Epistemic beliefs and prior knowledge as predictors of the construction of different types of arguments on socioscientific issues