Applying generalized beliefs about a group to individuals. The brain's compression algorithm — efficient but often wrong. In hiring: "Stanford grads are smart" functions as a useful prior until it overrides case-specific evidence. The candidate in front of you may be the exception. The stereotype doesn't care. Kahneman located stereotyping in System 1: fast, automatic, pattern-matching. Overcoming it requires System 2 — deliberate, evidence-seeking, slow. The tension is structural: stereotypes can be statistically accurate at the group level and unjust at the individual level. Base rates inform; they should not decide. The question is always whether you're using the base rate as a prior and updating with evidence, or letting the prior override the evidence entirely.
Amazon's Bar Raiser programme and structured interviews combat stereotyping by forcing evidence. Every candidate answers the same questions. Every interviewer scores against a rubric. The category — school, background, appearance — loses its monopoly on the verdict. The individual's actual performance enters the evaluation. The process feels mechanical. The outcomes are less biased. The mechanism is simple: structure slows the evaluation enough to move it from System 1 to System 2.
Bertrand and Mullainathan demonstrated the cost empirically in 2004. They sent identical resumes to employers, varying only the names — half received stereotypically white names (Emily, Greg), half received stereotypically Black names (Lakisha, Jamal). Resumes with white-sounding names received 50% more callbacks. The qualifications were identical. The compression algorithm — the stereotype activated by the name — changed the outcome before a single line of experience was read. The hiring managers did not believe they were discriminating. The stereotype operated upstream of intention.
The countermeasures exist and they work. Structured interviews — where every candidate answers the same questions in the same order, scored against a rubric — reduce the stereotype's influence by forcing evaluation on standardised criteria rather than gut-feel pattern matching. Blind resume review — stripping names, schools, and photos — removes the categorical triggers that activate the compression. Forced devil's advocates — assigning someone to argue against the prevailing assessment — surface the assumptions the stereotype installed silently. The interventions share a common design principle: they slow the evaluation process just enough to move it from System 1 (fast, automatic, stereotype-driven) to System 2 (deliberate, analytical, criteria-driven). The stereotype doesn't disappear. It loses its monopoly on the decision.
Section 2
How to See It
Stereotyping operates whenever a decision about an individual is driven by group-level beliefs rather than individual evidence. The evaluation feels like judgment. It is compression.
You're seeing stereotyping when a decision changes based on which category the individual is assigned to — and the decision-maker cannot articulate what individual evidence justified the different treatment.
Venture Capital
You're seeing stereotyping when investors fund founders who "look the part" and pass on founders who don't — while citing pattern-matching as due diligence. The archetype: technical, young, elite university. A 45-year-old founder with deep domain expertise receives scepticism that a 24-year-old Stanford dropout receives enthusiasm — regardless of traction or unit economics. The VC is matching the founder to a template. The template was built from a biased training set. The evidence is being ignored.
Hiring & Talent
You're seeing stereotyping when "culture fit" assessments produce teams that look, sound, and think alike. The hiring panel reports a vague feeling that the person "wouldn't fit." Pressed for specifics, they cite communication style or "vibe" — proxies for familiarity that the stereotype supplies. The candidate with the non-traditional background triggers a mismatch against the template of "people who succeed here." The template was built from the current team. The loop compounds.
Product Design
You're seeing stereotyping when user personas encode demographic assumptions rather than behavioural data. "Sarah, 32, marketing manager, shops at Whole Foods" — the product team designs for Sarah. Real users are messier. The persona compresses the user base into a template that feels precise and is actually reductive.
Executive Leadership
You're seeing stereotyping when a leader assigns projects based on demographic assumptions about capability rather than demonstrated skill. The technical project goes to the engineer with the CS degree, not the self-taught developer with a stronger track record. The leader believes they are making rational assignments. They are applying stereotypes about which categories of people perform which categories of work.
Section 3
How to Use It
Understanding stereotyping means accepting that you use stereotypes — automatically, constantly — and then building systems that catch the compression before it becomes the decision.
Decision filter
"Before any evaluation of a person — hiring, investing, promoting — ask: what category did my brain assign this person to, and what attributes did the category supply? If the attributes are doing the work instead of the evidence, the stereotype is making the decision."
As a founder
Your hiring process is a stereotype-management system whether you design it that way or not. Unstructured interviews are stereotype amplifiers. Structured interviews are stereotype suppressors — they force evaluation on predetermined criteria defined before the candidate walked in. Build the rubric before you meet the candidate. Score against the rubric. Compare scores across candidates. The process feels mechanical. The outcomes are dramatically less biased.
As an investor
Audit your portfolio for homogeneity that exceeds what deal flow would produce. If 90% of your founders are male, from three universities, and under 35, the pattern is not the market — it's the compression algorithm. The structural intervention: blind the initial screening. Evaluate the business — market size, traction, unit economics — before you know who the founder is.
As a decision-maker
Every performance review system that relies on manager judgment is vulnerable to stereotype contamination. Research consistently shows that identical work receives different evaluations depending on the perceived category of the person who produced it. Women receive more criticism about communication style. Minorities receive less benefit-of-the-doubt on ambiguous performance. The fix is structural: calibrate ratings across managers using standardised rubrics, require specific behavioural examples for every rating, and run demographic audits on performance distributions. If one category of employee consistently receives lower ratings despite comparable output metrics, the review system is measuring the stereotype, not the performance.
Common misapplication: Believing that awareness eliminates the bias. It doesn't. The implicit association test demonstrated that people who explicitly reject stereotypes still exhibit automatic stereotype activation. Structural interventions — blind review, structured evaluation, forced criteria — reduce it substantially. The first approach fights the compression with willpower. The second removes the inputs that trigger the compression.
Second misapplication: Using stereotyping awareness as a weapon rather than a diagnostic. The question is not whether stereotypes were present — they are always present. The question is whether the stereotypes determined the outcome. The diagnostic: would the decision have been different if the individual's category membership were different but their evidence were identical? If yes, the stereotype drove the decision. If no, the decision was evidence-based despite the stereotype's presence. The test is counterfactual, not accusatory.
Bezos embedded anti-stereotype mechanisms into Amazon's hiring architecture through the Bar Raiser programme. Every hiring loop includes a Bar Raiser — a trained interviewer from outside the hiring team whose explicit job is to evaluate whether the candidate meets Amazon's hiring bar independent of the team's needs or biases. The Bar Raiser has veto power. The design addresses stereotyping at the structural level: the hiring manager's pattern-matching is checked by an outsider who evaluates against standardised criteria (the 14 Leadership Principles) rather than team-specific templates. Bezos also mandated written interview feedback submitted before the debrief — preventing the senior interviewer's assessment from anchoring the room and allowing stereotypes to cascade through the group.
Hastings built Netflix's talent philosophy around evaluating the person in front of you, not the archetype they represent. The "keeper test" — would I fight to keep this person if they told me they were leaving? — is an individuation exercise. It forces the manager to evaluate the specific person's specific contribution rather than their match to a category template. Netflix pays top-of-market for each individual based on what they would command externally, rather than fitting people into bands that implicitly stereotype based on title, tenure, or background. The "adequate performance gets a generous severance" policy disrupts stereotype inertia — it prevents the accumulation of team members who were hired because they matched a template rather than because they were exceptional. The result is a team that is cognitively diverse not because Netflix mandates diversity but because the evaluation system strips away the categorical shortcuts that produce homogeneity.
Section 6
Visual Explanation
The top panel traces the compression pipeline. Raw input is categorised, a template is applied, and a verdict is delivered — all before individual evidence enters the evaluation. The individual's actual attributes are discarded at the categorisation step, replaced by the category's default attributes. The middle panel shows structural countermeasures: blind review removes the triggers that activate the compression, structured criteria force individuation by requiring evidence-based evaluation, and the Bar Raiser provides an external check that the hiring manager's pattern-matching cannot override. The bottom panel captures the core tension: base rates can be accurate at the group level but unjust at the individual level. The prior must be updated with evidence. The Bayesian discipline: start with the base rate, update with case-specific evidence. Stereotyping is what happens when the update step is skipped.
Section 7
Connected Models
Reinforces
Base Rate Fallacy
The base rate fallacy is the tendency to ignore statistical base rates when evaluating individual cases. Stereotyping is the inverse: over-relying on group-level base rates when individual evidence is available. The tension: base rates are useful priors when they inform the evaluation. They become harmful when they override case-specific evidence. The discipline is Bayesian — start with the prior, update with evidence. Stereotyping is what happens when the update step is skipped. "Stanford grads are smart" is a valid prior. The error is when the prior becomes the verdict without updating with the individual's actual performance in the interview.
Reinforces
Representativeness Heuristic
The representativeness heuristic is the specific mechanism through which many stereotypes operate. Kahneman and Tversky showed that people judge probability by resemblance — how closely does this person resemble my prototype of the category? A founder who "looks like" a successful founder is judged as more likely to succeed, regardless of base rates. Stereotyping provides the prototype. Representativeness provides the judgment mechanism.
Reinforces
Confirmation Bias
Confirmation bias protects stereotypes from disconfirming evidence. Once a category is assigned, the mind selectively attends to information that confirms the stereotype and discounts information that contradicts it. The stereotype becomes self-reinforcing — each "confirming" instance strengthens the pattern, while disconfirming instances are explained away or forgotten.
Reinforces
Section 8
One Key Quote
"The human mind must think with the aid of categories. Once formed, categories are the basis for normal prejudgment. We cannot possibly avoid this process. Orderly living depends on it."
— Gordon Allport, The Nature of Prejudice (1954)
Allport's observation denies the comforting narrative that stereotyping is something bad people do. It is something all people do — because the brain's architecture requires it. The operational consequence is that anti-stereotype training — which asks people to suppress a process Allport called unavoidable — produces limited, short-lived effects. The research on implicit bias training consistently shows that awareness-based interventions change attitudes temporarily but do not change behaviour durably. The interventions that work are structural: they don't ask the brain to stop categorising. They remove the category information from the decision environment (blind review), force the brain to individuate by requiring specific evidence (structured criteria), or introduce external checks that catch the stereotype's output before it becomes the decision (calibration, Bar Raisers, devil's advocates). You cannot stop the compression. You can audit the output.
Section 9
Analyst's Take
Faster Than Normal — Editorial View
Stereotyping is not a bug in human cognition. It is the core compression algorithm. The brain cannot process the world without shortcuts. Stereotypes are the shortcuts. They categorise, compress, and predict — faster than any deliberate analysis could. The problem is not that stereotypes exist. The problem is that the compression is applied to people, where the information discarded — individual capability, individual context — is the information decisions should be based on.
The venture capital industry is the most expensive demonstration of stereotyping in the modern economy. VCs fund pattern matches. The pattern was built from a biased training set and reinforced by survivorship bias. The founders who matched the pattern and failed are invisible. The founders who didn't match and would have succeeded were never funded. The opportunity cost is incalculable.
The "Stanford grads are smart" prior is useful until it overrides evidence. Base rates can inform. The question is whether you're updating with evidence or letting the prior decide. When the evaluation is driven by the category rather than the evidence in front of you, the stereotype is making the decision. Kahneman is right: stereotyping is System 1. Overcoming it requires System 2 — and System 2 requires structure.
The "culture fit" concept is stereotyping's most successful rebrand. When a hiring manager says a candidate "isn't a culture fit," they are typically saying the candidate doesn't match the template of the existing team. "Culture fit" sounds like a value judgment. It operates as a stereotype filter. The companies that have replaced "culture fit" with "culture add" have discovered that the discomfort of cognitive diversity produces better decisions.
The only reliable countermeasure is structural, not psychological. Implicit bias training does not work — not because the training is bad, but because it targets the wrong level. Stereotypes operate at the automatic processing level, below conscious control. Structural interventions — blind review, structured interviews, Bar Raiser programmes — work because they don't ask the brain to stop doing what Allport proved it cannot stop doing. They change the decision environment so the stereotype's output is checked before it becomes the decision.
Every unstructured evaluation is a stereotype delivery mechanism. Every time you ask "what's your gut feeling about this candidate?" you are asking for the stereotype's verdict. The gut feeling is the compression algorithm's output — fast, confident, and shaped by category rather than evidence. The question is not whether your gut uses stereotypes. It does. The question is whether you build systems that catch the gut's output before it becomes the hire, the investment, or the promotion. The organisations that do make better decisions. Not because their people are less biased. Because their processes are.
Section 10
Test Yourself
Pattern matching or stereotyping?
Scenario 1
A venture capital associate reviews two pitch decks. Founder A is a 28-year-old Stanford CS graduate with one year at Google and no prior startup experience. Founder B is a 42-year-old community college graduate with fifteen years in logistics and two previously profitable exits. The associate rates Founder A as '8/10 investable' and Founder B as '5/10 investable.' When asked to justify, the associate cites Founder A's 'technical depth' and Founder B's 'limited network.'
Scenario 2
A product team designs a fintech app for 'young professionals aged 25-35 who are comfortable with technology.' Six months after launch, their fastest-growing user segment is adults aged 50-65 switching from traditional banking — but the design creates significant friction for this segment.
Scenario 3
A hiring committee uses a structured interview with a standardised rubric. Candidate A has MIT, five years at a FAANG. Candidate B has no degree, twelve years of open-source contributions including maintainership of a widely-used library. Both complete identical interview questions. Candidate B scores higher on three of four rubric dimensions. The committee hires Candidate B.
The foundational text. Allport established that stereotyping is a cognitive inevitability — a consequence of the brain's need to categorise — rather than a moral failing. His framework distinguishes between categorisation (necessary), prejudgment (automatic), and prejudice (categorisation hardened into hostility).
The field experiment that quantified stereotyping's impact on hiring. Identical resumes with different names revealed a 50% callback gap. The study's design isolated the stereotype's effect with surgical precision.
Kahneman's dual-process framework explains why stereotyping resists conscious override. Stereotypes operate through System 1 — fast, automatic, and below conscious control. System 2 can override System 1 but only when engaged and resourced.
Neuroscience research identifies the neural dissociation between stereotyping and prejudice. The practical implication: structural approaches that remove the inputs triggering categorisation are more effective than approaches that target the evaluative response after categorisation has already occurred.
Bock's account of Google's hiring transformation provides the most operationally detailed case study of structured anti-stereotyping interventions at scale. Google's shift from unstructured interviews to structured interviews with standardised questions, rubrics, and calibration committees reduced stereotype influence while improving predictive validity. The book documents the data that drove the change and the resistance from hiring managers who preferred their "gut feel" — which was, as the data showed, the stereotype's output wearing intuition as a disguise.
Stereotyping — the brain's compression algorithm matches individuals to categories before conscious evaluation begins. The category supplies the prediction. The prediction substitutes for evidence.
Halo Effect
The halo effect is stereotyping applied to a single positive attribute. A candidate with an elite university on their resume carries a halo that colours every subsequent evaluation — communication, analytical reasoning, leadership potential. The halo generalises a single salient signal across dimensions. The mechanism is the same: category-based processing substituting for individuated evaluation.
Leads-to
Golem Effect
The Golem effect is the self-fulfilling prophecy of low expectations. When a person is stereotyped — assigned to a category that carries negative attributes — the evaluator's expectations drop. The person receives less opportunity, less feedback, less investment. Their performance declines to match the expectation. The stereotype created the outcome it predicted.
Reinforces
Availability Heuristic
The availability heuristic makes stereotyping stickier: the most vivid examples of a category dominate the mental template. If the most memorable "successful founder" in an investor's experience was young, male, and technical, that template becomes the default. The template is built from availability, not from base rates. It is then applied with the same automatic force as any stereotype.
Scenario 4
A hiring manager interviews two candidates. Candidate A attended Stanford and worked at Google. Candidate B attended a state university and worked at a mid-size company. Both complete an identical case study. Candidate B's solution is analytically superior. The manager rates Candidate A higher on 'strategic thinking' and rates them equal on the case study.
