There is nothing about having dependent variables that are CORRELATED that calls for a distinct type of analysis. This will commonly occur when the same independent variables predict the different dependent variables. However, if the prediction errors are correlated, then structural equation modeling can take account of those correlations. - David Greenberg, SOciology Department, New York University, New York, NY On Mon, 2 Nov 1998, Cheoleon Lee wrote:

> The term 'simultaneous regression analysis' is used to distinguish them from > hierarchical regression analysis, in which there exist causal hierarchy > among independent variables. So, the term simply means regular OLS > regression. > If you DVs are correlated perhaps structural equation modeling (LISREL) is > the way to go. If you can establish some type of causal relationship(s) > between the DVs, you can do a path analysis, recursive or non-recursive. > Good luck. > > CLee > > Cheoleon Lee, Ph.D. > Statistical Analyst > The Gallup Organization > One Church Street, Suite 900 > Rockville, MD 20850 > > > -----Original Message----- > > From: Nancy DaSilva [SMTP:ndasilva@BAYOU.UH.EDU] > > Sent: Sunday, November 01, 1998 11:18 PM > > To: SPSSX-L@UGA.CC.UGA.EDU > > Subject: Simultaneous Multiple Regression Analysis > > > > Hi. I am interested in running several regression analyses and since my > > DVs are correlated I was wondering if SPSS or any other stats package > > (aside from structural equation modeling) had a command which would run > > multiple regressions taking into account the correlation among the DVs but > > without creating some type of composite of the DVs. An article I was > > reading mentioned the term "simultaneous multiple regression analysis" but > > no reference was included. Any help/advice would be appreciated. > > > > nancy >