Team members: Debargha Dey, Nina Buchina, Serena Dorrenstijn
Advisor: Javed-Vassileios Khan, TU/e
Role: Lead Researcher

ABSTRACT

Empathy-induced altruism is likely to motivate people in a crowdsourcing environment to produce better quality work. However, there hasn’t been any considerable investigation regarding how empathy can be effectively conveyed through user interfaces (UI). I conducted a research to find the effects of introducing empathy in a UI using words, and investigate its effect on workers’ motivation. My team and I validated that empathy is perceived to have a positive effect for workers.

INTRODUCTION

Crowdsourcing platforms are quickly growing and are currently used with multiple purposes ranging from design and evaluation to problem solving and analysis. With these platforms on the rise, the question on what motivates people to work and contribute on these platforms is more important than ever. Research suggests that money, altruism, and practice of skills were the most important factors that motivate people to work on crowdsourcing platforms. With altruism being one of the main motivators for workers, and the knowledge that altruism arises out of empathy, we investigated whether empathy has an effect on the quality and quantity of the produced work.

The concept of Empathy Channel in Human Computer Interaction through visual feedback is not new. The importance of empathy in intercultural HCI design was also explored in previous research. The website for the Crystal Project also explores the idea of incorporating empathy in a user interface in the context of effective communication. However, investigation of empathy, how it is conveyed in user interfaces (UI), especially in crowdsourcing platforms has not been done yet. We chose to investigate empathy in the UI of crowdsourcing platforms by utilizing two specific platforms, Google Consumer Surveys, and Crowdflower. The availability of a diverse and readily available workforce made this approach realistic.

METHOD

Analytics of the experiment task in Crowdflower crowdsourcing platform

Due to the digital nature of crowdsourcing platforms, it is difficult to enable the workers to experience a full empathic response to the requester of the work. Workers are often not aware of the context, impact, or meaning of a task and know little to nothing about the requester’s background. We hypothesized that this mismatch in gaining an empathic “bond” to the project or requester might influence the quantity and quality of the work produced by workers on crowdsourcing platforms. To test this hypothesis we first investigated what role empathy plays in crowdsourcing according to the workers themselves. We then investigated how to incorporate empathy in crowdsourcing. Our contribution is twofold: how to utilize crowdsourcing for investigating empathy in the UI; and whether manipulating the task’s description is enough to affect the quantity and quality of work.

To gain insight into how workers on crowdsourcing platforms currently experience and encounter empathy in their work, we executed two surveys on two different crowdsourcing platforms (Crowdflower and Google Consumer Surveys). To validate the subjective claims we received from the previous step, we conducted an experiment in 2 iterations to try to objectively measure whether empathy in the UI would affect the quality and quantity of work performed on a crowdsourcing platform (i.e. had a behavioral impact). For that reason, we had one task with two formulations: one without empathic elements in it, and one with.

RESULTS & CONCLUSION

We qualitatively found from our exploratory surveys was that workers in a crowdsourcing context realize that empathy is an important aspect to aid them in their work. Following our experiment, we found marginal quantitative proof that objectively empathic elements in a user interface yield more productivity from crowdworkers. This methodology foundation can be used to perform further research and also incorporate the element of empathy in other fields that try to motivate user actions, such as recommender systems, e-commerce, and ERP platforms. The effect of expressing empathy in a crowdsourcing UI through words needed more investigation to yield statistically significant results. Ongoing work to replicate the study on varied platforms with different communication medium and larger sample sizes is under way, and is aimed towards a submission in CHI 2017.

Disclaimer: Full article available upon request