CAN MEANING ENHANCE PRODUCTIVITY? A CROWDSOURCED EXPERIMENT PARTNERING CITIZEN-SCIENCE

Abstract

As a form of compensating wage, meaning is an important intrinsic motivator for work. However, current papers on meaning-making often imperfectly quantify its impact on task performance. To augment previous papers’ experimental setup, we partnered with a citizen-science organisation to concretise the injection of meaning as well as utilised an innovative experimental design and payment structure. We employed Amazon’s Mechanical Turk to randomise crowdsourced workers into four different conditions (“no-context”, “payment”, “meaning”, “selection”) for an image-classification task. Adopting a difference-in-differences methodology, we then isolated each treatment effect. Contrary to Chandler & Kapelner (2013) and Rogstadius et al. (2021), we found that meaning has an insignificant effect on both quantity and quality of images classified. However, providing piecewise payment for classifying additional images had a significant effect on quantity but not quality of images classified. We contend that this discrepancy in results boils down to poor experimental design on their part.