Research

People evaluate their own wealth differently from how they evaluate the wealth of others. Across six experiments, we find evidence that people focus disproportionately on debt when thinking about their own (vs. another person's) wealth. In Experiments 1–3, participants predicted how wealthy they or someone else would be in one year, assuming they had the same amount in assets and debt today. While participants were generally optimistic about the future, they believed debt would shrink faster for themselves than for others. Participants focused more on paying down debt than growing assets when thinking about their own wealth. Further, when asked to consider what they would do with a windfall, they allocated more towards repaying debt than they believed others would. In Experiments 4 and 5, participants assessed their own wealth or that of another person after purchasing a car or a house and borrowing to do so. In every case, participants considered others (vs. oneself) as better off financially when holding the price and amount borrowed constant. As debt increased, the gap between self and others widened. In Experiment 6, a separate group of participants also reported their beliefs about how others might see them. When actively considering another person's perspective, people saw themselves as wealthier. We conclude by discussing the role that different evaluations of wealth might play in patterns of conspicuous consumption.

In this paper, we examine how specific features of language drive consumer behavior. Our contribution, however, lies not in testing specific hypotheses; rather, it is in demonstrating a data-driven process for generating them. We devise an approach that generates interpretable hypotheses from text by integrating large-language models (LLMs), machine learning (ML), and psychology experiments. Using a dataset with over 60,000 headlines (and over 32,000 A/B tests), we produce human-interpretable hypotheses about what features of language might affect engagement. We then test a subset of these hypotheses out-of-sample using two datasets: one consisting of 1,600 A/B tests and another containing over 5,000 social media posts. Our approach indeed facilitates discovery. For instance, we find that describing physical reactions significantly increases engagement. In contrast, focusing on positive aspects of human behavior decreases it. A third hypothesis posited that referring to multimedia (e.g., GIFs, videos) would influence engagement, and it does, only it significantly increases engagement in one domain while significantly decreasing it in another. This approach extends beyond a single application. In general, it offers a data-driven method for discovery that can convert unstructured text data into insights that are interpretable, novel, testable, and generalizable. It does so while maintaining a transparent role for both human researchers and algorithmic processes. This approach offers a practical tool to researchers, organizations, and policymakers seeking to aggregate insights from multiple marketing experiments.

This paper investigates "co-holding"–whereby consumers simultaneously maintain low-yield liquid assets and high-interest revolving debt–through the lens of mental accounting and payment method preferences. We propose that consumers' decision to co-hold is partly driven by how they spend. Using data from a large retail bank, we document that approximately one in five credit card users co-hold despite the economic costs. Analysis of over 2,000 customer surveys and 3 million transactions over 38 months reveals that co-holders systematically prefer debit cards for routine expenses and credit for larger purchases. Although co-holders are thought to carry too much debt, our finding suggests it is the preference to spend using debit that sustains the behavior. Three pre-registered experiments demonstrate that shifting consumers' preferences toward credit card usage for everyday purchases increases debt repayment, thereby reducing co-holding propensity. We identify pain of paying and perceived convenience as key psychological mechanisms mediating these effects. Wealth levels do not appear to moderate this relationship. Our findings suggest co-holding is a byproduct of spending decisions and indicate that policies targeting spending preferences may be an effective way to mitigate co-holding.

Why do consumers simultaneously maintain low-yield liquid assets and high-interest revolving debt? This behavior, known as "co-holding," affects 23% of credit card users in our sample from a major international bank and costs the typical co-holder hundreds of dollars annually in unnecessary interest charges. Our analysis of 38 months of detailed banking records reveals that co-holding is remarkably persistent, with typical co-holders maintaining this behavior for most months observed. Our analysis also shows that co-holders engage in both co-holding and co-spending–regularly depositing and withdrawing from asset accounts while continuing to use credit cards for new purchases. To test whether co-holding could be addressed through information disclosure we conducted a large-scale field experiment (n=125,328), providing clear information about co-holding behavior and its costs. Customers received targeted messages through their bank's mobile app, where they could easily transfer money from assets to pay down debt. Despite sufficient power to detect economically small effects, we found no meaningful changes in debt repayment amount, though customers did respond in other ways—making more frequent repayments and paying above required minimums. These results challenge explanations based on limited attention or information gaps and suggest that simple information disclosure, even when carefully designed and delivered through trusted channels, may not effectively address costly financial behaviors.

People do not always say what they mean. In everyday conversations, people regularly sense that they have not fully communicated what they had in mind—a subjective experience we refer to as “misarticulation.” Understanding the nature and causes of misarticulation is crucial, as it affects people’s decisions to communicate. We propose that it is not enough to know something to feel it can be communicated well; rather, misarticulation is also influenced by the way concepts are represented in one’s mind (e.g., the complexity of a concept or the typicality of its constituent objects). Studies 1a-b (n=600) examine when, where, with whom, and why people experience misarticulation. Studies 2-3 (n=1,341) then teach participants concepts of varying complexity and typicality, respectively, before tasking them with articulating the concepts. More complex and atypical concepts elicited stronger misarticulation. Overall, misarticulation is a common, consequential experience influenced by the structure of thought.

Who do we blame when a conversation goes awry, the speaker or the listener? Five preregistered experiments (N = 2,300 adults recruited from Prolific) reveal that people assign more responsibility for miscommunication to speakers than listeners. In observed (Study 1) and real-time incentivized (Study 2) communication tasks, participants held speakers (Study 1) and writers (Study 2) more responsible than listeners and readers, respectively, for bad outcomes. In recalled miscommunications, speakers were blamed more than listeners, although this result was primarily driven by listeners blaming their conversation partners (Study 3a-b). One reason for the tendency to hold speakers more responsible for miscommunication is people perceive speakers to have more control over conversations than listeners (Study 4). Yet this tendency persisted even in potentially unwarranted situations: when pairs failed “impossible” communication tasks (Study 1) and speakers’ messages were high quality (Study 5). Overall, speakers shoulder more blame for communication failures than listeners.