Worker Performance in a Situated Crowdsourcing Market

We present an empirical study that investigates crowdsourcing performance in a situated market. Unlike online markets, situated crowdsourcing markets consist of workers who become serendipitously available for work in a particular location and context. So far, the literature has lacked a systematic study of task performance and uptake in such markets under varying incentives. In a 3-week field study, we demonstrate that in a situated crowdsourcing market, task uptake and accuracy are generally comparable with online markets. We also show that increasing task rewards in situated crowdsourcing leads to increased task uptake but not accuracy, while decreasing task rewards leads to decreases in both task uptake and accuracy.