Compensating for Uncertainty Biases in Health Risk Judgments
Under funding from the National Library of Medicine authorized by the American Recovery and Reinvestment Act, Applied Biomathematics has undertaken a research project entitled Compensating for uncertainty biases in health risk judgments. The project is on-going but has already produced several websites and software programs to which there are links below. These websites and software programs are not intended to provide medical advice, counseling, diagnosis or treatment. They are preliminary products intended merely to demonstrate some features that might be important in medical counseling software in the future. Any opinions or views expressed in these websites or software programs do not necessarily reflect those of Applied Biomathematics, the National Library of Medicine or the National Institutes of Health or any employees of these organizations.
If a site below is not visible in your browser, or if you cannot edit it, please request access from jason(At)ramas(Dot)com.
Jacketing falsely precise estimates with uncertainty and extracting measurement uncertainty from distributions
Bayes interval calculator for medical tests
Compute the probability of having or not having a disease given a positive or negative test
Beyond the statistician's bag of marbles
How binary sampling data informs us when we can’t make the usual assumptions
Perspectives from biology and evolution inform risk analysis and communication
Characterizing distributions for risk analysis
Challenge problems for a review of methods to characterize inputs for quantitative risk assessments
Techniques for effectively communicating uncertainties and risks to humans
Comparing binomial rates with really poor data
Software tools to compare binomial process rates under poor data about successes and failures
Confidence boxes and computing with confidence
Review papers and software for risk analysis based on statistical confidence structures
Distribution-free risk analysis
Series of papers on propagating probabilistic uncertainty without assumptions about distribution shapes
Using expressions like "k out of n" to convey uncertainty about a probability
Can collecting imprecise data worsen the overall uncertainty of statistical estimates?
Occasional lectures and working groups on uncertainty at Applied Biomathematics
Intervalized logistic regression
Review paper on interval censoring in logistic regression
Loss aversions arises from collision of the ambiguity detector with pessimism
List of motivated and expounded research needs for the ARRA project
How do human language speakers express and understand uncertainty in numerical phrases using hedges such as 'about' or 'nearly'
When learning more information necessarily increases your uncertainty
Probabilistic risk analysis with hardly any data
Tutorial on developing probabilistic risk analyses with little or no empirical data
The engineering principle of balanced design extends to uncertainty and risk analyses
Source libraries for robust Bayes analysis based on numerical and conjugacy methods
Blog for developers and beta testers of pbox.r, a source library for probability bounds analysis in R
Questions and answers for users of RAMAS Risk Calc
Statistical approaches for interval-censored data
Critical review of traditional maximum likelihood, bounding, and modeling intervals as uniform distributions
Add-in for Microsoft Excel that allows it to represent and compute with uncertain numbers
Uncertainty in cost-benefit analysis
A related project concerning the proper expression and balance of uncertainties in cost-benefit analyses
Uncertainty perception
Implications of psychometric findings for best practices in communicating risks and uncertainties
Critique of Winkler & Smith's "On uncertainty in medical testing" (2004 Medical Decision Making 24: 654)
Wikipedia entry for Probability boxes
Wikipedia entry for Probability bounds analysis
Applied Biomathematics' paper archive
Applied Biomathematics' public uncertainty websites
Applied Biomathematics homepage
Related projects:
Risk imaging and uncertainty visualization (proposal)
Drug development decisions under uncertain risks (proposal)
Discovering patterns in medical data while preserving patient privacy ("Privacy", NIH)
Compensating for uncertainty biases in health risk judgments ("ARRA", NIH)
Outbreak detection: combinatorial tests for small samples ("Outbreak", NIH)
Uncertainty in early spacecraft design (NASA)
Strategies for risk communication: evolution, evidence and experience ("Montauk", NSF)
Uncertainty projection in engineered systems (Sandia National Labs)
Safe environmental concentrations under uncertainty ("Deconvolution", NIH)
Quality assurance methods for Monte Carlo risk analysis ("QAPRA" NIH)
Probabilistic risk assessment quality assurance ("Risk Calc", EPRI)
Exact statistics for detecting disease clusters in structured environments ("Cluster", NIH)
Software for determining ecological implications of toxicology data ("Ecotox", EPRI)
"I keep saying the sexy job in the next ten years will be statisticians."
- Hal Varian, Google economist, quoted in The New York Times
"...statisticians are the new sexy vampires, only even more pasty."