SmartPLS and unexplicable path weighting results

[resatsch]

Dear Forum members and admins,

I have a problem with a measurement model and shown results in SmartPLS.

I have 179*3=537 cases in total (179 measured at three different points of time) and want to use SmartPLS to research a technology acceptance research question.

With SPSS I made a datafile that was imported and marked valid in SmartPLS. Then I built the measurement model and created LVs in the GUI.

Problem is, that I get a very strange result when conducting a PLS analysis (Path-weighting scheme, Original data metric, Max Iterations = 300, Abort criterion/initial weights are left as default). The path weights between the LV seem to be similar to a quick test I made with SPSS linear regresssion, but the paths from the measurement model to the LV have values such as 0,000 to -0,000, whereas others have a 0,001 (it is a formative measurement model). This is likely to be an error (considering the results of another data set from another measurement point shows "normal" values 0,320, 0,552, 0, something).

So what is the best guess? Is the error in the underyling data set (maybe due to a wrong convertion to csv), did I miss to tick the right box, or what other possibilities are there (scale-wise?)?

I would be very happy to receive an answer, I can also provide a screenshot of the model.

Best,
Florian

PS: Despite using the forum search engine quite a time, I haven't found a helpful post on my question - if i missed it pls point me to it

[resatsch]

Hello everyone,

after a while of trial and error, I found the answer myself.

I can highly recommend to anyone to follow exactly the procedures for data import in Christian's post here: viewtopic.php?t=104&highlight=por

In my case, the import worked well, the file was valid, etc. But I skipped the .por export and some little bug must have been in the file.

Now it looks fine :)

Best,
Florian