Thanks for the followup. The problem really lied in the mediation. I hadn't included any mediation among the variables at all. All independent variable were causal for the dependent variable.
However, I contacted Dr Jamil Malik, who also happens to be on Researchgate through mail. He gave me quite good clues.