Statistics – correspondence analysis

 In Statistics & methods

This analysis is popular in market research, but I can see a value to use it in healthcare also.

So instead of showing a frequency crosstable, you can show a type of map between 2 variables.

In this example I show which attribute people thinks about for different car brands.

In this diagram we can see that “Volvo” is close to Safe and Comfortable, but also “Ford” is close to these 2 attributes. Boring is in the middle of the map so no brand was close to that. “ABC car” is connected to untrustable and cheap. “Ferrari” is far from the other car brands because many people has just answered “Fast” as an attribute.

This map or diagram is based of a crosstable, that you also get when you do the correspondence analysis, see below:

I have made red circles around the frequencies that is highest, and if you look at this table you can understand the correspondence analysis better.

This analysis is in the module “Categories” and the command is : Dimension reduction – Correspondence Analysis

Greetings from Gunilla

error: Content is protected !!