The point of research is to try and increase knowledge. It would be wrong to only do studies where people are pretty certain of what the results would be.
There is a big problem in science that journals tend to want to publish only positive results which means that if something is found to have no effect, we don't tend to hear about it. So, if 9 out of 10 studies show no effect and are not published but the 1 out of 10 studies that does show an effect IS published, it gives us the impression that there is a positive effect when actually it's much more likely that there isn't. I think it's really important that journals publish well-designed studies even if the results are not very exciting.