Before examining the effects of political commercials, it is useful to review their limits. First, ads cannot transform candidates from ordinary politicians into attractive, charismatic potential presidents. Political advertisements cannot make competent but dull campaigners into enthralling, soulful candidates. 1 No political commercial can turn capable but wooden candidates like Walter Mondale, Michael Dukakis, and Bob Dole into charismatic charmers. Smart consultants realize this, and like sculptors who recognize the strengths and limits of their clay, they take candidates as they are and try to polish and accentuate their strong points, while attempting to cover up and reframe their weaknesses.

Second, political advertisements cannot change strongly held partisan preferences. As I have noted throughout the book, people bring well-developed political attitudes to the electoral setting, and these attitudes insulate them from media effects. As Stephen Ansolabehere and Shanto Iyengar have concluded, "Advertisements induce few Republicans to vote Democratic and few Democrats to vote Republican. Exposure to an opposition candidate's advertisements, in fact, can sometimes strengthen a voter's loyalty to his or her party." 2 In addition, political spots have relatively few effects on Independent voters, many of whom are cynical about politics to begin with. 3

The good news about political ads is that people learn from them. There is substantial evidence that voters become better informed about candidates' stands on the issues from watching political ads. 4 This runs counter to the conventional wisdom. But it makes sense when you remember that political ads contain considerable issue information, that people can learn from the media when they pay attention to the message, and that voters are more motivated to pay attention to political communications during the presidential election than during most other periods.

Of course, there is debate about how much people learn from political ads. Critics argue that people learn only the sketchiest of information from ads since political spots tend to be vague and diffuse and contain only information that will benefit the candidate. Defenders note that 30-second spots can contain a good deal of issue information. In their book on political advertising, Edwin Diamond and Stephen Bates offer up an example of a 24-second ad that contains a considerable amount of policy issue information:

I believe that the question of abortion is one that ought to be reserved exclusively to a woman and her doctor. I favor giving women the unfettered right to abortion. I also favor the federal funding of abortions through Medicaid for poor women as an extension of that right to an abortion, and I oppose any statutory or constitutional limitations on that right. 5

Naturally, few candidates would make such a clear and unambiguous statement on abortion. But Diamond and Bates's point is that there is nothing in the technology of advertising that would inhibit voter learning if a candidate did produce such a spot.

Ads are designed not to teach, but to persuade. How persuasive are presidential campaign ads? I have already noted that they rarely convert strong partisans to the other party's candidate, and that party identification helps to mediate voters' response to ads. Nonetheless, ads do solidify the support of partisans, strengthening their commitment to the party's nominees. Ads also influence weak partisans -- for example, Democrats with a weak attachment to their party.

In one study, so-called weak Democrats (those who had abandoned the party in the previous election) were powerfully influenced by a Democratic advertisement, much more, as it turned out, than weak Democrats who saw a Republican ad or those who did not view an experimental commercial. 6 In addition, ads seem to have a greater effect on voters with little interest in politics. The investigators found that low-interest voters who viewed an ad from their own party's candidates shifted significantly in their voting preferences toward that candidate. As the researchers note, "These voters moved from being largely undecided or not voting back into their party's camp." 7

Since these studies are experiments, we need to be careful about generalizing the results to the real world. In addition, they do not tell us why weak partisans and low-interest voters are more influenced by ads than strong partisans and high-interest voters. When combined with other research, these studies suggest that political advertising is likely to have its greatest effect on the subset of voters who have a weak attachment to a political party and little knowledge about politics. As Darrell West notes, "ads are developed to stir the hopes and fears of the 20% to 30% of the electorate that is undecided, not the 70% to 80% that is committed or hopeless." 8 . . .