Document LJwYx0Q4NeOMK4o44aXBBXxr5

' @ Johnson Subject Manuscripts SubjecDtr. Wehner's Manuscripts Deposition Stude New Brunswick N.J. January 12 1977 File 157 File . TO Dr. B. Semple In my earlier letter to Mr. Buschbom I pointed out several suggestions to improve the presentation of the original paper For the paper on the Leaching of Radionuclides I offer some additional suggestions which would standardize the statistical terminology closer to its current usage in the statistical literature Some of these could also apply to the original paper The suggestions which follow are based largely on a distinction between the unknown constant parameters of a mathematical distribution and the estimates of these parameters which are calculated from the data and which vary because of sampling variability and possibly other causes 1 Current statistical practice usually restricts the use of unknown Greek characters such as or to represent the parameters of a distribution and Roman charactersusch as X and s to represent the estimates of the parameters which are computed from experimental data Many writers also use modified Greek characters such as and 8 hat or sigma to represent estimates of parameters 2 In the past the terms mean and average were frequently used interchangeably and to some extent this confusion still persists in the literature The best current usage how- ever restricts the term mean to represent the unknown parameter of a distribution actually the first moment of a distribution and the term average to the estimate of the parameter calculated from data 3 In the case of the parameter standard deviation actu- ally the square root of the second moment about the mean there is no recognized simple word to represent the estimate calculated from data therefore the most acceptable terms are the awkward expressions estimate of the standard de- viation or the sample standard deviation By consistently using s or 8 to represent the estimate one can get around the awkward expressions A. In the paper averages are frequently shown along with to or 20 to demonstrate the variability in the data Using the notation described above these should be - or 2s if the s represents the estimate of the standard deviation of individual observations On the other hand if the au- thors have computed the standard error of the average then JNJTALC000444527 cont 4 this should be so stated or the correct notation should be employed as follows sin S = Vn where standard S = the estimate of the deivation individual observations of sxn n = the standard error of the average or the deviation estimate of the standard of averages of size n = number of observations used to compute the average presentation Some authors use son instead smaller and the ever term is used it should of s because the former thereby looks better be so stated is much Which- jdm Heinze CC Mr. Mr. G. G. E. E. Heinze Heinze : Dr. G. Hildick Mr. G. Lee Dr. G. Lord Dr. R. Rolle Dr. W. T. Sherman DivDestian vestan J. Sivertson _/ RECEIVED RECEIVED JAN 17 1977 GEORGLEEE JNJTALC000444528