Associations between privacy, risk awareness, and interactive motivations of social networking service users, and motivation prediction from observable features

We investigate factors that influence private information disclosure activities.Factors for disclosure are dependent on disclosure scopes and openness scores.We construct models to predict motivations from public observable factors.Motivations can be predicted with high accuracy from 0.764 to 0.858. In this paper, we use survey data firstly to study how significantly motivations of using social networking service (SNS), risk awareness, and profile attributes affect user's disclosure activities. Openness score is calculated from the number of disclosed items. We study influential factors and their rankings at various disclosure scopes and openness scores. Our findings reveal that gender, profile photo, certain motivations, and risk awareness highly affect private information disclosure activities. However the ranking of influential factors is not uniform. Gender and profile photo have greater influence, however, their influence becomes lower and loses significance as openness is getting higher, falling behind motivations and number of friends. Secondly, we discuss constructing prediction models based on binary logistic regression to predict motivations from number of friends and profile attributes that are visible from the public. We classify motives into inward motive to interact with existing social networking service friends, outward motive to acquire via the SNS, and neutral motive which cannot distinguish whether user have inward or outward motive. Results show that the models can predict motives well and have good discrimination power. Using dimensional reduction, important predictors for optimum models are identified.