No statistical analysis at all applies if the relevant premises are false. People defend the statistical analysis using statistical metrics, rather than the premesis by which the stats are used (eg., the model type, the causal model, etc.).
Eg., if there is no causal connection between X, Y then any analysis of association between X, Y is not a causal one.
Entire fields of people have been taught that stats can justify science -- and hence the heaps of pseudoscience.
It is science which justifies the use of stats; not vice versa.
Eg., if there is no causal connection between X, Y then any analysis of association between X, Y is not a causal one.
Entire fields of people have been taught that stats can justify science -- and hence the heaps of pseudoscience.
It is science which justifies the use of stats; not vice versa.