Linking plants trait to community assembly is an important area of community ecology but quantitative studies are nearly absent from the grassland biome of North America despite the large extent and the high research activity in this region. There are numerous plant trait-performance studies from grasslands that generally report very weak results. I hypothesized that a competitive index based on three key traits (plant height, specific leaf area, plant density) would predict competitive success during community assembly. I tested this idea in a community assembly experiment that included manipulation of soil heterogeneity, seeding patterns, and patch size over a five-year period. I used patch data from treatments where a single species was sown into individual 20x20 cm patches and measured species performance between years 2-5 of the experiment. I found significant relationships between the competitive index in both presence probability and abundance patterns in each year of the analysis. Interestingly, single or combined direct measures of plant traits failed to explain abundance or presence probability. These findings suggest that a competitive index may be a more effective way to link plant traits and community assembly and may offer a promising new approach for trait-plant performance investigations.