Advanced Statistical Analysis of Measured Data
The Sample Statistics screen provides a useful tool for computing statistics for a selected subset of measured values.  This screen also provides an option for conducting Normality tests for sample sizes of 50 or more. The Sample Statistics screen is accessed via the Tools menu of the Measurement Data Entry Worksheets or the Statistics menu of the Error Source Worksheets.  Key screen features are described below.

Sample Distribution Statistics

The upper left portion of the screen displays key statistical information for the entire sample of data entered in the Measurement Data Entry Worksheet or the Error Source Worksheet.

Statistics for Selected Cells

This section of the screen allows the user compute statistics for a selected sample subset.  The desired subset is specified by selecting a range of rows of the sample data entered in Measurement Data Entry Worksheet or the Error Source Worksheet.

Test for Normality

The Sample Statistics screen also contains an option for testing the normality of samples of data, with a sample size of at least 50.  Normality testing is used to determine if the sampled data can be assumed to be normally distributed.  Normality testing can be conducted on the entire sample or on the selected cells only.

Normality testing in UncertaintyAnalyzer is done using three different tests:  A chi-square "goodness of fit" test, a "skewness" test and a "kurtosis" test.  of a sample in comparison with what is expected of samples from a normally distributed population.  The resulting Normality Test Statistics are displayed in the bottom portion of the screen.