Lab meetings

In our lab, we believe staying engaged with the field, whether it’s through paper discussion or learning about one another’s work, is essential to our development as researchers. Many ideas in I/O psychology are deeply interconnected: we could apply the same methodologies to solve different problems, or integrate a new topic idea with our focus of study to inspire something new… Especially because our lab has a special focus on modern technologies and research methodologies, continuous learning is crucial. Keeping up with emerging research helps us broaden our horizons, engaging in deeper discussions about the future of our field, and its broader implications for work and life. 

Each week, we center our discussion around a selected article or topic. Conversations typically focus on 1) what’s compelling about the paper, 2) areas the study could improve on, 3) methodologies used and their strengths or limitations, 4) practical implications in real-world settings, and 5) connections to our own research work and existing knowledge. Feel free to check out some selected topics below to have a glance into our lab routine and learning journey! 

# Practical issues and considerations in longitudinal studies/analysis (discussion leader: Muchen)

As intensive longitudinal designs become more common, researchers can move beyond between-person comparisons to study within-person dynamics and behavioral variability. Referring to McNeish’s work, Muchen walked us through how time and contexts serve as important factors, how individuals and situations interact, different approaches for modeling variability, trajectories and dynamic processes over time. We also discussed the challenges of modeling intensive longitudinal data, including missing data, time scales, overfitting and interpretation. This session underscores the importance of having theories to support selected methodologies and thoughtfully address many nuances involved in analyzing complicated multilevel data.

Article of Discussion -- (McNeish, 2025)

# Professional identity (discussion leader: Dr. Sun)

Building a meaningful research career requires more than simply following the latest trends - it begins with cultivating informed curiosity and discovering questions that genuinely sustain our interest over time. In this discussion, we explore Rozin’s perspective on finding and filling the “holes” in psychology. The paper encourages us to reflect on the projects we choose to pursue, the aspects of research that most capture our attention, and how our interests evolve and deepen through experience. As we immerse ourselves in the literature, engage in diverse research experiences, and exchange ideas with mentors and collaborators, how do these experiences shape and refine our research identity? As emerging technologies such as AI rapidly transform the research landscape, how can we distinguish between passing trends and opportunities for meaningful, lasting scientific contributions? 

Article of Discussion -- (Rozin, 2007)

# AI & Sustainability (discussion leader: Moana)

While AI may feel instantly accessible at our fingertips online, the infrastructure powering it often depends on large-scale data centers built through significant land use and, in some cases, the displacement of local communities. At the same time, the technological and financial benefits are often concentrated elsewhere. Unequal access to advanced AI systems, tools, and training can also deepen the digital divide, limiting students and workers from disadvantaged backgrounds in accessing the emerging opportunities AI creates. This raises important questions about how we can implement intentional policies that ensure AI does not reinforce existing socioeconomic inequalities.

Article of Discussion -- (Hammerschmidt et al, 2025)

# AI social chatbots & mental health (discussion leader: Xinyi) 

Research has found that companionship is the most primary reason why individuals interact with chatbots, especially among those with smaller social networks. However, these types of chatbot usage are consistently associated with lower wellbeing, particularly when usage is more intensive, involves higher levels of self-disclosure, and lacks weak human social support networks. While AI may fulfill certain emotional needs, recent cases also highlighted potential risks when individuals rely on AI for emotional support. These encourage us to think about the ethical futures of AI as a source of companionship: the implications of AI’s friendly but unreciprocal nature, concerns of data privacy, differences between AI and actual human relationships, and when human designers and institutions should intervene.

Article of Discussion -- (Zhang et al., 2025)

# Discussion and thinking about selection system design in 21st century (discussion leader: Pengda)

Selection systems are designed to provide valuable information for both individuals and organizations. They help answer questions such as whether a person’s abilities and skills align with the requirements of a role, whether their personal characteristics are associated with success in a particular domain, and which careers may be the best fit based on their interests and competency profiles. Ultimately, selection systems aim to predict future behavior and performance, making their design both scientifically rigorous and practically meaningful. As we think about designing selection systems for the 21st century, several foundational questions arise: How should we conduct job analyses? How do we ensure our assessments are reliable and valid? What constructs are we truly trying to measure, and what methods best capture them? 

Presentation slide by Pengda

# Multiverse analysis (discussion leader: Alexandra)

Research findings are often shaped not only by the data collected, but also by the many analytical decisions researchers make along the way. From data processing choices and inclusion criteria to variable dichotomization, statistical models, and covariate selection, different reasonable decisions can sometimes lead to different conclusions. This paper introduces multiverse analysis as a framework for increasing transparency by systematically examining how results change across a range of defensible data construction and analytic choices. Although multiverse analyses can be computationally and practically demanding, they encourage more transparent, reproducible, and trustworthy science, especially as research questions and datasets become increasingly complex.

Article of Discussion -- (Steegen et al, 2016)

# Value theoretical framework and goal setting (discussion leader: Dr. Sun)

Values are distinct from preferences. Working in an environment where our core values are fulfilled (having a high fit congruence) is associated with many positive outcomes, e.g., better performance, attitude and health. While it might be difficult to find a job that perfectly aligns with our values, tradeoffs in values are also often inevitable, value frameworks still provide a useful basis for self-reflection and decision making, helping us reflect on our core values and linking them to important long-term decision making, such as future career and the kind of life we want to live.

Article of Discussion -- (Schwartz, 2012)

# Structural Topic Modeling (discussion leader: Moana)

Structural Topic Modeling (STM) is a text analysis method that uncovers latent themes within large collections of documents while incorporating document-level metadata into the modeling process. Unlike traditional topic models, STM allows researchers to examine how topics vary across groups, utilize metrics such as semantic coherence and exclusivity, tools for selecting an appropriate number of topics, and a rich set of visualization techniques to facilitate interpretation. In this session, Moana walks us through the complete STM workflow in R on a harassment dataset. We can see how she prepares text data, estimates the model, evaluates competing topic solutions, and interprets the resulting topics through visualizations - providing insights to which the contexts and steps to apply STM.

Presentation slide by Moana

# Leisure interests (discussion leader: Christine)

Leisure interests are an important yet often overlooked aspect of individual differences, shaping how people spend their free time, recover from work, and develop meaningful experiences outside of their careers. Christine’s study walks us through how various leisure interest domains relate to vocational interests through the RIASEC framework. Beyond understanding these relationships, the study encourages us to think more broadly about the future of leisure interest assessment. As leisure activities continue to evolve alongside societal and technological changes, have people’s leisure interests changed substantially over the past decade or two and how can we capture these emerging interests? It might also be interesting to think about how leisure interests relate to broader interest constructs, including basic and vocational interests, and how our leisure interests influence work-life balance, psychological recovery, well-being, and other important life outcomes.

Presentation slide by Christine

# AI ethics (discussion leader: Xavier)

As AI continues to offer new possibilities for enhancing and streamlining the research process, it is critical to understand how to use these tools in an ethical, transparent, and responsible manner. APA and federal register copyright guidelines emphasize the importance of documenting AI use, appropriately acknowledging AI-generated content, and maintaining human accountability throughout the research process. While we can utilize AI in various research tasks after following ethical guidelines, its use still raises important ethical questions. How should researchers address biases embedded in AI systems that stem from biased training data? Would AI enhance human scholarship by allowing researchers to focus on higher-level thinking, or could it inadvertently diminish critical thinking and scientific rigor? As AI gets better and better, how will this change the research landscape? 

Presentation slide by Xavier