1. In clonal dioecious plants, the frequency and spatial distribution of flowering ramets contains information on the underlying genet sex ratio. These measures can also provide insight on potential ecological mechanisms causing variation and bias in sex ratios among populations. 2. We used a novel likelihood-based approach and spatial clustering model to estimate the genet sex ratios from flowering ramet data collected from 32 populations of dioecious Thalictrum pubescens, a clonal species from eastern N. America that occupies moist wetland and forested environments. We investigated sex ratios of seed families, clone size, patterns of flowering and plant height to determine potential causes of sex ratio bias. 3. Flowering ramet sex ratios varied considerably among populations but were significantly male-biased. Seed families grown to flowering also exhibited the same degree of male bias. Both models predicted close correspondence between ramet and genet sex ratios. The likelihood model revealed that gender differences in ramet production could not account for biased sex ratios. The spatial clustering model indicated that ramets were significantly clustered at two spatial scales and estimated similar cluster sizes and densities for both sexes. There was no evidence for spatial segregation of the sexes. Both sexes were equally likely to flower in consecutive years and repeated bouts of flowering had no effect on ramet height. 4. Synthesis. Our analyses suggest that the widespread occurrence of male-biased sex ratios in T. pubescens is unlikely to result from sexual differences in clonal growth or habitat preferences. The bias appears to become established early in the life cycle, perhaps at the seed stage as consequence of local resource competition.