Dear professor Mike
I remember that you seem to mention from one of your article that a performance of MASEM model with and without artifact correction is worthy of scholarly attention.
Am I right? By the way, It has been the topic i am always wondering about since i read your article.
Corrected correlation might actually cause systematic bias, to some degree, compared to raw correlation.
However, this bias has nothing to do with performance maybe because it is more likely to do with estimation method and sample sizes.
Performance Comparison between two model is feasible? then, what can be absolute criteria for that?
I would like to ask your opinion on that issue.
Printer-friendly version
Dear James,
It depends on who you talk to. Many people may think the other way---the observed correlations under-estimate the true correlations among the constructs.
If you are going to compare these two approaches using a computer simulation, then you need to define your criteria. Do you consider the correlation with measurement error or without measurement error as the true parameter? You may come up with a totally reverse conclusion depending on which one you use.
By the way, there are already methods to correct for unreliability in a simple correlation. You may check the literature in meta-analysis. I did not include correlation for unreliability in the metaSEM package because I have not figured out how to do it correctly in my two-stage SEM approach. It is in my TODO list.
Regards,
Mike
always thanks for a kind comment. :)