA design and can ultimately provide guidance on how to maximize MPA potential. Here, we examine associations between the degree of protection, duration of protection, and protected area size, with fish biodiversity inside of protected areas relative to that of sites nearby, but outside of protected areas. We quantitatively synthesize 116 published estimates of species richness from 72 MPAs and 38 estimates of Shannon entropy from 21 MPAs. We show that species richness is on average 18% (95% CIs: 10%–29%) higher in protected areas than in areas open to fishing; on average, Shannon entropy is 13% (95% CIs: −2% to 31%) higher within protected areas relative to outside. We find no relationship between the degree and duration of protection with the ratio of species richness inside versus outside of protected areas; both fully and partially protected areas contribute to the accumulation of species inside of protected areas, and protected areas of all ages contribute similarly on average to biodiversity conservation. In contrast to our expectations, increasing protected area size was associated with a decreased ratio of species richness sampled at sites inside versus outside of the protected area, possibly due, for example, to insufficient enforcement and/or low compliance. Finally, we discuss why meta-analyses such as ours that summarize effect sizes of local scale biodiversity responses, that is, those at a single site, can only give a partial answer to the question of whether larger protected areas harbor more species than comparable unprotected areas.
