Methods: Findings come from longitudinal research with 421 homeless adults moving into PSH in the Los Angeles area. Baseline interviews were conducted prior to or within 5 days of moving in and the first follow-up interview at 3 months after move-in. Using hierarchical linear modeling techniques, these analyses investigate the association between respondent-level (e.g., gender and race/ethnicity) and dyadic-level (e.g., substance use homophily, strength of network tie) characteristics and: 1) the likelihood of an alter (i.e., network member) being retained in a social network before and after PSH move-in, and 2) the likelihood of an alter newly entering a network after PSH move-in.
Results: On average, respondents sustained connections with one-third of network members between baseline and 3-month follow-up. Members most likely to be sustained after initial PSH move-in were relatives, romantic partners, staff (e.g., case managers, medical providers), and persons with whom the respondent shared a strong tie (tie strength is a continuous measure including frequency of contact, length known, and several indicators of emotional and tangible support). Members likely to be lost after PSH move-in were those originally met on the street, and those who lived nearby while the respondent was still homeless. Overall, African-American respondents and respondents who identified as “another race/ethnicity” (not white, African-American, or Latino/Hispanic) were more likely to lose network members, as were military veterans.
Models assessing which network members are likely to be gained (that is, newly included in the social network after PSH move-in) identified the strongest member characteristic correlate was living in the respondent’s building. African-American respondents and military veterans were more likely than other respondents to gain new network members, while male respondents were less likely to gain members.
Conclusions and Implications: These findings demonstrate volatility in social networks as persons move into PSH; in particular, respondents are likely to report the loss of persons met on the streets, and gains in those who live in a respondent’s building. We also see demographic disparities in network consistency, with African-American and veteran respondents reporting more network turnover during the transition to housing. Because social networks -- particularly networks comprised of members modeling pro-social behavior and with whom the respondent has strong ties -- can impact individual risk behavior, interventions focused on gaining and/or maintaining such members may be important for improving the health and wellbeing of persons living in PSH.