Abstract
<p>The wisdom-of-crowds phenomenon demonstrates that aggregating judgments across multiple individuals can produce remarkably accurate estimates. More recently, research has shown that dependent judgments in sequential collaboration, where contributors revise or maintain previous judgments, can equal or outperform traditional aggregation methods based on independent judgments. Recently, a computational model of sequential collaboration has been proposed, but its assumptions and predictions could only been tested qualitatively. To address this limitation, a Bayesian model of sequential collaboration was developed that relies on the same congitive processes and allows model parameters to be estimated directly from observed judgments and decisions. Simulation studies demonstrated accurate parameter recovery across a wide range of sample sizes and item sets. Applying the model to four empirical datasets showed that estimated parameters aligned with empirical findings for sequential collaboration. Model comparisons further revealed that accounting for response biases substantially improved model fit in homogeneous estimation tasks, whereas other model variations received no support. These findings extend the theoretical framework of sequential collaboration by enabling quantitative testing of cognitive mechanisms underlying dependent judgments. More broadly, the model integrates ideas from wisdom-of-crowds research, cultural consensus theory, anchoring, and group decision making, providing a foundation for future research on collective judgment and social influence.</p>