Transparency Series on Polling

The Brown Institute began it’s 2016-2017 Transparency Series with a day-long workshop on opinion polls. Harry Enten and Neil Paine from FiveThirtyEight introduced 25 students to the complexities of modern polling. Many organizations conduct polls. But they can provide different, often conflicting impressions of public opinion. Enten, a political writer, and Paine, a sports reporter and data journalist, introduced students to the broad strategies and technical tools for making sense of poll results. Part art, part science, their approach rarely considers a single poll in isolation, but is instead informed by deep knowledge of the polling organizations, historical precedents for specific relationships between polls, and computational methods combining poll numbers.

Enten has an impressive portfolio of stories and shared his experiences with students. This was not a dry, technical exercise. Enten’s work is computational, but he is a first rate journalist. He cautioned students not to dive into a data set without a basic question to answer. We followed his process with exercises using the Roper Center’s extensive archive of polling questions. This incredible resource provides a detailed view of the public’s opinion over decades. Enten shared his other “go to” data sources and data analysis strategies, using some of his reported stories as examples. Paine introduced R, an increasingly common tool among data journalists. (In the first Transparency Series session of 2015-2016, Amanda Cox of the New York Times demonstrated her use of R for creating visualizations.) Through a handful of basic constructions, we learned how to implement Enten’s strategies, and, in the process, experienced simple programming as a powerful tool for finding stories in data.

We can’t thank Enten and Paine enough for sharing their weekend with us. It was an incredible glimpse into an extremely timely topic.