Double-blind experiments are a tiny subset of the types of studies which happen in engineering and science. Yes, they do serve the useful purpose of mitigating 'fudging' of the results by the subjects and the researchers themselves, which is important when the chances of conflict of interest or bias are high (such as with drug trials).
As has been pointed out though, there are significant benefits to having researchers be more hands-on with their experiments. These range from making sure the experimental plan is being properly followed, sanity checks on the measured data (before weeks are wasted gathering bad data), revisions to the plans or equipment if required, coming up with additional tests in a slightly modified setup to verify any 'surprises' in the data, etc. You might be surprised at how often overlooked issues are discovered just by the person being there.
Of course, if you're one of those people who thinks every researcher is going to fudge data, compute a thousand different test statistics to fit their preconceived outcomes, modify the experiment in senseless ways until it gives them the outcomes they want, hide data that doesn't fit a certain model, etc., then I doubt any system of experimentation is going to be good enough for you.