Sorry for my long question but I'm sure it will be interesting for some of you:

My research is in the construction area. The dependent variable is time required to finish a construction task and several independent variables has been collected (crew size, job size, etc.). However the main dependent variable is consisted of several other activities; I mean:

Time to finish a construction task (dependent variable) = time for activity 1 + time for activity 2 + ....

And for each of activity 1, activity 2, etc, I have recorded the time as well with their own independent variables (affecting factors). So, some of the variables affecting activity 1, are not relevant to activity 2. etc. But all of the independent variables should be considered if the main task is considered as the dependent variable.

Therefore, I'm wondering how to model: one way is to enter all of the independent variables and put the total time as the main dependent variable. Another way is to split the model into several smaller models so for each of the activities do a regression model, but in this way can I add the outcomes of the models?

Another interesting issue is that some of the factors that have positive relationship with activity 1 as an example, become negative when I do a complete regression analysis and makes it difficult to interpret the results.

What are your ideas on this?

  • $\begingroup$ PLS regression can handle multiple response $\endgroup$
    – O_Devinyak
    Sep 26 '12 at 10:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.