This article illustrates that the task design and problem selection are of characteristic influence to evoke sense-reason-act programming (SRA) among primary school pupils when programming robots. Research shows that the task design influences the development of computational thinking (CT). The literature provides evidence that the context, the problem space, and the representation of the problem to apply SRA-programming require the programming task to be embedded in a dynamic context in which a programmable robot must use sensory information to anticipate changes in the environment. In order to ascertain whether the problem space and the task design influence the evocation of SRA-thinking, this article uses a research design comparing the differences between two programming conditions (static/dynamic). In these conditions, pupils use Lego EV-3 robots and Mindstorms software to solve programming problems. As a post-measurement, a Lego challenge is applied. In this article, it is shown that the integration of a dynamic task design to solve a programming problem is essential for a deeper understanding of CT skills. Furthermore, when pupils can immediately test the consequences of their program in a dynamic environment and, thus, the learning environment provides an appropriate problem, they gain a deeper understanding of the added value of sensors and will be better able to reason about complex problems. It is found that programming in a dynamic problem environment almost naturally evokes SRA-thinking, as opposed to programming in a static environment. The influence of SRA-programming as demonstrated identifies characteristics of CT.