While working on his PhD, Tang explored these questions from the quantitative perspective by looking into how predictive models can be used in an organizational context. His dissertation, as well as several of his other published papers, explored the concept of overfitting. In this modeling issue, analysis is rooted too deeply in one particular data set. In human resources analytics, overfitting can cause unique problems, since it leads to analysis that does not necessarily reflect the diversity of experiences in the workplace. As Tang discovered through his research, organizational analysis is especially prone to overfitting, since it’s tied so closely to the sample size. 

“The issue of overfitting is not new, but it is particularly relevant in an organizational context,” he explained. “Many organizations are small in size, so they typically don’t have the luxury of large data samples to build their models on.”  

More recently, organizations have been countering their data limitations by using machine learning–a relationship that greatly interests Tang. Since machine learning can utilize data beyond the workplace’s own sets, HR teams can use it to deal with overfitting and get more diverse insights on how other organizations ensure fair hiring practices and strong job satisfaction. Though machine learning does a good job of filling in these gaps, Tang warns that it can do more harm than good if there’s not an expert leading the way.  

“In an ongoing research project, we showed in an example that a lack of HR expertise when using machine learning to predict employee turnover may cause an issue called column-wise data leakage,” he explained. “This means that the predictive model demonstrated seemingly great accuracy the first time, but failed miserably to predict in subsequent applications.” Though it’s helpful to have more data to pull patterns from, it’s just as necessary as ever to have people on a team who can understand that data and apply it to real-world situations.  

Still, Tang is excited to see machine learning and other technological advances evolve to the point where they can support human resources efforts. Machine learning and its effects on HR analytics are rapidly developing, and both the benefits and drawbacks provide windows for further study.

"These aspects all present exciting research opportunities," Tang said.

As he begins his first semester teaching at Kogod, Tang looks forward to getting to know his colleagues and students. He’s joining a community full of perspectives and insights on the issues that matter to him, and he’s excited to have conversations and make connections that will help him learn and fine-tune his teaching style. On the research side, he’s also looking forward to studying how emerging technologies like language models and virtual reality interact with HR practices and organizational behavior. To best prepare himself and his students for the future of work, he wants to understand the role that these innovations can play in everything from hiring to what the workplace itself looks like.  

Ultimately, though, the questions Tang asks in his research come back to how to ensure that a workplace is one where all people can grow and meet their goals. From ensuring an equitable hiring process to improving diversity and inclusion in jobs, the field of management is in a position to ensure equity and fairness for employees from all walks of life.  

“I’m especially interested in studying how organizations can build more diverse and productive workforces through hiring,” he said, “so job applicants from diverse backgrounds can have equal and fair access to employment opportunities where they can realize their full potential.”