I want to perform path analyses on a dataset with just below 100 respondents. I found that there is some curvilinearity going on and potentially also some multicollinearity. I believe partial least square structural equation modeling is best for the type of analyses and with this (limited) dataset.
Previously, I used the package PLSpm to perform this type of analysis, but I read that it is deleted. Does anyone know why it was deleted? I did find it and was able to install it, but probably there is a valid reason why it was deleted? Is there a better option? I also found PLS (partial least squares and principal component regression) and PLSr (Pleasure - Partial least squares analysis with permutation testing), but the manuals seem to indicate that these packages have different aims to test other than what I want to test.
Also, previously, I compared the results of SMARTpls with the PLSpm package, and the result differed strongly, which is why I again wonder whether it is approved to use this package (PLSpm) and whether it does the job? Can I rely on this (old and deleted) package? It is very useful since it also visually presented the model visually, so I hope someone can confirm that it is still okay to use this package?
Thank you very much in advance. It is my first post here, so I hope that I follow all the rules.
Best regards, Ellen