Morris Hansen Lecture

Register for free to attend the 31st Annual Morris Hansen Lecture on Tuesday, Nov. 14, 2023, from 3:30 – 5:30 p.m. ET at USDA’s Washington, DC offices at 1400 Independence Ave. SW, Jefferson Auditorium, South Building. The topic of the event is "Model Selection and Its Important Roles in Surveys." A reception will follow from 5:30 – 6:30 p.m. ET on the Whitten Patio.

We are honored to have as Keynote Speaker Dr. Jiming Jiang, Professor of Statistics at University of California-Davis. The discussants will be Dr. Jane Meza, Professor of Biostatistics and Associate Vice Chancellor in the Office of Global Engagement, University of Nebraska Medical Center, and Dr. Andreea Erciulescu, Senior Statistician, Westat.

Keynote Abstract

The practice of statistics often engages challenges of conflicting interests. A classic example is the type I and type II errors in hypothesis testing. In the context of surveys, an important issue that one often must deal with is trying to obtain accurate information from the surveys and, at the same time, keep the cost within the budget. In modern data science, one often faces the dilemma between privacy protection and providing useful information. Such a challenge is set straight in the basic principle of model selection, that is, striking a (good) balance between model fitting and model complexity. Furthermore, criteria in model selection need to incorporate consideration of practical interest. We discuss the important roles of model selection in surveys. Next, we introduce a flexible class of model selection strategies, known as the fence methods. Features of the fence methods include a data-driven procedure of choosing a cut-off or tuning parameter, known as the adaptive fence, and the ability of incorporating practical considerations in the model selection criterion. Some examples in small area estimation are used to illustrate the methods.

Conference Details

Date: Nov. 14, 2023
Time: 3:30 p.m. - 5:30 p.m. ET

A reception at USDA’s Whitten Building (across Independence Avenue) will follow the event.

Registration Information
Previous Lectures

Last Modified: 10/18/2023