Gordon, Zettelmeyer, Bhargava, and Chapsky "found some evidence that PSA’s as controls would lead an advertiser to draw the incorrect conclusion about the true lift of an ad campaign."[1]

[1] You can't tell me not to cite! ;)