Nearly 400 statistical practitioners and data scientists convened in New Orleans February 27–29 to attend the 2024 Conference on Statistical Practice. The conference program included eight short courses, 52 poster presentations, and 21 concurrent sessions.
One of the hallmarks of CSP is it brings together applied statisticians who work in diverse settings to share new statistical methodologies and best practices in statistical analysis, design, consulting, and programming. By the end of the conference, participants had listened to practical and insightful advice, shared career advice, and established new friendships.
Opening Keynote Speaker
Cynthia Rudin, Earl D. McLean Jr. Professor of Computer Science at Duke University, opened the conference with the keynote address titled “Simpler Machine Learning Models for a Complicated World.” Rudin directs the Interpretable Machine Learning Lab with the goal of designing predictive models people can understand. She is the recipient of the 2022 Squirrel AI Award for Artificial Intelligence for the Benefit of Humanity from the Association for the Advancement of Artificial Intelligence (the “Nobel Prize of AI”).
She discussed that although the trend in machine learning has tended toward building more complicated (black box) models, such models have not shown performance advantages for many real-world data sets. Instead, simpler models (sometimes small enough to fit on an index card) can be just as accurate. However, the design of interpretable models for practical applications is challenging.
Rudin presented a new paradigm for machine learning that gives insight into the existence of simpler models for a large class of real-world problems and solves the interaction bottleneck. In this paradigm, machine learning algorithms are not focused on finding a single optimal model, but instead capture the full collection of good (i.e., low-loss) models, which are called “the Rashomon set.” Finding Rashomon sets is extremely computationally difficult, but the benefits are massive.
Concurrent Presentations and Panels
Concurrent presentations and panel discussions were based on the following four themes:
1. Career, Professional, and Leadership Development
2. Study Design and Data Management
3. Implementation and Analysis
4. Effective Communication
Within the Career, Professional, and Leadership Development theme, there were several panels on topics such as project management tools, ethics, statistical consulting training, and pathways and personal development for leadership. Concurrent session topics included reproducible research practices, building a strong collaborative workforce in biostatistics, and a comparative analysis of contractual risks in statistical consulting.
In the Study Design and Data Management theme, topics ranged from study design for generalized linear mixed modeling to data quality assessment.
Implementation and Analysis presentations described research on fisheries, mental health disparities, and health outcomes research.
Writing, publishing, building trust within a collaborative research group, communicating technical topics to nonstatisticians, and documenting collaborative efforts as statisticians in departments where leaders are nonstatisticians rounded out the Effective Communication theme.
Networking and Connecting
The more intimate structure of CSP facilitates interaction among attendees, which encourages discussions about conference programs. The steering committee made a concerted effort to facilitate networking—particularly among first-time CSP attendees.
Opportunities included themed dinners, guided networking, and interactive card games. The topical-themed dinners included guided discussions about leadership, early-career professionals, and ethics. Some session panelists hosted group discussions that continued the panel conversations at dinner. Many attendees joined these group dinners to explore the topics in depth while enjoying the food scene in New Orleans and meeting new colleagues.
At the last session break, guided networking promoted the interaction between participants and experts from academia, industry, and government. Jeremiah Aakre from the Mayo Clinic, Amy Crisp from the University of Florida, Beth Ely from the US Centers for Disease Control and Prevention, and Lauren Wilson from Sandia National Laboratory shared advice about topics such as imposter syndrome, career transitions, and leadership.
Closing Panel
CSP 2024 concluded with a panel discussion about the special issue of Stat titled Statistical Consulting and Collaboration. The panelists discussed the success metrics from the special issue and pathways for applied and collaborative statisticians to disseminate their work in the future.
Closing panelists included the following:
- Helen Zhang, University of Arizona
- Lee-Ann Hayak, National Museum of Natural History
- Xiaoyue Niu, Penn State University
- Ji-Hyun Lee, University of Florida
- Robyn Ball, The Jackson Laboratory
- Harry (Dean) Johnson, Washington State University
The Journal of Statistical Theory and Practice will host peer-reviewed manuscripts from selected authors who presented at CSP 2024. The topical collection is targeted for publication in spring 2025.
Student Poster Awards
The following three students were awarded ASA student membership for their outstanding poster presentations:
- Joshua Cook from the University of West Florida presented work about a comprehensive analysis of the statistical considerations involved in developing a consensus statement.
- Xingruo Zhang from The University of Chicago presented work that extended an existing framework of R-squared measures for regular mixed-effects models. The proposed framework applies to two specifications of the random location effects: random intercepts with covariate-influenced variances and random intercepts combined with random slopes of observation-level covariates.
- Maria Camila C. Mejia Garcia from The University of Texas Rio Grande Valley presented work applying imputation techniques for classification implemented in a simulation study of a data set from a Brazilian Hospital.
Travel Award Winners
Garrett Duncan was awarded the John J. Bartko award for financial support to attend the conference. Duncan completed his master’s program earlier this year at Brigham Young University and would like to pursue a PhD.
Xingruo (Summer) Zhang was selected to receive the Lester R. Curtin Award to travel to CSP. Zhang will graduate in June 2024 from The University of Chicago’s PhD program.
The Lingzi Lu Memorial Award was presented to Kristin Gaffney, who completed her Master of Public Health degree with a biostatistics concentration at the University of Nebraska Medical Center College of Public Health in 2022.
Read more about these student award winners on STATtr@k.
Future of CSP
The statistics landscape is constantly evolving, and we recognize the need to adapt and innovate to meet the changing needs and expectations of our community members. In 2025, we will be re-envisioning CSP and—to allow for a productive process—the ASA has decided not to host the conference. This is going to afford us the opportunity to consider how CSP might look going forward.
Your input and feedback are invited and especially valued since CSP has always been focused on attendees’ needs and wishes.
2024 CSP Steering Committee Members
Julia Sharp, National Institute of Standards and Technology (Chair)
Margaret “Maggie” Betz, Purdue University (Vice Chair)
David Aaby, Northwestern University
Caroll Co, DLH
Kimberly Cressman
Allison M. Florance, Novartis Pharmaceuticals
Emily Griffith, North Carolina State University
Elena Khusainova, Yale University
Mary Jeanne Kwasny, Northwestern University
Donna LaLonde, American Statistical Association
Christina Link, American Statistical Association
Anne Meibohm
Sudeshna Paul, Emory University
Ralph (Mac) Turner, Phoenix Data Analytics
Emma Zohner, Rice University
Leave a Reply