Ruth B. Noller Research Grant

The Creative Education Foundation (CEF) is proud to present the 8th Annual Ruth B. Noller Research Grant to PhD Candidate, Simone Luchini. Luchini is a PhD student in the Cognitive Neuroscience of Creativity Lab at Penn State University. The portion of his research supported by the Noller Grant focuses on the development of equitable automated creativity assessment tools.

Luchini asserts, “It is critical that educators should have access to quick and accurate tools for the assessment of creativity, so that the latter may be fostered in students, and they may be prepared for the future labor market.” Through his research, Luchini seeks to explore language-based bias in assessment tools that rely on English. He adds, “These tools should be proven to be algorithmically fair, meaning they should not be biased against diverse students. One major barrier to this goal remains the reliance of modern artificial intelligence technologies on mainly linguistically homogeneous training data (i.e., mostly native English speakers).”

Luchini theorizes that “[t]he consequence of training deep learning models on openly available text-data (i.e., the internet) is that they will learn indiscriminately from the data, paying more attention to whatever patterns they find to be most common within it. If most of the data a model learns from is generated by native English speakers, this may lead to the undesired effect of making the model “prefer” native English speech in the future. My research will attempt to shed light on whether this linguistic bias is present in modern automated creativity assessment tools.”

Luchini describes his methodology as follows:

I will collect responses to an idea generation task from a series of engineers, either native or non-native English language speakers. I will then test how large language models perform when predicting the creativity of each of these responses and see how they face off against human-judgments. This project will allow for recommendations to be made with regards to the proper use of current automated creativity assessment tools in education. This work will further serve to identify any room for improvement in these tools when it comes to their fairness towards linguistic diversity.

CEF is thrilled to support this cutting edge creativity research that exemplifies the spirit of creativity pioneer, Ruth B. Noller. We would also like to thank our Noller Grant Committee for their dedication and service: Liz Monroe-Cook, Committee Chair, Jean Bakk, Branko Broekman, Cyndi Burnett, Clare Dus, Tim Hurson, Karina Loera-Barcenas, and Walt Stevenson.

Dr. van Broekhoven’s research explores the extent to which long term interactions between group members influence the emergence of creativity over time. She observes that the “input-process-output” or IPO model is most typically used to study group productivity, but notes that this model is generally restricted to observing interactions between group members at a single point in time, rather than over time. She notes that “…the emergence of creativity via communicative patterns between group members over time remains underexplored, while this process shapes the production of ideas. As such, based on the sociocultural approach to creativity, we aim to identify blocking and enabling communication patterns that emerge from interactions between students, and to investigate how these communicative patterns affect group creativity over time.”

Dr. van Broekhoven’s research design focuses on five groups of undergraduate students in four sessions in which students identify and define a lifestyle problem, and develop an innovative solution. She will conduct a fine-grained, moment-to-moment interaction analysis where every turn of a participant will be coded. By coding every turn of a participant, she hopes to identify communicative patterns in how ideas are proposed, and responses of team members. Communicative patterns will be correlated to the quality of each generated idea to answer the question as to how blocking and enabling communicative patterns between group members influences the emergence of creativity over time.

Dr. van Broekhoven’s research will bring creativity scholarship further by investigating the moment-to-moment interaction between group members that shapes the production of creative ideas. She will test her assertion that group creativity can only be properly understood in the context of what other people say in the preceding and succeeding turns; what actually happens in the moment-to-moment group interaction over time remains an unexplored research path. As she states, “By doing this, I aim to open up this black box of group creativity, and how we can cultivate group creativity. Second, this research will not only use traditional scoring methods, but will also apply novel computational methods to assess creativity via natural language processing (i.e., semantic distance).”