Abstract
Non-symmetrical correspondence analysis (NSCA) is a very practical statistical technique for the identification of the structure of association between asymmetrically related categorical variables forming a contingency table. This paper considers some tools that can be used to numerically and graphically explore in detail the association between these variables and include the use of confidence regions, the establishment of the link between NSCA and the analysis of variance of categorical variables, and the effect of imposing linear constraints on a variable.
| Original language | English |
|---|---|
| Pages (from-to) | 55-76 |
| Number of pages | 22 |
| Journal | Journal of Classification |
| Volume | 26 |
| Issue number | 1 |
| Publication status | Published - 2009 |
Keywords
- contingency tables
- correspondence analysis (statistics)
- statistical analysis