Title: Small-world phenomenon and strategies for making friends on social networking sites in mobile environment: random and non-random
Authors: Bo-Chiuan Su; Tsz-Shiuan Yen
Addresses: Department of Information Management, National Dong Hwa University, No. 1 Sec. 2, University Road, Shoufeng, Hualien, Taiwan ' Department of Information Management, National Dong Hwa University, No. 1 Sec. 2, University Road, Shoufeng, Hualien, Taiwan
Abstract: The ubiquity of social networking sites (SNS) is astounding. As of 2016, Facebook had over 1.65 billion users. SNS is known for challenging conventional connotations of friendship by lumping all of one's social connections, including remote acquaintances, into one uniform friend category. Research suggests that social search is the future of online search and social search is seen as the online search for persons or information on persons. Based on only the structure of SNS, this paper devises a couple of exploration techniques to predict what strategies users are going to make friends on SNS in mobile environment. Essentially, there are two strategies for making friends on SNS in mobile environment: random and non-random (friends of friends) strategies. The non-random strategy means that new friends are selected in a friends-of-friends manner. Each SNS user can select randomly chosen target user, or neighbourhood search is done to develop a social network. This study develops logit models to capture the characteristics of these choice strategies. We conduct a series of experiments related to Facebook membership. Visualising how users are linked together or not is accessible via the Facebook application programming interface (API), so that we could crawl their Facebook accounts and extract their friends' data and access all their information under their agreement. In this study we obtained the users' permission to install an application within their accounts. The study analyses 226 Facebook users who totally have 37,571 friends and 607,802 relationships. The results significantly show that the number of individuals' friends, the value of closeness function, the betweenness value, and the clustering coefficient are the four important factors when an individual makes friends on Facebook. The accuracy of our models predicting the random or non-random strategies for making friends on Facebook is 84.5%. In addition, this paper significantly verifies the small-world phenomenon or six degrees of separation on Facebook. Implications for practitioners and academics are discussed.
Keywords: Facebook; hierarchical logistic regression; mobile communications; small-world phenomenon; social networking sites.
International Journal of Mobile Communications, 2017 Vol.15 No.4, pp.355 - 371
Accepted: 29 Jun 2016
Published online: 28 Apr 2017 *