In this work, we used a model assay system (polyclonal human IgG-goat antihuman IgG) to elucidate some of the key factors that influence the analytical performance of bioassays that employ metal-enhanced fluorescence (MEF) using silver nanoparticles (NPs). Cy5 dye was used as the fluorescent label, and results were compared with a standard assay performed in the absence of NPs. Two sizes of silver NPs were prepared with respective diameters of 60 +/- 10 and 149 +/- 16 nm. The absorption spectra of the NPs in solution were fitted accurately using Mie theory, and the dipole resonance of the 149-nm NPs in solution was found to match well with the absorption spectrum of Cy5. Such spectral matching is a key factor in optimizing MEF. NPs were deposited uniformly and reproducibly on polyelectrolyte-coated polystyrene substrates. Compared to the standard assay performed without the aid of NPs, significant improvements in sensitivity and in limit of detection (LOD) were obtained for the assay with the 149-nm NPs. An important observation was that the relative enhancement of fluorescence increased as the concentration of antigen increased. The metal-assisted assay data were analyzed using standard statistical methods and yielded a LOD of 0.086 ng/mL for the spectrally matched NPs compared to a value of 5.67 ng/mL obtained for the same assay in the absence of NPs. This improvement of approximately 66x in LOD demonstrates the potential of metal-enhanced fluorescence for improving the analytical performance of bioassays when care is taken to optimize the key determining parameters.