Essentially all researchers use statistical methods to determine whether observed results provide evidence to support their hypothesis or could be due to chance alone. Functionality and ease-of-use are key factors dictating researchers? choice of statistical software. Software packages using frequentist statistics have dominated the marketplace. Current trends in software development include a focus on user-friendliness, often by incorporating drop-down menus and point-and-click options. JMP and SPSS are successful examples of this approach in the frequentist statistics market. Bayesian analysis is another powerful branch of statistics that is gaining popularity as it offers researchers a rigorous and robust approach to incorporate prior knowledge into research in a quantitative manner. Unfortunately, currently available Bayesian software packages require an intensive amount of programming expertise, such as manual specification of the mathematical form of statistical models, something many investigators lack. Few user-friendly, point-and-click Bayesian analysis software packages currently exist. Rapid advances in statistics have increased journal reviewers? expectations regarding sophistication in statistical analyses applied to support research findings. A recent survey among statisticians working in medical product development showed that ?lack of appropriate software? is a major hurdle to implementing Bayesian methods, and that having a ?user friendly software [package]? is needed. This need is likely higher among researchers without formal training in statistics. Bayesic Technologies, LLC, aims in this Phase I STTR proposal to demonstrate the feasibility of creating a user-friendly, point-and-click, Bayesian software of limited scope, BSTAT 1.0, to make Bayesian statistics more accessible for researchers and clinicians. This prototype will include the three primary capabilities common to all statistical software: i) import and process data, ii) compute descriptive statistics and generate plots, and iii) perform inferential estimation and hypothesis testing. The overall goal of this research is to design, develop, and commercialize software that will enable researchers to conduct Bayesian statistical analysis without requiring them to have significant programming skill.
Aim 1 : Develop accurate and valid functionality. This prototype will allow users to i) specify a prior distribution and ii) output a posterior distribution after quantitatively incorporating the researcher?s own data. The scope of inference will be limited to generalized linear regression models.
Aim 2 : Develop a user-friendly graphical user interface. Users will be able use point-and-click/dropdown menus to perform Bayesian analyses. Colleagues from public health programs will be asked to beta-test the software.
Aim 3 : Develop touch-screen functionality to allow users to specify hand-drawn distributions. We will incorporate state-of-the-art technology so users can draw a variety of distributions, which will then be converted to an internal representation that satisfies the laws of probability and is suitable for further numerical analysis.

Public Health Relevance

The analysis, summary, and interpretation of data is fundamental to biomedical and population health research, surveillance, and most other activities. The goal of this STTR Phase I project is to prove the feasibility of developing a user-friendly, next-generation software package that enables the application of Bayesian statistical methods to help public health practitioners and medical researchers meet biostatistics- related core competencies as prescribed by the American Schools and Programs of Public Health. These analytic methods and tools contribute to the improvement of medical R&D and human healthcare outcomes.

National Institute of Health (NIH)
National Institute of General Medical Sciences (NIGMS)
Small Business Technology Transfer (STTR) Grants - Phase I (R41)
Project #
Application #
Study Section
Special Emphasis Panel (ZRG1)
Program Officer
Brazhnik, Paul
Project Start
Project End
Budget Start
Budget End
Support Year
Fiscal Year
Total Cost
Indirect Cost
Bayesic Technologies, LLC
Oklahoma City
United States
Zip Code