Does anyone know how Stata calculates the Standard Error in logistic
regression (command logit) from the "coefficient" into "OR" form?
The estimated coefficient are trasformed to OR with the formula (e^b
rather than b).
Empirically this is not true for SE.
Which is the trasformation for Standard Errors?
Thank You in Advance
Paolo Grillo
--
* * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * *
Dott. Paolo Grillo
Specialista in NEFROLOGIA
Specialista in STATISTICA SANITARIA
Dipartimento MEDICINA del LAVORO
Centro Ricerca Epidemiologia Occupazionale, Clinica e Ambientale (EPOCA)
Tel 02/50320148
Tel 02/57992637
Fax 02/50320126
E-Mail [email protected]
