Frequency and duration of self-initiated task-switching in an online investigation of interrupted performance

in HCOMP-2014: Human Computation and Crowdsourcing, Works in Progress and Demonstrations, Journal article, Palo Alto, CA

Abstract

Online experimentation has been shown to be a reliable method for collecting data quickly and easily. Nevertheless, it is a method of data collection that can have a negatively affect confidence in results because of the loss of experimental control that comes with moving experiments online. To relieve these concerns, researchers have developed a number of tools and techniques for evaluating the performance of online participants and crowdworkers. In this paper we develop a measure of participants' attentiveness during an online data-entry experiment. We explored participants' propensity to switch away from the experiment to other tasks, and the duration of time that they spent away from the experiment during each switch. Our results show that switching is prevalent, even when participants are asked not to switch to other tasks.