# Interpret the key results for Simple Correspondence Analysis

Complete the following steps to interpret a simple correspondence analysis. Key output includes principal components, inertia, proportion of inertia, quality, mass, and several graphs.

## Step 1: Determine the number of principal components

Use the proportion of inertia to determine the minimum number of principal components, also called principal axes, that account for most of the deviation from the expected values in the data. Retain the principal components that explain an acceptable proportion of total inertia. The acceptable level depends on your application. Ideally, the first one, two, or three components account for most of the total inertia.

If the minimum number of principal components needed does not match the number of components that you entered for the analysis, repeat the analysis using the appropriate number of components.

## Step 2: Interpret the principal components

Use the quality values to determine the proportion of the row inertia or column inertia represented by the components. Quality is always a number between 0 and 1. Larger quality values indicate that the row or column is well represented by the components. Lower values indicate poorer representation. The quality values for the rows and columns can help you interpret the components.

Use the contribution values for the rows and/or columns assess which column and row categories contribute most to the inertia of each component. To visually interpret the components, use a row or column plot.

## Step 3: Examine relationships among categories

Examine calculated inertia values for the row and column categories and look for possible associations. Categories with stronger associations have a higher inertia value, which indicates they contribute more to the total chi-squared value.

You can also use an asymmetric row or column plot to visually examine possible relationships. For a row plot, the closer a row profile is to a column vertex, the higher the row profile is with respect to the column category. For a column plot, the closer a column profile is to a row vertex, the higher the column profile is with respect to the row category.

By using this site you agree to the use of cookies for analytics and personalized content.  Read our policy