Crowdsourcing Safety Perceptions of People: Opportunities and Limitations

摘要

Online crowdsourcing has successfully been used as a paradigm to collect large amount of perceptions about our cities quickly and cheaply, enabling social scientists to quantitatively test urban theories at scale. While doing so, researchers have not focussed on getting answers from specific demographics, relying on a self-selected crowd instead. However, existing theories suggest that knowing who the respondents are is crucial for understanding safety perceptions about people (instead of, for example, about the built environment). In this case to quantitatively validate theories, it is not just the amount of data that matters, but also what demographics participate (or not). In this paper we investigate to what extent online crowdsourcing can be used for the specific case of safety perceptions about people. We built an image-based online crowdsourcing platform, collected safety perception ratings and background information from more than 700 people and used them to quantitatively evaluate established theories based on qualitative research. On one hand, we show in this paper that online, image-based crowdsourcing can be used to gather perceptions about people too, not just architecture, confirming established theories based on qualitative work. Furthermore, we are able to uncover detailed interactions that would be challenging to grasp using qualitative methods. On the other hand, we show limitations of using crowdsourcing as a method. By not controlling who makes up the crowd, we were not able to investigate all theories as we did not reach all user groups that have been discussed in qualitatitve research.