slug: bias-in-high-stakes-science format: portrait dimensions: 2400x3200px color_palette: background: "#0D1117" # deep navy/charcoal primary_accent: "#E8C547" # warning amber secondary_accent: "#4FC3F7" # cool blue (science/data) danger_accent: "#EF5350" # red for strike markers text_primary: "#F0F0F0" text_secondary: "#A0AEC0" section_bg: "#1A2332" # slightly lighter panel zones: - id: header position: top, full-width height: 12% elements: - type: title text: "When Brilliant People Believe Wrong Things" style: bold, 72px, primary_accent source_id: non-factual (design element) - type: subtitle text: "Cognitive Bias in High-Stakes Science" style: 44px, text_primary source_id: non-factual (design element) - type: tagline text: "A case study from quantum computing research, 2018–2026" style: 28px italic, text_secondary source_id: microsoft_strike_count - id: central_brain_illustration position: upper-center height: 18% elements: - type: illustration_direction text: > Split-brain silhouette (left = rational, right = emotionally-biased). Left side: clean grid of data points, a balance scale. Right side: same data but filtered through a distorting lens; mismatched pieces. Amber glow on the right brain indicating heat/pressure. Do not include text in this zone. source_id: non-factual (illustrative metaphor) - type: caption text: "The brain under competitive pressure processes the same evidence differently" style: 24px italic, text_secondary, centered below illustration source_id: non-factual (synthesis caption) - id: bias_grid position: middle layout: 2x2 grid height: 38% cells: - id: sunk_cost position: top-left icon: hourglass sinking into coins header: text: "THE SUNK COST TRAP" style: bold, 32px, primary_accent body: - text: > Once large investments are made, abandoning them feels like accepting a loss — and losses feel ~2× more painful than equivalent gains feel rewarding. source_id: loss_aversion_ratio - text: "(Kahneman & Tversky, Econometrica, 1979)" style: 20px, text_secondary, italic case_anchor: text: "Case: ~16 yrs · est. $1B+ in topological qubits" style: 22px, secondary_accent source_id: sunk_cost_escalation biology_note: text: "Loss aversion is encoded in the amygdala — it fires harder for potential losses than equivalent gains" style: 20px, text_secondary, italic - id: confirmation_bias position: top-right icon: funnel/filter showing only matching items passing through header: text: "CONFIRMATION BIAS" style: bold, 32px, primary_accent body: - text: > "The seeking or interpreting of evidence in ways partial to existing beliefs." Contradictory data is not suppressed deliberately — it is simply perceived as noise. source_id: confirmation_bias_definition - text: "(Nickerson, Review of General Psychology, 1998)" style: 20px, text_secondary, italic case_anchor: text: "Case: Favorable data presented; contradictory data omitted" style: 22px, secondary_accent source_id: confirmation_bias_definition biology_note: text: "The brain's pattern-matching systems prioritize confirming signals — a feature, not a bug, until the stakes are high" style: 20px, text_secondary, italic - id: motivated_reasoning position: bottom-left icon: trophy / competitor arrow warping a data graph header: text: "MOTIVATED REASONING" style: bold, 32px, primary_accent body: - text: > Competitive goals shape which results researchers pursue and how they interpret ambiguous findings — while the reasoner maintains an illusion of objectivity. source_id: motivated_reasoning - text: "(Kunda, Psychological Bulletin, 1990)" style: 20px, text_secondary, italic case_anchor: text: "Case: IBM & Google advancing → intense pressure to show progress" style: 22px, secondary_accent source_id: motivated_reasoning biology_note: text: "Dopamine reward pathways reinforce finding what we expect to find — making motivated perception feel like honest discovery" style: 20px, text_secondary, italic - id: prestige_as_proof position: bottom-right icon: journal cover with star/badge, magnifying glass misreading it header: text: "PRESTIGE AS PROOF" style: bold, 32px, primary_accent body: - text: > Institutions use top-journal publication as a signal that results are validated. But peer review is a filter, not a guarantee — and >80% of published papers report positive results. source_id: publication_positive_rate - text: "(Fanelli, PLOS ONE, 2010)" style: 20px, text_secondary, italic case_anchor: text: "Case: Nature publication cited as proof; coding errors not caught" style: 22px, secondary_accent source_id: publication_positive_rate biology_note: text: "Authority signals bypass critical evaluation — a cognitive shortcut that works until the authority is wrong" style: 20px, text_secondary, italic - id: timeline position: lower-middle height: 14% layout: horizontal timeline, 3 nodes header: text: "THREE STRIKES: THE MICROSOFT MAJORANA PATTERN" style: bold, 30px, primary_accent, centered nodes: - year: "2018" label: "RETRACTED" detail: "Majorana signatures paper withdrawn from Nature" color: danger_accent source_id: microsoft_strike_count - year: "2021" label: "RETRACTED" detail: "Quantized Majorana conductance paper retracted" color: danger_accent source_id: microsoft_strike_count - year: "2026" label: "CHALLENGED" detail: "Legg Nature critique: coding errors, flawed protocol, omitted data" color: primary_accent source_id: microsoft_strike_count connector: dashed amber line, left to right - id: footer_lesson position: bottom, full-width height: 10% background: primary_accent elements: - type: pull_quote text: > "Prestigious publication is a starting point for scientific evaluation — not an endpoint. The same biases that shape individual decisions operate at institutional scale." style: bold, 28px, dark (#0D1117), centered source_id: non-factual (synthesis) - type: source_line text: > Sources: Kahneman & Tversky (1979) · Arkes & Blumer (1985) · Nickerson (1998) · Kunda (1990) · Fanelli (2010) · Legg, Nature (2026) DOI: 10.1038/s41586-026-10567-8 style: 18px, text_secondary - id: course_credit position: bottom-right corner height: 3% text: "A Skeptic's Guide to Quantum Computing · dmccreary.github.io/quantum-computing" style: 16px, text_secondary non_factual_design_elements: - brain illustration (metaphor only) - section icons (sinking hourglass, funnel, trophy, journal badge) - color coding and background - connector lines and layout grid