Abstract
The present paper articulates the many ways social network researchers conceptualize and operationalize network influences on people's behaviors. Four categories of network influences are described: (1) personal network, (2) positional, (3) network-level, and (4) individual network-level interactions. Personal network effects are based on data from the individual's direct and indirect contacts. Positional effects are derived from the individual's position in the network such as being central or peripheral. Network-level effects are measured using network-level indicators such as centralization or clustering. Interaction effects occur when there is consideration of both the individual and network level measures such as understanding the influence of being in a central position in a centralized or decentralized network. These various network effects are contrasted with contagion, which is the most frequently used mechanism for diffusion. One conclusion drawn from this review is that when we invoke contagion explanations, we give agency to the product, whereas when we invoke interpersonal influence explanations, we give agency to people and social systems.