Read the original article here.

---

Digital Manipulation: How They Use Data to Control You

Understanding Confirmation Bias: A Core Vulnerability

In an age dominated by digital information, understanding human psychology is crucial to recognizing how data can be used to influence and manipulate us. One of the most significant cognitive biases exploited in the digital realm is confirmation bias.

Confirmation Bias (also Confirmatory Bias, Myside Bias, or Congeniality Bias): The tendency to search for, interpret, favor, and recall information in a way that confirms or supports one's prior beliefs or values.

Essentially, people tend to seek out information that reinforces what they already believe, ignore or downplay evidence that contradicts their views, and interpret ambiguous information in a way that fits their existing perspective. This bias is particularly strong for emotionally charged issues, deeply entrenched beliefs, and when a desired outcome is involved.

This inherent human tendency makes individuals susceptible to manipulation, especially in digital environments where data about their beliefs and preferences is readily available and can be used to curate their information diet.

How Confirmation Bias Manifests: The Psychological Mechanisms

Confirmation bias is not a single phenomenon but a result of biased processes occurring during information processing. These include:

1. Biased Search for Information

People don't neutrally search for evidence; they often frame their searches and questions to confirm their existing hypotheses.

Positive Test Strategy: A heuristic (mental shortcut) where individuals test a hypothesis by examining cases where they expect a property or event to occur, rather than actively seeking out cases that would disprove it (falsification).

While positive tests can sometimes be informative, relying solely on this strategy, especially in complex real-world situations, can lead to overlooking contradictory evidence.

Examples and Explanations:

• Wason's Rule Discovery Task: In early experiments, participants were given a sequence (e.g., 2, 4, 6) and told it fit a rule. They proposed triples to test the rule. Instead of testing triples that would break their hypothesized rule (e.g., testing (3, 5, 7) if they thought the rule was "add 2 each time" vs. testing (1, 2, 3) which would falsify that), they often tested triples that fit their hypothesis (e.g., (8, 10, 12)). They sought confirmation rather than falsification.
• Introvert/Extrovert Interview Study: Participants interviewing someone believed to be an introvert chose questions like, "What do you find unpleasant about noisy parties?" (presumes introversion). If the person was thought to be an extrovert, questions like, "What would you do to liven up a dull party?" were preferred (presumes extroversion). These questions made it hard for the interviewee to demonstrate characteristics contrary to the hypothesis.
• Digital Relevance: Search engines and social media algorithms learn about your existing beliefs and interests based on your past searches, clicks, likes, and shares. They then prioritize content (articles, videos, posts) that is likely to agree with those beliefs, effectively engaging in a biased search on your behalf and limiting your exposure to dissenting views. If you frequently search for information supporting a particular political viewpoint, the algorithm will show you more of that, reinforcing your "positive test strategy" by providing an endless stream of "confirming" results.

This biased search is amplified by selective exposure, where individuals actively choose to consume information sources that align with their personal beliefs (e.g., reading news outlets known for a specific political slant).

2. Biased Interpretation of Information

Even when presented with the same information, individuals prone to confirmation bias will interpret it in a way that supports their pre-existing views.

Disconfirmation Bias: The tendency to set higher standards of evidence for hypotheses that go against one's current expectations compared to those that align with them.

Examples and Explanations:

• Capital Punishment Study: Participants strongly for or against capital punishment read descriptions of two fictional studies with mixed results on its deterrent effect. Regardless of the study's presented conclusion, participants rated studies supporting their existing view as more well-conducted and convincing than those that contradicted it. They found flaws or ambiguities in the contradictory studies while accepting the supportive ones at face value.
• Political Statement Interpretation: During the 2004 US election, participants evaluated contradictory statements from political figures. They were much more likely to interpret statements from the candidate they opposed as genuinely contradictory and hypocritical compared to statements from their favored candidate, for whom they found ways to rationalize the apparent inconsistencies. Brain scans during this experiment showed activity in emotional centers when participants evaluated contradictory statements by their favored candidate, suggesting an active process of reducing cognitive dissonance rather than neutral evaluation.
• Digital Relevance: When you encounter a news article or social media post, your existing beliefs influence how you evaluate its credibility and meaning. If the headline confirms your view, you might accept it with less scrutiny. If it challenges your view, you are more likely to look for reasons to dismiss it (e.g., "This source is biased," "The data is flawed"). Digital platforms, by feeding you content confirming your biases, train you to be less critical of such information, while content that challenges your views might be immediately flagged as "fake" or "biased" simply because it doesn't align. Data about your past interactions (comments, shares, reactions) is used to predict and reinforce these interpretive biases.

3. Biased Recall of Information

People may selectively remember evidence that reinforces their expectations, forgetting or downplaying information that contradicts them.

Selective Recall (Confirmatory Memory or Access-Biased Memory): The tendency to remember evidence selectively to reinforce one's existing beliefs or expectations.

Psychological theories differ on whether this is because matching information is easier to store/recall (schema theory) or because surprising information is more memorable. Both have some experimental support.

Examples and Explanations:

• Librarian vs. Salesperson Study: Participants read a description of a woman with both introverted and extroverted traits. When told she was being considered for a librarian job (stereotypically introverted), they recalled more examples of her introversion. When told it was for a sales job (stereotypically extroverted), they recalled more extroverted behaviors.
• O.J. Simpson Trial Recall: Participants who reported their feelings about the verdict over time found that their current opinions influenced how they recalled their past emotional reactions and certainty about his guilt or innocence. Memories of past feelings seemed to be reconstructed based on current beliefs.
• ESP Belief Study: Believers in extrasensory perception (ESP) who read evidence not supporting ESP recalled significantly less information than disbelievers or believers who read supporting evidence. Some even incorrectly remembered the non-supportive results as actually supporting ESP.
• Digital Relevance: After scrolling through a personalized news feed filled with posts confirming your political views, you are more likely to remember the "facts" and anecdotes that support that view and forget or discount the information that challenged it, even if you briefly saw it. Data about your engagement (how long you looked at something, whether you clicked or scrolled past) can inform algorithms about which information is likely to stick and reinforce your biases, further tailoring future content.

Why Does Confirmation Bias Occur? Explanations

Various theories attempt to explain why we are susceptible to confirmation bias:

• Cognitive Limitations: Our brains have limited capacity to process complex information. Heuristics (mental shortcuts) are used, including the positive test strategy, which are generally efficient but can lead to bias, especially when dealing with broad possibilities. It's difficult for the mind to hold and test multiple conflicting hypotheses simultaneously.
• Motivational Factors: We prefer positive thoughts and strive for consistency in our beliefs and attitudes. This desire can make us demand less evidence for preferred conclusions and more evidence for undesirable ones ("Can I believe this?" vs. "Must I believe this?"). Wishful thinking plays a role.
• Cost-Benefit Analysis: From an evolutionary or pragmatic perspective, it might sometimes be more "rational" to avoid costly errors rather than seeking objective truth. For instance, in social interactions, asking questions that confirm a perceived personality (e.g., "Do you feel awkward in social situations?" to someone who seems introverted) can come across as more empathic, even if it reinforces a potentially incorrect assumption.
• Exploratory vs. Confirmatory Thought: Most people default to "confirmatory thought," seeking to justify a specific viewpoint, rather than "exploratory thought," which neutrally considers multiple perspectives. We are more likely to engage in critical, exploratory thinking only when we anticipate needing to justify ourselves to a well-informed, neutral audience whose views are unknown – a rare scenario.
• Make-Believe: Some theories suggest roots in childhood coping mechanisms where make-believe can evolve into adult self-deception and illusion, making rationalization of false beliefs habitual and unconscious.
• Optimal Information Acquisition (Economic View): Under conditions of limited time and costly information, seeking confirmatory evidence can be seen as a rational, optimal strategy. If you strongly believe something, seeking information likely to confirm it might build confidence more efficiently. Data manipulators exploit this by making "confirmatory" information readily available and "disconfirming" information harder to find or less appealing, altering the perceived "cost" of seeking diverse information.

Confirmation Bias in the Digital Age: Data, Algorithms, and Control

The digital environment, particularly social media and personalized content platforms, acts as a powerful amplifier of confirmation bias. This is where the connection to "How They Use Data to Control You" becomes most explicit.

Filter Bubbles and Algorithmic Editing

Filter Bubble: An intellectual isolation that can occur when websites use algorithms to selectively guess what information a user would like to see based on information about the user (such as location, past clicks, and search history) and, as a result, users become separated from information that disagrees with their viewpoints, effectively isolating them in their own cultural or ideological bubbles. Algorithmic Editing: The curation of information presented to individuals by algorithms, often based on their perceived preferences and biases, leading to the selection and prioritization of certain content while excluding others.

Social media platforms, news aggregators, and search engines use vast amounts of data collected about users – their browsing history, location, demographics, expressed opinions, who they follow, what they like, share, and comment on – to build detailed profiles. These profiles are then used by algorithms to predict what content a user will find engaging. Because people are naturally drawn to information that confirms their biases, algorithms are optimized to feed this preference.

How Data Enables Manipulation:

1. Data Collection: Every interaction online leaves a data trail. This data paints a picture of your beliefs, values, interests, and likely biases.
2. Profile Building: Data is aggregated and analyzed to create detailed user profiles, identifying your tendencies towards specific viewpoints (political, social, consumer, etc.).
3. Algorithmic Targeting: Algorithms use these profiles to select content (news, posts, ads, videos) that is most likely to align with your existing biases. Content that challenges your views is less likely to be shown.
4. Reinforcement: Repeated exposure to confirming information strengthens existing beliefs and makes them more resistant to change. This creates echo chambers, where individuals primarily hear and interact with like-minded people and information.
5. Exploitation: This creates opportunities for manipulators (advertisers, political campaigns, state actors) to target specific groups with tailored messages designed to resonate with and exploit their confirmation biases for various goals (selling products, influencing votes, spreading propaganda, inciting division). Data allows them to identify exactly who is susceptible to what kind of message.

This algorithmic filtering means individuals are increasingly exposed only to information they are likely to agree with, while opposing views are excluded. This isn't just a passive reflection of preference; it's an active process driven by data and designed to maximize engagement (clicks, shares, time spent on the platform), which in turn reinforces the platform's business model.

Confirmation Bias and Fake News

The rapid spread of fake news in the digital age is heavily reliant on confirmation bias. False or misleading information is often crafted to appeal directly to existing beliefs and prejudices within specific groups.

Fake News: False or misleading information presented as credible news from a seemingly reliable source.

Confirmation bias makes people less likely to critically evaluate information that confirms what they already suspect or believe, especially when it comes from a source they perceive (or is presented algorithmically as) trustworthy or aligned with their group. The combination of a biased search (algorithm feeds confirming info), biased interpretation (info aligning with bias is judged as true), and biased recall (confirming info is better remembered) creates a fertile ground for misinformation to spread and persist.

Data enables the precise targeting of fake news narratives to individuals most likely to believe and share them based on their demonstrated biases.

Digital Nudging as a Potential Countermeasure (or another form of influence?)

In response to the harms of filter bubbles and fake news, some social media sites are exploring "digital nudging":

• Nudging of Information: Providing disclaimers or labels questioning the validity of a source.
• Nudging of Presentation: Exposing users to information or viewpoints they might not have sought out, potentially challenging their biases.

While intended to combat confirmation bias, these nudges are themselves interventions based on data and algorithmic decision-making, highlighting the pervasive role of data in shaping the online information landscape, whether for good or ill.

Real-World Impacts of Confirmation Bias (Amplified Digitally)

While confirmation bias exists outside the digital world, the scale and speed of digital communication, coupled with data-driven personalization, significantly amplify its effects in various domains:

• Politics: Leads to extreme attitude polarization, difficulty in finding common ground, and susceptibility to politically targeted disinformation campaigns exploiting pre-existing partisan biases. Data is used to identify and target voters with highly specific, often emotionally charged messages that confirm their existing political identity and biases.
• Science and Research: Can lead scientists to favor data that supports their hypotheses and be overly critical of contradictory evidence, potentially slowing down scientific progress or perpetuating flawed theories (e.g., the file drawer effect, where non-confirming results aren't published). Digital access to vast amounts of pre-filtered or curated research can make it easier for researchers (or those citing research) to selectively find data confirming their views.
• Finance: Causes investors to become overconfident, seeking out news or analysis that confirms their investment choices while ignoring warning signs. Financial news feeds can be personalized based on investment positions, reinforcing this bias.
• Medicine and Health: Can lead doctors to prematurely diagnose based on initial impressions and seek only confirming evidence. Patients, influenced by information confirming their self-diagnosis (often found online), may pressure doctors for specific tests or treatments. Misinformation about health (e.g., vaccines) spreads rapidly online, exploiting confirmation bias, and data can be used to target individuals based on their health concerns or beliefs.
• Law and Policing: Investigators may focus on confirming evidence once a suspect is identified, neglecting potentially exonerating information. In court, jurors may interpret complex evidence through the lens of early opinions formed.
• Social Psychology: Influences self-verification (reinforcing existing self-image by favoring confirming feedback) and self-enhancement (seeking positive feedback). Social media interactions provide constant feedback loops, and algorithms prioritize content that elicits positive engagement, potentially reinforcing existing self-perceptions or desires, whether accurate or not.
• Mass Delusions and Paranormal Beliefs: Confirmation bias helps perpetuate mass delusions (like witch trials or mass hysteria) and paranormal beliefs (like ESP or numerology). Individuals selectively seek and interpret information to fit the narrative, ignoring overwhelming contradictory evidence. The internet facilitates rapid spread of anecdotes and selective "evidence" that fuels these beliefs, often within dedicated online communities that reinforce each other's biases.
• Recruitment: Interviewers may favor candidates who confirm their initial positive or negative impressions, overlooking objective qualifications, contributing to lack of diversity.

Associated Effects and Outcomes of Confirmation Bias

Confirmation bias is linked to several other cognitive phenomena that are pertinent to understanding digital influence:

• Attitude Polarization: When people with opposing views are exposed to the same mixed evidence, they often become more extreme in their original views rather than finding common ground. This is fueled by biased interpretation and selective recall. Digital echo chambers exacerbate this by presenting highly curated, biased information streams to opposing groups, pushing them further apart.
• Persistence of Discredited Beliefs (Belief Perseverance): Beliefs can persist even after the original evidence for them has been shown to be false or retracted. Individuals may trust their initial interpretations or recollections over subsequent corrections. This explains why debunking misinformation is so difficult; the false belief, once confirmed, is sticky. Data-driven repetition of false narratives can cement them before corrections have a chance to take hold.
• Preference for Early Information (Irrational Primacy Effect): Information encountered early in a sequence is often given more weight, influencing the interpretation of subsequent information. In digital feeds, the first few pieces of information seen can set a frame that influences how everything else is perceived.
• Illusory Association (Illusory Correlation): The tendency to perceive a correlation between two events or situations when none exists or it is much weaker than believed. This often happens by focusing only on instances where both events occur together and neglecting contradictory cases. Online communities can collectively reinforce illusory correlations through shared anecdotes and selective focus, often amplified by algorithms that group like-minded people.

Historical Recognition

It is important to note that confirmation bias is not a new phenomenon unique to the digital age. Philosophers and thinkers throughout history, from Thucydides and Francis Bacon to Dante and Tolstoy, observed this human tendency to favor information aligning with one's existing views. However, the advent of data collection and algorithmic filtering provides tools that allow this ancient bias to be exploited on an unprecedented scale.

Understanding Discovery and Explanations

Early psychological experiments by Peter Wason in the 1960s, particularly the rule discovery task, highlighted the tendency to seek confirmation rather than falsification. While initially interpreted as a direct confirmation bias, later work by Klayman and Ha suggested the "positive test strategy" as a more nuanced explanation – a heuristic that is sometimes informative but prone to leading to bias. This evolution in understanding underscores the complexity of human reasoning and the subtle ways biases manifest.

Individual Differences

While a universal human trait, the degree of confirmation bias varies between individuals. It doesn't necessarily correlate with intelligence but is more linked to "active open-mindedness" – the willingness to actively seek out reasons why one's initial ideas might be wrong. Individuals who value objective, fact-based thinking might paradoxically exhibit stronger myside bias if their definition of a "good argument" is simply one supported by facts (selectively gathered facts, due to the bias). This suggests that training in balanced argumentation and critical evaluation can potentially mitigate the bias.

Conclusion: Navigating the Data-Driven Landscape

Confirmation bias is a fundamental cognitive tendency that influences how we search for, interpret, and remember information. In the digital age, the pervasive collection and analysis of data enable platforms and actors to create highly personalized information environments. By identifying and exploiting our pre-existing beliefs and biases, algorithms curate content that primarily confirms our views, creating filter bubbles and echo chambers. This makes us more susceptible to misinformation, polarizes opinions, and limits our exposure to diverse perspectives.

Recognizing confirmation bias in ourselves and understanding how data is used to feed it is the first step in navigating the digital landscape more critically. Actively seeking out diverse sources, questioning information that strongly aligns with our views, and being aware of the emotional pull of confirmation are essential skills in an era where our data is constantly being used to predict and influence our thoughts and behaviors.