Teaching computing in primary school: create or fix?

We describe the setup of an experiment conducted in the spring of 2018. The experiment aims to identify the differences between teaching computing through creating programs as opposed to fixing them. Quantitative and qualitative data are collected in the forms of pre- and post- knowledge tests, pre- and post- attitude tests, programming environment logging data and student interaction video's. The logging data shows that starting from incorrect program leads to less interactions with the code. However it leads to more testing in simulation and on the physical system.