Shoe-Leather Pol­i­tics

The Vic­tory Lab

In The Vic­tory Lab: The Se­cret Sci­ence of Win­ning Cam­paigns (SSWC), the closed cur­tains typ­i­cally shrouding the inner-work­ings of cam­paigns are drawn open, pro­viding an in­sid­er's view on voter ac­qui­si­tion and the or­ga­ni­za­tional topogra­phies that feed these in­tense mo­bi­liza­tion ef­forts. With some ex­cep­tional first-person ac­counts from leading fig­ures in po­lit­ical sci­ence and cam­paign man­age­ment, Sasha Is­senberg de­tails an in­ti­mate nar­ra­tive of how these elec­tion ma­chines have evolved. What is par­tic­u­larly in­ter­esting is how po­lit­ical vet­erans are adapting old "shoe-leather pol­i­tics" to in­cor­po­rate modern and in­ter­dis­ci­pli­nary strate­gies in data an­a­lytics, voter mi­cro­tar­geting, and pre­dic­tive al­go­rithmic mod­eling.

No day after the elec­tion

In the small town where I grew up, there was a trav­eling car­nival that would setup rides for a week or two each summer. For someone with no prac­tical ex­pe­ri­ence in pol­i­tics, the elec­toral race feels some­what com­pa­rable— an ex­trav­a­gant kind of po­lit­ical pa­rade that pops up in the col­lec­tive Amer­ican con­scious­ness every four years. In much of the lit­er­a­ture, the cam­paign is por­trayed as a high-stakes and nimble game. Is­senberg, quoting an un­named of­fi­cial, em­pha­sizes the fierce drive of the cam­paign quite frankly:

"There is no day after the elec­tion."¹

Gar­nering sup­port and pop­ular recog­ni­tion can take years of prepa­ra­tion and or­ga­ni­za­tion; and once those bal­lots are cast, that's the end of it. Win or lose, the vol­un­teers go home and the field of­fices close up shop.

The In­ternet and Po­lit­ical Cam­paigns

During dis­cus­sion in Nicco Mele's The In­ternet and Po­lit­ical Cam­paigns, I've been taken aback by the amount of tools and in­fra­struc­ture that is often left by the way­side. The dif­ferent sys­tems and soft­ware for tracking voter data, pre­dicting voter turnout, or strate­gi­cally an­a­lyzing and lever­aging un­de­cided votes— de­vel­oped and de­ployed with im­pres­sive turn­around— can simply dis­ap­pear. Some­times data makes its way back to the can­di­date's re­spec­tive na­tional com­mittee; but, by and large as these teams dis­solve, so too will the tools and in­fra­struc­ture.²

Modern Po­lit­ical Sci­ence

The in­tro­duc­tion of the po­lit­ical sci­en­tist seems to mark a sig­nif­i­cant change in this cul­ture of dis­pos­ability— studying the ex­isting struc­ture and methods in order to apply the same tac­tics to fu­ture so­cial or po­lit­ical mo­bi­liza­tion. In keeping with what Daniel Kreiss prof­fers in Pro­to­type Pol­i­tics, dis­cussed in my last post, po­lit­ical sci­en­tists— drawing from an amal­ga­ma­tion of tra­di­tional mar­keting tech­niques, so­cial sci­ences and be­hav­ioral psy­chology— seek to gen­erate con­crete and re­pro­ducible models of which strate­gies work and why.

On the tech­nical side, this ac­cu­mu­la­tion of data some­times man­i­fests in post-elec­tion con­sul­tan­cies— re­jecting the dis­so­lu­tion of months or years of work by re­de­vel­oping cam­paign soft­ware and ser­vices as pri­vate-sector busi­nesses. What po­lit­ical sci­en­tists bring to the table is a new level of rigor and pre­ci­sion, proving or dis­proving even long-standing cam­paign methods with care­fully con­structed tests and analysis. For the vet­eran cam­paigner, inviting a po­lit­ical sci­en­tist or tech­nical vir­tuoso within the walls of the inner circle— un­set­tling deeply en­trenched con­ven­tions on how to dis­tribute pre­cious time and funds— takes a lot of con­vincing. How­ever, for the prag­matic Dave Carney— in­doc­tri­nated into the busi­ness of cam­paigning since 1980— it be­came a simple ques­tion of cost:

We spend a lot of money on mail and phones... If it's not working, let's spend it on things that do work, or don't spend it... There was more to it than doing the same old thing over and over.³

Car­ney's shrewd­ness to reach be­yond the walled garden of pol­i­tics, cre­ated an op­por­tu­nity for fruitful ex­change— putting some shine back in that old shoe-leather.

Changing Tech­nology

Of course, this isn't an iso­lated in­ci­dent. The rise of com­puting has marked an in­cred­ible par­a­digm shift that is still trans­forming many pro­fes­sions. During my un­der­grad­uate in ar­chi­tec­ture, this was ev­i­dent in the fierce ide­o­log­ical re­jec­tion of com­puter-aided de­sign (CAD) by those who had tra­di­tion­ally worked on hand-drafted draw­ings— now re­garded as a la­bo­rious and an­ti­quated prac­tice among most modern de­signers. Dig­ital ren­dering, Adobe Pho­to­shop, para­metric mod­eling, vir­tual re­ality tours, and other tech­nical util­i­ties for rep­re­senting or ex­pressing a space have be­come ubiq­ui­tous to the ar­chi­tec­tural firm. Yet some of these tools weren't even in­vented until the past decade!

Sim­i­larly, the long-standing prac­tices of po­lit­ical cam­paigns, though not en­tirely re­placed or sub­sumed by the latest in tech­nical in­no­va­tion, are bound to be in­formed— even aug­mented— by these dra­matic changes. Nev­er­the­less, ar­chi­tects haven't stopped sketching— and can­vassers haven't stopped knocking on doors. How­ever, since the in­tro­duc­tion of po­lit­ical sci­en­tists onto the cam­paign trail, those ac­tiv­i­ties that had once been taken for granted have in­deed faced re­newed scrutiny— re-eval­u­ated to test their rel­e­vance against new and changing con­texts, de­mo­graphics and media.

One of the more re­mark­able changes that re­sulted from this in­ter­dis­ci­pli­nary ex­change is the in­flu­ence be­hav­ioral psy­chology has had on cam­paign strategy:

Elec­tion scholars had ig­nored large swaths of modern psy­chology, which was in­creas­ingly iden­ti­fying ways in which people were nei­ther so­cially pre­pro­grammed to­ward cer­tain at­ti­tudes nor walking cal­cu­la­tors able to make per­fectly ra­tional choices.⁴

Cam­bridge An­a­lytica

How­ever, there is at least one com­pany, the data an­a­lytics firm Cam­bridge An­a­lytica (CA), that claims to have found an ef­fec­tive way for cam­paigns to cap­i­talize on voter psy­chology. The role CA played in the 2016 cam­paign has only been fur­ther ob­fus­cated since the elec­tion; how­ever, in the frenzy to cap­i­talize on the golden ques­tion of "How Trump Won the Cam­paign", CA is an easy target for pointing fin­gers.

Psy­cho­graphic Mod­eling

In the months leading up to their part­ner­ship with the Trump cam­paign, CA had been pub­licly pro­moting their use of the OCEAN model for "psy­cho­graphic" pro­filing; but, Brad Parscale, the cam­paign's dig­ital media di­rector (now as­signed to the post of cam­paign man­ager for 2020) has since un­der­played the role CA had in the cam­paign's out­come.⁵ Ad­di­tion­ally, only about 4 months after Trump's vic­tory, The New York Times pub­lished an ar­ticle which states that "Cam­bridge ex­ec­u­tives now con­cede that the com­pany never used psy­cho­graphics in the Trump cam­paign."⁶

In April of 2016, WIRED Mag­a­zine in­cluded Alexander Nix, CA's CEO, in 25 Ge­niuses Who Are Cre­ating the Fu­ture of Busi­ness, stating:

For too long... de­mo­graphics and pur­chasing be­havior have been the pri­mary guide­posts of the mar­keting in­dustry, used to guess what a target au­di­ence might want. Now Nix's com­pany... can pro­vide psy­cho­log­ical pro­filing to help ad­ver­tisers tailor their mes­sages to spe­cific per­son­ality types.⁷

But if CA isn't using de­mo­graphics or pur­chasing be­havior to build these models, what are they using? Ad­di­tion­ally, how are they get­ting this data? This has proven to be a dif­fi­cult ques­tion. CA might ac­tu­ally have a le­git­i­mate legal de­fense in claiming that the re­lin­quishing of data, or any spe­cific de­tails about how this data in­forms its per­son­ality pro­files, would com­pro­mise their busi­ness model. As a re­sult, CA's "se­cret sauce" is ren­dered ir­re­solv­able.⁸ In fact, if they'd just kept their mouths shut, this black magic cam­paigning might look like a sure­fire ad­van­tage to fu­ture can­di­dates.

Is­senberg, how­ever— of­fers a quote by the econ­o­mist Richard Thaler to em­pha­size the dif­fi­culty of im­ple­menting an ef­fec­tive strategy drawn from be­hav­ioral psy­chology:

The sci­ence half of po­lit­ical sci­ence is to some ex­tent a bit of a mis­nomer. At least no one has been quite ready to agree on what the sci­ence part of it is.⁹

Brent Seaborn— man­ager of an­other po­lit­ical data con­sul­tancy, Tar­get­Point— cor­rob­o­rates this sen­ti­ment in the same ar­ticle from The Times, saying of CA:

They've got a lot of re­ally smart people; but it's not as easy as it looks to tran­si­tion from being ex­cel­lent at one thing and bringing it into pol­i­tics. I think there's a big ques­tion about whether we think psy­cho­graphic pro­filing even works.¹⁰

Un­re­mark­ably, people do want to know whether or not these tac­tics work, and Trump's vic­tory has es­pe­cially peaked both public and pri­vate in­ter­ests into the inner-work­ings of these al­go­rithmic "black boxes" and how they're de­ter­mining vi­able tar­gets— and po­ten­tially shaping cam­paigns. Is­senberg ex­plic­itly em­pha­sizes how po­lit­ical sci­en­tists, and a cul­ture of rig­orous and pre­cise testing, is re­ally where this un­der­standing emerges.

How­ever, Is­senberg lends an­other van­tage— that of po­lit­ical sci­en­tist Samuel Popkin, pre­vi­ously a Har­vard sta­tis­tics pro­fessor. Pop­kin's at­ti­tude, as in­ter­preted by Is­senberg, was that "even those with col­lege diplomas could never gather all the in­for­ma­tion nec­es­sary to weigh the en­tire set of costs and ben­e­fits at­tached to each issue or can­di­date."¹¹ De­spite the im­passe this con­tin­gency cre­ates, Is­senberg's book is ti­tled "The Vic­tory Lab"; and not "We Don't Know What the Hell is Going On". And I think, overall, Is­senberg still main­tains that the true value in po­lit­ical sci­ence is de­pen­dant on a process of crit­ical de­ci­sion-making in­formed by rig­orous testing— per­haps, drawing from an in­dul­gence of in­formed but in­tu­itive feel­ings about where to di­rect these ef­forts without dis­counting old methods of cam­paigning.

The Rea­soning Voter

Is­senberg quotes Popkin— com­menting on the po­lit­ical cam­paign in his book, The Rea­soning Voter:

These con­tests are com­monly crit­i­cized as tawdry and point­less af­fairs, full of dirty pol­i­tics, dirty tricks, and mud­sling­ings, which ought to be cleaned up, if not elim­i­nated from the system. In their use of san­i­tary metaphors, how­ever, many of these cri­tiques con­fuse judg­ments of Amer­ican cul­ture with aes­thetic crit­i­cisms of Amer­ican politi­cians... They do not look closely at how voters re­spond to what they learn from cam­paigns, and they do not look closely at the people they wish to san­i­tize. If cam­paigns are vulgar, it is be­cause Amer­i­cans are vulgar.

Popkin elab­o­rates his forth­right con­cep­tion of the elec­torate in an ar­ticle for the Wash­ington Post en­ti­tled We Need Loud, Mean Cam­paigns.¹² Coming more from the side of the "san­i­tizers", I would def­i­nitely feel un­com­fort­able with some of the eth­ical de­ci­sions made over the course of a cam­paign. Reading about voter sup­pres­sion tech­niques, I was pretty dis­gusted that this was an in­te­gral part of cam­paigning. But, as Popkin points out, this sen­ti­ment isn't going to pre­vent the op­po­si­tion from em­ploying the same tac­tics; nor will it en­courage and en­liven de­bate:

Re­formers say they want to turn down the volume, dis­cuss more im­por­tant is­sues and turn out more voters— worthy goals, but also con­tra­dic­tory. Deco­rous cam­paigns will not raise more im­por­tant is­sues. Nei­ther will they mo­bi­lize more voters nor over­come off­stage mut­ter­ings about race and other so­cial is­sues... If gov­ern­ment is going to be able to solve our prob­lems, we need bigger and noisier cam­paigns to rouse voters.¹³

In the 2016 elec­tion— ar­guably the biggest and nois­iest yet— Trump seems an ideal can­di­date to test this theory. And Popkin isn't wrong! The amount of ex­pres­sion that I've per­son­ally wit­nessed, es­pe­cially on­line, is as­tounding. The elec­tion has dis­in­te­grated friend­ships; thrust ques­tions of racism, sexism, im­mi­gra­tion, na­tion­alism and equality to the fore­front of public dis­course; and po­lar­ized par­tisan pol­i­tics to an ex­treme.

What Pop­kin's model of voter be­havior does not guar­antee, how­ever, is that these loud, mean cam­paigns will re­sult in pro­duc­tive change. While pop­ular in­volve­ment in po­lit­ical dis­cus­sion and mo­bi­liza­tion have ex­pe­ri­enced a re­vi­tal­iza­tion un­like any other elec­tion in years, the new methods of com­mu­ni­ca­tion— though pro­viding an ubiq­ui­tous channel for ag­o­nistic speech— have ef­fec­tively clois­tered op­po­si­tional opin­ions and shrouded those "echo cham­bers" in a layer of static that has proven in­cred­ibly dif­fi­cult to pen­e­trate.

Pop­kin's ar­ticle was pub­lished in 1991, and I'm cu­rious to know whether or not he main­tains his opin­ions even now... The pas­sionate lifeblood, the "brass bands" and ex­cite­ment, has cer­tainly found its way back into the po­lit­ical cam­paign; but I don't think we have the tools to mit­i­gate this stim­u­la­tion.¹⁴ I think Pop­kin's in­ten­tions are good; but as the cam­paign be­comes louder and meaner, so must the old shoe-leather pol­i­tics evolve to con­front this noise— lever­aging pop­ular par­tic­i­pa­tion with modern means of po­lit­ical ac­tion and so­cial re­form.