I would like to compare the mean values of petrophysical properties for the same stratigraphic interval at different borehole locations. These properties are determined using empirical regression equations and are assumed to be valid and applicable. Usually, if the preconditions are met, I would apply, e.g., an independent t-test to compare the means of both groups. However, by doing so I am not able to consider the uncertainty in the determination of the single group values (resulting from the application of the regression equation that should allow prediction of the target properties with a mean error ranging between 5 to 15 %). I want to avoid having a false positive result in the comparing-mean test (that means, seeing a significant difference between both means although this results from the uncertainty of the prediction equation and not from real significant variations between both groups). How can I circumnavigate this problem? Does anyone have an idea? Raw data and all statistical parameter of the applied regression equations are available to me. Big thanks in advance.

More Sven Fuchs's questions See All
Similar questions and discussions