How to do a pairwise deletion for TSSEM1 and 2

Posted on
No user picture. verdendi Joined: 05/05/2024
Forums
Hello everyone,

I hope you are doing well.
I am trying to do meta analysis for SEM papers, using TSSEM1 and 2 functions in MetaSEM packange.
It would be highly appreciated if you answer how we can use a pairwise deletion approach with TSSEM 1 and 2 functions.

The scope of literature for our meta analysis is quite broad and there are many missing values for pairs of the variables. I cannot use a normal way to pool the correlation matrix by using TSSEM 1. So I decided to do a pairwise deletion approach.

However, I am now sure how to use TSSEM1 or 2 functions with the pair-wise deletion approach. As a try, I used TSSEM1 for each pair of the variables and calculated the pooled correlation value for each pair using TSSEM1. Then, I combined those values in a matrix and tried to use it in TSSEM2 function. As TSSEM2 requires a TSSEM1 object, I had an error message. So I tried to make my pooled matrix as TSSEM1 object by using it in TSSEM1 function as a only input matrix. This does not work as TSSEM1 requires more than 2 matrices as input. It would be really appreciated if you can tell me how to do pairwise deletion using TSSEM1 and 2 functions.

Best,
Insung Hwang

Replied on Mon, 05/13/2024 - 10:15
Picture of user. Mike Cheung Joined: 10/08/2009

Simulation studies have demonstrated that the TSSEM approach is effective even when studies only provide one correlation coefficient. This means that there is no need to perform pairwise analysis manually.

However, if you prefer to do it yourself, you can use the wls() function. Just keep in mind that you will need to handle the details yourself.