Eyewitness testimony is shockingly unreliable.  How unreliable?  Ask Elizabeth Loftus.

Boot et al show that it is important for psychology studies to have active controls.  To exclude placebo effects, the control should include the expectation of change without the actual manipulation.