Biotic homogenization, with its emphasis on invasions, extinctions, and convergence in taxonomic similarity, provides an important framework for investigating changes in biodiversity across scales. Most research thus far has focused on documenting patterns of biotic homogenization; less emphasis has been placed on identifying drivers. Through their selective foraging, large populations of white-tailed deer are altering population sizes, driving extirpations, and facilitating invasions of plants throughout the eastern United States. I tested the hypothesis that they can also drive biotic homogenization in forest understory communities. I studied the effects of 18 years of deer exclusion in the ground layer vegetation and composition of a hemlock-northern hardwood stand in N Wisconsin. ANOVA reveals that percent cover is four times higher in deer-excluded plots, and diversity is greater. Within blocks, average pairwise similarity in species composition (based on Jaccard and Bray-Curtis metrics) is highest among transects within deer exclosures and lowest among transects in treatment and control areas. This seems to imply deer exclusion facilitates biotic homogenization. Examining patterns among blocks, however, reveals control plots exhibit higher pairwise similarity values than deer excluded plots. This reflects biotic homogenization. These findings suggest deer exclusion reverses biotic homogenization at the stand level, but it is not clear how these patterns might change at different spatial or temporal scales.