Background/Question/Methods: Contact between individuals is well established as a key driver of disease transmission. However, contacts occur within a spatial environment, the context of which is likely to be important in structuring contact patterns. Many methods of social network analysis either ignore or control for spatial patterns, yet this may obscure social behaviour or infer relationships where ones don’t truly exist. First, we used a spatio-temporally explicit hierarchical model to explore whether spatial or social processes were a better predictor of exposure probability to a virus in a wild rodent species. I will also discuss some of the inherent difficulties in assessing association behaviour in small animals, and introduce a new technology we have been developing to answer some of these challenges. Results/Conclusions: We found that the probability of exposure was negatively correlated with total home range overlap (a proxy for contact), but positively correlated with contacts with other positive individuals. There was no evidence of spatial factors influencing exposure, suggesting that transmission in our study system is primarily through direct contact. Using an approach that accounts for spatial processes is useful when an indirect or environmental component to transmission is possible, or when the environment is highly heterogeneous. However, truly accounting for contact behaviour in small animals is extremely challenging and will require the adoption of novel technology, therefore I will also give some preliminary insights into a new proximity sensing tool we have developed.