Experimental blinding is a very important part of a good robust study. Ideally both the researchers and the participants would be "blind" - i.e. not aware whether the medication in a trial is the genuine drug or the placebo. This is not always possible due to the methodology required for a certain study, but is the ideal that should be aimed for.
If there is no blinding then the study may allow the bias of the researchers or the participants to impact on the results. A double-blind study is usually a good indicator of robust evidence, but must be taken into consideration with other parts of the study such as sample size and selection.