| Research Resources Branch |
Intramural Program Investigators ^ Home ^ |
| Larry J. Brant, Ph.D., Staff Scientist Chief, Statistics and Experimental Design Section |
 Dr. Larry J. Brant received his B.S in Mathematics in 1968 from Frostburg State College, Frostburg, Maryland. He received his M.A. in 1972 in Mathematics from the Pennsylvania State University, University Park, Pennsylvania, and his Ph.D. in 1978 from The Johns Hopkins University, School of Hygiene and Public Health, Baltimore, Maryland. Research Interests: Development of Statistical Methods (in particular, multiple comparisons), Development of Models for Biological Processes, Longitudinal Studies, Aging, Health Screening, Epidemiology of Circumpolar Health, and Combinatorics. The Statistical and Experimental Design Section is responsible for providing statistical and experimental design expertise appropriate to studies of aging and gerontology. Statistical methodology, including the use of Bayesian, maximum likelihood, and numerical computing methods, is applied and developed for longitudinal studies and other studies of aging. A major emphasis is on the development and application of methods that provide cogent, yet easily understood results. |
| The research and development of the Section currently focuses on several types of statistical models. These include 1) longitudinal multi-level models, which use empirical Bayesian methods to analyze the repeated measurements for all individuals in the study population as a function of the between- and within-subject variance estimates, 2) mixture models for describing and identifying high risk or preclinical disease groups of patients based on the distribution of changes in biological markers over time, 3) survival analysis techniques for studying risk factors in follow-up studies, 4) multiple comparisons for addressing the issue of multiplicity in the testing of group differences in experimental or observational designs, and 5) issues of power, sample size, and other experimental design issues. |
| Recent efforts in longitudinal data analysis include the development of an imputation method using estimates from a linear mixed-effects model to correct for measurement error bias in traditional risk factor analyses in both logistic regression and proportional hazards regression models. Also, methods for the prediction or classification of preclinical disease states are being developed using longitudinal measurements of biological markers and multilevel models. Methods developed by the Section have been applied in studies of prostate cancer, pulmonary function, cardiovascular science, long-term caloric restriction in rats, and genome-wide mapping in mice. |
| Contact Information: Research Resources Branch Biomedical Research Center, 04B116 251 Bayview Boulevard, Suite 100 Baltimore, MD 21224-6825 Phone 410-558-8148 Fax 410-558-8333 E mail brantl@grc.nia.nih.gov |
| IRP Home What's New Contact Us Accessibility Disclaimer Privacy Site Search Site Map NIA Home |
 |
 |
 |