Introduction
The spotlight on psychographic profiling and targeting in political campaigns really hit the headlines because of Cambridge Analytica, a now-defunct company specializing in political data analytics and behavior change. Whistleblowers from within Cambridge Analytica have claimed that the company carried out psychological operations (“psy-ops”) during election campaigns worldwide (Cadwalladr, 2018; Kaiser, 2019b; Wylie, 2019). This paper takes a theoretical and practical look at those claims. It explores whether combining psychographic profiling with big data in election campaigns counts as “psy-ops,” focusing especially on a key example-the role of Cambridge Analytica in the UK’s 2016 Brexit Referendum about leaving the European Union.
Psychographic research actually started back in the 1960s to better understand consumer behavior rather than politics, moving beyond simple demographics into areas like personality traits, hobbies, opinions, needs, values, and attitudes to gain fresh insights from large groups of people (Wells, 1975). Over the past decade or so though, psychographics have been paired with neuromarketing tools alongside data tracking techniques to measure how advertising emotionally affects audiences and how to craft political messages tailored to their psychological profiles (Chester and Montgomery, 2017; Bakir and Mc Stay, 2018, 2020). The European General Data Protection Regulation defines “profiling” as any automated processing of personal info used to analyze or predict aspects such as work performance or personal preferences (Privacy International, 2020). According to Alexander Nix-Cambridge Analytica’s CEO-it’s this mix of psychographic targeting combined with big data that helps marketers figure out personalities behind target audiences so they can customize messages. Nix said this when describing their work on Ted Cruz’s campaign for the Republican nomination in 2016. He said big data lets campaigners know what message will persuade whom on what issue before even making it; plus addressable advertising tech makes it possible to deliver highly personalized ads (Nix, 2016).
Psy-ops is actually a part of broader “Information Operations,” which is another way of saying “information warfare.” That umbrella covers psy-ops along with electronic warfare, computer network ops and deception (Briant, 2015, p.23). Simpson (1994) points out that psy-ops has long been a tactic used either in war or social struggles-it appears in military manuals and communist writings alike. He defines psy-ops as mass communication tied directly to selective violence tactics like sabotage or assassination aimed at achieving ideological or political goals by exploiting cultural-psychological traits of a group plus its communication channels (Simpson, 1994 p.11). So yes-psy-ops involves forceful coercion through violence according to Simpson-but coercion can also happen without outright violence: by limiting people’s choices through manipulating how they feel or act. In line with this idea Briant describes psy-ops as propaganda designed specifically to stir certain emotions and behaviors among targeted groups (Briant, 2015 p.12-13). Under this way of thinking if psychographic profiling plus big data leads people into constrained choices by controlling what info they see-that fits right into psy-ops territory.
To dig deeper into whistleblower claims about psy-ops here we place Cambridge Analytica’s profiling efforts within academic studies around deception, coercion and influence drawn from propaganda theory along with psychology & marketing research. These fields wrestle with figuring out when persuasion crosses over into manipulation; show how digital marketing tools lend themselves toward deceptive or coercive influence; and explore what that means for democracy overall. Using case study methods we follow up with an empirical look at whether Cambridge Analytica’s psychographic profiling was coercive or deceptive during Britain’s crucial Brexit referendum campaign in 2016 on EU membership. Drawing on national surveys from US/UK/worldwide sources we wrap up discussing public rejection of these tactics; measures put forward since then; plus urgent topics needing more research.
Literature Review
Persuasion vs. Influence
When exactly does persuasive messaging become manipulative propaganda? And when might such shifts be okay under democratic norms? Researchers studying propaganda & policy-making have tackled these tough questions.
Propaganda studies offer several frameworks for telling apart ethical persuasion from propaganda based mostly on honesty about who sends messages-and whether messages are truthful (Jowett & O’Donnell ,2012). But Bakir et al.,(2019) argue those models miss some parts by only focusing on deception without capturing all types of propaganda properly nor helping us clearly separate it from persuasion altogether. They suggest that ethical persuasive communication should always rest upon informed consent-which needs three things: enough info so people can judge properly; no deception involved so consent isn’t built on lies; and freedom without pressure like threats-people must give consent voluntarily not under force (Bakir et al.,2019; see also Chappell ,2017).
Decisions made freely after being fully informed are disabled too when influence tactics aim just at changing behavior instead of attitudes-as often happens for military/government uses too. Briant notes British military increasingly blends propaganda with deliberate environment manipulation aimed at influencing decision-making-for example offering jobs where bomb-making was common so people switch away from violent acts but don’t necessarily change their views about insurgency itself(Briant ,2015, p64). Influence activities aren’t limited just to defense agencies either-they’re central even inside British government: The Behavioural Insight Team (“nudge unit”) launched back in 2010 became partly employee-owned by around mid-decade while working globally by applying behavioural science insights across multiple countries aiming “to inform policy & improve public services” by nudging citizens towards better choices(The Behavioural Insights Team ,2020).
Nudges refer basically any changes made within choice contexts that steer people predictably without forbidding options nor changing incentives much-and crucially must remain easy/costless for folks wishing not to comply(Thaler & Sunstein ,2008, p6 ). Mols et al.(2015, p84) narrow down “nudges” further emphasizing interventions exploiting human psychology where automatic shortcuts override deliberate thinking -like going with easiest options or following social norms. Facebook’s get-out-the-vote button introduced first time USA2008 then Scotland’s ref’14 exemplifies this-a simple reminder shown once per user reminding them voting day is happening + showing friends who’ve voted + polling locations link-users could easily ignore it yet Facebook found small bump voter turnout(Bond et al., Jones et al., dates cited ).
Public policy scholars critique nudges for various reasons including doubts over long term effects but mainly because nudges short-circuit reflection/deliberation depriving voters chances consider pros-cons thoroughly(Hausman & Welch , 10 ). Nudges may secretly trick citizens towards behaviors deemed desirable by elites(Facebook execs/politicians/expert planners)-making them inherently elitist choice-limiting tools(Mols et al., 87 emphasis added ). Whether outcomes end positive socially doesn’t erase concerns: limiting choice + blocking thought = coercion worrisome whoever wields power. Facebook’s nudge experiments raise red flags since users got different versions silently based algorithm optimization/no prior notice + not everyone uses Facebook + demographic differences mean uneven impact = foreign platform potentially meddling elections abroad(Grassegger ,18).
So broadly speaking both propaganda research & policy studies agree persuasion turns manipulative once deception/coercion dominate via threats/choice constraints/reflection blocks leading target groups’ opinions/actions no longer truly informed nor freely chosen. Governments argue sometimes such actions yield national good-from reducing insurgencies improving services-but such logic weakens greatly applied inside elections seeking votes. Campaigning using deceit/coercion discouraging critical thought contradicts democratic ideals where voters deserve free/informed decisions over leadership.
Political Campaigns Meet Digital Marketing Tools: The Rise of Influence Activities
Politicians nowadays increasingly rely on digital marketing methods prone towards influence operations-and sadly these sometimes involve clear deception/coercion. In recent years traditional electoral techniques centering around demographics/opinion polls/targeted ads/direct mail have gradually shifted towards harnessing big data analytics driven by complex corporate webs involving data brokers/analytics firms. Broad voter lists get merged commercially sourced info creating detailed granular profiles(Bartlett et al., Nadler et al., Privacy International ). This fuels digital influence campaigning enabling delivery online optimized adverts fine-tuned per audience segment.
One popular technique is A/B testing comparing two variants quickly often powered now via AI. The Obama ’12 campaign ran hundreds boosting donations/signups significantly(Formisimo ), while Trump’s ’16 team reportedly tested tens-of-thousands daily ads using Facebook’s Dynamic Creative tool optimizing engagement(Beckett ; Bartlett et al.).
Other sophisticated tools include psychographics mixed with emotional/neuromarketing combined tracking users’ reactions tailoring effective appeals precisely. Chester/Montgomery, Kimetal. Mc Stay document evolution whereby previously core emotional testing enters vast new territory thanks big-data enabling far deeper psychological/political inference(Tufekci ). Kosinski created an algorithm predicting Big Five personality traits automatically based solely off Facebook likes(Gosling/Mc Crae ) extending predictions further politically/religiously etc. Kosinski’s findings attracted Cambridge Analytica(Federal Trade Commission ). Subsequent meta-analysis supports reliability across social media footprints(Azucaretal.) albeit caution needed given limited languages/samples. These advances suggest covert behaviour-change campaigns achievable via multistream data access/micro-targeted ads customized psychologically(Kaiser , Federal Trade Commission reports, Wylie ).
Digital marketing firms also help find new supporters via “lookalike” modeling linking voter records emails cookies social handles demographics interests effectively expanding targets. Political directors like Brad Parscale leveraged Facebook platforms growing base matching similar non-supporters(Green/Issenberg); UK’s post-’15 campaigns likewise use analytics identifying persuadables swing voters(Electoral Commission), delivering location/time matched messaging(Kaiser ). Leave. EU trademarked “behavioural microtargeting” combining psychological analysis creative message tailoring(Kaiser ).
Although these developments promise improved engagement responsiveness digital marketing unfortunately also enables deceptive/coercive tactics. Deception has always existed democracies’ elections(Herbst ; Perloff ), but second decade twenty-first century shows rising manipulation digitally. For example bots fueled fake grassroots support (“astroturf”) inflating candidates spreading false stories during US midterms/Massachusetts special(Ratkiewicz ; Metaxas), amplified low credibility news articles(Shao ); micro-targeted divisive anonymous accounts spread misleading inflammatory content shaping marginal district votes(Intelligence Community Assessment, Bakir/Mc Stay/Kimetal). Investigative reports state CA provided Trump campaign detailed statistical models isolating likely supporters enabling precise Facebook ad targeting(Green/Issenberg, Kaiser, Wylie ). In last days ‘16 campaign attempts suppress turnout among key Clinton demographic groups occurred(Kaiser, Wylie, Banks shipman accounts). Such manipulations control information exposure constraining individual choices thus constituting coercion.
Democratic Implications of Digital Deception, Coercion, and Influence Activities
The evidence above illustrates how deceptive/coercive influence activities increasingly shape digital political campaigns. Psychographics add another dimension claiming personality prediction accuracy fed by massive datasets. Against hype skepticism remains regarding CA’s uniqueness/impact(González)and media effect challenges measuring true voting shifts(Aral/Eckles ; Marchal ), but enough empirical findings prompt serious democratic reflections.
One strand looks at algorithmic filter bubbles shaping content consumption selectively hiding opposing views reducing diverse dialogue(Pariser, Douglas, Zollo, Bessi etc.). Conspiracies thrive confined inside agreeing communities generating echo chambers(Douglas); dissent ignored/polarization worsened(Zollo); misinformation correction efforts largely fail. Bubbles enable micro-targeted political messaging invisible wider publics hindering fact-check/counteractions.
Another strand analyzes fake news dissemination(Vosoughi ) finding falsehoods spread faster/broader/deeper driven mainly humans. Guess reveals older Americans disproportionately share fake news even accounting ideology. Conspiracy-oriented users promote rumors(Bessi), dissent met resistance increasing polarization(Zollo ) highlighting dangers injected disinformation poses democratically undermining trust participation.
Nadler identified a powerful ‘digital influence machine’ exploiting behavioral science/social division leveraging identity threats coalition divides behavioral nudges weaponized politicization(Nadler, Kimetal). Rather than belief changes focus lies amplifying resentments sow distrust subtly shifting behavior e. g. vote mobilization suppression. Knowledge/data-driven identification vulnerable audiences typical marketing practice-not rogue alone. Numerous applications raise normative worries regardless impact size. Marchal notes experiencing such covert microtargeting damages perceived electoral fairness weakening democratic confidence(Marchal).
Given democracy risks outlined next sections examine if CA’s UK Brexit referendum involvement showed similar deceptive/coercive patterns.
MATERIALS AND METHODS
CASE STUDY METHOD
This paper uses qualitative case study method aiming identify significant features(deception/coercion/influence)within complex opaque digital campaigning preserving rich details. Gerring says case study suits exploratory deep holistic understanding real world phenomena well(Gerring).
This case critical because(a)CA pioneered military-style psy-op adapted politically(b)whistleblowers/journalists/regulators uncovered practices giving rare insight(c)social acceptability debate enabled.
Difficulties arise due lack direct access creators/messages/audience post collapse NDA/lawsuits restrict interviews regulators block(Io C, F. T. C) Journalistic inquiry hindered leave. EU legal action. Archives incomplete(brexit predates ad transparency archives), audience recall unreliable. Non-branded ephemeral social media make tracing exposure difficult. Still document-rich public sources available.
Sources include UK Parliament DCMS Inquiry extensive hearings/witnesses/documents(Regulatory bodies Electoral Commission/Data Protection Office FTC Legal actions Books investigative journalism statements whistleblowers ex-party officials directors promotional materials(Wylie, Kaiser, Nix, Banks, Wigmore, Soutter, Dennemarck) Statements critically assessed noting self-promotion strategic PR spin notably Nix/Wigmore admissions.
Placing CA statements/actions against broad multi-source archive clarifies contradictions sheds light secretive deceptive coerced digital campaigning brexit referendum context.
RESULTS
Cambridge Analytica – Rise And Fall
CA had complicated corporate setup active UK(us dates vary parent companies involved(Siegelman, Kaiser UK Court judgement)(see fig1).
SCL Group Ltd Corporate Structure And Subsidiaries
SCL Elections executed psy-ops worldwide mostly emerging democracies(Channel4 News DCMS Kaiser Wylie Briant refs). Whistleblower Kaiser notes rather than broad political ads goal was creating conditions making desired actions likelier(Kaiser). Parent SCL Group also defense sector deploying psy-op via SCL Strategic Ltd(Tatham Wylie refs)(fig1). Kaiser cites brochure boasting ‘psy-op use defence/humanitarian’. Wylie testifies CA funded Mercer aimed adapting IO tactics militarily applied politically incl. uk. US(Wylie). Data misuse drove downfall. Wylie’s FB user exploitation allegations contradicted CEO denials(CA/Nix testimonies). Undercover sting aired March18 C4 showed Nix bragging bribery/honey traps/media manipulation global elections(Channel4 News DCMS refs). Post collapse ICO fined for serious Data Protection Act breaches related US presidential links(IC Office report Nov18) FTC levied $5bn FB fine July19 also filed complaint against US arm CA harvesting FB user personal info illegally(Federal Trade Comm reports Wylie refs) Data collected deceptively via GSRApp harvested millions FB users incl. friends violating privacy rules(Nix Kogan Federal Trade Comm documents Schroepfer FB disclosures ) CA trained algorithms scoring personality matched scores US voter files employed targeting(Federal Trade Comm Wylie Kaiser refs) Database included multi-sourced info proprietary client owned estimated thousands datapoints per adult USA(Kaiser) Coercively deployed. Wylie testified strategy stoked paranoia, bigotry tested reaction provoking discomfort re Mexican immigrant marriage question exposing triggers advancing conspiratorial mindsets(Wylie Senate testimony DCMS ) Psych research confirms conspiracies affect attitudes variably sometimes reduce voting/trust depending context(Douglas Kim Cao refs) Regardless actual impact described approach is covert coercive manipulation altering information environment posing choice limits affecting voters exposed messages questioning same method used Brexit?
The Brexit Vote – Deceptive Coercive Digital Campaigning
The June23rd ’16 Brexit vote decided narrowly Leave vs Remain. Leave pushed immigration control, reduced terrorism risks, economic sovereignty opposed Remain economic arguments. Leave Remain identity polarization intense enduring years(Hobolt references). Both sides misused stats produced falsehoods fact-checkers noted(Kirk Full Fact ref ) Leave. EU unofficial Leave group led Banks/Tice notable profiled targeted deceptive messaging linked closely CA involvement examined here. Leave. EU Did Not Declare Expenditures With Electoral Regulator On Ca Despite Claims Of No Work By Nix. Banks Admits Some Ukip Data Used For Scoping Work Which Was Then Unpaid. Kaiser’s Whistleblower Evidence Contradicts This Claim Showing Early Engagement With Leave. EU And Other Groups Through Data Analysis And Modeling Of Persuadable Voters Including Four Key Groups Named(Eager Activists Young Reformers Disaffected Tories Left Behinds)(Kaiser submissions ) Former Ukip Officials Also Confirm Database Access Shared Between Parties Following Financial Trouble. Leave. EU Used Modeling To Identify Regions Most Concerned About EU Immigration For Nigel Farage Target Visits(Wigmore Interview Evidence)Using Social Media Look-alike Techniques Or Modeling Copies Compliment Large Scale Poll Phone Surveys Help Fine Tune Messages(Banks Shipman citations ) Questions Persist Over Actual Datasets Used Behind Political Profiling Also Murky Details Around Covert Coercions Emerging From Psychographically Based Messaging Combined With Real World Visits Designed To Maximize Exposure To Specific Framing. Coercively Speaking Military Psy-Ops Tool Target Audience Analysis Developed By Defence Arm Scl Group Employed Digitally Politically Offering Behavior Change Via Psychological Manipulation Tactics Part Of Pitch Offered Voter Suppression Included Leaving Out Opposition From Outreach(Covert Suppression Control Strategy )(Cambridge Analytical Docs Kaiser’s Testimony Tatham NATO Ref.). Pitch Emphasizes Personalized Target Messaging Cutting Waste Increased Focus Swing Voters Support Mobilization(Likely True Also Applied In Practice Since Farage Personal Appearances Confirm Intentions To Rally Immigrant Concerned Groups)(Documents Quotes Above ) Deceptively Leave. EU Spread Harsh Misleading Anti Immigrant Content Via Social Media Including Fake Videos Billboards Comparing Migrants To Threat Imagery Provoking Fear Hatred Often Picking Up Media Attention Creating Viral Effects(Caesar Banks Shipman Channel4News Reports)(Social Media Ads Tailored Hard-To-Challenge Given Micro Target Hidden Nature ) Whistleblower Kaiser Notes Personality Traits Such As High Neuroticism Used To Trigger Fear Appeals Within Profiles Like “Left Behinds”(Kaiser’s Book Email Docs Referenced Above ) Public Polls Show Majority Oppose Deceptive Political Ads Micro Target Techniques Knight Foundation Gallup Surveys ORG You Gov Confirm Significant Unawareness Among Electorate Despite Growing Use Especially Dark Ads Still Less Than Half Aware Public Demands Greater Transparency Controls. US Google Twitter Attempt Limits But Facebook Continues Allow Advanced Micro Targeting Undermining Confidence Democracy(Citations Various Above Discuss Limitations Policies Changes Etc.) Discussion Follows Exploring Ethical Democratic Issues Around These Practices Their Legacy Post-Cambridge Analytica Collapse Current Regulatory Efforts Research Needs.
Discussion
CAmbridge Analyitca offered Leave EU campaign advanced profiling exhibiting both deception(covert unauthorized personal data use undeclared expenses misleading content propagation )and coercion(target audience analysis aiming covert behavior modification voter suppression strategies coordinated tailored hostile immigration messaging reinforced face-to-face outreach )features.(Regulatory fines lawsuits whistleblower testimonies confirm major breaches facts stated above.)While cause-effect direct vote shift impossible prove media effect challenges exist ongoing deployment raises fundamental issues about democratic integrity. Some progress seen platform policies removing explicit false ads restricting granularity except facebook still lags allowing exploitative targeting risking erosion public trust especially globally where protections weak. Regular independent multidisciplinary scrutiny urged given AI evolution complicating detection enforcement. Surveys confirm majority electorate dislike hidden behavioral steering subverting informed free choice essential democracy. Psychological profiling combined mass communication systems using optimized personalized messaging meets many criteria defining modern form psychooperations coerce manipulate unseen create knowledge asymmetries limit deliberation reflection. This threatens transparent shared decision-making foundational democratic process. Limitations include inability verify actual effectiveness despite documented attempts. Notwithstanding failures understanding intent scale important. The balance between efficient communication persuasion legitimate engagement versus manipulatory opaque restrictive practices demands close watch awareness education reform. Highlighting technological practice gaps socio-political consequences necessary starting point future improvements. Sharing knowledge internationally critical prevent worsening exploitation futures. Digital politics transforming fast requiring ethical frameworks evolved accordingly. Microtargeted deceit remains pressing challenge calling stronger legal controls transparency informed citizenship commitment open accountable governance. Technology won’t solve alone multisector collaboration essential. In sum ca brexit episode shows potential harms unchecked sophisticated psy ops integrated bigdata psychoanalytics must address urgently.
Source link
