In the USA, thousands of volunteers are engaged in tracking plant and animal phenology through a variety of citizen science programs for the purpose of amassing spatially and temporally comprehensive datasets useful to scientists and resource managers. The quality of these observations and their suitability for scientific analysis, however, remains largely unevaluated. We aimed to evaluate the accuracy of plant phenology observations collected by citizen scientist volunteers following protocols designed by the USA National Phenology Network (USA-NPN). Phenology observations made by volunteers receiving several hours of formal training were compared to those collected independently by a professional ecologist. Approximately 11,000 observations were recorded by 28 volunteers over the course of one field season. Volunteers consistently identified phenophases correctly (91% overall) for the 19 species observed. Volunteers demonstrated greatest overall accuracy identifying unfolded leaves, ripe fruits, and open flowers. Transitional accuracy decreased for some species/phenophase combinations (70% average), and accuracy varied significantly by phenophase and species (p < 0.0001). Volunteers who submitted fewer observations over the period of study did not exhibit a higher error rate than those who submitted more total observations. Overall, these results suggest that volunteers with limited training can provide reliable observations when following explicit, standardized protocols. Future studies should investigate different observation models (i.e., group/individual, online/in-person training) over subsequent seasons with multiple expert comparisons to further substantiate the ability of these monitoring programs to supply accurate broadscale datasets capable of answering pressing ecological questions about global change.