UPCOMING WEBINAR: From Thousands of Tables To One Clear Story: using AI in survey analysis Register now >>

Articles in this section

How to Do a Bartlett's Test of Sphericity

  • Updated
Follow

The Bartlett Test of Sphericity is used to check that the correlation matrix of the variables in your dataset diverges significantly from the identity matrix. An identity matrix is a matrix in which all of the values along the diagonal are 1 and all of the other values are 0. If the p-value from Bartlett’s Test of Sphericity is lower than your chosen significance level (common choices are 0.10, 0.05, and 0.01), then we can assume that the variables in the correlation matrix are not orthogonal.

mceclip0.png

Requirements

A Displayr document with a data set that contains two or more variables that have a Structure set as Numeric or Numeric-Multi. Note that binary variables can also be used as inputs, but will be converted to numeric when calculating the correlation coefficients. If you use categorical or ordinal variables, they will be coerced to numeric based on their values for the purposes of running the test.

Method

  1. From the toolbar, select Anything 26UI_Anything.png > Advanced Analysis > Test > Bartlett Test of Sphericity, or from the Report tree, go to + > Advanced Analysis > Test > Bartlett Test of Sphericity.
  2. In the object inspector 26UI_ObjectInspector.png, specify the variables to use under Data > Input Variables.
  3. OPTIONAL: Select Missing data to change the way missing values are handled
    • Select Use partial data (pairwise correlations) to treat each pair of variables separately and only include observations that have valid values for each pair in the data set.
    • Select Error if missing data if you do not want the test performed if there is any missing data
    • Select Exclude cases with missing data if you only want to use cases with valid values on all the variables
    • Select Imputation (replace missing values with estimate) to request that imputation techniques be used to replace missing values with estimates
  4. OPTIONAL: Select Variable names to display variable names in the output instead of the variable labels
  5. OPTIONAL: Select More decimal places to display numeric values with 8 decimal places.

Next

How to Do Principal Component Analysis in Displayr

Related articles