We designed a sediment bioassay using 25% growth inhibition of Hyalella azteca as the end point.Hyalella azteca exhibits size-specific fecundity, so growth is a surrogate of reproductive production. We investigated density effects on growth to address whether crowding could affect test interpretation; amphipods in 14,000/m² exposures were 16 to 20% smaller than those at 7,000/m². Using power analysis, we found that 20 to 25 samples are required to determine significance when α = 0.10 and 1 − β = 0.90. To minimize the need for laboratory resources, we designed a two-step (screening and confirmatory) bioassay, which we tested with field-collected sediments. The screening bioassay compared 11 sediments to a reference. Three sediments were “toxic” (significant growth inhibition when 1 − β = 0.66 and n = 5), five sediments were “nontoxic” (>90% of reference), and three sediments were “possibly toxic” (growth inhibition was insignificant). In the confirmatory bioassay, three possibly toxic and two nontoxic samples were reevaluated. Two were toxic (1 − β = 0.91 and n = 20), and the remaining four samples were nontoxic. In summary, five sediments were toxic and six sediments were nontoxic. The two-step analysis used minimal laboratory resources but maximized statistical power, where needed, to discriminate growth effects.