16Nonparametric Bootstrap and Small Area Estimation to Mitigate Bias in Crowdsourced Data: Simulation Study and Application to Perceived Safety

David Buil‐Gil1, Reka Solymosi1, and Angelo Moretti2

1University of Manchester, Department of Criminology, Manchester, UK

2Manchester Metropolitan University, Department of Computing and Mathematics, Manchester, UK

16.1 Introduction

Open and crowdsourced data are shaping a new revolution in social research methods. More social science researchers are applying crowdsourcing techniques to collect open data on social problems of great concern for governments and societies, such as crime and perceived safety (Salesses 2009; Salesses, Schechtner, and Hidalgo 2013; Solymosi and Bowers 2018; Solymosi, Bowers, and Fujiyama 2018; Williams, Burnap, and Sloan 2017). Crowdsourcing techniques are defined here as methods for obtaining information by enlisting the services of large crowds of people into one collaborative project (Howe 2006, 2008). Data generated through people's participation in these (generally) online platforms serving a variety of functions allow for analyzing social problems, examining their causal explanations, and even exploring their spatial and temporal patterns.

Such data already offer many advantages over traditional approaches to data collection (see Brabham 2008; Goodchild 2007; Haklay 2013; Surowiecki 2004). Some are highlighted later in this chapter (e.g. reduced cost of data collection, spatial information). Crowdsourced ...

Get Big Data Meets Survey Science now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.