i am testing a model with one dependent variable (4 items), and 7 independent variables i have a persistent discriminant validity issue between the dependent and one of the seven independent variables. the independent variable in question measures the objective behind a given strategy choice ( the items being the different objectives which can lead to strategy choice: social, commercial, growth, double objective). the dependent variable stands on efficiency, effectiveness, long termism and adaptativeness of the strategy. There seems like there is a theoretical resemblance between the two constructs, even if they are logically distinct (the objective (independent variable) is a premise to strategy choice.
actually discriminant validity is reduced if i omit the majority of the indicators in the dependent variable. the variable then becomes a mono-item construct. is there another solution? or can we just not bother solving it since it concerns an dependent variable?