Iowa has had two high-profile political tech failures in the last week, but a lower-profile political tech success shows a path forward.
Last Saturday, the highly regarded “Des Moines Register” poll was scrapped at the last minute because a call center employee failed to read Pete ButtigiegPeter (Pete) Paul ButtigiegSanders campaign to request partial Iowa recanvass Trump under pressure to renew last nuke treaty with Russia Buttigieg takes delegate lead in Iowa after updated results released MORE’s name while conducting the poll over the phone. Then on election night, Democratic caucus chairs were unable to submit their results through a newly built mobile app, delaying vote reporting for days.
These debacles illustrate two distinct approaches to political technology that are both dangerous for Democrats heading into the most important election of our lifetimes.
In one case, we have the old timer. Everything is done the “gold standard” way, the same way it’s been done for decades. There are no shiny new technologies or creative approaches because the old way should work just fine.
Then there is the gearhead. He is always coming up with something new. There’s a hand-built, custom app for everything he does, each one more clever and complicated than the last. He’s tried teaching his colleagues how to use the apps, but they’re really, really hard to use.
The last week in Iowa has shown how these two very different approaches to political technology can produce the same result: less timely, accurate, and trusted data about our elections.
In anticipation of this crucial election, the Iowa Democratic Party was charged with providing three different vote counts in an effort to increase transparency about a caucus process some believe to be antiquated and undemocratic.
To meet this goal, they said yes to the Gearhead approach. They hired a small, new technology company to build a mobile application instead of using simple off-the-shelf tools like Typeform or Google Forms; many reports suggest they didn’t have the time or money to test and train the volunteers to use the new apps. With new but untested technology came new avenues for errors and inefficiencies — the opposite of what technology should provide.
The Des Moines Register faced different challenges. They paid a calling house to conduct a survey over the phone — a no-brainer decision in 1996 but a questionable one today. With phone response rates declining and at-capacity calling houses relying on lightly trained and low-paid interviewers, the well-documented issues with phone polling aren’t getting any better.
The solution cannot be changed at the margins at calling houses or polling firms that make expensive polls even more expensive. Instead, the answer must be to adopt well-tested technology that reduces errors and costs in survey research.
The right technology is at the center of Iowa’s success story, and part of the path forward.
Change Research employs a very different approach from legacy pollsters, recruiting a representative sample online through a proprietary approach called Dynamic Online Sampling. Change’s methodology, refined over 1000+ polls since 2017, doesn’t rely on training third parties to use the software, minimizing the risk of human error and making it faster, more affordable, and more reliable. And according to The Economist’s G. Elliott Morris, this methodology was also the most accurate in Iowa.
Applying the right technologies is crucial in an environment eager for fast, accurate, and abundant data about voters’ preferences. Campaigns rely on data to shape their strategies and operations, to play the expectations game, and, if they are fortunate, to declare victory. The electorate itself is increasingly data-literate and data-hungry. And when news consumers exhibit skepticism about both polling and the integrity of our elections, it is more important than ever for political data to be communicated accurately, clearly, and in a timely fashion.
We believe that everyone involved in both sagas in Iowa had the best intentions and was trying to do the right thing. But as an industry, and as Democrats, we need to continue to strike the right balance between the Old Timer and the Gearhead, applying the most effective technologies quickly and decisively so we can defeat Trump and Republicans in 2020.
Mike Greenfield was the first data scientist at both PayPal and LinkedIn. He co-founded Change Research with the goal of bringing together leaders from the world of startups, data science, politics, and social change to re-imagine both the techniques and the uses of public opinion polling.
Nancy Zdunkewicz is a senior pollster and strategist to campaigns, committees, labor unions, and nonprofits. Prior to joining Change Research, she was the Director of Polling and Modeling for the Democratic Congressional Campaign Committee and Managing Director of the polling nonprofit, Democracy Corps.