## You are here

3 posts / 0 new
Offline
Joined: 06/10/2019 - 15:24

Hi Dr. Cheung and colleague,

I am trying to mimic your codes in your metaSEM article (cheung, 2018). I construct a fake dataset and try the tssem1 as following

a<-matrix(c(1,0.2,0.4,1),ncol=2)
b<-matrix(c(1,0.6,0.4,1),ncol=2)
c<-matrix(c(1,0.3,0.7,1),ncol=2)
d<-matrix(c(1,0.9,0.8,1),ncol=2)
e<-matrix(c(1,0.2,0.5,1),ncol=2)
f<-matrix(c(1,0.36,0.34,1),ncol=2)
g<-matrix(c(1,0.3,0.8,1),ncol=2)
h<-matrix(c(1,0.9,0.2,1),ncol=2)
i<-matrix(c(1,0.2,0.4,1),ncol=2)
mats<-list(a,b,c,d,e,f,g,h,i)
n<-c(5,7,20,39,6,29,30,78,29)

fixed1 <- tssem1(mats,n, method = "FEM")

That is a list of 2*2 var-cov matrices with some arbitrary sample sizes. However, I got the error message “Error in if (all.equal(covMatrix, t(covMatrix))) { : argument is not interpretable as logical’’ when running the tssem1.
I am wondering what is the problem. Is it because of the data format of my fake data and your example data (Becker, 2009)?

Thanks for your time and patience!

Offline
Joined: 10/08/2009 - 22:37

Dear Xin Lin,

The error message means that your covariance/correlation matrices are not symmetric. For example,

a<-matrix(c(1,0.2,0.4,1),ncol=2)
a
[,1] [,2]
[1,]  1.0  0.4
[2,]  0.2  1.0

Best,
Mike

Offline
Joined: 06/10/2019 - 15:24
Dear Dr. Cheung,

Dear Dr. Cheung,