Serendipity and Diversity in Professional Social Matching: Towards diversity-enhancing recommendation strategies

Professional Social Matching (PSM) is the practice of building and maintaining connections in the context of knowledge work. Various people recommender systems and social matching applications have been designed to facilitate PSM by finding relevant others among numerous options. However, conventional recommendation approaches have been found to support algorithmic and human biases, disrupting knowledge flow and social networking, which is vital for PSM. This dissertation focuses on two central concepts: diversity and serendipity. Diversity refers to the importance of exposing individuals to different perspectives, backgrounds, and experiences to foster productive and creative knowledge work. Serendipity, on the other hand, pertains to the occurrence of unsought yet valuable connections that can lead to unexpected and fortunate encounters.

The research questions driving this dissertation revolve around the role of diversity and serendipity in PSM tools and the manifestation of these concepts in recommendation strategies. The research process involved a series of five publications. The first two publications employed online surveys to investigate social serendipity and the processes in making valuable connections in online and offline realms. The third publication entails a literature review with a specific emphasis on the conceptual framework of Big Social Data (BSD), as its comprehension holds significant relevance for the domain of user modeling within recommender systems. The last two publications experimented with diversity-enhancing recommendation strategies and examined the alignment between subjective perceptions and objective measures of recommendation relevance.

The findings uncovered diverse insights into the characteristics and antecedents of social serendipity, highlighting the necessity for identifying novel mechanisms to foster serendipity experiences in PSM. The results also revealed consistent and significant differences in subjective perceptions of the proposed diversity-enhancing strategies, thus indicating their preliminary effectiveness. Participants showcased the ability to identify relevant others at all levels of similarity and structural network positions, despite the inherent bias in selection. The research contributions lie in elucidating the proactive and reciprocal sense-making involved in PSM, identifying qualities that foster serendipitous encounters, exploring the potential of Big Social Data, and developing and evaluating recommendation mechanisms that promote diversity in professional social networks.