]]>

]]>

]]>

]]>

]]>

Dear Group Members,Thanks in advance for any help or advice that you can give!I am currently working with 10 cross-sectional datasets and would like to use this data to explore the relationships between different latent factors using SEM. However, because the datasets were collected annually over an 11 year period and the surveys were not consistent across years (some items were missing certain years or measured using different scales), I can't just directly aggregate the datasets. As such, I was thinking of using MASEM as a way to integrate the different datasets, so that I can both account for the effects of time as well as account for the years in which some data was missing. Would it be appropriate to use MASEM this way? If so, in order to do a two-step MASEM, would I would just need to calculate the correlation/covariance matrix for each dataset, then calculate the pooled matrix and fit the SEM model? Does it matter which statistical program I use to calculate the matrices? Also, it seems like either covariance or correlation matrices can be used for MASEM. However, is there a preferred matrix to use when the observed items are not measured on different scales across different datasets, as is my case?Again, thank you for your time and help!See More

A technique to synthesizing correlation/covariance matrices for the purpose of fitting structural equation models.

]]>

]]>

A technique to synthesizing correlation/covariance matrices for the purpose of fitting structural equation models.