Pearson’s Chi-square for Canonical Correlation Analysis

From Q
Jump to navigation Jump to search

Tests for independence between two sets of Categorical questions, containing [math]\displaystyle{ p }[/math] and [math]\displaystyle{ q }[/math] variables or categories respectively. The tests statistic is:

[math]\displaystyle{ V = \sum^m_{i=1} nr^2_i }[/math]

where

[math]\displaystyle{ n }[/math] is the Effective Sample Size,
[math]\displaystyle{ p \approx \Pr(\chi^2_{pq} \ge \Lambda) }[/math].

Note that when preparing the data Q will modify [math]\displaystyle{ p }[/math] and [math]\displaystyle{ q }[/math] as required (explanations are provided in the test outputs).

See also

Multivariate Tests