If the structure of an individual's waking-life social network influences the occurrence of people in dreams, the author would expect kin to be central in a dream social network. Another way to think about whether dream social networks are like waking-life social networks is to consider who is joined to whom. Because of these individual differences, assortativity values do not pigeonhole the entire class of dream social networks as similar or dissimilar to waking-life social networks. The second sort of random network is made by starting with the two-mode network for the people and the dreams of an individual. The author concludes that people are not generated in dreams completely at random but by a process that is systematic albeit unknown. More generally, the random walk is on all of the dreamer's associative network, not just the subnetwork for people, which the author focuses on here.