Empirical work suggests that spatial sorting, a mechanism of evolutionary change fueled by phenotype-dependent dispersal, may lead to phenotypic shifts on ecological timescales. However, we currently lack a quantitative framework to measure the strength of spatial sorting. To address this gap, we present a quantitative genetics model of spatial sorting and identify an evolutionary parameter in the model to measure the strength of spatial sorting. This parameter, referred to as the standardized sorting gradient, is structurally akin to the standardized selection gradient, which is commonly used to measure the strength of natural selection. To show the utility of our approach, we analyzed binary wing-morphology data of soapberry bugs (Jadera haematoloma) recolonizing flooded habitats extirpated by a hurricane. We found that non-random recolonization resulted in a standardized sorting gradient of 0.90 {+/-} 0.08, which ranked in the top 5 percentile compared to standardized selection gradients documented in the scientific literature. Genetic crosses further revealed that the binary trait, short and long-wing forms, is heritable h2 = 0.65 {+/-} 0.13) and yielded a standardized evolutionary response of 0.58 {+/-} 0.13. Our results underscore that, like natural selection, spatial sorting, too, can yield rapid evolution, providing quantitative evidence that non-random dispersal may produce novel eco-evolutionary dynamics.