Welcome to Pollapalooza, our weekly polling roundup.

The oldest poll I could find on Taylor Swift was from 2010. Even then, it was clear she was on her way to being a pop-culture phenomenon. When CBS News/60 Minutes/Vanity Fair asked American adults which of five musicians they would most like to have dinner with, 22 percent said Swift — more than Jay-Z, Susan Boyle and Lady Gaga, and second only to Paul McCartney. And among respondents aged 18-29, she was No. 1.

Thirteen years and several hit albums later, Swift has cemented her place as one of music’s biggest superstars. According to a recent Morning Consult poll, 53 percent of American adults identified as fans of Swift. And last weekend, she kicked off her widely anticipated — and logistically messy — Eras Tour, a retrospective revue of her musical evolution. (If anyone has a spare ticket to one of the shows this weekend in Las Vegas, DM me.)

But when you ask Swift fans which era they liked the most, you get 10 different answers. To celebrate the start of the Eras Tour, Morning Consult and another pollster, YouGov, asked people to identify their favorite Swift album, and they found plenty of disagreement. But those disagreements can teach us something valuable about polling.

In the YouGov poll, 15 percent of people who like or love at least one of Swift’s albums identified “Lover” as their favorite, followed by 13 percent of people who identified “1989” as their favorite. But in Morning Consult’s version, a plurality of self-identified “avid” Swift fans preferred “1989,” while “Lover” had only 5 percent support. Swift’s debut album, “Taylor Swift,” was second in the Morning Consult poll with 14 percent, and “Fearless” was third with 12 percent. And the results look weirder still when you look at Morning Consult’s results among all adults, not just avid fans. There, “Fearless” is tied with “1989” at No. 1 with 7 percent, followed by “Taylor Swift” at 6 percent. 

These polls may seem all over the place, but there’s good reason for them to disagree. First, most of these differences are within the polls’ margins of error. Basically, whenever you poll only a small sample of a larger population, some polling error is inevitable — usually enough to explain minor differences between various polls. Since the margin of error in Morning Consult’s poll of avid fans was ±5 percentage points, the actual number of fans whose favorite album is “1989” could be anywhere from 10 percent to 20 percent. That range includes the 13 percent who picked “1989” in the YouGov poll, so the two aren’t necessarily contradictory. 

Second, the population being polled matters. It’s not surprising that the entire population of American adults has different tastes from people who like or love one of Swift’s albums and avid Swift fans. The latter groups may have additional insight into some of these albums that help them appreciate them more. Political polls can disagree (without disagreeing) in the same way: A poll of adults may have different results from a poll of likely voters. So when conducting political analysis, we look at polls that are right for the context (e.g., when forecasting elections, we look at surveys of likely voters).

Third, how the pollster asks its questions is important. Look closely at these two polls, and you’ll notice that Morning Consult asked about the “Taylor’s Versions” of “Fearless” and “Red” separately from the originals, which could be affecting the toplines. If you add the two together, 19 percent of avid Swift fans prefer one of the two versions of “Fearless,” and 17 percent prefer one of the versions of “Red.” That’s more than the 15 percent who preferred “1989” (though again, still within the margin of error)! 

Similarly, the headlines of political polls can hinge — sometimes unfairly — on a pollster’s choices. For example, in 2018, Quinnipiac University found that 32 percent of voters found Democrats in Congress responsible for the recent government shutdown, 31 percent found then-President Donald Trump responsible, and 18 percent found Republicans in Congress responsible. Some of the news coverage of the poll focused on the fact that a plurality of voters blamed Democrats. But as this column pointed out, when you totaled the results by party, voters blamed Republican politicians over Democratic ones, 49 percent to 32 percent.

When poring over polls, it’s important to bear these guidelines in mind and not jump to conclusions. Of course, that goes equally for something serious like predicting the next presidential election and something fun like determining a music legend’s most popular album. (And by the way, the answer should obviously be “1989.”)

Other polling bites

• Morning Consult’s demographic breakdown of Swift fans is worth a deep dive too. Millennials and women are her “base,” so to speak, with 58 percent of the former and 56 percent of the latter identifying as Swift fans. Interestingly, 62 percent of Democrats are Swift fans, but only 48 percent of Republicans are. That’s a shift from that 2010 CBS News survey when Republicans were much more likely than Democrats to say they wanted to have dinner with Swift. That could reflect Swift’s evolution from a country artist to an outspokenly liberal pop star.
• Back to politics: According to Ipsos/Reuters, 54 percent of Americans think Trump’s potential indictment in connection with an alleged hush money payment to porn actor Stormy Daniels is politically motivated. In comparison, 38 percent think it isn’t. However, Americans weren’t necessarily prepared to jump into action. Seventy-seven percent of adults said they would do nothing if he was arrested, while 6 percent said they would protest, 6 percent said they would donate to his legal defense fund, and 4 percent even said they would take up arms. But this is probably an example of expressive responding — people responding emotionally to a poll question without literally meaning it. Some studies have found that polls can overestimate the number of people willing to engage in political violence.
• Florida Gov. Ron DeSantis, a likely Republican presidential candidate, recently stirred up some intraparty dissent when he expressed skepticism about aiding Ukraine in its war against Russia. Turns out, the disagreement among Republican elites on foreign policy also extends to voters — but most of them are on DeSantis’s side. According to a Morning Consult poll conducted last week, 46 percent of potential GOP primary voters thought supporting Ukraine is not a vital U.S. interest, while 37 percent thought it is. That’s in keeping with a long-term shift among Republicans toward isolationism.
• President Biden recently upset progressives by approving a new oil-drilling project in Alaska, but a Morning Consult poll shows that more American adults approve of it than disapprove, 48 percent to 27 percent. Interestingly, support for the project is roughly equivalent among Republicans (54 percent) and Democrats (48 percent). However, news of the project doesn’t seem to have reached everyone; 25 percent of adults had no opinion to share.
• Eight states have passed legislation banning gender-affirming care for children under 18, and several more are considering doing so. However, American adults oppose such bills, 53 percent to 41 percent, according to a new Selzer and Co. poll for Grinnell College.

Biden approval

According to FiveThirtyEight’s presidential approval tracker, 42.7 percent of Americans approve of the job Biden is doing as president, while 52.8 percent disapprove (a net approval rating of -10.1 points). At this time last week, 43.7 percent approved and 51.5 percent disapproved (a net approval rating of -7.8 points). One month ago, Biden had an approval rating of 43.2 percent and a disapproval rating of 51.7 percent, for a net approval rating of -8.5 points.