The Society for Computation in Psychology’s 2024 conference – SCiP 2024 – featured a theme of Digital Cognitive Science examining the impact of technology and computation on cognitive psychology. The schedule included a keynote presentation by James Pennebaker examining the impact of simple word count software on the exploration of emotion and other social constructs in text writing. Sessions included diverse topics on artificial intelligence, data science, modeling, and the intersection of technology with social sciences. The society emphasizes opportunities for graduate students to present their work, creating an environment where individuals can exchange ideas and insights to shape the future of computation in psychology.
Abi M. Sipes Undergraduate Researcher at Butler University, USA.
Can you tell us a bit about yourself and your research interests?
I am currently wrapping up my undergraduate studies in Computer Science, Psychology and Neuroscience, and I intend to graduate school to pursue a PhD in a field related to Computational Neuroscience. I currently work with an agent-based model, and hope to expand the types of computational models I work with in graduate school while maintaining a focus on psychological and neuroscience research.
What first interested you in this field of research?
My interest in psychological and neuroscience research is from a fascination with how we interact with the world around us, how we learn from experience, and how we make decisions. I have also always had a curiosity about neural changes that occur in individuals with mental health disorders and neurodegenerative disorders. With a background in computer science, I discovered many possibilities of how to explore these topics which is why I want to pursue a career in computational neuroscience, and I am excited to see what specific research questions this leads to.
Can you briefly explain the research you presented at SCiP 2024?
I presented an oral presentation titled “The Digital Despondency: A Computational Model of Depression” summarizing my research project utilizing an agent-based model called Sugarscape. This simulation is representative of a human society with a robust number of metrics to give agents human attributes and experiences where the main objective is to gather resources and survive. Using this model to explore a psychological topic, I implemented Major Depressive Disorder (MDD) into the simulation to analyze how it would change the experience of individual agents, their interaction with one another, and the society as a whole. Our measures to analyze this are the population size and how many time steps the population lasts, the average wealth an agent has, the average age an agent is when they die, and agent happiness. Each measure is done for the overall population, the depressed population, and the non-depressed population. To investigate the impact, these results are either stratified across timesteps or across depressed percentages of the population spanning from 0% to 100%.
How will you continue to build on this research?
The results I presented at SCiP 2024 are from our initial implementation of MDD, so we are beginning work on making this implementation more sophisticated. Now that we know this implementation is possible in the Sugarscape model, we can focus on expanding the implementation and expanding our analyses to validate our implementation and explore more nuances in the results. We want to ensure that we are effectively modeling MDD without having overly declarative code so that the results are organic and come from the functionality of the simulation.
Zhen Xu PhD student at Teachers College, Columbia University, USA.
Can you tell us a bit about yourself and your research interests?
I recently graduated with a master’s degree in Learning Analytics and have now begun my PhD study at Teachers College, Columbia University. Before this, I studied Learning Science and STEM teaching pedagogy at Beijing Normal University. My research interest focuses on education in AI, particularly in applying data science methods to rich educational data to derive empirical insights and inform evidence-based interventions.
What first interested you in this field of research?
As a graduate student, I experienced how technology was transforming the way we learn and teach – from the rapid shift to online education during COVID-19 to the release of GPT in 2022. While Edtech developers were passionately pushing the boundaries of innovation, it is still challenging for educators to balance potential benefits with practical challenges. This prompted me to explore the role of technology in education, particularly in this era of rapid change. Driven by the belief that discussions about these transformations should be grounded in solid, data-driven evidence, my goal is to leverage data science to uncover how AI is reshaping education, address emerging challenges, and inform evidence-based interventions that support both learners and educators in this new landscape.
Can you briefly explain the research you presented at SCiP 2024?
This work was conducted in the AEQUITAS Lab at Columbia with my supervisor, Renzhe Yu, and a computer science PhD student, SKY Wang. We investigated real-world disparities in students’ online communication and examined how these inequality patterns evolved with the introduction of LLM tools. Using computational methods, we analyzed 785,655 forum post-response pairs created by 35,401 students across 2,075 courses at a minority-serving public university in the United States to identify systematic inequalities in forum interactions through an intersectional lens of race and gender. Our findings reveal how minority students could face further marginalization in online communication and how multiple minority identities may exacerbate these inequalities. Moreover, most of the disparity patterns we identified persist after the introduction of LLM tools, with new forms of inequality also emerging. These results could expand our understanding of inequities in students’ online interactions and shed light on the increasingly complex dynamics of student interactions in the LLM era. Our findings could also serve as a foundation and call for further research across fields such as education, psychology, computational social science, and HCI to explore the underlying causes, mechanisms, and actionable solutions to address these disparities.
How will you continue to build on this research?
Our next step is to expand the scope of our analysis to include multiple institutions and examine whether the patterns we identified are consistent across different types of institutions, such as more selective universities versus community colleges. We are currently preparing to process data from other institutions provided by partners in our lab to broaden this analysis. Additionally, we aim to investigate how different instructional contexts might mitigate these inequalities, offering actionable insights for instructors to design more inclusive and equitable learning environments.