This paper presents an empirical investigation into whether the decision to form a link with a node takes into account how well connected that node is. Given data in the form of a random sample from a network, this paper proposes a novel way to measure the degree of a node to adjust for mismeasurement and also controls for the endogeneity of this variable. It is shown that in fact the probability of forming a link with a node is increasing in the links received by that node but decreasing or unaffected by the links made by that node.