WHERE ARE THE VALUES OF BETA WHEN FOLLOWING SITUATION OCCOUR

[Tabinda]

I am hoping someone can give me some advice and references about testing a multiple mediation model using SMART PLS with the following model design.
In SMART PLS I HAVE 3 IVs, an intervening variable and 3 DVs. I have not drawn/linked the direct relationship from IV to DV as i assume that SMART PLS has the capability of calculating Beta and t-vlaue for the direct relationship between IV and DV, as the total effect significance value is calculated by smart pls. Now my question is where is Beta value for the direct path. Is i calculated automatically if so then where it is given in SMART PLS 2 or SMART PLS 3.
THE MODEL LOOKS AS UNDER
Please also enlarge the diagram and read the attached not on it.
I will be very thankful to you all.

[jmbecker]

You also need to draw the direct relations if you want to calculate the direct effect. Otherwise the total effect is just the indirect effect.

[Tabinda]

Thank you very much for replying back. I was also expecting the same answer. But can you please further clarify , In the same model should I draw all the direct relationship once between IVs and DVs and run the model or I must do it one by one and run the model every time i add a new path and exclude the last one. Which is the right way in PLS.

In other words should i first draw the direct path from IV1 to DV1, DV2, Dv3 and check for mediation then remove these paths and re run the model by drawing path from IV2 to DV1, DV2, DV3 and check for mediation and so on or draw all the paths once and run the model in one go...Which one is more preferable and does he results differ as well...

Thanks and waiting for your reply

[jmbecker]

Yes, you need to connect all IVs with all DVs, if your hypothesis is that M is mediating all relations. You should not test them all separatly, because that will lead to omitted variable bias.