To strengthen the connection between the statistical community and National Science Foundation, we continue the series introduced in the May 2023 issue that poses questions to NSF program officers and awardees. If you have questions or comments for the program officers, send them to ASA Director of Science Policy Steve Pierson.

For the program officer Q&A this month, we have questions and responses from both Cheryl Eavey of the Social, Behavioral, and Economic Sciences Directorate and Yulia Gel of the Division of Mathematical Sciences in the NSF Directorate for Physical and Mathematical Sciences.

The awardee responses are from Dan Kowal of Rice University, who won an award from the Social, Behavioral, and Economic Sciences Directorate.

Program Directors

Cheryl Eavey is program director of the Methodology, Measurement, and Statistics Program in the Division of Social and Economic Sciences at the National Science Foundation. She has been at the foundation since 1993. Additionally, Eavey has been involved in several cross-program and cross-directorate NSF activities, including Harnessing the Data Revolution and Cyberinfrastructure for Sustained Scientific Innovation. She has a bachelor’s degree in mathematics and political science from Valparaiso University and a master’s and PhD in political science from Michigan State University. Eavey has served on the faculty of the political science department at Florida State University, the business school at Washington University in St. Louis, and the US Business School in Prague, Czech Republic.

What is the Methodology, Measurement, and Statistics Program?

The Methodology, Measurement, and Statistics Program supports the development of innovative analytical and statistical methods and models for the social, behavioral, and economic sciences. It seeks proposals that are methodologically innovative, grounded in theory, and have potential utility for multiple fields within the social, behavioral, and economic sciences.

Does the program support statisticians?

Yes, the program supports relevant work from the statistics community. It also supports economists, psychologists, political scientists, geographers, sociologists, and others—all working on the development of methods for the social, behavioral, and economic sciences.

Does a statistician need to collaborate with a Social, Behavioral, and Economic Sciences Directorate scientist to be considered by the program?

No, a collaboration is not required. That said, the Methodology, Measurement, and Statistics Program welcomes proposals that involve collaborations between statisticians and Social, Behavioral, and Economic Sciences Directorate researchers.

Does a proposal to the program need to include applications?

Technically no, a proposal does not need to include specific applications to be in scope for the program. At a minimum, however, a proposal should demonstrate the value of the proposed methods for the social, behavioral, and economic sciences community. And proposals that include applications with real-world data in areas supported by the social, behavioral, and economic sciences are most welcome.

What are the differences between the Methodology, Measurement, and Statistics Program and the Statistics Program in the Division of Mathematical Sciences?

If we think of statistical research on a continuum from basic to applied, one could characterize the Statistics program as more on the basic end and Methodology, Measurement, and Statistics as more on the applied end. Methodology, Measurement, and Statistics–supported projects would be considered basic from the perspective of the social, behavioral, and economic sciences, but not perhaps from the perspective of the Statistics Program. That said, there is some overlap in the missions of the programs. Because of that, Methodology, Measurement, and Statistics and the Statistics Program have a long history of co-review.

Is there anything else we should know about the Methodology, Measurement, and Statistics Program?

Since 1999, Methodology, Measurement, and Statistics has partnered with a consortium of federal statistical agencies to further research related to the production and use of official statistics. Over the years, we have had to pause this activity to accommodate administrative procedures. Currently, the activity is on hold, but we are working to reactivate the partnership.

Yulia Gel is on leave from The University of Texas at Dallas. This is her third year as a rotator program director of the Statistics Program of the Division of Mathematical Sciences in the NSF Directorate for Physical and Mathematical Sciences.

If I move to a different university within the United States, what shall I do with my NSF award?

Typically, when the principal investigator moves to a new organization and both their original and new organizations agree, the grant can be transferred to the principal investigator’s new institution, subject to NSF approval. The decision to approve such a transfer request is based on the following three conditions:

• The principal investigator’s original institution agrees to relinquish the project before its expiration
• The principal investigator’s new institution offers facilities and resources that are sufficient for the successful continuation of the project
• There are no significant changes in the research scope and budget with respect to those described in the originally approved project

Each transfer request must be initiated by either the principal investigator or the principal investigator’s organization. As the first step, the principal investigator shall notify his/her cognizant NSF program officer about the expected change of institutions and request preliminary approval for the transfer. Then, if the NSF and both institutions agree, the principal investigator shall work with his/her sponsored projects office to initiate the formal notification of transfer. Such a formal request should be electronically initiated by either the principal investigator or the principal investigator’s organization through Research.gov. Upon receipt of the request, an NSF program officer will review it and, if it is approved, a new grant number at the new organization will be established.

It is important to note that the principal investigator’s original institution has the prerogative to agree to the transfer, nominate a substitute principal investigator, or request that the grant be terminated and closed out. For details, see Chapter VII.B 2(f).

I have a CAREER award and plan to move. Are there additional requirements?

If you plan to move to a new institution that is CAREER-eligible, the process is the same as above, except you need to secure a new departmental letter. The new departmental letter must document support for the project goals as described in the original proposal or in a revised scope and provide a plan for mentoring the principal investigator.

If you plan to move to a new institution that is not CAREER-eligible (e.g., outside the United States), your CAREER award shall be relinquished; however, funding may be extended to allow the students or postdoctoral researchers to continue to be supported on the award for the remaining funded year. You should contact your cognizant program officer as soon as you know you will move to inquire about this possibility.

Awardee

Dan Kowal is the Dobelman Family Assistant Professor in the department of statistics at Rice University. He works on Bayesian models and algorithms for large and dependent data, with primary applications in public health, epidemiology and environmental justice, and economics/finance.

Kowal was recently funded through the Methodology, Measurement, and Statistics Program in the NSF Directorate for Social, Behavioral, and Economic Sciences. He has experience submitting proposals to the NSF, but this was his first attempt outside the Division of Mathematical Statistics.

What NSF entity funded or contributed, and how will funding be used?

The proposal “Adaptive Dependent Data Models via Graph-Informed Shrinkage and Sparsity” seeks to improve Bayesian modeling and computation for dependent data such as time-ordered data, high-resolution images, and spatially-referenced measurements. Models that appropriately capture and leverage the dependencies in the data can extract clearer signals, provide more accurate imputations of missing data, and communicate more precise uncertainty quantification.

The proposal was funded by the Methodology, Measurement, and Statistics Program in the Directorate for Social, Behavioral, and Economic Sciences for $287,536. The award includes support for the principal investigator’s summer salary, PhD student funding, undergraduate summer research stipends, and conference travel.

What will the proposal accomplish?

The primary innovation of this proposal is the targeted deployment of the data dependence structure along multiple stages of the Bayesian modeling pipeline:

• The model for the signal to provide smoothness and regularization
• The accompanying shrinkage or sparsity prior for enhanced local adaptivity
• The computational and numerical strategies for scalable posterior inference

The data dependence is encoded as a graph that links observational units, allowing broad applicability for important problems in local elections and redistricting; inflation modeling and forecasting; spatial pattern extraction for economic, health, and urban data; and functional regression for monitoring and exposure data.

Describe your approach to the Methodology, Measurement, and Statistics Program.

The idea to target the Methodology, Measurement, and Statistics Program was suggested by a senior colleague, Marina Vannucci, a few years ago. She mentioned the Methodology, Measurement, and Statistics Program often funds statistics proposals—often quite generously. I read through the recent awards granted by the Methodology, Measurement, and Statistics Program on the NSF website and found many were interesting and similar in nature to the statistical problems and application areas in the proposal I wanted to write. I took this as a good sign and decided to go for it!

What advice do you have for others applying for NSF funding?

First, participate in an NSF review panel. You will see a variety of proposal styles, scopes, and topics and learn how panelists discuss and evaluate proposals. It’s also important to see firsthand that many good proposals (with names you recognize) don’t get funded!

Second, talk to others about the grant process—not just senior mentors, but also peers and colleagues. There’s great value in a two-way exchange of stories, successes, and frustrations.

Finally, explore the continuum of “ambitious new project” vs. “logical extension of your work.” There’s no one-size-fits-all solution—different projects can fall in different places on this spectrum.