# Definition variables for means not working -- minimal working example library( OpenMx ) ### Continuous outcome, continuous covariate # Generate data N <- 1000 x1 <- rnorm(N,0,1) y1 <- -.2*x1 + rnorm(N,0,1) dat <- data.frame( y1,x1 ) summary(lm(y1~x1,data=dat)) # Model linmod <- mxModel( name='LinearReg' , mxMatrix('Full',1,1,free=T,values=1,name='M'), mxMatrix('Full',1,1,free=T,values=2,name='eV'), mxMatrix('Full',1,1,free=T,values=.2,name='BX'), mxMatrix('Full',1,1,labels='data.x1',name='X1'), mxAlgebra( M+BX%x%X1 , name='eM'), mxData( dat , type='raw' ) , mxExpectationNormal( means='eM' , covariance='eV' , dimnames='y1' ), mxFitFunctionWLS( type='WLS' , allContinuousMethod='marginals' ) ) # Is not working linmodFit <- mxRun( linmod ) linmodFit <- mxTryHard( linmod ) summary(linmodFit) # This works fine linmod$fitfunction <- mxFitFunctionML() linmodFit <- mxRun( linmod ) summary(linmodFit)