Current Research
Visual Attention to AI-Generated Images Vs. Images of Real People
Consumer attention to visual stimuli in advertising and signage plays a critical role in shaping purchase intentions and brand perceptions, particularly in the horticultural sector. Since the COVID-19 pandemic, interest in plants has surged, with nearly 38% of U.S. households engaging in flower gardening and over 35% in indoor plant care (Whitinger & Cohen, 2024). Despite this increased interest, signage in independent garden centers (IGCs) and home improvement centers (HICs) often emphasizes plant imagery over human representation, primarily to showcase a plant’s mature appearance. However, visual endorsers can have a significant influence on consumer motivation and decision-making. Endorsers act as persuasive agents that trigger motivational processing (Alhabash et al., 2020. Endorsers in ads can reduce perceived risk and increase trust by reinforcing brand identity (Wallace, 2001; Sheth & Parvatiyar, 1995). As the horticultural sector explores new advertising formats, understanding how consumers visually attend to different types of endorsers is vital for enhancing persuasion.
Concurrently, advances in AI have enabled a new wave of advertising personalization. Generative AI tools can quickly create adaptive, data-informed content, allowing brands to tailor their messaging in real-time (Wiredu, 2023). While this increases the potential for personalization, it may also lead to mixed consumer reactions. Some individuals embrace AI-generated content as innovative, while others express skepticism or disappointment when personalization feels superficial or rigid. This response—termed algorithmic disillusionment—reflects a sense of disenchantment with profiling systems that rely on basic demographic data such as age, gender, or location (De Graaf et al., 2017; Ruckenstein & Granroth, 2020). These tensions raise important questions about how consumers cognitively and behaviorally process ads featuring AI-generated versus real human endorsers, especially in sectors such as horticulture, where human models are scarce. The present study explores whether the type of endorser (AI vs. human) influences visual attention and purchase intent.
Status: Data collection in progress
PIs: Juan Mundel, Patricia Huddleston, Jing Yang, Bridget Behe
This research is generously funded by the Horticultural Research Institute
Archived Research
Psychophysiological Responses to Alcohol Advertising via Social Media
In collaboration with Dr. Anna McAlister, Dr. Elizabeth Taylor Quilliam and Dr. Jef Richards, Saleem Alhabash conducted a number of studies focusing on attitudinal and behavioral responses to social media advertisements and promotion of alcohol. Our findings showed that exposure to alcohol ads on a social media site like Facebook significantly increases the desire and actions related to alcohol consumption. In addition, expressing intentions to like, share and comment on alcohol ads or status updates explains around half of the variance in expressing intentions to consume alcohol.
Status: Completed
PI: Saleem Alhabash
What Makes Us Click?
Psychophysiological Precursors of likes, shares and comments on Facebook. "There are 4.5 billion Facebook likes every day!" (Wishpond, 2015). In an information-based economy, interactions via social media, such as likes, shares, and comments - also known as viral behaviors - become the currency of measuring online effectiveness for advertisers and marketers. Moreover, we, as individual users, also use these metrics as means of self-assurance and -assessment. So far, we know what kind of content gets viral on social media, and possibly why, yet we don't quite understand what drives these types of behaviors on the biological level. This study employs a novel method of collecting psychophysiological responses to Facebook use in an organic way - meaning, there are no controls other than asking individuals to press on the like button, share something, comment on something, and write their own status update. All of this is happening while we're collecting physiological data, which will enable us to understand the threshold physiological responses to predict each type of viral behavior.
Status: Completed
PI: Saleem Alhabash