UM Today | Asper School of Business
December 20, 2024 —
What if scholars could capture how a country collectively feels and thinks about a subject, even in an age of information overload?
Wenxi Pu (Associate Dean, Assistant Professor, The Associates Fellow in Innovation) has spent about eight years working to do just that in a piece now published in the FT50-ranked Strategic Entrepreneurship Journal.
Pu and co-authors’ work, “Shifts in national entrepreneurial culture: The promise of linguistic cultural artifacts and machine learning analysis,” looks at national entrepreneurial culture, asking how cultural attitudes have shifted over time.
“One of the things I’m most proud of is how this study offers new ways to capture culture about entrepreneurship at the national level,” says Pu.
With a dataset of close to 700,000 articles published about entrepreneurship and related topics in over 100 regional and national US newspapers (two decades’ worth, spanning 1996 to 2016), Pu argues that a linguistic analysis of a nation’s media production, supported by machine learning algorithms and AI, can reveal and capture cultural attitudes about entrepreneurship.
In other words, researchers can infer how we think and feel about a topic over time by analyzing huge amounts of written data.
When it comes to entrepreneurship in the United States between the mid nineties and 2010s, the general trend is that positivity bias toward entrepreneurship has increased.
“We found the emotional tone trending up and the analytical thinking trending down in the news articles about entrepreneurship that we analyzed, suggesting that we are culturally more emotional while less analytical about entrepreneurship over time,” Pu explains.
Pu’s study found that positivity bias toward entrepreneurship increased the most in accounts of entrepreneurial aspirations and journeys—the starts of startups and the entrepreneurial path.
The also team found a correlation between this rise in positivity about entrepreneurship with the quality of entrepreneurial ventures taking place at the same time. As positive attitudes about entrepreneurship increase, more ventures got started, but they also had lower growth potential.
“Further analysis suggested that this positivity bias might have encouraged entrepreneurs to create new ventures but might have limited the growth potential for those start-ups, so we need to strike a balance,” Pu says.
For Pu, this study has been generative, showcasing innovative ways to imagine culture and opening up a slew of questions to pursue.
“As a method paper, this study contributes more than just these findings about shifts in national entrepreneurial culture. It also allowed us to test a method using machine learning and AI to make this process of qualitative analysis far more efficient,” he says.
As a researcher and the Associate Dean of Teaching and Learning at the Asper School of Business, Pu encourages his colleagues and students to tap into the ways that AI allows us to do what we haven’t been able to do before and the ways that AI allows more time for truly creative, innovative pursuits.
The future is exciting to him, and he looks forward to seeing how AI transforms businesses and our daily lives.
Reading Pu’s work, it is hard not to imagine new applications for both the AI-supported methodology and for the questions the study raises about entrepreneurship itself. It’s certainly generative; more than that, it is creative.
—
The Asper School of Business aims to expand global knowledge and engage in intellectual exploration to advance teaching, learning, and research. Our researchers’ scholarly work is regularly published in internationally renowned publications.
Be part of this flourishing research culture and learn more about research programs in management (MSc and PhD) at the Stu Clark Graduate School here.
link
