Statistical software testing (SST) is an important means for the quantification of software dependability. The concept of SST is that the software is executed on statistically generated test-cases that simulate the software's operating environment. As opposed to SST tests, coverage tests are engineered to execute each statement, decision, Boolean expression etc. at least once. Coverage testing is required by many standards, SST is not required by the standards. Coverage testing however cannot quantify reliability. In this paper, we want to investigate how to link these two important testing strategies. We want to maintain the features of SST, which allow us to quantify dependability, but combine them with a view towards code-coverage. The aim is to not only perform SST, but perform SST on the full code and thus achieve a dependability estimate that is attached to all code-parts. We demonstrate on a software example, taken from the protection system of a heavy water reactor, how to achieve this, by creating a link between the structure of the software input-space and the structure of the code.