Many spatial social networks have the property that nearby nodes are more likely to be connected than are nodes that are farther apart. We develop a characteristic of spatial graphs that captures whether or not shorter distance ties are preferred over longer distance ties, and the degree to which this edge length bias occurs. This allows us to estimate what is far and what is close ï€ what we call neighborhood radius ï€ for any randomly generated spatial network. Results from Monte Carlo Markov Chain simulations presented similar distribution of edge length bias to data from personal networks from a neighborhood in New Orleans, Post-Katrina, although the latter presented greater variation.
