Introduction to multivariate analysis (AskiaAnalyse)

Multivariate analysis can be used to identify patterns and relationships in many variables simultaneously, and it can be used in situations where the number of variables you want to analyse is too great to be included in a standard cross-tab (which is also known as bivariate analysis). By running a multivariate analysis, you can understand the dependence or the interdependence of the variables being analysed. You can learn:

• Why the relationship between the variables exists: what are the mechanisms and processes by which one variable is linked to another?
• The nature of the relationship: is it causal or non-causal?
• How general the relationship is: does it hold for people in general, or is it specific to certain subgroups?

Askiaanalyse offers several types of multivariate analysis, in two categories (explanatory methods and descriptive methods):

By variable type, as indicated in the diagram above, the available multivariate analyses are as follows (click the name of an analysis type for further information):

• Numeric or scaled response variables (as both explanatory and explained variables): linear regression

• Numeric or scaled response variables: PCA correlation matrix

• Numeric or scaled response variables and closed variables: typology

• Closed or or scaled response variables: specificities

To perform a multivariate analysis, open the analysis menu and select multivariate analysis. A special analysis definition opens, similar to the standard general tab, but containing options relevant to multivariate analysis. At the top of the general tab, select the appropriate analysis type from the list above. To learn more about a particular analysis type, follow the appropriate link above.