Many scientists I know are probably better coders than you are. Again... To them, it's a tool. To do their jobs they have to know their tools well.
I've never said that computer science cannot be applied as a tool. As a computer scientist I regularly apply math, physics, psychology, bio-mechanics and linguistics as a tool. That does not mean they are not valuable research fields on their own or that I would not value the opinion of an expert in e.g. psychology when setting up a user test.
Computer science does not equal coding; it is NOT a programming course. In fact, the computer science bachelor+master program at our university offers only 3 programming courses, each covering a different programming paradigm. Most subfields of computer science require very little or no coding (AI, theoretical computer science, information retrieval, computer vision, ...). Even in software engineering, coding is just only one step in the software development process. Other steps include requirement specification (and figuring out what your clients really want, which is often surprisingly hard), planning, architecture design, documentation, testing, deployment. If you build small programs and you are the only user of such programs, as your scientist probably do, then you can indeed often get away with collapsing all of these processes into a single coding process. For larger projects (say the linux kernel or the software used by your bank) that's probably not a good idea.
Btw, interesting read on the role of computer science:
http://www.nature.com/nature/journal/v440/n7083/pdf/440419a.pdf