If Naïve Researchers and Participants Alike Are Prone to Try

If naïve researchers and participants alike are prone to try to figure out what a study is trying to investigate, how do we know the extent to which participants' responses are due solely to our experimental treatments rather than to some sort of informal cues by researchers and good guesses by participants? Even if participants are unable to articulate their suspicions regarding what they thought the study was about, does that mean they were not actively trying to figure it out while participating? Does the process of random assignment solve, or at least help to address, this problem, and if so, why and how? Do biased responses by "good participants" essentially get balanced out by biased responses by "negative participants?"