• Statistical Design & Analysis
• Experimental Design
• Longitudinal Analysis
• Modeling & Simulation
• Program Evaluation
• Psychometrics
• Qualitative Analysis
• Complex Data Analysis
• Biostatistics
• Statistical Genetics & Bioinformatics
• Environmental Statistics
• SUDAAN

Contacts

• 
  919-787-8264
• 
  770-407-4907
• 
  301-770-8209

Experimental Design

RTI researchers embed experimental designs into sample surveys to estimate differences in response rates, differences in alternate levels of incentives, and differences in measurement errors between alternative survey instruments and/or alternative modes of interviewing. We also develop and analyze data from experimental designs for clinical trials, toxicology experiments, and other observational studies in epidemiology and public health and to characterize the performance of analytical laboratory procedures.

Methodologies/Techniques

• Blocking or stratification
• Randomizing experimental units to treatments
• Randomizing the treatment order in cross-over clinical trials
• Confounding and aliasing higher-order effects to minimize the number of treatment combinations in a multifactor experimental setting
• Applying specialized methods of analysis for a variety of designs involving continuous, categorical, and time-to-event data; and for complicated designs, such as cluster-correlated and repeated measures designs

Applications

Clinical Trials

• Early phase dose-ranging and dose-finding clinical trials
• Phase III randomized double-blind clinical trials
• Clinical trials with factorial designs
• Clinical trials with Bayesian designs
• Development of trial design to estimate dose-response relationships
• Investigation of calibration design properties (spacing and replication of standard concentrations) aimed at optimal estimation of detection limits
• Clinical studies
  • Computation of sample sizes and optimal randomization methods
  • Consultation on optimal analytical methods for a variety of outcomes from complicated study designs—for example, cluster-correlated and repeated measures designs

Observational and Cohort Studies

• Case control designs
• Prospective cohort designs
• Longitudinal designs
• Registries
• Vaccine efficacy studies

Laboratory Experiments

• Toxicology studies: Specialized analyses of data from teratology, developmental toxicology, neurotoxicology, carcinogenicity, and other studies involving administration of single and multiple compounds (dose-interaction studies)
• Assessments of factors affecting the performance of an analytical technique (for example, measuring cyanide in waste water)

Field Studies

• Assessments of optional data collection modes
• Assessments of effects of different participant incentives
• Estimations of population-level impact from introducing a treatment
• Identifications and assessments of the magnitude and characteristics of measurement error and examinations of how these vary temporally

Projects

• Data Coordinating Center for the NICHD Neonatal Research Network. We provide statistical leadership to this multisite clinical research network, designing various types of multiphase randomized clinical trials and observational studies, long-standing registries, and longitudinal studies of children and their caregivers.
• RTI Sleep Study. Designed, conducted, analyzed, and published papers on a cross-over clinical study to determine whether (1) mattress firmness affects sleep quality and (2) the mattress that optimizes sleep quality varies from person to person.
• Developmental Toxicity Testing and Research for the National Toxicology Program of NIEHS. Developed, documented, and implemented optimal methods of analysis for a variety of developmental toxicology studies. Developed methods of analysis for dose-interaction studies involving multiple compounds and cluster-correlated data, with continuous and binary outcomes. Wrote the statistical analysis plans for three dose-interaction studies with incomplete factorial designs, and reviewed all project protocols and final reports for accuracy of interpretation of statistical methods and results.
• Manuscript Preparation for AIDS Drug Interaction Studies for the National Toxicology Program of NIEHS. Provided statistical consulting in the preparation of four manuscripts from previously completed drug-interaction studies in pregnant mice (above project). All four of these manuscripts have been accepted for publication in Birth Defects Research, Part B: Developmental and Reproductive Toxicology.