To an ever-increasing extent, pathologists are being required to use statistics in their practice. In clinical pathology or laboratory medicine, statistics are a fundamental requirement for the evaluation of the reliability of quantitative results for values of serum and, in general, body fluid analytes such as electrolytes, glucose, blood urea nitrogen (BUN), creatinine, critical enzymes, etc. and for the analysis of the correlation between the results generated on different analyzers, all of which are used for quantitative determination of the same analytes. Correlations between measurements of parameters that allow categorization of tumors such as correlation of nuclear grade with pathological stage require knowledge of statistical methods in anatomic pathology. Correlation of the staging of different cancers with survival involves another major use of statistics in both anatomic and clinical pathology. Often, pathologists utilize statistical methods without knowledge of the physical and mathematical basis that underlies the particular statistics that they are using. This can give rise to erroneous conclusions. For example, many, but certainly not all, quantitative analyses for analytes follow so-called Gaussian statistics, an example of parametric statistics, with a known mathematical form for the distribution of values that gives rise to the “bell-shaped curve,” called the normal distribution. This involves computation of means, standard deviations, confidence intervals for means, and a number of other parameters. However, these methods cannot be used for analyte values that do not follow Gaussian statistics which requires that the distribution of values for a given analyte be distributed in what is termed a normal distribution as represented by the so-called bell-shaped curve. This can affect determinations such as the reference ranges for analytes based on values determined from presumed normal or well individuals. If the distribution of values is assumed to be Gaussian, the range would be computed as the mean of the values plus or minus two standard deviations from the mean. However, if the values actually do not follow a Gaussian distribution, serious errors can be made in establishing the reference range which may be too narrow or too wide. Not infrequently, the use of non-parametric statistics rather must be used in establishing reference ranges. We are currently living in what has been termed “the age of metrology.” This means that, to an increasing degree, statistics govern most aspects of laboratory medicine including whether or not values can be accepted as being “true” or v