Intentions are one of the most widely used constructs in consumer research. We review over 50 years of research that has helped us understand what intentions are, their antecedents and consequences, and how best to measure and use them as a proxy for or predictor of behavior. We define intentions and differentiate them from other closely relatedly psychological constructs. We review several psychological theories where intentions play a central role and highlight what is known about the strength of the intentions-behavior relationship, and factors that moderate the strength of that relationship. We also review more methodological research and discuss what is known about how to best measure intentions and use them to predict behavior. Finally, we suggest opportunities for continued research on intentions and discuss their continued relevance in a world of big data.

In a randomized field experiment with the education charitable giving platform DonorsChoose.org (N = 30,297), we examined email personalization using a potential donor’s name. We measured the effectiveness of matching potential donors to specific teachers in need based on surname, surname initial letters, gender, ethnicity, and surname country of origin. Full surname matching was most effective, with potential donors being more likely to open an email, click on a link in the email, and donate to a teacher who shared their own surname. They also donated more money overall. Our results suggest that uniting people with shared names is an effective individual-level approach to email personalization. Potential donors who shared a surname first-letter but not an entire name with teachers also behaved more generously. We discuss how using a person’s name in marketing communications may capture attention and bridge social distance.

  2021 Bocconi Junior Research Grant

Recently, gender-ambiguous (non-binary) voices have been added to voice assistants to combat gender stereotypes and foster inclusion. However, if people react negatively to such voices, these laudable efforts may be counterproductive. In six preregistered studies (N = 2,424) we find that people do react negatively, rating products described by narrators with gender-ambiguous voices less favorably than when they are described by clearly male or female narrators. This is due to the voices creating a feeling of unease related to difficulty understanding the gender of the narrator, what we call social disfluency, that spills over to affect evaluations of the products being described. These effects are best explained by low familiarity with voices that sound ambiguous. Thus, initial negative reactions can be overcome with more exposure.

  2018 Center for Global Economy and Business Grant

Voice assistants often present choices where consumers listen to product options. But do consumers process information differently when listening compared to reading? Bridging theories on evaluability and memory, six experiments, including one conducted in consumers’ homes on Alexa voice speakers, demonstrate that consumers listening to speech utilize higher-evaluability product information (which can be understood without making comparisons to other options) to guide their judgments and choices relatively more than consumers reading the same text. A difference in memory drives this tendency. This is because (1) due to its ephemeral nature, processing speech requires greater reliance on memory and (2) information higher in evaluability is more easily remembered. Thus, higher-evaluability information is likely to be remembered regardless of presentation mode (speech vs. text), while memory for lower-evaluability information is likely to favor text, leading to the observed effect. The findings speak to the evaluability, memory, and auditory information processing literatures, and underscore that marketing managers presenting choices via speech will do well to highlight favorable highly-evaluable information about products such as recommendations, sales ranks, or descriptions such as “like new.” Substantively, a new format for presenting information is demonstrated which may improve voice-based sales.

Choosing something improves a person’s attitude toward it, a classic example of behavior affecting attitudes. Three studies re-examine the causal role of behavior in this “post-choice spreading of alternatives” phenomenon, demonstrating that neither the behavior of choosing nor the self-perception of having made a choice is required for it to occur. A rationalization process similar to the one that follows from actively choosing can occur whenever someone merely accepts an outcome as a true state of affairs. When a person accepts an externally assigned outcome, the negative features of that outcome no longer seem as important, allowing attitudes to improve. People normally accept outcomes they have personally chosen, but they may also accept outcomes they neither chose nor could reject. Thus, we question the causal role of behavior in a classic phenomenon. This finding contradicts explanations for post-choice spreading based on self-perception theory, where people learn their attitudes from their own voluntary behaviors. Though future work would be needed to confirm, we also discuss potential implications for cognitive dissonance theory, suggesting that agency over choice may not be prerequisite for dissonance, as previously believed. Relaxing this prerequisite expands the scope of phenomena to which dissonance theory could potentially apply.

With the advent of voice technology, consumers increasingly interact with technological devices orally (i.e., by speaking) rather than manually (i.e., by typing, clicking, and touching). This roundtable aims to examine how this shift from manual to oral communication with technology influences consumer behavior. We will discuss fundamental conceptual differences between oral and manual communication, methodological challenges of studying cross-modality effects in interactions with technology, and avenues for future research.

Melzner, Johann, Andrea Bonezzi, Jonah Berger, Christian Hildebrand, Mansur Khamitov, Anne-Kathrin Klesse, David Luna, Shiri Melumad, Vicki G. Morwitz, Kurt P. Munz, Demi Oba, Massimiliano Ostinelli, Aner Sela, and Ana Valenzuela

  Best Student Presentation

Contrary to previous findings on the association between disfluency and information processing, three preregistered studies demonstrate that disfluency causes consumers to simplify difficult-seeming decisions by using heuristics. The results demonstrate that disfluency leads to a greater reliance on “heuristic” reasoning where consumers utilize information such as a brand name or a recommendation to make their choice. This tendency to rely on heuristic information in the face of disfluency is more pronounced among decision-makers high in Need for Cognition.

Does using a gender-neutral voice affect information processing and product judgment? Processing Fluency literature says the difficulty in information processing leads to the negative judgment of the information and product presented in that information at the metacognitive level. While research in this area has involved textual information, scant research attention on fluency has been paid to the social nature of communication. We investigate a novel concept, “social-processing fluency”, related to the ease or difficulty of identifying the demographic information about the source of voice information (i.e., gender) used to determine social categorization.

Download Poster

Six experiments demonstrate that choosing from options presented by voice (versus text) increases the cognitive burden on consumers (due to difficulty making comparisons), leading them to choose recommended items more often but also to defer choice at higher rates. Auditory consumers focus on context-independent “evaluable” product features to guide judgment.

Six experiments demonstrate that information presented by voice is more difficult to process than the same information in writing. As a result, auditory choosers are less able to differentiate choice options, leading them to choose recommended items more often, but also to defer choice at higher rates. Difficulty making comparisons contributes to these effects, reducing the impact of certain context effects. Voice presentation also negatively affects evaluations of single items when compared against a salient reference in memory, and thus may have important ramifications for decision makers using digital voice assistants for shopping.

Choosing an option leads to more favorable attitudes toward that option compared to before choice. Three studies demonstrate that
this “post choice spreading of alternatives” may not require choice at all. Spreading depends on accepting an outcome, rather than on
the behavior of choosing or self-perception of having chosen. People normally accept the outcomes of their own choices, but they can
also accept outcomes they did not have personal agency to choose or the freedom to reject. Higher outcome acceptance predicts
greater post-outcome attitude change.

In a randomized email field experiment with DonorsChoose.org (N = 30,297), donors who shared a surname with a teacher were more likely to open, click, donate, and donated more to the teacher’s classroom. Different-surname donors were also more generous when they shared a first-letter with the requesting teacher. We quantify various components of similarity including matches on full name, initial name letters, gender, ethnicity, and country of origin.

Download Poster

Choosing an option leads to more favorable attitudes toward that option compared to before choice. Three studies demonstrate that this “post choice spreading of alternatives” may not require choice at all. Spreading depends on accepting an outcome, rather than on the behavior of choosing or self-perception of having chosen.

Is losing better resolved quickly, or does holding onto hope for a positive outcome improve an otherwise negative experience? In three lab studies and one field study, consumers preferred to learn that they would lose later in a game compared to winning, but changed their preference after playing the game.

In a randomized field experiment with DonorsChoose.org (N = 30,297), potential donors who shared a surname with a teacher were more likely to open, click, donate, and donated more to the teacher’s classroom in response to an email request. We highlight how overtly personalizing an email to highlight a recipient’s identity can be effective to persuade them to donate to charity. Controlling for ethnicity, we also find that different-surname donors were more generous when they shared a surname first-letter with a requesting teacher.

Choosing a product leads to more favorable attitudes toward it (and more negative attitudes toward rejected options) compared to before a choice. This “post-choice spreading of alternatives” has been explained in terms of cognitive dissonance theory. Researchers have recently claimed that only a perception of having made a choice (vs. actual choice) is required for this spreading effect. We demonstrate in three experiments that even this perception is not necessary: spreading of alternatives can occur absent choice or a perception of having chosen. Thus, self-perceiving choice agency may not be prerequisite for dissonance, as previously believed.

In a randomized email field experiment with DonorsChoose.org (N = 30,302), donors who shared a surname with a teacher were more likely to open, click, donate, and donated more to the teacher’s classroom. Different-surname donors were also more generous when they shared a first-letter with the requesting teacher.

Sorting by a product attribute can diminish the importance weight of that attribute. When choosing is difficult, consumers may treat sorting as screening. Once options are sorted, consumers may form a consideration set comprising the options at the top. Because these options are more homogeneous with respect to the sorted attribute, consumers pay less attention to the sorted attribute in favor of a second attribute. This attentional shift emerges in a subsequent conjoint analysis, with less weight placed on the sorted attribute and more weight on a second attribute.

Download Poster