Roost switching is a common occurrence in bats, yet the causes and consequences of such behavior are poorly understood. In this study we explore the ecological correlates of roost fidelity in the tent-making bat Artibeus watsoni, particularly focusing on the effect of sex, reproductive status, and roost availability using a three-factor general linear model (GLM). We estimated roost fidelity of radio-tracked individuals and found that the GLM was significant (R2 = 0.72, F10,34 = 8.91, p < 0.001). Significant interaction terms were observed for relative roost availability and sex (F4,34 = 16.96, p < 0.001), and relative roost availability and reproductive status (F6,34 = 7.62, p < 0.001), indicating that variation in roost fidelity among males and females, and among individuals under different breeding conditions, depended on relative roost availability at the site where they were radio-tracked. Individuals in areas of high roost availability exhibited lower roost fidelity than those sampled in areas of lower roost availability. Females exhibited less roost fidelity than males for all roost availability categories, but the difference between males and females was only significant at high roost availability. The general pattern of decreased roost fidelity as roost availability increased was also prevalent among individuals in different breeding conditions. Additionally, satellite males exhibited higher roost fidelity than resident males in areas of low roost availability, and lactating females had higher roost fidelity than non-breeding females in areas of medium roost availability. Our study thus demonstrates that sex, reproductive status, and roost availability all affect roost fidelity in the tent-making bat A. watsoni, and also suggests that roost availability is the most important factor influencing roost fidelity in this bat, providing the first quantitative evidence that roost fidelity is correlated with roost abundance in a single species.