I’m trying to figure out how to get individual scores for the shared variance of 3 variables in my dataset. I thought about using CFA with equal loadings and then somehow getting the measurement error as scores for each item and every datapoint. It’s kinda like how you can get factor scores, but for the error part instead.
I’ve been searching online but can’t find any clear instructions on how to do this in Lavaan. It doesn’t seem to be a common thing people do.
Has anyone done something like this before? Any tips on how to approach it or what functions to use in Lavaan? I’m kinda stuck and would really appreciate some guidance!
Your question got me really curious about this CFA approach. I’ve never tried extracting shared variance scores like that before, but it sounds super interesting!
Have you considered looking into residual variances instead of measurement error? I wonder if that might be a way to get at what you’re after.
I’m really intrigued by your idea though - what made you want to look at the shared variance specifically? Are you hoping to see how much overlap there is between your variables?
It’d be awesome to hear more about your project and what you’re hoping to discover with this analysis. Maybe explaining a bit more about your research goals could help us brainstorm some other approaches too?
Keep us posted on what you figure out! I’d love to learn more about this technique if you end up cracking the code.
yo CreativeChef15, sounds like a tricky problem! have u tried the lavPredict() function? it might give u what ur looking for. another idea - maybe check out the residuals() function in lavaan. it could help u get at those error scores. good luck with ur project, let us know how it goes!
I’ve encountered a similar challenge in my research. While Lavaan doesn’t have a built-in function for extracting shared variance scores, you might be able to adapt the residual.cov() function to your needs.
This approach involves fitting your CFA model, then using residual.cov() to obtain the residual covariance matrix. From there, you could potentially calculate the shared variance component.
Another option is to explore the lavPredict() function. It allows you to extract various types of predicted values from your fitted model, which might include the information you’re seeking.
Keep in mind that interpreting these scores can be complex. Make sure you have a solid theoretical basis for using shared variance in your analysis. If you’re unsure, consulting with a statistical expert might be worthwhile to ensure your approach aligns with your research goals.
