Exploring Freelancer Attributes with Peer Endorsements

Abstract

Online freelancing markets connect buyers with workers globally to assign various categories of tasks. A worker’s quality on the platform could be assessed by their past performance (reputation systems), skills, and experience. However, these methods cannot provide insights about the worker quality in case of newly acquired skills or new workers participating on the platform. An endorsement system can be particularly instrumental in such cases by gathering and sharing endorsements of skills given to a worker by other workers. We investigate how endorsements received by a worker from their peers relate to their attributes. Drawing on social value orientation theory, we determine worker category from the past endorsement activity. Next, we apply the decision tree induction method to extract the worker attributes that are related to endorsement decisions across all worker categories. Finally, we frame our propositions based on the extracted rules.