This paper reports on preliminary thoughts which have been conducted in designing an empirical experiment to assess the comprehensibility of a visual notation in comparison to a textual notation. The paper sketches shortly how a corresponding hypothesis could be developed. Furthermore, it presents several recommendations that aim at the reduction of confounding effects. It is believed that these recommendations are applicable to other experiments in the domain of MDE, too. Finally, the paper reports on initial experiences that have been made while formulating test questions.