Test populations for normality
 
  Skip Navigation Links
 

Origin: Normality Tests

Normality tests are for testing whether the input data is normally distributed. It is required by some complicated statistical tests such as Student's t-test, one-way and two-way ANOVA, because they make assumptions that data comes from a normally distributed population, and if such assumptions are not valid, the results of the tests will be unreliable.

Origin provides several normality test to determine whether or not a sample of values follows a normal distribution. The sample size N, the mean, the standard deviation, the SE of Mean, the statistic, a p-value, and a decision rule are output to a Report Sheet.

Customers can set the following features while performing normality tests:

  1. Normality test method
    • Shapiro-Wilk method
    • Kolmogorov-Smirnov method
    • Lilliefors
    • Anderson-Darling
    • D'Agostino-K squared
    • Chen-Shapiro

    Note: if the Kolmogorov-Smirnov method is chosen, there are two options for you to compute the statistics which are Estimated (use the Mean and Variance estimated by sample data) and Specified (use the Mean and Variance specified by customers).

  1. Plot type: Histogram and Box Chart

A sample result is shown below. Besides the Normality Test results, the report sheet also includes a Descriptive Statistics table, including Sample Size, Mean, Standard Deviation, and SE of Mean.

Origin includes the ability to automatically recalculate the analysis results of the Normality Test operation any time you change the parameters or update your source data. In addition, the settings for this analysis routine can be saved to an analysis theme for use later with similar data.

 

Skip Navigation Links.
Copyright © 2012 OriginLab Corporation. All rights reserved.
20+ years serving the scientific and engineering community