11:28am | 02/18/2018
In the previous article, I took a look at David Karpf's Analytic Activism— examining how political campaigns have reacted to a hybrid media environment.1 In that article, I also imagined a public citizen, a parrhesiastes, to participate in the establishment of a digital public domain for vibrant political speech and discussion beyond social media's "echo chambers"— disparate information bubbles filtered of ideologically oppositional content.2 Karpf's collective analysis of how media and analytics function within the modern political arena is incredibly insightful and engaging— but how did we get here? Additionally, how can past political campaigns function as models for the strategic implementation of emerging technologies in future races and activist movements?
Daniel Kreiss, in Prototype Politics: Technology-Intensive Campaigning and the Data of Democracy (TICDD), does well to address these questions— analysing the historic adoption of the Internet and the evolution of its involvement in political races.3 Kreiss looks specifically at the transfer of knowledge from one campaign to the next, and at how individual campaigns can function as explicit prototypical models for future elections— especially as their relationship with Karpf's hybrid media environment continues to mutate. This involves a critical examination of the political party (as well as third-party political data consultancies) as a vessel for carrying information from one campaign to the next:
During and after elections, particular campaigns are transformed through meaning-making processes into prototypes for some actors, a model for future campaign practice, and a set of claims about the world that are actionable for practioners.4
The bridging of this information— how analytic programming, organizational knowledge and constituent data moves between races and other political movements— is a key theme throughout TICDD. Lastly, Kreiss breaks down the pursuit of technical teams in identifying the whole citizen. This is the concept of how disparate pieces of information are captured from the spectrum of hybrid media to produce a more predictable, or wholistic, model of an individual voter:
Practitioners seek to be able to represent whole citizens through data as a way of relating to and leveraging their media use, psychological dispositions, and social relationships for electoral ends.5
In César Hidalgo's class: Networks, Complexity and Their Application, we've been taking a look at the mathematical models describing network topography and behavior.6 Our readings cover a breadth of professions and topics, from viruses and epidemiology to digital and social networks; however, what is quite incredible about network science is that it creates very accurate models that are applicable across a broader range of disciplines. On that point, I think it's interesting to understand the internal organizational networks of campaigns in addition to their relationship with voter data and the constituents themselves— specifically, how they organize and behave as networks.
There are a few interesting properties in regards to the "scale-free" network model— a network in which the degree (connectivity) of nodes follows a power law distribution.7
1. One principle of a scale-free network is that new links that are added will naturally show "preference" for more highly-connected nodes, or hubs if you will. Put another way, the existing nodes with more links are more likely to afford connectivity to newly added nodes. This is part of the theory of preferential attachment.
This makes a lot of sense if you imagine Google or Facebook as highly-connected hubs of the WorldWideWeb. If anyone is putting a website onto the Internet— perhaps connecting for the first time; then, they are much more likely to connect to these already highly-connected and active hubs. Similarly, websites that don't receive a lot of attention are unlikely to form connections with new nodes.
2. Another quality of the scale-free network is that (assuming some constant of growth) the age of the node has been found to correlate with its number of connections. That would indicate that the sooner a node has been established within a given network, the greater chances it affords for further connections.
It's remarkable how relevant these two simple principles are when applied to political campaigns— particulary, given that some of this research in network science had been established decades prior. Looking at the broader social and digital networks of information and people that are created over the duration of a political campaign, there appears a likely candidate for topographic and behavioral analysis. In fact, I would argue that these networks are the essence of the political race— which is especially remarkable considering the rapid rise and fall of this infrastructure from election to election.
In this sense, you might say that the campaign itself is really a machine for leveraging this networked space— of information, people, and skills. An interesting aspect of conceptualizing the campaign in this way is in the emergence of a set of universal principles that govern its behavior like in network science. In a way, this it how Kreiss writes Political Prototypes— looking not only at the winners, but finding correlations in the tactical methods, hierarchies, and efficacy of campaigns (along with other behavioral qualifications).
The models Kreiss focuses on— Howard Dean's presidential race, McCain, Obama's 2008 and 2012 campaigns, and Hillary's run in the 2008 primary— all exhibit unique and evolving interfaces with media, constituents and their data. Recall the second predictive principle of scale-free networks. We would expect that the sooner a campaign is able to begin developing their marketing strategy— building a brand, a narrative, and a database of people, the better its chances of reaching a larger audience (forming more connections).
Additionally, as I noted earlier, the Democratic and Republican Committees play substantial roles as data vaults for their respective candidate. Considering the DNC or the RNC as a highly-connected hub— and recalling now both principles of the scale-free network, it would seem that these user profiles should play a tremendous part in the outcome of a race. The problem, which Kreiss repeatedly emphasizes, is that the data often necessitates a time-intensive translation as it's handed off from one organization to the next— each with its own unique workflow and developer stack.
In an interesting network science article, The Strength of Weak Ties (1973), the author— Mark Granovetter — wrote:
Some have indicated that early innovators are marginal, that they 'underconform to norms to such a degree that they are perceived as highly deviant'. Others find that those named more frequently adopt an innovation substantially earlier.8
This is given against the "perceived risks of adoption of a given innovation"— a struggle central to the political race.9 Kreiss, as well as Karpf, makes it clear that the political prototype can be used to more strategically draw conclusions from previous campaigns. In this way, it can serve as a basis for the evolution of different digital social media platforms and the force of their involvement in the political race. Despite this analytic insight, its extremely difficult to agilely adapt to this changing environment. If, for example, you look at Google's unsuccessful challenge to Facebook— Google Plus, it can be a bit of a gamble with the attention and time of the technical team versus the potential success of the platform.
Of course, this is a risk that has won races. Obama's nickname, the first "YouTube" President— or the "Facebook" President— came about because of his relative mastery in adapting to these media, adopting them to popularize his campaign (and himself as a candidate). Going a step further, these network theories could even be used to analyze the viability of these potential platforms for effectively disseminating information on the progress of the campaign, the candidate, and their views on policy.
Marketing groups employ actors, musicians, and athletes to endorse a product because they are essentially highly-connected nodes that are now connected to that product— and will encourage further connections through this promotion. In the same way, presidential candidates must push to become these highly-connected nodes. Celebrity politicians, like Reagan or Trump, already have a powerful advantage through their presence and entanglement in this network and its diverse media. Others will dedicate years of work to building these connections— embedding themselves across different social circles, and constructing physical and digital mailing lists.
Of the more scientific— perhaps extended to the algorithmically derived— models of networks and voter behavior, the political writer Sasha Issenberg quotes the conservative opinion of Princeton's then-president, Woodrow Wilson:
I do not like the term 'political science'. Human relationships are not, in any proper sense, the subject matter of science. They are the stuff of insight, and sympathy, and spiritual comprehension.10
Throughout the majority of arguments and analysis on political campaigning, it seems, the discussion of ethics and morality have been absent. I think Wilson brings up an interesting point. Even in the vernacular languange of campaigns— "whole citizen", or "persuadable voters"— there is an objectification of the constituent. Value is afforded to citizens only if they're able to provide one of two things: a donation, or a vote. In fact, I found it disturbing to realize that it's common practice for many campaigns to suppress the vote of people who are undecided, but unlikely to vote in the candidate's favor— they'd rather they didn't vote at all. It seems as though, in the rush of technological innovation and algorithmic decision-making, campaigns have diminished the "whole citizen" to a set of statistics.
On the other side of the spectrum, working in contemporary trading and economics, Ray Dalio has imagined that qualifying people into something like a meritocracy would create enormous efficiency in the workplace— encouraging expert management and production. Along with Alexander Nix of Cambridge Analytica— which has developed a similar psychographic modeling technique for identifying persuadable voters (in the Brexit Leave campaign, and admittedly in the 2016 presidential election, here in the US)— this lack of ethical consideration appears insensitive to diversity and personal uniqueness. Additionally, this same model has been used as a powerful demographic surveyor for more effective propaganda. These same techniques have been developed for "pyscho-social" warfare: We're seeing it now in the US, in the uproar over the Russian "trolls". I think the way in which this information is so publicly transparent— published articles on the SCL website— makes it seem as though this is non-lethal. However, as information about the role of foreign interference reaches further audiences, it seems clear that some major reform is needed.11
The Guardian put out an article last summer with a great quote given by Alex Younger, head of MI6, in December of 2016. I think coming from his role of authority, it should be taken quite seriously:
“The connectivity that is the heart of globalisation can be exploited by states with hostile intent to further their aims.[…] The risks at stake are profound and represent a fundamental threat to our sovereignty.”12
N O T E S
1 Karpf, David. "Analytic Audiences." Analytic Activism, 2017, 93-122. doi:10.1093/acprof:oso/9780190266127.003.0004.
2 Karpf details these echo chambers in his book (cited in note 1).
3 Kreiss, Daniel. Prototype politics: technology-intensive campaigning and the data of democracy. New York, NY: Oxford University Press, 2016.
4 See note 3
5 See note 3
6 MAS.581 Networks, Complexity, and Their Application, Cambridge, MA. Class, MIT Spring 2018. Professor Cesar Hidalgo.
7 Power law distribution is a distribution graph (compare it to more regularly referenced "bell curve") that exhibits higher distribution toward the y-axis asymptote (extremity). One example is the distribution of wealth in the United States. A 2015 article published by Fortune Magazine cited a study estimating that "the wealthiest 160,000 families have as much as the poorest 145 million families."
8 Granovetter, Mark S. "The Strength of Weak Ties." American Journal of Sociology 78, no. 6 (1973): 1360-380. doi:10.1086/225469.
9 See note 8
10 Issenberg, Sasha. The victory lab: the secret science of winning campaigns. New York: B \ D \ W \ Y, Broadway Books, 2016.
11 SCL, Strategic Communications Laboratories. A relative of Cambridge Analytica serving "governments and military organizations worldwide" (meta-description, https://sclgroup.cc/home); however, many of the client-target relationships have been wealthy first-world to volatile third-world.
12 Cadwalladr, Carole. "The Great British Brexit Robbery: How Our Democracy was Hijacked." The Observer, May 7, 2017. Accessed February 26, 2018. https://www.theguardian.com/ technology/2017/may/07/ the-great-british-brexit-robbery-hijacked-democracy.