No consensus on framing? Towards an integrative approach to define frames both as text and visuals