An Empirical Analysis of System-generated Data in Location-based Crowdsourcing

This paper develops a research model explaining how task location and incentives affect the take up and, for those tasks that are processed, the time to start. For an empirical analysis, we use the system-generated data of all 1860 location-based crowdsourcing tasks in Berlin available on the Streetspotr platform within one year. The results indicate that while the population density of the task location does not influence the probability that some crowdworker will eventually process the task, a task located in a more densely-populated area tends to be taken up more quickly. Moreover, the take-up probability is expected to increase as the monetary and non-monetary incentives are raised. However, both increasing the monetary incentives and lowering the non-monetary incentives tends to shorten the time to start. This suggests that high non-monetary incentives with which unattractive tasks are endowed do not entice the crowdworkers to quickly set about processing these tasks