Using a PLS model with hypothetic questions

[stefanperras]

Hi,

In my research I try to identify project management success factors that influence the behavior of the customer.
The problem in my current research is that I cannot measure any facts from past industry projects, as these projects went terrible wrong or are long time ago. Addtionally my total industry population is only around 40.

My idea is now to ask the industry population (project managers) about their opinion on potential critical success factors (which I collected from customer interviews and literature) like
"What of the following potential success factors do you regard as relevant and want to employ in your future projects?
- Success factor 1 (Not relevant at all ... Absolut essential)
- Success factor 2 (Not relevant at all ... Absolut essential)
...

These success factors load formative on different exogen latent variables. These latent variables then all finally load on the endogen latent variable "customer behavior"

To measure the potential customer behavior, I plan to ask the industry population (I cannot ask the customers as there is no concrete project to link customers and industry companies) how they estimate the customer behavior in future projects like
"How do you think will the customer behavior look like in your future projects?"
- Customer behavior 1 (Not true at all ... Very true)
- Customer behavior 2 (Not true at all ... Very true)
- Customer behavior 3 (Not true at all ... Very true)
...


Is such a double hypothetical approach acceptable? I am fully aware that causal relations are as robust as they would be as one measure objective facts, but considering the unavailabilty of concrete projects, it is the only way I currently see as possible solution. Any other ideas are highly appreciated.

Many thanks

Stefan