Method Selection GuideΒΆ
The following table shows which correlation method to use based on your variable types:
Variable X |
Variable Y |
Method |
|---|---|---|
dichotomous (discretized) |
dichotomous (discretized) |
|
polytomous (discretized) |
polytomous (discretized) |
|
continuous |
polytomous (discretized) |
|
continuous |
dichotomous (discretized) |
|
continuous |
dichotomous |
Where:
dichotomous variable: An ordinal variable with exactly two categories (e.g., Yes/No, 0/1).
polytomous variable: An ordinal variable with more than two categories (e.g., Likert scale with 5 options).
discretized: Indicates that the variable is assumed to originate from an underlying continuous latent distribution, and that observed categories result from applying thresholds to this latent variable.