A multilevel analysis of factors influencing local social interaction

Social interaction is an important aspect of people’s quality of life. Since social networks are becoming more spatially spread as a result of the development of transportation and communication technology, it is interesting to investigate to which extent local social interactions still take place. This paper therefore analyses to what extent local social interactions are affected by characteristics of the residential environment as well as personal and mobility attributes. The analyses are based on data collected in 2008 in Eindhoven and a number of smaller surrounding villages in the south of the Netherlands among 747 respondents. The data collection instrument consisting of a two-day social interaction diary was used to gather detailed information about the respondents’ interactions and the persons they contacted. Using these data a multilevel binomial logit model is estimated analyzing whether or not a social interaction took place with a local tie (someone within 1 km distance). The results indicate that the likelihood of interacting with a local tie is affected by personal characteristics (age, gender, household size, work), mobility characteristics (number of cars, commuting time), size of the community and the presence of facilities (supermarket, sports, public transport) in the residential area.