Nate Silver is someone with whom I usually agree. His research is impeccable. He contends that negative campaign ads don't really work. And I'll probably look like an uninformed fool for saying this, but there have been a number of negative ads in my memory that did actual harm to the targeted candidate. First off, the infamous "Willie Horton" ads run by the George HW Bush campaign against Michael Dukakis in 1988. That series of ads showed how Dukakis let a dangerous criminal out of jail, only to have the guy commit more heinous crime. Second -- the Swift Boat campaign against John Kerry in 2004, calling Kerry's version of events into question, making him out to be less patriotic because he criticized US involvement in the Vietnam War after he was discharged from the Navy. Finally -- in 1990, Senator Jesse Helms ran for re-election against Harvey Gantt, an African American, beating him by only 5% (contrast that with his previous campaign in 1984, when he beat a white opponent by only 4% at the height of the Reagan revolution). Towards the end of the 1990 campaign, Helms ran an ad that depicted the hands of a white man ripping up an employment rejection notice from a company that gave the job to a "less qualified minority." Helms took a lot of heat for that ad, but I'm fairly sure that it generated more white votes for him. Perhaps this is where Silver would say, "Yes, but did that ad turn an election loss into an election victory?" I'm not sure; probably not. Hard to beat an incumbent like Helms. But I think where negative ads really work is in pushing the envelope of what's acceptable ground in an election campaign. I'm certain that the Atwater ad for Bush41 and the Helms ad made it more acceptable to attack a decorated war hero like Kerry, or for Bush43 to attack McCain in 2000 as someone who was gay, who had illegitimate children, and a drug-addicted wife.