Hotline wait-times, complicated returns, odd charges on a bill, cumbersome applications: negative service experiences are common. And consumer reactions carry major implications for firms. Drawing on the premise that language reflects unspoken cognition, this research adopts a mixed-methods paradigm to investigate whether a subtle linguistic feature in complaints (i.e., the use of passive voice): (1) reveals consumers’ perception of responsibility for a bad service experience, and (2) predicts escalation. To this end, study 1 applies natural language processing to more than 160,000 real complaints filed with the U.S. Consumer Financial Protection Bureau. We find that consumers who write with greater passive voice are significantly more likely to dispute offers of resolution, even after extensive controls. Studies 2A–3B use correlational and experimental methods to demonstrate that passive constructions increase as perceived fault (for a poor service experience) shifts from self to service provider. Study 4 extends this finding by showing that, when consumers attribute responsibility for an incident to the firm, they in turn: (i) use more passive voice to complain and (ii) report stronger escalation intentions—an effect that persists after controlling for individuals’ baseline writing-style. Conceptually, this research uncovers linguistic structure, an understudied aspect of language, as a diagnostic window into consumers’ state of mind. Managerially, we suggest an automated, cost-effective approach to identify early which dissatisfied customers are likely to escalate. Such insight allows in turn for more targeted service-recovery interventions.

What makes cultural products such as edutainment (i.e., online talks) successful versus not? Asked differently, which characteristics make certain addresses more (vs. less) appealing? Across 12 field and lab studies, we explore when, why, and for whom the information load carried in TED talks causes them to gain (vs. lose) popularity. First and foremost, we uncover a negative effect whereby increases in the number of topics broached in a talk (i.e., information load) hurt viewer adoption. The cause? Processing disfluency. As information load soars, content becomes more difficult to process, which in turn reduces interest. Probing process further, we show this effect fades among audience members with greater need for cognition, a personality trait marking a penchant for deep and broad information processing. Similarly, the effect fades among edutainment viewers favoring education goals (i.e., cognitive enrichment) whereas it amplifies among those favoring entertainment (i.e., hedonic pleasure). Our investigation also documents the counterintuitiveness of our findings (i.e., how individuals mispredict which talks they would actually [dis]like). From these results, we derive theoretical insights for processing fluency research and the psychology of cultural products adoption (i.e., we weigh in on when, why, and for whom fluency has favorable vs. unfavorable downstream effects). We also derive prescriptive insights for (a) players of the edutainment industry whose very business hinges on curating appealing content (e.g., TED, Talks@Google, The Moth, Big Think, Spotify) and (b) communicators of all creeds wishing to broaden their reach and appeal (e.g., professors, scientists, politicians, journalists, bloggers, podcasters, content editors, online community managers).

The academic study of grammatical voice (e.g., active and passive voice) has a long history in the social sciences. Passive voice, for example, has been used to identify victim blaming in traumatic events, false versus truthful speech patterns, and levels of construal. Most evaluations of passive voice are experimental or small-scale field studies, however, and perhaps one reason for its lack of adoption is the difficulty associated with obtaining valid, reliable, and replicable results through automated means. In this paper, we introduce an automated tool to identify passive voice from large-scale text data, PassivePy. With minimal computational overhead, this package achieves 97% agreement with human coded data for grammatical voice as revealed in two large validation studies. In this paper, we discuss why passive voice is an important social and psychological construct, how PassivePy works, and conclude with pathways to apply this package in everyday psychological research.

The advent of internet shopping brings countless options for consumers. In response, online retailers aim to facilitate search by attaching descriptive terms to their inventory (e.g., product genres or categories). Called “tagging,” such a practice has been shown to increase product visibility/exposure in an otherwise limitless sea of options. Across five distinct platforms covering a gamut of consumer domains (i.e., food recipes, mobile games, books, videos, and journal publications), we find that the benefits of tagging for product visibility/exposure come with unexpected costs. Indeed, though assigning more tags to a product does indeed boost its visibility/exposure online, doing so also hurts said product’s perceived quality. This effect emerges regardless of proxies used to quantify (i) visibility/exposure or (ii) product liking. We draw on categorization and learning research to explain these results. While multi-tagging increases the likelihood of a product to appear in consumers’ search, it also increases the likelihood of a mismatch between what consumers really sought and what said product can actually deliver.

The proliferation of peer‐to‐peer fundraising platforms (e.g., GoFundMe, Rally, Fundly) poses conceptual and substantive challenges for behavior scientists and fundraisers. This article explores how fundraisers should craft their appeals to maximize their chance of success. Four field‐ and laboratory‐studies find that direct appeals (i.e., narratives written in the first person by the intended recipient) raise less money than otherwise‐identical indirect appeals (i.e., narratives written in the third person, seemingly by a third party on behalf of the intended recipient). The cause? Prospective donors ascribe lesser (greater) credibility to direct (indirect) appeals, which in turn curtails (increases) their giving. Since the narrative voice (direct vs. indirect) in which appeals are crafted is often discretionary (i.e., adjustable), our findings offer prescriptive guidelines for fundraisers.

Meta-analytic evidence suggests that verbal patterns of emotion betray deceit, but it is presently unclear whether the location of maximum emotion in lies and truths matters to reveal deception. We contribute to the deception literature by offering analyses at the sentence level to locate where emotion is most pronounced in deceptive versus truthful texts. Using two public data sets—news articles (Study 1) and hotel reviews (Study 2)—we found that maximum emotion occurs toward the beginning of deceptive texts while maximum emotion appears later for truthful texts. In addition to demonstrating the effect across diverse settings, we used two different measurements for emotion and separated the results by valence, replicating the maximum emotion effect each time. The predictive nature of maximum affect ranged from 54% to 56% across data sets, a rate consistent with most deception studies using 50-50 lie–truth base rates. Implications for future research and deception theory are discussed.

Supplementary_Materials_V2_SUBMIT for The Location of Maximum Emotion in Deceptive and Truthful Texts by Amir Sepehri, David M. Markowitz and Rod Duclos in Social Psychological and Personality Science