Mean replacement and weighting vector

[JannikOr]

Dear all,

I am currently working with a data set, where we have sampling weights and missing data. To address this issue, I used mean replacement and selected the sampling weights in the weighting vector. I was expecting the mean replacement to also use the sampling weights. However, if I am not mistaken, smartPLS does use unweighted means for mean replacement (i.e., the mean replacement does not take the sampling weights into account). I was wondering if there is a reason to favor unweighted or weighted means in the mean replacement?

Best,
Jannik

[jmbecker]

That is a good question. Thank you for raising this issue. I am not sure anymore whether it had a substantive reason or whether it is just a design issue. We will check and also discuss this in the next meeting with the software designer.