Persuasion and the other thing: A critique of big data methodologies in politics | Ethnography Matters

Molly Sauter; Persuasion and the other thing: A critique of big data methodologies in politics; In Ethnography Matters; 2017-05-24.

tl;dr → 3026 words. Big Data (which so is very big) is bad. The sphere is problematized. A problematic which situates the hegemons is synthesized via the dialectic. A mode of resistance is posited.

<ahem>… and by way of brief rebuttal: The Computers and The Establishment that owns & operates The Computers, their work inuring to the mutual benefit of them both, individually and severally, are smarter than all that (c.f. the trivial use of “grep -v”), and also the suggested modality of dissent violates the T&C which was previously freely given and binds & constrains individual future actions; its unilateral repudiation makes the performer at once dishonest, conflicted, and an outlaw who deserves no quarter; not in theory, not in practice, or under the reigning jurisdictional supervision (c.f. 18 U.S.C. Section 1001, as opined).</ahem>

Previously filled.


  • Cambridge Analytica
  • Donald Trump
  • Brexit Campaign
  • Facebook
    • “likes”
    • targeted nudges
  • Mother Jones
  • The Guardian
  • SCL Group
  • Apple (Computer) Inc.


  • There is not enough consent (from the subjects)
  • <quote>Democracy shifts from a form of governance at least theoretically concerned with public debate and persuasion to one focused on private, opaque manipulation and emotional coercion.</quote>


The obfuscation schemes, taxonomized in Brunton & Nissenbaum:

  • noisy bots
  • “like-farming,” i.e. spamming.
  • TrackMeNot
    a browser extension which generates abusive search query engines.
  • AdNauseam
    a browser extension which generates abusive click streams.
  • FaceCloak
    Something about storing data “off Facebook,” yet performing the data “on Facebook.”
  • Bayesian Flooding … sounds fancy; it means creating profile- & page- spam entries on Facebook.


Unless otherwise noted persons are credited as “an activist.”

  • Finn Brunton
    with Helen Nissenbaum
  • Michal Kosinski, a bad guy in the pantheon
    with et al. as David Stillwell, Thore Graepel
  • Helen Nissenbaum
    with Finn Brunton (for symmetry)
  • Kelly Oliver
  • Molly Sauter
  • Zeynep Tufekci
  • Sara Marie Watson


… sounds fancy, and more than a little dangerous (<quote> cacklingly evil</quote>).  In rough order of appearance.

  • psychographics
  • algorithmic nudging
  • entitlements (<quote> held by advertisers, tech firms, and researchers who deploy big data analytics in support of political campaigns or other political projects </quote>
  • sense of entitlement
  • subjectivity (something about having agency; being such is good)
    objects (data objects); something about not having agency; being such is bad.
  • obfuscation
  • sabotage
    <quote>sabotaging the efficacy of the methodology in general, to resist attempts to be read, known, and manipulated.</quote>
  • emotional contagion; c.f. Facebook, an ”experiment,” 2014
  • nudge (contra shove)
  • algorithmic modeling → “opinions embedded in mathematics” [page 21, O'Neil].
  • otherness
  • knowability
  • digital shadow-selves
  • a paradoxical problem
    wow man, dig it … a paradox, a problem with a paradox, that’s like a paradox2.
  • data broker
  • entitlement of inference
    <quote>a certain entitlement of inference</quote>
    <quote>the entitlement of inference on display</quote>
  • influence techniques
    secret or opaque influence techniques
  • consent of the governed
    meaningful consent
  • inferential modeling collectssynthesizes non-disclosed information
  • opting out
    social media abstinence
  • data doppelganger
  • pervasive surveillance and modeling systems
  • obfuscation
    <quote>creates noise, either at the level of the platform or the individual profile</quote>


Comments are closed.