School of Integrated Technology, university of Taipei
Abstract
Online social network have come out as effective crowdsourcing platforms in recent days. However, issues about how to effectively exploit them have not been adequately addressed yet. Here, we investigate the reliability and effectiveness of multimedia-involved behavioral testing via social network-based crowdsourcing especially focused on Facebook as a medium to recruit participants. We conducted a crowdsourcing-based experiment for a music recommendation problem. It is shown that different methods for advertisement yields different degrees of efficiency and there exist significant differences in behavioral patterns across different genders and different age groups. Also, we compare our experiment with other multimedia-involved crowdsourcing experiments built on amazon mechanical turk, which suggests that social network-based crowdsourcing experiments can achieve comparable efficiency. Based on the analysis results, advantages and disadvantages of social network-based crowdsourcing experiments and suggestions for successful experiments are also discussed. We will conclude that social networks have a potential to serve as appropriate platforms for behavioral tests to gather in-depth data for even long-term periods of time.
Introduction
Nowadays, online crowdsourcing has been one of the widely used methods to conduct behavioral experiments with a large number of people. In comparison to the traditional way of performing offline experiments: crowdsourcing on the Internet can collect large data with less efforts to recruit participants, set up the experimental environments, and run the experiments. (Wu et al., 2013) In addition, it has been shown that the quality of experimental results from crowdsourcing is almost the same to that from offline-recruited participants (Casler et al., 2013).