The statistical analysis of data from Normal distributions, with particular reference to small samples.
Abstract:ESDU 91041, which supersedes ESDU 68014, explains the analysis of samples of less than 30 to predict the behaviour of the population. The analysis includes the prediction of mean values, standard deviations, percentiles and coefficients of variation. Also included is a method for comparison of samples which is particularly useful if it can be used to show two small samples may be combined into one larger sample so improving the accuracy of predictions. It is also useful when comparing the influence of design changes. Tables of the required statistics are included and equations are provided that allow less commonly used statistics to be estimated to an acceptable engineering accuracy. A fully worked example illustrates the major facets of the analysis of small samples. A Fortran program, ESDUpac A9141, is provided. It accepts a sample and outputs the mean and the standard deviation; it also performs a goodness-of- fit test to a chosen significance level. It then calculates, to chosen confidence levels, the population mean and standard deviation, a chosen population percentile and the population coefficient of variation.
- Degrees of Freedom
- Point Estimates
- Standard Deviation
- Standard Normal Distribution
|Data Item ESDU 91041|
This program is only available to subscribers.