Reply to Paul Smith's Comment


  • Bernard A. Silverman Consulting Meteorologist Centennial, CO, USA


Smith stated that he is skeptical about taking point estimates of the seeding effect literally.  That is a valid concern. They should not be taken literally. After all, point estimates are statistical estimates with an inherent degree of uncertainty that is quantified by the standard error of estimate.  A point estimate with its standard error of estimate is generally used in null hypothesis testing to infer if there is any effect at all. The best estimate of the strength of a seeding effect is given by its confidence interval because it infers a range within which the true effect lies at the specified level of confidence. That is why I emphasized the use of confidence intervals in the Kern program evaluation using a 90% level of confidence.




Technical Notes and Correspondence