Book Review/Heather Seggel

The Story in the Numbers

Newscasts are so densely freckled with polling data it can sometimes be hard to tease out the intentions behind its use. Anthony Salvanto has worked at asking the right question and interpreting the results for decades; as the Elections and Surveys Director at CBS News, he mans the network’s “Decision Desk” on election nights, and follows up with voters to see how they view the consequences of their actions afterwards. In his first book, Where Did You Get This Number? A Pollster’s Guide to Making Sense of the World (Simon & Schuster), he explains the polling process and how those seemingly dry numbers can tell a detailed story if you just know how to read them.

One of the questions Salvanto is frequently asked, or complained to about, is “Why didn’t you pick me?” (This very complaint about a poll turned up in my Facebook feed while I was reading.) He explains how a sample of less than one thousand respondents can unfairly favor one demographic “side” over another, but once you get that one thousand on the record, asking for more input offers very few benefits. Pro tip: They’re unlikely to pick you if you aggressively announce your availability, since that suggests you’re bringing an agenda with you. Wait for the phone call and your time will probably come. The evolution of polling in the age of cell phones has been bumpy, but now you’re as likely to be interviewed via cell or even online in a medium that was once almost entirely landline based.

Salvanto begins his first chapter recalling how he pulled up the secret box on his computer monitor and called the November 2016 election for Donald Trump; he then looks back at media outlets who failed to see Trump as a viable candidate and shows how his polling data made it clear that it could happen.

Running down the night hour by hour with quick cutaways to discuss how the data was gathered and interpreted was probably the best way to open the book, but for this reader it felt like being kicked in the teeth, then asked to read about how costly the dental repairs will be. Thankfully it gets better.

A chapter about guns and voters opens with a brief discussion of attributes versus benefits. During the 2016 campaign even those planning to vote for Trump generally agreed that Clinton was more qualified for the job. Salvanto asked which candidate made respondents feel safer and Trump consistently edged out Clinton; those same respondents also voted in record numbers, enabling Trump to win the electoral college. This swings into a discussion of “single issue” voters and the ways gun rights advocates think, and act. In a survey conducted in late 2017, two thirds of gun owners said they believe gun control groups were out to take away all guns, period (a patent falsehood). Three-fourths of those opposed to stricter gun laws believed that was the intent behind the laws.

Conversely, non-gun owners tend to think more broadly about things; for starters, they believe stricter laws are a public health issue. We’ve heard before that gun owners are who we should all model our activism after, and it’s borne out by the numbers: 53% will donate to a candidate or contact their Congressional representative about the issues that matter to them, as opposed to a rather sleepy 23% of non-owners. When it comes to literally voting around that one issue the gap is even wider: 74% of gun owners won’t support a candidate who disagrees with them on gun laws, versus a mere 17% of non-owners. This is surprising when you consider the things both sides actually agree on: There’s accord about some of the causes of gun violence, broad support for background checks, and agreement that mental health screening and access to treatment are important. The data show places where both sides can work together, but it doesn’t stop them from talking past one another, which keeps us stuck.

Whether you’re distrustful of polls or think they see all, “Where Did You Get This Number?” makes a good case for thinking critically about what they mean. Salvanto is open about his bias in favor of polling versus the aggregate prognostication that the internet has popularized, but he backs up his preference with—surprise!—data, and is cheerfully open about it being an imperfect science. If you’ve wondered about the “mix of gumption and calculation” that goes into extrapolating the national mood from an artfully chosen sample, start here.

Heather Seggel is a freelance writer based in Northern California, specifically the part that’s on fire. Oh wait, that’s most of them. Email heatherlseggel@gmail.com.

From The Progressive Populist, October 1, 2018

Populist.com