Selected Research

The Safety Illusion: How the Mere Presence of a Dislike Button Shapes Belief in Social Media Content (Job Market Paper)

Huang, Yvonne, Minzhe Xu, Yang Yang, and Shunyuan Zhang, Under Review.

Prior research shows that visible engagement metrics, such as counts of likes, dislikes, and shares, can influence perceived content credibility. The present research examines whether the mere presence of a negative engagement affordance—a dislike button—can shape credibility judgments even when it provides no direct social information about the content. Classic belief-formation theories suggest that such a negative affordance should prompt skepticism and lower perceived credibility. However, across one large-scale analysis of real-world social media comments (N = 40,458) and four controlled experiments (N = 1,845), the findings show that the presence of a dislike button increases perceived credibility. This effect occurs because consumers interpret the button as a signal that the platform can monitor, filter, or regulate low-quality content, creating an illusion of platform-level protection from misinformation. Consistent with this mechanism, the effect is attenuated when consumers are explicitly informed that the platform takes steps to reduce misinformation. This research advances understanding of consumer belief formation in digital environments and reveals how seemingly minor interface features can alter the way consumers judge the credibility of social media content.

People Are More Likely to Believe and Share Fake News on Smartphones (vs. PCs)

Xu, Minzhe, Zhenqi (Jessie) Liu, Yvonne Huang, and Yang Yang, Invited for 2nd round review at the Journal of Consumer Research.

While smartphones have surpassed PCs (laptops or desktops) to become the dominant device for news consumption, little is known about whether the device a consumer uses to access online information affects their ability to discern misinformation from true information. Analysis of a large Twitter dataset (N = 103,567) and four experiments (N = 1,617) find that people who access information on smartphones are less discerning—they are more likely to believe and share misinformation but not true information—than people who access information on PCs. We identify two mechanisms. First, smartphones enable consumers to access information almost anywhere, including in distracting environments, and distractions increase the difficulty of detecting misinformation. Second, holding the environment constant, certain characteristics of smartphones, such as smaller screens and association with leisure activities, make people less deliberative and thus less discerning. Theoretically, this research enriches the misinformation literature and the smartphone literature and makes a unique contribution to both. Practically, it highlights the need to alert users, especially smartphone users and users in distracting environments, to misinformation.

Beyond Exposure: How Personalized Recommendations Shape Consumption Variety

Huang, Yvonne, Xiang Wang and Yang Yang, Under Review.

Personalized recommender systems shape what consumers encounter and consume across digital platforms. Prior research has largely examined how personalization influences consumption variety through an exposure-based lens, focusing on how recommender systems alter the breadth and composition of the options consumers see. We identify a complementary pathway: personalization can also change how consumers choose among the same options. Across four studies spanning video, music, and news, we show that when consumers believe a recommender system is personalized, their choices more strongly reflect their underlying variety preferences, even when content exposure is held constant. Specifically, personalization belief strengthens the association between variety preference and choice, such that choices shift toward favorite-category content among consumers lower in variety preference and toward non-favorite-category content among consumers higher in variety preference. This effect is driven by consumers’ intention to shape future recommendations through their current choices. These findings suggest that consumers are not merely passive recipients of personalized recommendations; they are active participants who adjust their choices when they believe those choices will guide future recommendations. The research contributes to work on personalized recommender systems and consumer meta-preferences

Designing More Effective Carbon Labels: The Role of Relative Carbon Emission Misestimation

Huang, Yvonne, Felipe M. Affonso, Wenbo Wang, and Yang Yang, Invited for 2nd round review at the Journal of Public Policy & Marketing.

Carbon labels are increasingly used to encourage sustainable consumption, yet evidence of their effectiveness remains mixed. This research examines when and why carbon labels influence choice. A field experiment in a restaurant (N = 24,505) reveals that labels with both directional and magnitude information (e.g., “chicken emits only 10% as much carbon as beef”) reduced beef orders by 14.1% compared to directional-only labels (e.g., “chicken emits less carbon than beef”). However, laboratory experiments (N = 2,890) show that this advantage is contingent on the objective emission difference between products. While specifying the magnitude increases carbon-friendly choices when the lower-emission option is substantially better (e.g., chicken vs. beef), it provides no additional benefit when the difference is modest (e.g., fish vs. chicken). We further document a systematic pattern in consumers’ perceptions of relative carbon emissions: consumers overestimate low emission ratios and underestimate high emission ratios, compressing estimates toward the middle. This misestimation pattern correlates with actual food consumption and helps explain when and why adding magnitude information provides incremental benefits beyond directional information alone. Our findings identify the relative emission difference as a key contextual factor shaping carbon-labeling effectiveness and offer actionable guidance for firms and policymakers designing carbon labels.

A Blessing and a Curse: The Diverging Impact of Personalized Recommender Systems on News Consumption

Huang, Yvonne, Xiang Wang and Yang Yang, Data Collection in Progress.

Personalized recommender systems increasingly shape the political information consumers encounter. A growing concern is that these systems may deepen polarization by creating “filter bubbles” and exposing consumers more often to attitude-consistent information than to attitude-inconsistent information. We propose, however, that the impact of personalization depends on consumers’ attitudinal variety preferences—their underlying desire to read news that aligns with or challenges their existing views. When consumers believe a news platform is personalized, they treat their current reading selections as signals that shape future recommendations. As a result, personalization strengthens the relationship between consumers’ preference for viewpoint diversity and their selection of attitude-consistent versus attitude-inconsistent news. When recommendations are believed to be personalized, consumers who prefer seeing news from different viewpoints select more attitude-inconsistent news, whereas consumers who prefer news aligned with their existing views select even more attitude-consistent news. These findings suggest that personalized news systems may be both a blessing and a curse, broadening exposure for some consumers while narrowing consumption for others.

No Filter: The Smartphone Self-Congruence Effect

Song, Camilla Eunyoung, Aner Sela, and Yvonne Huang, Working Paper.

This research reveals that smartphones enable more self-congruent choices—decisions and opinions that are more consistent with individuals’ dispositional traits, self-views, and beliefs. Four experiments demonstrate this smartphone self-congruence effect across diverse domains, including traits (Studies 1A-1B), political orientations (Study 2), and specific self-views (Study 3), consistently showing greater disposition-behavior alignment among smartphone users compared to PC users. Two key findings illuminate the underlying process: the effect emerges for actual but not ideal self-congruence, and manifests only when using one’s own smartphone but not someone else’s device. These results suggest that personal smartphones uniquely activate private knowledge related to the actual self, leading to behaviors that more authentically reflect one’s true dispositions. By revealing that device type systematically shapes how individuals’ dispositions manifest in choices, this research opens new theoretical paths for understanding mobile and digital consumption behaviors. The findings have practical implications for digital marketing, survey methodology, and consumer research, highlighting the critical role of device type in analyzing and predicting consumer behavior in the digital age.

Page updated

Google Sites

Report abuse