International Choice Modelling Conference, International Choice Modelling Conference 2017

Font Size: 
Maintaining social contacts over time: the role of life-course events, personality traits and tie strength
Chiara Calastri, Stephane Hess, Juan Antonio Carrasco, Charisma Choudhury, Andrew Daly

Last modified: 28 March 2017


Socialising is a basic human trait that prompts people to interact with different social contacts and in different forms. The study of social interactions is gaining interest as their impact on different decisions is being recognized (e.g. Grønhøj  et al. 2011; Sunitiyoso et al. 2011). The study of social networks and their impact is highly cross-disciplinary. Our research is indeed aimed at gaining a deeper understanding of network dynamics, which could then be used in further social networks studies as well as in work investigating social influence on behaviour.

The understanding of social network characteristics and of the processes by which people establish social ties has mainly been studied in sociology (Kossinets & Watts 2006), social network analysis (Hammer 1979) and recently within a random utility framework (Arentze et al., 2013). The evolution of social network structures has been investigated in a more limited number of studies (as appropriate datasets are rare) using egocentric network datasets including repeated observations (e.g. Wellman et al. 1997; Sharmeen et al. 2016). Wellman et al. (1997) and Sharmeen et al. (2015) associated long-term changes such as relocating or getting married to changes in the composition of personal networks.

In the present study, we aim to fill in this research gap by developing a hybrid choice model to predict the maintenance and loss (removal) of members in the social network using information on life-course events, demographics, personality traits (latent) as well as  strength of the relationship (latent).
The Chilean Communities in Concepciòn dataset includes a “name generator” and a “name interpreter”, instruments through which respondents (“egos”) are asked to report the names and socio-demographics of their network members (“alters”). Both waves of the survey (collected in 2008 and 2012) contain social network information, data on frequency of interaction as well as ego socio-demographics and a time-use diary. 105 people took part in both waves, of which 94 completed the name generator in both occasions and have some degree of overlap in the network. We investigate the changes in the social network over 4 years by testing the behavioural framework illustrated in the PDF version of the abstract.

Each “alter” reported in the first wave can be either maintained or lost in the second wave: this is the dependent variable in our model. We test the effect of observed “ego” life-course events and characteristics of the “alters” on the utility of maintaining a contact. We also assume that there are unobservable elements involved in this decision: sociability or extraversion is considered to be one of the five main personality traits (John & Srivastava, 1999). This trait, representing the extent to which a person is sociable, talkative, NOT reserved or shy (Gosling et al., 2003) has been measured through a 6-item scale. We use the responses to these items as indicators of sociability and test its impact on utility.

We also test the hypothesis that the maintenance of relationships over time depends on the strength of the contact: if two people are emotionally close, their relationship will more likely be maintained over the years. The concept of “strength of a tie”, though, cannot be objectively measured. Although egocentric network surveys (like the present one) ask people to assess the strength of their relationships, this concept might be subjectively interpreted and have different meanings for different people. For this reason, we introduce “tie strength” as a latent construct in the model.
The literature found emotional closeness, frequency and duration of contact, breadth of discussion and confiding (Marsden & Campbell, 1983), as well as exchange of favours and help (Wellman et al. 1997) to be indicators of tie strength. We use stated emotional closeness, mode-specific communication and the mutual dependence (i.e. whether help is granted or received to/by each alter) as indicators of tie strength and attempt different model specifications to distinguish between indicators and predictors of tie strength. We also acknowledge that some measures of homophily may affect interactions directly as opposed to through tie strength, for example distance.

Differently from previous work, we suggest an integrated framework which examines the decision to maintain a social contact and the causal role of both life-course changes and latent personality and tie strength constructs. This comprehensive representation of the decision process is expected to give a picture of the dynamics of network composition able to accommodate the complexities of real-life behaviour.


Arentze, T. A., Kowald, M., & Axhausen, K. W. (2013). An agent-based random-utility-maximization model to generate social networks with transitivity in geographic space. Social Networks, 35(3), 451-459.

Gosling, S. D., Rentfrow, P. J., & Swann, W. B. (2003). A very brief measure of the Big-Five personality domains. Journal of Research in personality,37(6), 504-528.

Grønhøj, A., & Thøgersen, J. (2011). Feedback on household electricity consumption: learning and social influence processes. International Journal of Consumer Studies, 35(2), 138-145.

John, O. P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and theoretical perspectives. Handbook of personality: Theory and research2(1999), 102-138.

Kossinets, G. and Watts, D. J. (2006). Empirical Analysis of an Evolving Social Network. Science, January 2006, 88-90

Marsden, P. V., & Campbell, K. E. (1984). Measuring tie strength. Social forces, 63(2), 482-501.

Hammer, M., Predictability of social connections over time, Social Networks, Volume 2, Issue 2, 1979, Pages 165-180.

Sharmeen, F., Arentze, T. and Timmermans, H. (2015) Predicting the Evolution of Social Networks with Life Cycle Events. Transportation.42 (5): 733-751 Special issue: Social Networks and Transport.

Sharmeen, F., Chávez, O., Carrasco, J.A., Arentze, T. and Tuleda, A. (2016). Modelling population-wide personal network dynamics using a two wave data collection method and an origin-destination survey. Presented at the 2016 Annual Meeting of the Transportation Research Board, Washington.

Sunitiyoso, Y., Avineri, E., & Chatterjee, K. (2011). The effect of social interactions on travel behaviour: An exploratory study using a laboratory experiment. Transportation Research Part A: Policy and Practice, 45(4), 332-344.

Wellman, B., Wong, R. Y. L., Tindall, D., & Nazer, N. (1997). A decade of network change: Turnover, persistence and stability in personal communities. Social networks, 19(1), 27-50.

Conference registration is required in order to view papers.