"It's all about making systems work better."

Professor Roger McHaney started his career in the 1980s building cars, preparing airplanes for take off, and making sure newspapers were hot off the press – or, more precisely, analyzing the big, complicated manufacturing systems behind those things in order to make them as efficient as possible.

Imagine an automated vehicle delivering airplane parts or newsprint across a giant factory floor, for example. Optimizing the path that vehicle takes improves its battery life. Longer battery life means quicker work. Quicker work reduces costs. Reduced costs equal success for the company.

As the Daniel D. Burke Chair for Exceptional Faculty in Kansas State’s College of Business Administration, McHaney’s work now focuses on more human aspects of those big, complicated systems in places like hospitals and classrooms. But, rather than optimizing for efficiency, he’s often more interested in optimizing the quality of the human experience.

“It’s all about making systems work better – about giving better experiences to people. We’re trying to find ways to optimize the resources that exist to make things work better,” he said.

A recent study with colleagues at Ariel University and San Francisco State University considered the social networks formed by students in an online professional MBA course. Students took part in asynchronous discussions throughout the course. After completing the class, they were asked to create a scoring system for what was valuable within the discussion. Then they rated anonymized comments from their discussions based on that rubric.

The study has been accepted for publication in the International Journal of Business Information Systems.

McHaney and his collaborators used social network analysis and sentiment analysis techniques to study the results of the students’ assessments of their peers’ comments, with an emphasis on the impact of gender. They found that men responded more negatively to discussion threads initiated by women. Women, meanwhile, tended to take what McHaney called a “let’s figure it out together” approach.

The team also found that a few people developed into a “central hub that everyone revolves around,” while other people had lots of interactions with those people who are central hubs without making a lot of meaningful contributions to the content of the course. Gender did not have a significant impact on those sorts of behaviors, according to their results.

“By understanding how different people interact we can make the classroom better for everyone,” McHaney said. The research “exposed what was really going on in these classrooms online. In a regular classroom, the professor has this really nice oversight where they can hear what people are saying and watch who is talking to who and so forth. But, in an online social network many times, you’re not privy to what’s happening.”

McHaney frequently uses this approach of taking people’s qualitative assessments of their subjective experiences and applying a quantitative lens to them using computational analysis. Despite the introduction of and importance of these qualitative elements, the work reflects back to what he was doing in factories.

“Developing a social network is very similar to modeling manufacturing systems. You’re essentially creating nodes that connect, and those nodes have information flowing between them,” he said.

“You have a production center, and its output goes to another location. Or you have a social network with information flowing through a system. [In either case, you’re] breaking it into components, how those components interact, and how information or products move between those nodes.”

In another collaboration with Ariel University, the team analyzed nearly 18,000 reports written by doctors and nurses after adverse medical events took place – everything from a patient having a mental health episode in the emergency room, to a negative interaction between a doctor and patient, to a surgery that went badly. Taking advantage of social network analysis techniques and natural language processing techniques, they identified differences in how doctors and nurses perceive the patient experience of these events. This study appeared in the International Journal of Risk & Safety in Medicine.

Most strikingly, it found that doctors perceived that the patients were less harmed than nurses did.

“We’ve logged these vast amounts of data in different medical areas and those data are not necessarily getting utilized to their fullest,” McHaney said. “Here, we were looking at how perspectives differ and which one would be most useful to the hospital in terms of correcting problems.”

The team was able to use the data to study patient flow, and to explore how mobile apps might be used to train radiological technicians to assess patients’ treatment needs after they have X-rays taken.

All of these studies capture McHaney’s perspective on merging the qualitative and the quantitative – which is currently gaining so much attention with tools like ChatGPT, Google Bard, and Microsoft’s Bing AI.

“We’re just on the cusp of a new era for understanding this approach. You’ve got all these AI tools coming out. They’re going to reveal new and better ways of approaching problems that we can’t fully understand,” McHaney said. “We have to apply them wisely to develop insights. And we hope our research is contributing to that while also improving people’s experiences in places like classrooms and hospitals.”