SUBMITTED 12/4/15 to sunset.texas.gov

Comments/Concerns

The necessity for thorough statistical evaluations by independent, certified credentialed statisticians evolved from our Coastal Bend Sierra Club Group's Executive Committee’s studying various Railroad Commission of Texas’ (RRC) actions re permit applications from industries applying for the right to explore for uranium and also from disposal companies applying to build facilities to dispose of oil and gas production waste in the Eagle Ford Shale area.  In particular, the analyses pertaining to collecting, analyzing, and interpreting data were deficient in using statistical methodology.  Indeed, we believe that professional statisticians would have questioned the validity of some of these practices.

 Clearly, without statistical validity, there can be no scientific credibility.  Many of the technical decisions made by RRC scientists and engineers require a thorough understanding of mathematical statistics—not simply applied statistics which relies on software to carry out manipulation of data. 

 This statement is not a criticism of RRC’s scientific staff.  Rather, the point is this:  Just as it is unreasonable to expect mathematicians and statisticians alone to make decisions on matters grounded in science or engineering, it is equally unreasonable to expect scientists and engineers alone to make decisions on matters grounded in mathematical statistics.

Members of the Coastal Bend Sierra Club Group appeal to you to find a strategy that will translate into a policy that will provide necessary statistical support for TRRC technical decision makers.  They need and deserve no less.

Proposed Solution

A thorough statistical evaluation by independent, credentialed statisticians should become an integral part of all evaluation of the Railroad Commission of Texas’ (RRC) regulations, permit applications, and summary reports involving collection, manipulation, analysis, or interpretation of data.  (Note:  Data includes assumed or hypothetical values used in mathematical modeling as well as actual measured values.)