Honestly I feel like this is an even bigger problem in the natural sciences. In my experience journals want strong and interesting positive results. They rarely want negative results unless it is debunking a previous result that seemed interesting and positive. I have worked on projects before, that while interesting and on a solid theoretical base as far as approximations go, were ultimately incorrect for what ever reason. I should be able to publish and report on this so no one else wastes their time on it or can find a glaring flaw in it and build upon it. While the initial idea may be a good one the implementation of that idea is not always a good one. It can be useful to other scientists to know these things.