Zongxia Li, Andrew Mao, Daniel Stephens, Pranav Goel, Emily Walpole, Alden Dima, Juan Fung, Jordan Boyd-Graber. Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers). 2024.

Collecting truthful subjective information from multiple individuals is an important problem in many social and online systems. While peer prediction mechanisms promise to elicit truthful information by rewarding participants with carefully constructed payments, they also admit uninformative equilibria where coordinating participants provide no useful information. To understand how participants behave towards such mechanisms in practice, we conduct the first controlled online experiment of a peer prediction mechanism, engaging the participants in a multiplayer, real-time and repeated game. Using a hidden Markov model to capture players' strategies from their actions, our results show that participants successfully coordinate on uninformative equilibria and the truthful equilibrium is not focal, even when some uninformative equilibria do not exist or are undesirable. In contrast, most players are consistently truthful in the absence of peer prediction, suggesting that these mechanisms may be harmful when truthful reporting has similar cost to strategic behavior.

Designers of human computation systms often face the need to aggregate noisy information provided by multiple people. While voting is often used for this purpose, the choice of voting method is typically not principled. We conduct extensive experiments on Amazon Mechanical Turk to better understand how different voting rules perform in practice. Our empirical conclusions show that noisy human voting can differ from what popular theoretical models would predict. Our short-term goal is to motivate the design of better human computation systems; our long-term goal is to spark an interaction between researchers in (computational) social choice and human computation.

Paid and volunteer crowd work have emerged as a means for harnessing human intelligence for performing diverse tasks. However, little is known about the relative performance of volunteer versus paid crowd work, and how financial incentives influence the quality and efficiency of output. We study the performance of volunteers as well as workers paid with different monetary schemes on a difficult real-world crowdsourcing task. We observe that performance by unpaid and paid workers can be compared in carefully designed tasks, that financial incentives can be used to trade quality for speed, and that the compensation system on Amazon Mechanical Turk creates particular indirect incentives for workers. Our methodology and results have implications for the ideal choice of financial incentives and motivates further study on how monetary incentives influence worker behavior in crowdsourcing.

We present studies of the attention and time, or engagement, invested by crowd workers on tasks. Consideration of worker engagement is especially important in volunteer settings such as online citizen science. Using data from Galaxy Zoo, a prominent citizen science project, we design and construct statistical models that provide predictions about the forthcoming engagement of volunteers. We characterize the accuracy of predictions with respect to different sets of features that describe user behavior and study the sensitivity of predictions to variations in the amount of data and retraining. We design our model for guiding system actions in real-time settings, and discuss the prospect for harnessing predictive models of engagement to enhance user attention and effort on volunteer tasks.

The flourishing of online labor markets such as Amazon Mechanical Turk (MTurk) makes it easy to recruit many workers for solving small tasks. We study whether information elicitation and aggregation over a combinatorial space can be achieved by integrating small pieces of potentially imprecise information, gathered from a large number of workers through simple, one-shot interactions in an online labor market. We consider the setting of predicting the ranking of $n$ competing candidates, each having a hidden underlying strength parameter. At each step, our method estimates the strength parameters from the collected pairwise comparison data and adaptively chooses another pairwise comparison question for the next recruited worker. Through an MTurk experiment, we show that the adaptive method effectively elicits and aggregates information, outperforming a naive method using a random pairwise comparison question at each step.