Thursday, January 28, 2016

Data-driven politics: what is it? what is it good for?

If, like me, you are interested in the nuts and bolts of U.S. political campaigns -- not just ostensible issues but the mechanics -- you'll appreciate a two part interview with Daniel Kreiss of the University of North Carolina issued as a podcast by the geeks at 538.

The discussion left me with lots of thoughts, some of which I'll throw out here.
  • Kreiss dates data-driven politics to the late 1890s when disgust with corruption in elections (think city bosses and patronage) led political reformers to try to aggregate potential voters as individuals who could be persuaded by issues, rather than attracting them with social events, parades and booze. He only mentions in passing that the old "social" regime of boisterous elections created the highest percentage turnout in our history. We may have gotten more serious, but a lively element was drained from democracy with the clean up.
  • The professor describes how William Jennings Bryan built a card file of some 200,000 voters who had written him letters over a 30 year career in politics. In the mid-19th century, hand written letters (snail mail sent by post) were a central medium of campaigns, The reason that there are so many surviving letters from Abraham Lincoln is that he sat in Springfield and wrote literally thousands of notes urging on supporters before and during the 1860 campaign.
  • Kreiss stressed that early and mid-20th century campaigns always worked with some sort of voter file, with lists of names, addresses, party affiliations and sometimes other information. What they didn't have was computers, so upkeep was terribly laborious.
My mother maintained this system of data file cards for the 20th District, 20th Ward Republicans for decades.
  • He pinpoints the 1970s and 80s as eras when parties were relatively weak and voter lists became commodities provided by commercial political vendors. I can testify that these lists were lousy! To mount a campaign, you bought from the best vendor you could find and accepted that huge numbers of people, addresses and phone numbers included on them would be just plain wrong. This was also the era when campaign consultants made TV the center of electioneering -- a very lucrative strategy for them and probably initially effectual.
  • Kreiss credits the research of Donald P. Green (and Alan Gerber and Lisa Garcia Bedolla) with prompting a turn away from TV ads and toward using improved computerized party-maintained databases to facilitate personal contacts with voters in contemporary campaigns. Technology can help target, facilitate, and capture the results of those contacts.
  • He also observes that, for all its growing technical sophistication, much of this painstaking data collection still tends to get lost at the end of each campaign. If the data entry wasn't completed before election day, final contacts get tossed away along with useless door hangers and other paper debris. Nobody is paying to capture this final trove once the votes are cast.
  • I wish Kreiss and the 538 interviewer had talked about the effects of various mail ballot and early voting systems on campaigning. Managing that well is a new campaign frontier.
Kreiss and the 538 people remain agnostic on how much effect all this technology really has on the vote. It's hard to believe it doesn't accomplish something, but there still aren't good quantitative measures of how many votes are won by how much money and how many person hours through data-targeted campaigning.

Mostly these interviews made me feel sympathetic toward Iowans and New Hampshire residents this month. They are not going to be able to turn on any media or answer their phones or respond to a knock at the door without encountering someone from a campaign. Those of us who live in big states with late primaries in which party preferences are set in stone have little idea of how intense primaries in a year like this must feel at ground zero.

A History Of Data In American Politics (Part 1): William Jennings Bryan to Barack Obama ; A History Of Data In American Politics (Part 2): Obama 2008 To The Present

No comments: