Planet Four: Craters—Optimizing task workflow to improve volunteer engagement and crater counting performance. (7th April 2019)

• Record Type: Journal Article
• Authors: Sprinks, J.; Houghton, R.; Bamford, S.; Morley, J. G.
• Is Part Of: Meteoritics & planetary science. Volume 54:Number 6(2019)
• Page Start: 1325
• Page End: 1346
• Publication Date: 2019-04-07
• Subjects: Meteorites -- Periodicals; Planetology -- Periodicals; 523.4
• DOI: 10.1111/maps.13277
• Languages: English
• ISSNs: 1086-9379

Abstract

Abstract: Virtual citizen science platforms allow nonscientists to take part in scientific research across a range of disciplines, including planetary science. What is required of the volunteer can vary considerably in terms of task type, variety, judgment required, and autonomy—even when the overall goal is unchanged. Through analysis of our live Zooniverse Planet Four: Craters citizen science platform, the effects of task workflow design factors including volunteer autonomy, task variety, task type, and judgment required on volunteer engagement and crater marking performance were investigated. Website analytics showed volunteers using the Full interface (most autonomy and variety) were more likely to return to the platform, although the amount of time spent per visit was unaffected by the interface used. However, analysis of performance suggested that how this time was used did differ. The interface involving the least complex task resulted in the greatest amount of data and rate of collection, although this also coincided with a greater number of false positives when compared with the expert. Performance in terms of agreement, both between participants and with the expert judgment, was significantly improved when using the Stepped interface for crater position and the Ramped (Mark) when measuring diameter—interfaces that both directly measured the metric with a specific, delineated task. The implications for planetary scientists considering the citizen science route is that there is a balancing act to perform, weighing the importance of volunteer engagement with scientists' data needs and the resources that can be committed to data validation.