From “Do Political Ads Even Work? 2022 is set to be the most expensive midterm election in history. But the political science research is murky on how much that matters” by Walter Shapiro at The New Republic:

For anyone living in a media market featuring contested political races—especially places such as Phoenix, Las Vegas, and Atlanta—these are the jackhammer months. Never in an off-year election have so many campaign spots been aired. TV spending will more than double from 2018 levels, according to estimates by the analytical firm AdImpact. The Wesleyan Media Project, which studies campaign commercials, calculated that more than two million ads had aired on broadcast television by early August—long before campaigns began their fall offensive….No candidate will willingly stop spending on TV ads to test whether they are really effective.

Is all this spending justified? Convincing data on the cost/benefit ratio is scarce, according to political scientists.

….But as the political science professors John Sides, Lynn Vavreck, and Christopher Warshaw concede in an article in the May issue of the American Political Science Review, “There are significant limitations to what we know about the effects of televised campaign advertising on election outcomes.”….There are so many ads that it is impossible to find evidence that any single spot—no matter how emotionally powerful—made a measurable difference in voter sentiment. When it comes to campaign spots, volume, not artistry, is what makes a difference.

And the effectiveness of political ads varies according to the offices being sought:

Many academic studies of TV ads have focused on presidential races. But the effects of ads, while still small, are significantly larger in races that are not at the top of the ticket, such as a House election. Vavreck and her two co-authors used polling and other data to assess the potency of TV ads from 2000 to 2018. Their conclusion: “Despite increasing partisanship in the electorate, there are still persuadable voters that respond to television advertising—especially in down-ballot elections, where voters have less information about candidates.” As Vavreck summarized when we spoke, “The effects of advertising are small and go away quickly. But small does not mean inconsequential, especially in a close race.”

Regarding the ways political attitudes and behavior are shaped by ads vs. other campaign investments, Shapiro writes:

The study also came to the surprising conclusion, partly based on an analysis of voter files, that the limited power of TV ads lies mostly in the realm of changing perceptions of candidates rather than in motivating people to turn out. John Sides, who teaches at Vanderbilt and is a co-author of the study, told me, “If you want to mobilize voters, it’s much better to do it with personal contact than TV ads.”

On the impact of positive vs. negative political ads:

….Travis Ridout, a professor of political science at Washington State University and a co-director of the Wesleyan Advertising Project, said, “According to the best studies, negative ads are not more effective than positive ads.” Yale political scientist Alexander Coppock, who studies persuasion, is also dubious. “I have a low opinion of the focus-group approach,” he told me. “You don’t have to remember or like an ad for it to be effective. It can still work if you hate it.”

How much is enough?

….A prime example of uncertainty is whether in free-spending races there is a saturation point when an additional commercial fails to have any effect. It seems logical, but proving it is akin to finding the great white whale. “We didn’t find a point of diminishing returns,” Sides said, but he theorized that it probably exists. In the context of a campaign, there is always pressure to put more money on TV. Part of it is a competitive instinct and part of it, frankly, is that many consultants are paid on the basis of the size of the media buy. “Clearly, this is an industry where nobody knows what’s effective,” Coppock said. “And everyone involved says, ‘You have to do more.’”

Looking at one of the marquee U.S. Senate races, Shapiro notes,

….But what is intriguing is the situation in Ohio where Democratic Senate candidate Tim Ryan dominated the airwaves over the summer while his GOP rival, J.D. Vance, did not air a single spot until August after winning a bruising May primary. Over a four-week period before Labor Day, Ryan aired 5,503 TV ads, according to the Wesleyan Media Project, dwarfing Vance’s modest buy. But will Ryan’s early advertising advantage matter in November? “I’m inclined to think that Ryan’s early advantage isn’t worth nothing,” Sides said, “but it’s not worth very much.” Ridout, however, makes the point, “We know from psychology that early impressions stick. And if you get the impression that Tim Ryan is a good guy fighting for the working class, it will take more negative ads to dislodge it.”

Shapiro concludes, “And most of all, don’t assume that campaign ad decisions are based on impeccable research. In truth, campaigns are flying blind just like the rest of us.” Here’s another take on the effectiveness of political ads from one of our recent articles. And it looks like there is a lot of room for quality research into how different ads work with particular constituencies.