UTSA researcher launches software to help navigate statistics

UTSA’s H. Paul LeBlanc III recently unveiled the newest version of his web-based, statistical tool, Stat Tree.

OCTOBER 7, 2024 — Five years after receiving funding from the National Science Foundation (NSF), a UTSA communication professor has produced software that could save months of work for researchers, statisticians and data analysts around the world. UTSA Professor of Communication H. Paul LeBlanc III recently unveiled the newest version of his web-based, statistical tool, Stat Tree.

While its development was funded to serve as “market software that reduces the time corporations and organizations use to train data and analytics employees,” Stat Tree has evolved into a free tool that users in all fields can use to select the appropriate statistical test for a given hypothesis or research question.

The tool encompasses 35 different parametric and non-parametric statistical tests. Users are guided through a series of questions and presented with demonstrations of statistical tests in various languages, including SPSS, R, SAS, and Stata. Each test includes a video demonstration, sample output and script for running tests. All scripts can be copied and pasted by users to support their projects.

LeBlanc was first motivated to create the tool in the form of a handout for his students in 2001.

“This platform can benefit many learners and researchers who need straightforward guidance with statistical tools for analysis.”

“Coming from the humanities as an undergraduate, I struggled with learning statistics as a graduate student,” said LeBlanc. “Most difficult for me was connecting the statistical procedures needed for the particular research questions and hypotheses developed within my discipline. When I started teaching undergraduate research methods at UTSA in 2001, I was motivated to show students how to more directly connect theory and methodology to their particular application. My first attempt was a two-page, two-dimensional decision tree which I handed out in class.”

In Fall 2014, LeBlanc began developing Stat Tree as an online interactive tool.

The following spring, he presented the tool at the Innovations in Online Learning Conference.

The project received a $50,000 NSF grant in spring, 2019 for the NSF Innovation Corps, or I-Corps™ program. The software was made publicly available in August of 2023.

While continually developing the tool, LeBlanc went on to present the concept at the Southwest Regional I-Corps conference and subsequently at the National I-Corps conference.

“We asked our peers: What is your biggest pain in choosing a statistical test for your research question?” said LeBlanc.

The UTSA researcher and his team found that some statisticians and data analysts spent up to six months training new staff. Interviewees also said that learning new languages to conduct different types of analysis was another major pain point. Organizations and educators tended to choose one language consistently, meaning that researchers must adjust to a new language when they begin a new job or a new class. LeBlanc believes his tool can dramatically expedite that adjustment period, allowing researchers to quickly translate their research questions into their chosen language.

“This platform can benefit many learners and researchers who need straightforward guidance with statistical tools for analysis,” said Seok Kang, professor and director of digital initiatives for the UTSA College of Liberal and Fine Arts. “Our students need exposure to coding-based statistical analysis as the tools are open-source and offer various visualization options.”

The latest release, version five, introduces enhanced functionality including demonstrations for six tests (Chi-Square, Paired and Independent Samples t-Tests, One-way ANOVA, Pearson Correlation, and Simple Linear Regression) in both Python and Julia. It also offers demonstrations for univariate descriptive statistics in Python and Julia.

“Any discipline that uses a variety of statistical measures will find this tool invaluable,” said Leslie Doss, former associate professor of practice. “With the addition of Julia and Python — the most widely used statistical languages by many elite universities — the program can now be used to properly select and use the correct statistical method, whether users are doing basic analysis or post-graduate research.”

LeBlanc hopes that the tool will be widely adopted by researchers around the world, in academia and industry alike.

“I want to ease the pain of students, researchers and data analysts globally, and help them answer their research question without learning a new programming language,” added LeBlanc.

— Audrey Gray