Causal analysis reveals that public access to land can shape our view of species' geographic distributions

•Aim: Citizen science data are widely used to study species' distributions. Using butterfly observations from South‐West England, we sought to better understand what shapes the spatial footprint of recording.
•Location: South‐West England.
•Methods: We combined directed acyclic graphs (DAGs) with piecewise path analysis to quantify the causal effects of public access and habitat (using land cover as a proxy) on the number of records for the 39 butterfly species that occur in this region.
•Result: The length of public footpath per unit area emerged as the strongest driver of local recording outcomes, especially for common species, while human population density and area of public access land show weaker or more variable effects. The total effect of public access features was larger than the total effect of habitat for 22 of the 39 species.
•Main Conclusions: Overall, our findings indicate that current butterfly occurrence data largely reflect where citizen scientists can go, rather than solely where species occur. Incorporating accessibility metrics can therefore improve the design and interpretation of citizen science recording schemes and surveys.