Gylling Data Management
Providing an Industry Standard Software for Research

Statistical Verification

To provide transparency of ARM calculations on reports, we provide examples that can be replicated by clients in other statistical packages (such as SAS or R) to verify the reported values in ARM.
 
The below Verification Packages address one or more ARM features. Each link below opens a page describing the theory and calculations performed by ARM. These pages may contain links to the following types of files:
  • ARM Report pdf - the report generated in ARM, with the verified values highlighted.
  • Report sets - the ARM report set used to generate Report.pdf in ARM.
  • SAS file - a SAS script for duplicating the analysis in the SAS program.
  • SAS Report pdf - the report generated from SAS used to verify the ARM report.
 
 

 
 
Feature Description
Levene's Test Verifies Levene's test for homogeneity of variance, found on the AOV Means Table report.
Repeated Measures Verifies missing data cases for the Repeated Measures report.
Moments Discussion on Skewness and Kurtosis, common tests for normality of data.
Non-Parametric Statistics
Details the non-parametric statistics that are run when the AR Action Code is used.
This includes the Kruskal-Wallis, Friedman's, and Aligned Rank tests.
Model Selection Criteria Describes the various criteria that can be used when identifying and selecting spatial models for analysis, such as logLik, AIC, and BIC.

Column
Diagnostics

Outliers - Describes how outliers are calculated on the Column Diagnostics panel.
Measures of Spatial Correlation - Examples of how Moran's I is used to identify spatial autocorrelation.
Tukey's 1 DF - Describes the test for interaction between treatments and blocks.
Spatial Biplots - Illustrates how to use the spatial biplot diagnostic plots for spatial modeling.