Ontogenetic changes in gregariousness by pre-reproductive animals, like that observed in juvenile Caribbean spiny lobsters (Panulirus argus), may be adaptive and reflect size-specific changes in the effectiveness of aggregation in promoting survival. Alternatively, aggregation may simply result from changes in the distribution or availability of suitable habitat structure, or from other behaviors that enhance survival. There are currently two hypotheses explaining the potential benefits of gregarious behavior in juvenile spiny lobsters, both of which focus on increasing survivorship by reducing predation pressure: the group benefit hypothesis and the guide hypothesis. The group benefit hypothesis argues that aggregations of juvenile lobsters reduce individual susceptibility to predators because groups are better able to fend off attackers or benefit by dilution of risk. The guide hypothesis suggests that aggregation is a consequence of shelter seeking behavior, in which individuals searching for shelter follow conspecific odors, thus reducing the time they spend in the open exposed to higher predation rates. The guide mechanism should be most effective in areas of low shelter density. We used an individual-based, spatially-explicit model describing recruitment of juvenile spiny lobster in the Florida Keys to compare behavioral models incorporating a guide effect and group benefit under conditions of high and low shelter densities. We found that the guide effect significantly enhanced survival only under the most extreme circumstances where shelter was scarce, the risk of predation highest, and the effective distance of the guide effect strongest. In contrast, small increases in direct group benefit led to significantly higher population abundances under a wide range of conditions.