We conducted a two-wave personal network study in a rural Romanian community, interviewing the same participants (n = 68) using two name generators. Wave 1 employed a fixed-choice generator (n = 25) focused on emotional closeness; Wave 2 used a free-choice generator based on frequent interaction. We compared tie characteristics and assessed retention across waves. Alters who were kin, co-residents, or emotionally close were more likely to be retained, regardless of generator type. These findings underscore the role of relational attributes in personal network stability and highlight design considerations for network studies in resource-limited, culturally distinct settings.