Likewise , as we argued in describing the resemblance fallacy , making assumptions that the internal representation is a mental model or image-like may simply give the illusion of solving the processing-internal representation-external representation riddle. (p. 209)
Do we understand how visualizations work, if at all? Work on visualization seems to be premised on the intuition that “a picture is worth a thousand words”. External Cognition: How do Graphical Representations Work? (PDF) by Scaife and Rogers (Int . J . Human – Computer Studies (1996) 45 , 185 – 213) is a metastudy that questions what we really know.
In their conclusion they point to interactivity as what needs to be studies. How dow people interact with graphics – annotate them, draw them, and manipulate them.
In sum , we propose a new agenda for research into graphical representations that is based on an analysis of interactivity and , thus , considers the relationship between different external and internal representations . Such an approach should help us to better understand , design and select graphical representations – be they “old fashioned” or technologically advanced – which are appropriate for the learning environment , problem-solving task or entertainment activity in question. (p. 210)