The study addresses the long-standing debate about the rationality of human information-seeking. A prevailing view suggests that humans are not well-equipped to navigate the vast information landscape, often engaging with irrelevant or harmful information due to inherent irrationality. This perspective stems from philosophical traditions and cognitive research demonstrating that the maximization of instrumental value doesn't fully explain information-seeking. While acknowledging that clear, immediate goals (like test preparation) can guide information selection, it's argued that unconstrained information-seeking is driven by curiosity—an intrinsic drive for knowledge irrespective of utility. However, the concept of curiosity as a goal-independent drive faces challenges. Studies show that curiosity predicts better learning, aligning with James's description of curiosity as an 'impulse toward better cognition.' Neural correlates of processing instrumental and non-instrumental information also show striking similarities. The researchers propose that the apparent irrationality in information-seeking is explained by motivation, a third variable overlooked in previous research. They posit that information's utility varies depending on the individual's motivational state, similar to how food's utility varies with hunger. By measuring motivational state, it becomes possible to infer information-seeking goals and assess whether behavior is cost-benefit rational given those goals. The COVID-19 pandemic provided a naturalistic setting to study this, with its inherent motivational relevance and individual differences in the personal relevance of information.

The literature review explores existing theories of information-seeking, highlighting the conflict between the view of humans as rational information processors and the evidence suggesting irrationality in information choices. It discusses the role of curiosity as a driving force behind information seeking and its association with improved learning. However, the authors question the idea of curiosity as a completely goal-independent drive, citing evidence suggesting a connection between curiosity, learning, and goal attainment. The review also lays the groundwork for the researchers' hypothesis that motivation is a key factor influencing information-seeking behaviour by impacting the perceived utility of information.

The study used a large sample (n=5376) of US participants recruited via Amazon Mechanical Turk, collecting data twice a week between March 11th and May 7th, 2020. Epistemic behavior was measured using a validated waiting task. Participants chose whether to wait for answers to questions (COVID-19 related and general), rated their satisfaction with answers, and recalled answers a week later. Question usefulness was rated as a proxy for utility expectations. Motivational state was assessed using questionnaires measuring COVID-19 concern (specific motivation) and non-specific anxiety (general affect). The study employed a waiting task where participants chose to wait for answers to questions (COVID-19 related or general) for varying durations (4-16 seconds). They then rated the usefulness of answers and their satisfaction. A week later, participants took a memory test. The researchers used multilevel regression models to analyze the data, testing for interactions between COVID-19 concern, question type, usefulness judgments and waiting choices, satisfaction ratings, and memory. Mediation analysis examined the relationship between COVID-19 concern and waiting via usefulness judgments. Finally, they explored how prediction errors (difference between actual satisfaction and expected usefulness) influenced subsequent waiting choices. Data were analyzed using Bayesian methods implemented in R, Stan, and Julia. Various exclusion criteria were applied to ensure data quality, focusing on language fluency, application interactions, response consistency, and mean response time. The measures of COVID-19 concern and non-specific anxiety were validated through Bayesian Principal Component Analysis and by examining their correlations with real-world experiences such as job loss, income reduction, self-isolation, and social distancing behaviors.

The study found evidence supporting both a directing and an energizing effect of motivation on information-seeking. First, higher COVID-19 concern was associated with greater information-seeking for COVID-19 related questions compared to general questions (interaction effect). This directing effect was mediated by higher perceived usefulness of COVID-19 related information. Second, higher COVID-19 concern was associated with increased information-seeking for both COVID-19 related and general questions (main effect). This energizing effect was also mediated by higher perceived usefulness of information overall. These findings were consistent across measures of information-seeking (waiting choices), satisfaction, and memory. Participants with higher COVID-19 concern reported greater satisfaction with all answers, especially COVID-19 related ones. However, memory was better for COVID-19 related information but worse for general information in high COVID-19 concern individuals. The influence of prediction errors (the difference between expected and actual utility) on subsequent information seeking supported the use of an average value of information in decision-making. Control analyses demonstrated that the effects of COVID-19 concern were independent of non-specific anxiety. While COVID-19 concern and non-specific anxiety were correlated, they had opposite effects on information-seeking; high anxiety was associated with reduced information-seeking overall.

The findings provide strong support for the hypothesis that motivation plays a dual role in information-seeking, both directing and energizing behavior. The directing effect aligns with rational theories of motivation, where motivational states influence the perceived utility of information within a specific domain. The energizing effect suggests that acute motivation can broadly enhance information-seeking, even for unrelated content. This is consistent with reward-based learning theories, demonstrating that increased motivation can enhance overall engagement. The results challenge the view of curiosity as a purely non-instrumental drive, suggesting that all information-seeking is goal-directed and driven by expectations of utility, even for seemingly useless information like trivia. The use of prediction errors in guiding subsequent information-seeking demonstrates a dynamic, adaptive process of information valuation.

This study demonstrates that motivation significantly influences information-seeking behavior, both directing it towards relevant domains and energizing it across domains. The findings highlight the importance of considering motivational factors when studying information-seeking, particularly during crises like the COVID-19 pandemic. Future research could explore causal manipulations of motivation in controlled settings and extend the findings to other significant life events that induce similar shifts in motivational states. This framework offers a more unified understanding of human information-seeking by applying principles from reward-based learning and emphasizing the role of utility expectations.

The study's correlational nature limits causal inferences. While the mediation analyses support the proposed model, it remains challenging to definitively establish causality between motivation, utility expectations, and information-seeking. The use of self-reported measures for both motivation and utility expectations introduces potential biases. The reliance on a convenience sample recruited from Amazon Mechanical Turk may limit the generalizability of findings to other populations. The specific context of the COVID-19 pandemic might also limit the generalizability of results to other situations.