Fifteen books kept off on purpose.

The ‘law of attraction’: thoughts emit frequencies that attract matching events, wealth, and health.

pseudoscience: no mechanism, no evidence, and the physics invoked is a misappropriation of quantum terminology. The New York Times Book Review called it pseudoscience and an ‘illusion of knowledge’; Scientific American (2007) rejected its account of brainwave activity. The claim is also unfalsifiable by construction: failure is always attributed to insufficient positivity or hidden negative thoughts, which is a textbook pseudoscience marker.

Illness, poverty, and disaster result from the sufferer’s own thoughts; disease can be cured by belief.

debunked and actively harmful: Mary Carmichael and Ben Radford (Committee for Skeptical Inquiry) named the flipside: if you get sick or hurt, it is your fault. The book contains explicit discouragement of medical attention and implies victims of mass tragedies attracted their fate. Barbara Ehrenreich’s Bright-Sided documents the harm this framing does to cancer patients.

Positive visualization improves outcomes.

partially inverted by the research: Gabriele Oettingen’s two decades of work shows that pure positive fantasizing about outcomes REDUCES energy and attainment; what works is mental contrasting plus implementation intentions (WOOP), which requires confronting obstacles, the opposite of the book’s instruction to never entertain negative thoughts.

The ‘rich dad’ was a real mentor whose lessons the book transmits.

almost certainly fabricated: Kiyosaki has never identified him, and after years of press questions has given shifting, evasive answers, at one point suggesting the figure was composite or fictional (‘is Harry Potter real?’). The book is marketed as memoir; its central character appears not to have existed.

Overall factual accuracy and financial competence.

extensively documented as false: real-estate author John T. Reed’s chapter-by-chapter analysis catalogues large numbers of factual errors, impossible timelines, and legally wrong statements, concluding the book contains ‘much wrong advice, much bad advice, some dangerous advice, and virtually no good advice.’ Kiyosaki has never substantiated the real-estate track record the book’s authority rests on.

Specific advice: your house is not an asset, debt is good, avoid saving, corporations let you deduct almost everything, invest in what you ‘know’.

dangerous where not merely wrong: the good-debt/leverage advice given without risk framing is how retail investors get wiped out; the tax claims are inaccurate as stated; the anti-diversification and anti-index-fund posture contradicts essentially all of finance. Personal-finance writers (the White Coat Investor among many) note the book gives motivation but no executable method. Deliberately so, since the method is sold separately.

The business behind the book.

documented: Learning Annex won a $23.7M judgment against Kiyosaki’s Rich Global LLC in 2012, which then filed Chapter 7 claiming $1.8M in assets despite $45M in seminar royalties between 2007 and 2010. The Rich Dad seminar chain has been the subject of repeated high-pressure-upsell complaints and investigative reporting. The book functions as the top of a funnel.

Residual value: ‘financial literacy matters, learn how money works, assets vs liabilities’.

true but generic: available from any competent source without the fabrication and the funnel.

ADHD is caused by early childhood attachment disruption and parental stress, not by heredity; it is ‘a way of coping’ with a wounded environment.

contradicted by the strongest evidence base in psychiatric genetics: twin and family studies place ADHD heritability at roughly 74–88% across the lifespan (Faraone & Larsson’s meta-analysis of 37 twin studies gives ~74%; Larsson et al. and subsequent work converge on 70–80%). Demontis et al.’s GWAS (2019, expanded 2023) identified dozens of genome-wide significant loci. ADHD is among the most heritable conditions in psychiatry, comparable to height.

Maté’s dismissal of twin studies as methodologically invalid.

rejected by the field: his rejection is assertion, not analysis. Adoption studies, sibling studies, and molecular-genetic (SNP-heritability, polygenic score) methods that do not rely on the equal-environments assumption converge on the same high heritability, which is precisely the triangulation his critique cannot survive.

Adverse childhood experience causes ADHD (inferred from the ACE-ADHD correlation).

causally reversed / confounded: the correlation is real but the arrow is contested. Children with ADHD elicit more chaotic environments; ADHD parents (highly likely, given heritability) generate more household instability; and the same genes load onto both. Scott Barry Kaufman (‘ADHD Isn’t a Trauma Response’, Psychology Today, Feb 2025) and Russell Barkley have both argued this publicly. A 2023 Conversation piece by University of Melbourne researchers reviewed Maté’s broader trauma-causes-everything claims (cancer, autoimmune disease, ADHD) and found the evidence does not support them.

Credentialing note.

context: Maté is a family physician, not a researcher; he has published no peer-reviewed research on ADHD etiology, and his theory rests on clinical intuition and self-observation. He is also not shy about extending the trauma explanation to cancer and autoimmune conditions, which is where the overreach becomes most visible.

Each person has one primary love language.

failed: Emily Impett, Haeyoung Gideon Park & Amy Muise (2024, Current Directions in Psychological Science) reviewed the literature and found people consistently endorse ALL five as meaningful ways of giving and receiving love. There is no evidence for a dominant single channel.

There are exactly five love languages.

unsupported: the taxonomy came from Chapman’s pastoral counseling notes, not from factor analysis or any systematic derivation. The five categories have no established discriminant validity and factor structures do not reliably reproduce them.

Couples are more satisfied when partners match / speak each other’s primary language.

failed replication: Bunt & Hazelwood (2017) tested the matching hypothesis directly and found no relationship-satisfaction benefit for matched couples. Impett et al. found no study supporting it. This is the book’s core actionable prescription and it is the claim with the clearest null result.

Provenance and framing.

note: Chapman is a Baptist pastor writing from an explicitly Christian marriage-counseling frame; the original text assumes heterosexual marriage and includes advice patterns that later editions softened. This limits universality independently of the evidence problem.

Grit is a distinct trait that predicts success better than talent or IQ.

overclaimed / construct-redundant: Marcus Credé, Michael Tynan & Peter Harms (2017, JPSP), meta-analysis of 88 samples / 66,807 people, found grit correlates with conscientiousness at ρ = .84, above the conventional threshold for redundancy. Schmidt et al. (2020) titled their paper ‘Grit and conscientiousness: another jangle fallacy’. Grit is largely conscientiousness with a new name.

Grit predicts outcomes (West Point retention, National Spelling Bee, GPA) with meaningful effect size.

effect sizes far smaller than the book implies: Credé et al. found grit’s correlation with performance/retention modest, and Credé specifically accused Duckworth of describing effect sizes in a way that made them sound misleadingly large. Grit accounts for a low single-digit percentage of variance in academic performance.

Both facets matter: perseverance of effort AND consistency of interest.

half-debunked: nearly all predictive power comes from perseverance of effort (r ≈ .20); consistency of interest adds close to zero incremental variance (r ≈ .08). Confirmed in Lam & Zhou (2022), 137 studies. The two-factor structure that makes grit ‘new’ is the part that does not work.

Grit can be taught / grown, and schools should cultivate it.

contested and politically criticized: no strong intervention evidence that grit is trainable at scale. Critics including Credé and education writers argue that grit talk shifts blame for structural disadvantage onto individual children. Duckworth herself has publicly distanced from grit-based school accountability measures.

10,000 hours of practice is the threshold for world-class expertise.

debunked by the source researcher: Anders Ericsson, whose 1993 Berlin violinist study Gladwell drew on, stated flatly “there never was a 10,000 hour rule” and called it a misinterpretation, writing the essay ‘The Danger of Delegating Education to Journalists’ in response. 10,000 was a group AVERAGE at age 20, not a threshold; variation between individuals was enormous. Ericsson also objected that Gladwell dropped the word ‘deliberate’. Quantity of practice without focused, feedback-driven, uncomfortable practice does little.

Practice essentially explains expert performance; innate talent is a myth.

failed / substantially overstated: Macnamara, Hambrick & Oswald (2014, Psychological Science), 88 studies, found deliberate practice explains ~26% of variance in games, ~21% in music, ~18% in sports, ~4% in education and <1% in professions. Practice matters, but nowhere near deterministically. Hambrick’s later work found starting age and working-memory capacity add independent predictive power.

The Beatles’ Hamburg hours and Bill Gates’s computer access as illustrations.

contested as storytelling: the Beatles’ Hamburg hour counts have been disputed (Mark Lewisohn’s research gives far fewer hours than Gladwell’s figure), and the cases are selected post hoc, which cannot establish a rule.

IQ accounts for about 20% of success, so EQ accounts for the other 80%.

debunked, and semi-retracted by the author: the inference is a non sequitur (the residual variance is not one construct). Goleman walked this back in the 10th-anniversary introduction, acknowledging his earlier writing had encouraged the 80% conclusion. The number still circulates in corporate training two decades later.

EI predicts job and leadership performance better than IQ.

overclaimed: Joseph & Newman (2010) meta-analysis found EI measures add roughly 6–10% incremental variance to job performance over IQ and the Big Five. Real, respectable, comparable to conscientiousness, and an order of magnitude below the book’s implied claim. Mixed-model/trait EI measures (including Goleman’s own competency inventories) perform worse than ability-based EI (Mayer-Salovey-Caruso) on construct validity, and are substantially contaminated by personality.

Construct legitimacy of ‘emotional intelligence’ as Goleman defines it.

contested: Peter Salovey and John Mayer, who coined the term, have publicly distinguished their narrower ability construct from Goleman’s expansive popular version. Critics (Edwin Locke, Gerald Matthews, Moshe Zeidner, Richard Roberts in ‘Emotional Intelligence: Science and Myth’) argue Goleman’s version lacks theoretical cohesion, shifts definitions, and bundles motivation, empathy, and social skill into a pseudo-unitary trait.

Amygdala hijack and the neuroscience chapters.

dated and simplified: built largely on Joseph LeDoux’s fear-conditioning work, which LeDoux himself has since substantially revised (he now argues against equating the amygdala circuit with the felt experience of fear). The popular ‘amygdala hijack’ shorthand outlived its neuroscience.

The habit loop (cue → routine → reward) as the general mechanism of human behavior.

partly replicated, heavily overextended: the cue-response-reward architecture is real and comes from solid basal-ganglia/striatal work (Ann Graybiel at MIT, whose research Duhigg popularizes). What is not supported is treating it as a master key for organizations, movements, and societies, which is what Parts 2 and 3 attempt.

“Keystone habits”: one habit that cascades into transforming other domains.

unsupported as stated: ‘keystone habit’ is Duhigg’s coinage, not a research construct, and there is no controlled evidence for reliable cross-domain spillover. Behaviorally, habit generalization is weak and domain-specific.

Willpower is “the single most important keystone habit” and is a muscle you strengthen (the Baumeister ego-depletion chapter, plus the Starbucks willpower-training story).

failed replication: rests on ego depletion, which collapsed in Hagger et al. 2016 and Vohs et al. 2021. This chapter is the book’s weakest link and its most repeated takeaway.

Narrative case studies (Febreze, Target’s pregnancy-prediction, Alcoa, Rosa Parks, the gambler/sleepwalker legal cases).

contested as method: the marketing facts are broadly accurate, but reviewers and several of Duhigg’s own cited scientists (as he acknowledges in the notes) said he over-concluded, simplified dangerously, or built rules from ancillary findings. Standard Gladwell-tier rigor: entertaining reverse-engineered stories, not evidence.

“The WHO has declared a sleep loss epidemic throughout industrialized nations.”

debunked: Alexey Guzey (guzey.com, Nov 2019) traced the citation to a National Geographic documentary that never mentions the WHO or any epidemic declaration. No such WHO declaration exists. Walker has since conceded in an interview that the body he had in mind was the CDC rather than the WHO (reported via Gelman’s statistical-modeling blog, Nov 2019), which is a different and much weaker claim; the printed text is unchanged as of 2026.

“Sleeping less than six or seven hours demolishes your immune system, more than doubling your risk of cancer.”

overclaimed/contradicted: a 2018 meta-analysis (~1.5M participants) found neither short nor long sleep duration associated with increased cancer risk. Walker cited no source supporting a doubling. Andrew Gelman (Columbia) amplified the critique on his blog; Walker never directly rebutted it.

“The shorter your sleep, the shorter your life span”: monotonic dose-response.

contested/wrong shape: mortality-vs-sleep is U-shaped (both short and long sleep associate with higher mortality, optimum ~7h). Walker presents a monotonic relationship the epidemiology does not support. Also correlational, not causal.

Sleep-deprivation harms graph in Chapter 6.

credibly contested as data manipulation: Guzey showed Walker reproduced a published figure with the 5-hour bar removed; that bar showed LOWER injury risk than 6 hours, contradicting his narrative. UC Berkeley’s review conceded the omission but called it and all other findings “minor” and said it did not alter conclusions. Walker’s public blog reply answered “reader questions” and did not address Guzey’s specific points; he said errors would be fixed in a future edition. As of 2026 no corrected edition has shipped.

“No species reduces sleep without severe harm” / sleep is universally non-negotiable at 8h.

overclaimed: counterexamples exist (e.g. cetaceans, migrating birds, some human short-sleeper genotypes such as DEC2/ADRB1 carriers). Individual sleep need is more variable than the book’s 8-hour prescription implies.

Lobster serotonin/octopamine hierarchies show that dominance hierarchies are ancient and biologically inevitable in humans (Rule 1).

the lobster neurochemistry is roughly right; the inference to humans is not: serotonin raises aggression and dominance posture in lobsters but is broadly associated with reduced aggression and increased affiliation in humans, so the analogy inverts. Evolutionary biologist P.Z. Myers attacked the argument as ignorant of the phylogenetic distance (~350 million years, with the trait not conserved through the intervening lineages). Shared neurotransmitters across distant taxa are not evidence of shared function. Note that critics also argue the naturalistic-fallacy structure (hierarchies are ancient, therefore current hierarchies are legitimate) does not follow even if the biology held.

Psychological content: self-authoring, incremental exposure to difficulty, taking responsibility, telling the truth, comparing yourself to your past self.

broadly defensible: the clinical core is recognizable CBT/exposure practice from a licensed clinical psychologist and former Harvard/Toronto professor with a real publication record in personality psychology. This is the part that holds up.

Jungian archetypes, chaos/order symbolism, the DNA-in-the-double-serpent reading, alchemical and biblical interpretation.

not science: presented interleaved with genuine psychology, which makes it hard for a general reader to tell where the evidence stops. Jungian archetype theory has no empirical standing in contemporary psychology.

Universality for this list.

seriously limited: the book is addressed to a specific reader (young, Western, male, disaffected), leans on Judeo-Christian scripture as its primary moral vocabulary, and carries positions on gender that many readers will reject before reaching the useful parts. The author is also a live political figure, which means recommending the book carries signaling costs unrelated to its content.

Frey was a violent, drug-charged fugitive who struck an officer with his car, fought several officers and served eighty-seven days in an Ohio jail.

fabricated at the centre: The Smoking Gun’s January 2006 investigation checked police and court records and interviewed the arresting officers. The Granville record describes an intoxicated but polite and cooperative driver, no assault or crack charge, a $733 bond and no more than five hours in custody. On Oprah, 26 January 2006, Frey admitted that the jail term was false. The invented ordeal supplies both his outlaw authority and the book’s final legal stakes.

A girl called Michelle was Frey’s close friend, the town blamed him after her fatal train collision, and the guilt helped drive his self-destruction.

fabricated appropriation: The Smoking Gun’s January 2006 investigation identified the victims as Jane Hall and Melissa Sanders and used the 1986 police report plus interviews with families and former classmates to find no role for Frey and no close friendship of the kind narrated. On Oprah, 26 January 2006, Frey admitted that he had not been close to Sanders. The book converts two families’ real deaths into its narrator’s trauma.

The treatment narrative is factual, including two root canals performed without anaesthetic because the clinic prohibited pain medication.

admitted uncertainty inside a supposedly witnessed scene: on Oprah, 26 January 2006, Frey said he could not know whether Novocain had been used, despite pages of precise sensory description. The treatment centre disputed the episode, while privacy rules prevented independent checking of every patient and incident. The clinic drama is therefore a mixture of documented invention and unresolved testimony, not a reliable recovery record.

The altered events were minor embellishments that left the memoir contract intact.

contradicted by the author and legal record: Frey’s 2006 Note to the Reader concedes that he changed events for dramatic effect and created a tougher version of himself. In re A Million Little Pieces Litigation, approved by the US District Court for the Southern District of New York in May 2007, provided refunds to 1,729 readers and required new author and publisher disclosures. The inventions create the criminality, trauma, physical ordeal and legal jeopardy that make the story exceptional.

Addiction is a weakness, and recovery rests on deciding not to use whenever craving appears.

contradicted and unsafe as general guidance: NIDA’s Drugs, Brains, and Behavior (revised 2018) describes addiction as a chronic, treatable disorder and says relapse signals a need to resume or modify treatment, not treatment failure. NIAAA’s current clinical guidance likewise treats alcohol use disorder as a medical condition with multiple evidence-based routes, including medication, behavioural care and mutual support. One dramatic, materially unreliable testimony cannot establish a treatment mechanism.

Thirty-seven percent of British workers believe their jobs make no meaningful contribution to the world, confirming that bullshit jobs are widespread.

unsupported at the claimed scale: YouGov’s August 2015 poll did report 37 percent answering no when asked whether their job made a ‘meaningful contribution to the world’, but that is a higher and different test from Graeber’s definition of work so pointless, unnecessary or pernicious that even its holder cannot justify it. Robert Dur and Max van Lent (Industrial Relations, January 2019), using approximately 100,000 workers across 47 countries, found about 8 percent calling their work socially useless and another 17 percent doubtful. Magdalena Soffia, Alex J. Wood and Brendan Burchell (Work, Employment and Society, October 2022) found only 4.8 percent of EU employees in 2015 saying they rarely or never did useful work. Different questions prevent a clean pooled estimate, but none licenses Graeber’s conversion of 37 percent into an objective census.

Bullshit jobs are rapidly proliferating under managerial feudalism.

contradicted by the available trend test: Graeber supplies no longitudinal measure of either bullshit jobs or managerial feudalism. Soffia, Wood and Burchell (2022), analysing representative European Working Conditions Survey waves from 2005, 2010 and 2015, found perceived uselessness declining rather than increasing. Their cross-sectional job-quality measures cannot prove an alternative cause, but the only direct trend evidence runs opposite to the book’s central historical claim.

Pointless work clusters among managers and professionals in finance, law, administration and related white-collar occupations, while socially useful workers generally know their work matters.

contradicted in Europe, partially supported in the United States: Soffia, Wood and Burchell (2022) found legal, business and administrative professionals among those least likely to report useless work, while refuse workers and cleaners were more likely to do so; graduates were also less likely than non-graduates to report uselessness. Simon Walo (Work, Employment and Society, October 2023), using the 2015 American Working Conditions Survey, N = 1,811, found significantly higher perceived uselessness in four of Graeber’s five nominated groups after controls, with legal occupations the exception. Walo also found alienation, social contact and public-service motivation mattered. The US result rescues part of the occupational pattern, not the book’s global claim or an objective verdict on those occupations.

Workers who say their jobs are useless are normally correct about the jobs’ objective social value.

not measured: Graeber makes subjective judgment definitive because he argues that no one is better placed to know, but every quantitative study here measures perceived usefulness. Dur and van Lent (2019), Soffia, Wood and Burchell (2022), and Walo (2023) identify associations with management quality, autonomy, social interaction, public-service motivation and other working conditions. Those findings make testimony important evidence about lived work while leaving the counterfactual question, what would happen if the job disappeared, unanswered. The book repeatedly crosses that gap without an independent measure.

Performing work one believes useless causes ‘spiritual violence’ and serious psychological harm.

association supported, direction unresolved: Soffia, Wood and Burchell (2022) tested five core hypotheses and rejected the four structural propositions about scale, growth, occupational location and graduate supply, but confirmed the fifth, a strong association between perceived uselessness and poor psychological wellbeing. Their data are cross-sectional, so they cannot distinguish uselessness causing distress from distress changing judgment or poor management producing both. This is the book’s strongest surviving empirical claim and still not the causal result its rhetoric implies.

Students with a growth mindset achieve substantially more than students with a fixed mindset.

overclaimed: Sisk, Burgoyne, Sun, Butler and Macnamara (Psychological Science, 2018) synthesized 273 studies and 365,915 participants and found only a weak association between mindset and academic achievement, r = .10. Correlation does not establish that the belief caused the result, since prior achievement, teaching, resources and school context can shape both.

Teaching a growth mindset produces meaningful gains in academic achievement.

near zero on average: Sisk et al. (Psychological Science, 2018) found d = .08 across intervention studies. Macnamara and Burgoyne (Psychological Bulletin, 2023), reviewing 63 studies and 97,672 students, found d = .05 overall, no significant effect after correction for publication bias, and d = .02 in the six studies with the strongest designs. Burnette et al. (Psychological Bulletin, 2023) reached the more favorable estimate of d = .14 for targeted groups under high-fidelity delivery, but its prediction interval included zero. The honest verdict is not that belief never matters; it is that the general-purpose intervention was sold far beyond its average effect.

Large real-world trials validate the book’s intervention promise.

small and conditional, or null: Yeager et al. (Nature, 2019), a preregistered national experiment with 12,490 US ninth-graders, found a 0.10-grade-point benefit among lower-achieving students, standardized d = .11, with larger effects in schools whose peer norms supported the message. The independently evaluated Changing Mindsets effectiveness trial (Education Endowment Foundation and National Institute of Economic and Social Research, 2019) randomized 5,018 pupils in 101 English schools and found no additional progress in literacy or numeracy, including for pupils eligible for free school meals.

The same fixed-versus-growth mechanism explains champions, great companies, healthy relationships and effective parenting.

unsupported at the book’s level of confidence: Dweck’s sports, business and relationship chapters lean on selected biographies, public outcomes and retrospective contrasts. Those stories cannot isolate mindset from talent, coaching, resources, incentives, power, selection or survivorship. The concept travels much further than the causal evidence.

After failing on K2 in 1993, Mortenson became lost, stumbled into Korphe, was nursed back to health there and promised to build the village a school.

substantially contradicted: 60 Minutes (CBS, 17 Apr 2011) interviewed Mortenson’s two porters from the K2 expedition, who said he did not become lost or visit Korphe on that descent; three further sources supported their account. Jon Krakauer’s Three Cups of Deceit (Apr 2011) placed the first Korphe visit in 1994 and concluded that the book converted later events into a creation story. Mortenson disputed the 1994 date but told Outside (18 Apr 2011) that the narrative involved “omissions and compressions”; he later told Outside that he had spent only hours in Korphe and made the school promise on a return visit, not as the book describes.

Mortenson was abducted and held for eight days by Taliban kidnappers in Waziristan in 1996.

contradicted by the people depicted: 60 Minutes (CBS, 17 Apr 2011) located four men present when the book’s photograph was taken; they denied being Taliban and denied kidnapping Mortenson. Mansur Khan Mahsud, pictured among the alleged abductors, told Krakauer in Three Cups of Deceit (Apr 2011) that Mortenson had been an honoured guest. Mortenson maintained in his Apr 2011 CBS statement that his passport and money were taken and that he was detained against his will. The central participants’ incompatible accounts, with no independent support supplied by the book, make the episode unusable as nonfiction evidence.

The Central Asia Institute’s school tally and stories provide a documented record of projects and show that building schools can defeat terrorism.

part fact, part unsupported promotion: 60 Minutes (CBS, 17 Apr 2011) reported claimed schools that did not exist, had been built by others, stood unused or no longer received CAI support. Later reporting by Outside (Alex Heard, ‘The Trials of Greg Mortenson’, 2012) also documented operating schools and beneficiaries, so the evidence does not support the opposite claim that all of CAI’s work was fictitious. The book supplies neither an auditable project list nor outcome evidence for its schools-versus-terrorism slogan. Real educational work cannot validate inflated outputs or a monocausal foreign-policy claim.

The book’s promotion and Mortenson’s speaking work were straightforward extensions of CAI’s charitable mission.

officially found deficient: the Montana Attorney General’s investigative report (5 Apr 2012) found that Mortenson and CAI’s board failed their nonprofit duties around book purchases, royalties, promotion, speaking travel and personal expenses. Mortenson had been removed as executive director in late 2011, and under the settlement agreed to repay more than $1 million, with credit for $420,000 already paid. The report also called CAI’s mission admirable and worth saving, found no basis for a criminal referral and did not investigate the books’ factual accuracy. These are documented financial and governance failures, not official corroboration of Krakauer’s narrative findings.