An Impact-driven Model for Quality Control in Skewed-domain Crowdsourcing Tasks

TitleAn Impact-driven Model for Quality Control in Skewed-domain Crowdsourcing Tasks
Publication TypeConference Paper
Year of Publication2016
AuthorsMaarry, K. E., and W. - T. Balke
Conference Name8th ACM Conference on Web Science
Date Published05/2016
PublisherACM
Conference LocationHannover, Germany
Abstract

Not only do the highly-distributed digital crowdsourcing solutions surpass both borders and time-zones, but they materialize the vision of impact sourcing, by tapping into new labor markets in developing countries. Unfortunately, crowdsourcing is associated with severe quality issues.  To that end, many countermeasures have been designed to detect spammers, except in practice, also honest, yet not perfect workers will often be exposed and deprived of much-needed earnings. Here, we argue for the need of an impact-driven quality control measure, especially for skewed-domain tasks. Such a measure should ensure high quality results, while simultaneously fulfilling the social responsibility aspect of crowdsourcing.

AttachmentSize
Camera Ready_Extended Abstract_69.pdf447.09 KB