The outcome of predation interactions between growing, size-structured predator and prey cohorts is difficult to predict. We manipulated the food resources available to juvenile spot subject to predation from southern flounder in a 60-day replicated pond experiment to test the hypothesis that spot growing slowly would experience higher predation mortality and stronger selection against small individuals than those growing rapidly. A nearly threefold difference in average growth rate between fast- and slow-growth treatments led to twofold higher predation mortality of slow-growing spot. Relative to no-flounder controls, larger spot were overrepresented at the end of the experiment in both treatments, but the magnitude of flounder size selection was much greater in the slow-growth treatment. The experimental results agreed qualitatively, but not quantitatively, with predictions from a prior size-dependent foraging model. In particular, the model significantly underestimated observed shifts in spot size structure to larger sizes. We hypothesized that competitive release and associated increases in spot growth due to thinning by flounder might reconcile this difference, and extended the model to incorporate this process. We then used the model to estimate the relative contribution of these two confounded predator effects (size-selective predation and thinning) to observed shifts in spot size structure. Model simulations indicated that the combined effects of size-selective predation and thinning could account for nearly all of the observed shift in spot size structure, but that thinning was the more important process. Our results highlight the utility of combining experimental and modeling approaches to unravel the complexities underlying interactions between growing, size-structured predator and prey cohorts.