Fracking Correlated with Higher Crude Death Rate in North Dakota
Analysis of annual data clearly reveals higher crude death rates in North Dakota since the development of horizontal slickwater fracturing for shale gas extraction in 1998.
Chicago, IL, February 26, 2013 --(PR.com)-- Optimal Data Analysis, a research and development LLC, today announced the findings of their initial comparison of crude death rates in North Dakota, before versus after the development of horizontal slickwater fracturing for shale gas extraction. Company president, Paul Yarnold, PhD, said: “Analysis of crude death rate data available from 1937 to 2005 revealed a highly reliable and elegantly parsimonious statistical model:
· “If the annual crude death rate is less than or equal to 8.89%, then predict 1997 or earlier; and
· “If the annual crude death rate is greater than 8.89%, then predict 1998 or more recent.
“This model is accurate, capturing 65% of the improvement above chance which is theoretically attainable (50% is considered a strong result): it correctly predicts 77% of years before 1998, and 88% afterward. Follow-up research known as jackknife validity analysis indicated the model should generalize into the future: predictions of higher annual crude death rates are expected to be 75% accurate—consistent with accuracy achieved prior to 1998, and the model is expected to capture 52% of attainable improvement above chance. Model performance and prediction accuracy can both be further increased if additional correlates of mortality are identified and included in the model. Even though there remains more to learn, the initial finding clearly reveals higher crude death rates in North Dakota since 1998, which emphasizes the urgent need for research aimed at identifying human health correlates of this new technology.”
· “If the annual crude death rate is less than or equal to 8.89%, then predict 1997 or earlier; and
· “If the annual crude death rate is greater than 8.89%, then predict 1998 or more recent.
“This model is accurate, capturing 65% of the improvement above chance which is theoretically attainable (50% is considered a strong result): it correctly predicts 77% of years before 1998, and 88% afterward. Follow-up research known as jackknife validity analysis indicated the model should generalize into the future: predictions of higher annual crude death rates are expected to be 75% accurate—consistent with accuracy achieved prior to 1998, and the model is expected to capture 52% of attainable improvement above chance. Model performance and prediction accuracy can both be further increased if additional correlates of mortality are identified and included in the model. Even though there remains more to learn, the initial finding clearly reveals higher crude death rates in North Dakota since 1998, which emphasizes the urgent need for research aimed at identifying human health correlates of this new technology.”
Contact
Optimal Data Analysis LLC
Paul R Yarnold, PhD
619-550-5861
OptimalDataAnalysis.com
Contact
Paul R Yarnold, PhD
619-550-5861
OptimalDataAnalysis.com
Categories