Appendix A: Key Studies Summary

Landmark Research in Luck Science, Probability, Network Theory, and Serendipity

The Science of Luck: Statistical Thinking, Network Theory, Serendipity Engineering, Opportunity Recognition, and the Psychology of Chance

This appendix summarizes 27 landmark studies across the major themes of the textbook. For each study, we provide the core finding, methodology, replication status (where known), and its specific relevance to luck science. These summaries are designed for efficient review before exams, during research projects, and when evaluating news reports that cite this work.

How to use this appendix: The summaries are intentionally condensed. For full context, consult the original papers (full citations in the Bibliography, Appendix 3) and the relevant textbook chapters. Where replication status is noted as "contested" or "mixed," the textbook chapter provides nuance on what specifically has and hasn't replicated.

Theme 1: Luck, Skill, and Chance in Outcomes

Study 1: Talent vs. Luck Simulation

Researchers: Alessandro Pluchino, Alessio Emanuele Biondo, Andrea Rapisarda Year: 2018 Published in: Advances in Complex Systems

Core finding: In a computational simulation in which 1,000 agents have normally distributed talent (with most near the mean) and encounter randomly distributed lucky and unlucky events over a 40-year career, the most successful agents at the end are NOT the most talented — they are those who combined moderate talent with the greatest number of lucky events. The most talented individuals frequently end up with mediocre outcomes due to insufficient luck.

Methodology: Agent-based simulation. Agents were assigned talent scores drawn from a normal distribution. Each year, agents randomly encountered "lucky" events (which amplified their talent) or "unlucky" events (which reduced their success probability). After 40 simulation years, wealth distributions were compared to talent distributions.

Key finding detail: The wealth distribution of "successful" agents followed a power law (a small number of agents capture most resources) while talent was distributed normally — meaning the distribution of outcomes is much more extreme than the distribution of underlying ability. If rewards perfectly tracked talent, a normal distribution of rewards would be expected.

Replication status: The specific simulation has not been formally replicated, but the qualitative findings are consistent with a large body of empirical literature on income inequality, academic citation distributions, and cultural market dynamics.

Relevance to luck science: Provides a formal model demonstrating why acknowledging luck is not just philosophically honest but empirically required. Even in a world where talent is real and consequential, luck effects dominate outcome distributions. Directly relevant to Chapter 2 and the meritocracy discussion in Chapter 18.

Study 2: Wiseman's UK Luck Study

Researcher: Richard Wiseman Year: 1993–2003 (decade-long program) Published in: Summarized in The Luck Factor (2003); partial data in academic publications

Core finding: Self-identified lucky people differ from self-identified unlucky people in four consistent behavioral patterns: (1) they create and notice chance opportunities through open body language, large social networks, and relaxed attention; (2) they make good luck-related decisions using intuition; (3) they create self-fulfilling prophecies through positive expectations; (4) they adopt a resilient attitude that transforms bad luck.

Methodology: Wiseman recruited 400 volunteers who self-identified as very lucky, very unlucky, or neutral. Over 10 years, he interviewed participants, administered personality and behavior questionnaires, conducted behavioral experiments (e.g., how much money they notice on the street, whether they notice opportunities in a fake newspaper), and tracked outcomes.

Key finding detail: In one experiment, participants were given a newspaper and asked to count the photographs. On page 2, a large notice read: "Stop counting — there are 43 photographs in this newspaper." Lucky people noticed it; unlucky people generally did not. The difference lay in open vs. narrowly focused attention — unlucky people were too task-focused to notice opportunities outside the task.

Replication status: The large-scale study has not been formally replicated in academic literature. The findings are consistent with related research on attention, social behavior, and dispositional optimism, but Wiseman's methodology has been critiqued for self-selection bias (people who think they're lucky may systematically differ from the population in ways that confound the findings).

Relevance to luck science: The primary empirical foundation for Part 3 (Psychology of Luck). Establishes that luck-relevant behaviors are measurable, distinguishable, and modifiable. Directly drives Chapters 12 and 16.

Study 3: Salganik, Dodds, and Watts — Music Lab

Researchers: Matthew J. Salganik, Peter Sheridan Dodds, Duncan J. Watts Year: 2006 Published in: Science

Core finding: In an artificial music market experiment, early random variation in which songs received initial downloads created dramatically different success hierarchies across parallel "worlds," even when song quality was held constant. Social influence amplified small early advantages into large outcome differences. Unpredictability was high; quality predicted success in the independent-choice condition but much less so in social influence conditions.

Methodology: 14,341 participants were randomly assigned to one of two conditions: (1) an independent condition in which they rated and downloaded songs without seeing others' choices, or (2) one of eight parallel "social influence worlds" in which the current download counts were visible. The same 48 songs were available in all conditions.

Key finding detail: Songs that happened to get downloaded early (by chance) in a social influence world accumulated further downloads through social reinforcement. The same songs performed very differently across the eight parallel worlds — the world with the most downloads of one song was essentially arbitrary. The correlation between quality rank (determined by independent condition) and market rank (in social influence conditions) was positive but weak.

Replication status: Well-replicated in spirit; the general phenomenon of social influence amplifying early random variation is robust and widely observed. The specific effect sizes are context-dependent.

Relevance to luck science: One of the cleanest experimental demonstrations that cultural market success is substantially determined by luck (early random variation + social amplification). Directly relevant to Chapters 3, 22, and the general argument that viral content involves irreducible luck.

Study 4: Barnsley Relative Age Effect

Researchers: Roger H. Barnsley, A. H. Thompson, Paula Barnsley Year: 1985 Published in: CAHPER Journal

Core finding: Players in elite Canadian hockey leagues are dramatically overrepresented among those born in the first three months after the January 1st age-group eligibility cutoff date. Children born in January, February, and March make up roughly 40% of elite players; children born in October, November, December make up roughly 10%.

Methodology: Examined birthdates of players in the Ontario Hockey League (major junior hockey), Quebec Major Junior Hockey League, and New England Division III college league. Compared birthdate distributions to expected equal distributions.

Key finding detail: The mechanism is not biological but social: at age 8–10, a child born in January is developmentally ~11 months older than a child born in December of the same calendar year. This small difference produces measurable differences in physical size, coordination, and maturity. The larger, more mature child is identified as "talented" and receives more coaching, practice time, and better team assignments. These enriched environments then produce actual skill differences that persist into adulthood.

Replication status: Highly replicated across sports (soccer, baseball, basketball, cricket), academic tracking, and gifted-and-talented programs. One of the most robust findings in sports sociology. Directly documented in academic settings by Malcolm Gladwell (popularized) and multiple academic follow-up studies.

Relevance to luck science: Definitive example of constitutive luck at work — birth timing, a factor entirely outside individual control, substantially shapes elite athletic outcomes. Directly cited in Chapter 18 as a case study in structural luck.

Study 5: Great Gatsby Curve

Researchers: Miles Corak (primary analyst); conceptualized and named by Alan Krueger Year: Published in series of papers; popularized in Krueger's 2012 Council of Economic Advisers speech Published in: Journal of Economic Perspectives (2013) and related publications

Core finding: Across a wide range of developed countries, higher income inequality is strongly associated with lower intergenerational income mobility. Countries with high inequality (like the United States) show stronger parent-child income correlations than low-inequality countries (like Denmark and Norway).

Methodology: Cross-national comparison of intergenerational income elasticity (the correlation between parent and child earnings, where 0 = complete mobility and 1 = complete immobility) against Gini coefficients (measure of income inequality). Data drawn from tax records, surveys, and longitudinal studies across approximately 15 countries.

Key finding detail: The United States has an intergenerational income elasticity of approximately 0.47 — meaning that about 47% of income advantage is transmitted from parent to child. Denmark's figure is approximately 0.15. The United States is less economically mobile than most peer nations while also being more unequal.

Replication status: The cross-national correlation is robust and replicated using multiple data sources. The mechanisms through which inequality reduces mobility remain actively researched (education access, network stratification, and neighborhood effects are all implicated).

Relevance to luck science: The primary empirical foundation for the structural luck argument in Chapter 18. The curve demonstrates that "pull yourself up by your bootstraps" narratives are statistically more accurate in low-inequality societies than in high-inequality ones.

Theme 2: Cognitive Biases and Probability Misperception

Study 6: The Hot Hand Fallacy

Researchers: Thomas Gilovich, Robert Vallone, Amos Tversky Year: 1985 Published in: Cognitive Psychology

Core finding: Analyses of NBA shooting data, field goal data, and controlled shooting experiments with Cornell players found no statistical evidence for the hot hand: a player's probability of making a shot after a series of makes was not higher than after a series of misses, once shooting percentage was accounted for.

Methodology: (1) Analysis of sequential shooting records for all players in the 1980–81 Philadelphia 76ers season. (2) Survey of NBA fans and players on hot hand beliefs. (3) Controlled shooting experiment with Cornell men's and women's varsity players. (4) Analysis of Boston Celtics free throw data.

Key finding detail: Even though the hot hand was believed by 91% of fans and by players themselves, statistical analysis found no evidence that making a shot increased the probability of the next shot. Sequences that appeared as streaks were consistent with the binomial distribution.

Replication status: Partially revised by Miller and Sanjurjo (2018) — see Study 7. The perception of the hot hand clearly exists (replicated); whether any real hot hand effect exists is now an open question.

Relevance to luck science: Foundational demonstration that humans perceive non-random patterns in genuinely random sequences. A core example for Chapters 4 and 7 on cognitive bias and small samples.

Study 7: Hot Hand Reanalysis — The Hot Hand Is Real (Sort Of)

Researchers: Joshua B. Miller, Adam Sanjurjo Year: 2018 Published in: Econometrica

Core finding: Gilovich, Vallone, and Tversky's analysis contained a subtle but significant mathematical bias: when measuring whether a hit follows a hit in a finite sequence, there is a built-in negative bias that artificially makes the data appear streak-free. Correcting for this bias reveals a small but statistically real hot hand effect.

Methodology: Mathematical analysis of the statistical properties of the original methodology; reanalysis of GVT's data with bias correction; new analyses of NBA three-point contest data.

Key finding detail: The bias arises because in a finite sequence, looking at the probability of a hit given the previous k hits necessarily excludes the last possible hit from the denominator. This creates a systematic downward bias in the conditional probability estimate. When corrected, the shooting data shows a positive (real) hot hand effect. However, the effect size is small.

Replication status: The mathematical error has been independently confirmed. The broader question of hot hand effects in real game conditions (where opponents adjust) remains debated.

Relevance to luck science: One of the textbook's primary examples of scientific self-correction — a landmark finding substantially revised by careful reanalysis. Relevant to Chapter 4 and Appendix B (Research Methods Primer). The appropriate lesson is not "the hot hand is real" but "statistics is hard and even experts make subtle errors that require ongoing scrutiny."

Study 8: Invisible Gorilla — Inattentional Blindness

Researchers: Daniel J. Simons, Christopher F. Chabris Year: 1999 Published in: Perception

Core finding: Approximately 50% of observers watching a video of people passing a basketball, while counting the number of passes, fail to notice a person in a gorilla suit who walks through the scene, stops to face the camera, and walks off.

Methodology: Participants were randomly assigned to count passes by either white-shirted or black-shirted players (varying difficulty). After viewing, participants were asked whether they noticed anything unusual. Awareness of the gorilla was recorded.

Key finding detail: The effect was not explained by looking away; eye-tracking confirms that many participants looked directly at the gorilla without consciously registering it. The effect is driven by attentional load: the more demanding the counting task, the less likely participants are to notice the unexpected event.

Replication status: Well replicated in multiple laboratories with variations (different unexpected objects, different contexts, fMRI studies confirming gaze without awareness). The basic phenomenon is robust.

Relevance to luck science: Demonstrates that what we notice is determined by what we're attending to — the primary mechanism behind why lucky people notice more opportunities (they maintain broader attentional aperture). Directly relevant to Chapter 16 (noticing problem) and Chapter 30 (opportunity recognition).

Study 9: Pygmalion Effect in the Classroom

Researchers: Robert Rosenthal, Lenore Jacobson Year: 1968 Published in: Holt, Rinehart and Winston (as a book; original data published in 1966)

Core finding: When teachers were told (falsely) that certain randomly selected students had recently been identified as "intellectual bloomers" likely to show significant academic growth, those students showed significantly greater IQ gains over the following year than control students.

Methodology: All students at an elementary school received an IQ test. Teachers were told that the test identified specific students as likely to show intellectual growth in the coming year. In fact, the "bloomers" were randomly selected — they did not differ from control students on any measure. IQ tests were readministered at year end.

Key finding detail: The effect was strongest for younger students (grades 1–2) and was observed in both verbal and reasoning IQ subscales. The mechanism appears to be that teachers provided more challenging material, warmer interactions, and more feedback to labeled students.

Replication status: Mixed. The original specific effect sizes have not robustly replicated, and some meta-analyses find small or null effects. However, the general phenomenon of teacher expectations affecting student performance has broad empirical support; the debate is about effect size and moderating conditions.

Relevance to luck science: Classic demonstration of self-fulfilling prophecy — positive expectations, even arbitrarily assigned, can produce real outcomes through behavior change. Directly relevant to Chapter 14.

Study 10: Prospect Theory

Researchers: Daniel Kahneman, Amos Tversky Year: 1979 Published in: Econometrica

Core finding: People evaluate outcomes relative to a reference point (not in absolute terms), and they weight losses more heavily than equivalent gains — approximately twice as heavily. This loss aversion leads to risk-seeking behavior in loss domains (gambling to avoid certain losses) and risk-avoidance in gain domains (taking certain small gains over risky large gains).

Methodology: Series of hypothetical choice problems presented to participants, documenting systematic departures from expected utility theory predictions. For example: most people prefer a certain $50 gain over a 50% chance of $100, but prefer a 50% chance of losing $100 over a certain $50 loss.

Key finding detail: The value function is concave in gains (diminishing sensitivity) and convex in losses (risk-seeking to avoid losses), with a steeper slope for losses than gains. The probability weighting function overweights small probabilities and underweights large ones.

Replication status: Among the most replicated findings in behavioral economics. Kahneman received the Nobel Prize in Economics in 2002 partly for this work (Tversky died in 1996). Some specific parameters have been revised, but the core phenomena are robust.

Relevance to luck science: The primary theoretical framework for Chapter 15. Loss aversion explains why people systematically avoid positive expected value opportunities, miss lucky breaks by fixating on downside risk, and behave differently when situations are framed as losses vs. gains.

Theme 3: Learned Helplessness and Resilience

Study 11: Learned Helplessness

Researchers: Martin E. P. Seligman, Steven F. Maier Year: 1967 Published in: Journal of Experimental Psychology

Core finding: Dogs exposed to inescapable electric shocks in a shuttle box subsequently failed to escape shocks even when escape was made possible. Control dogs that had not experienced inescapable shocks escaped immediately when given the opportunity.

Methodology: Three groups of dogs: (1) escapable shocks (could press a panel to stop the shock), (2) inescapable yoked shocks (same shocks as group 1 but no control), (3) no shock control. All were later placed in a shuttle box where they could escape shocks by jumping a barrier.

Key finding detail: 72% of group 2 dogs failed to escape in the shuttle box — they passively endured the shocks. Seligman termed this "learned helplessness": the animals had learned that their behavior was ineffective in controlling outcomes and applied this learning even when control was restored.

Replication status: Well replicated in animal models. Human analogs (using uncontrollable noise or unsolvable puzzles) are also well replicated, though the mechanisms differ somewhat from the animal model. Seligman later developed the learned helplessness model for human depression.

Relevance to luck science: The foundational experimental model for understanding how repeated bad luck (uncontrollable negative events) can extinguish the initiative-taking that would produce better luck in the future. Directly relevant to Chapter 13's treatment of locus of control and Chapter 17's resilience chapter.

Study 12: Post-Traumatic Growth

Researchers: Richard G. Tedeschi, Lawrence G. Calhoun Year: 1996 (foundational conceptual paper); 2004 (key review) Published in: Journal of Traumatic Stress (1996); Psychological Inquiry (2004)

Core finding: A significant proportion of people who experience highly challenging life events (cancer, bereavement, natural disaster, serious accident) report meaningful positive psychological changes as a result — including greater personal strength, new possibilities, improved relationships, greater appreciation for life, and spiritual deepening. This phenomenon is distinct from simply "bouncing back."

Methodology: Survey and interview studies of survivors of various traumas, using the Post-Traumatic Growth Inventory (a validated scale). Longitudinal studies tracking change over time. Qualitative interviews exploring the phenomenology of growth.

Key finding detail: PTG does not imply the absence of suffering — it often co-occurs with ongoing distress. The mechanism involves cognitive processing (making meaning of the experience, revising one's worldview) rather than suppression of negative emotion. Social support and disclosure appear to facilitate the process.

Replication status: The phenomenon of self-reported growth after trauma is reliably observed. The mechanisms and the extent to which it represents genuine versus illusory growth are actively debated in the literature.

Relevance to luck science: Directly supports Chapter 17's argument that resilience is not passive survival but active transformation — a luck asset that can be cultivated. PTG represents the possibility that bad luck, properly processed, can produce durable increases in life competence.

Theme 4: Network Effects and Weak Ties

Study 13: Strength of Weak Ties

Researcher: Mark Granovetter Year: 1973 Published in: American Journal of Sociology

Core finding: Professional information (particularly job opportunities) flows predominantly through weak ties (acquaintances, occasional contacts) rather than strong ties (close friends). In a study of 282 professional and technical workers who had found jobs through personal contacts, 83.4% had seen the contact "occasionally" or "rarely" — not "often."

Methodology: Interviews with 282 professional and technical workers in the Boston suburb of Newton who had recently changed jobs. For those who found jobs through personal contacts, participants described their relationship with the contact who provided the job information.

Key finding detail: The theoretical explanation: strong ties (close friends) inhabit the same social circles and possess largely redundant information. Weak ties (acquaintances) bridge different social clusters and carry non-redundant, novel information. Job opportunities, market information, and social innovations spread through the sparse bridges that weak ties provide.

Replication status: One of the most-cited papers in sociology; the general pattern has been replicated across job types, countries, and time periods. Some modern studies using digital trace data (LinkedIn, Facebook) find more complex patterns — very weak ties (near strangers on social media) may not carry the benefits that Granovetter observed in personal acquaintances. The "optimal tie strength" may be moderate rather than very weak.

Relevance to luck science: The primary theoretical foundation for Chapter 19 and a cornerstone of the network theory section. Explains why building diverse networks of acquaintances produces more luck opportunities than deepening existing close relationships.

Study 14: Small-World Networks

Researchers: Duncan J. Watts, Steven H. Strogatz Year: 1998 Published in: Nature

Core finding: Formalization of the "small-world" network property: networks can simultaneously have high local clustering (most of your neighbors know each other) AND short average path lengths (any two nodes are connected by surprisingly few steps). This combination arises when a small number of random "shortcut" links are added to an otherwise highly clustered network.

Methodology: Mathematical modeling and simulation. Watts and Strogatz began with a regular lattice (high clustering, long path lengths) and progressively added random "rewired" links. A small number of rewirings dramatically reduced average path length while barely affecting clustering coefficient.

Key finding detail: The mathematical result explains how six degrees of separation is possible in social networks. In a pure lattice (only local connections), reaching distant nodes would require many steps. The few "long-range" weak ties (in Granovetter's terms) provide the shortcuts that collapse global distances.

Replication status: The mathematical model is well established. Empirical small-world properties have been documented in power grids, neural networks, and social networks. The theory provides the mathematical foundation for Milgram's empirical observations.

Relevance to luck science: Explains the mechanism behind "it's a small world" luck stories — why unexpected connections appear at apparently improbable distances. Directly foundational to Chapter 20.

Study 15: Structural Holes and Good Ideas

Researcher: Ronald S. Burt Year: 2004 Published in: American Journal of Sociology

Core finding: Managers whose networks bridge structural holes (gaps between otherwise disconnected groups) generate ideas that are rated as significantly more valuable by senior executives, and those managers are promoted faster than managers embedded in dense, closed networks.

Methodology: Survey of managers at a large electronics firm. Network data collected on who discussed work-related issues with whom. Network structure analyzed to identify managers bridging structural holes vs. those in dense clusters. Managers submitted suggestions for improving supply chain. Ideas were evaluated blind by senior executives.

Key finding detail: The effect was not explained by individual ability, experience, or seniority. Bridge-position managers had access to more diverse information environments and were exposed to more varied problem-solving approaches — their ideas were rated as better because they effectively synthesized insights from multiple disconnected knowledge domains.

Replication status: The general finding linking bridge positions to performance has been replicated in multiple organizational contexts. The mechanism (information diversity) is consistent with the weak ties literature and has independent empirical support.

Relevance to luck science: The empirical foundation for the structural holes concept in Chapter 21. Demonstrates that network position (a partially engineerable form of luck) has measurable effects on performance outcomes above and beyond individual ability.

Study 16: Milgram's Small World Experiment

Researcher: Stanley Milgram Year: 1967 Published in: Psychology Today (popular); more formally in Travers & Milgram (1969) in Sociometry

Core finding: Letters sent through chains of acquaintances from Nebraska to a target contact in Boston arrived in a median of about 6 links — the origin of the "six degrees of separation" concept.

Methodology: Milgram mailed packages to 296 randomly selected residents of Nebraska (and some in Boston), asking them to forward the package to a named target person in Boston. They could only mail to someone they knew personally by first name. The number of links was recorded for packages that reached the target.

Key finding detail: Only 64 of 296 packages (22%) reached the target. The median path length for those that did was 5.5 (approximately 6 steps). Milgram interpreted this as demonstrating surprisingly short social distances; contemporary analysts have noted the selection bias introduced by incomplete chains.

Replication status: The "six degrees" phenomenon has been validated computationally using Facebook data (Backstrom et al., 2012, found an average of 4.74 degrees among 720 million Facebook users) and email data. The specific number varies by medium and definition of "acquaintance," but the qualitative finding of surprisingly short paths is robust.

Relevance to luck science: The empirical genesis for Chapter 20's treatment of small-world networks and their implications for luck — the world is far more interconnected than it appears, creating serendipity opportunities that seem improbable.

Theme 5: Survivorship Bias and Statistical Illusions

Study 17: Abraham Wald's Bullet Hole Problem

Researcher: Abraham Wald Year: 1943 (declassified from military use) Published in: Statistical Research Group (1943); discussed in various retrospective analyses

Core finding: During World War II, Wald was asked to recommend where to add armor to returning bomber aircraft based on observed bullet hole patterns. The conventional analysis proposed reinforcing the most-hit areas. Wald correctly argued the opposite: the areas that showed no hits should be reinforced, because planes hit in those areas did not return — they were the fatal hits that eliminated those planes from the sample.

Methodology: Statistical analysis of bullet damage patterns in returning aircraft. Wald applied basic survivorship bias reasoning to identify the inverse conclusion from the observed data.

Key finding detail: The returning planes were a biased sample: only planes that survived long enough to return were observed. The observed bullet holes represented non-critical hits. The blank areas on returning planes represented areas where hits were likely fatal.

Replication status: Not applicable — this is a historical decision analysis, not an experiment. The mathematical reasoning is uncontroversial.

Relevance to luck science: The definitive illustrative example of survivorship bias and is introduced as such in Chapter 9. Teaches the fundamental skill of asking "who is in my sample?" before drawing conclusions from observed data.

Study 18: Silicon Valley Survivorship Bias Study

Researchers: Multiple, including Gompers, Kovner, Lerner, & Scharfstein (2010); Eesley & Roberts (2012)

Year: Various Published in: Journal of Financial Economics; Management Science

Core finding: While a handful of high-profile startups founded by college dropouts (Apple, Microsoft, Facebook, Dell) are frequently cited as evidence that formal education is unnecessary for entrepreneurial success, systematic analysis shows that ventures founded by more educated entrepreneurs have higher survival rates and greater revenue at comparable stages.

Methodology: Gompers et al. (2010) analyzed 16,000+ venture-backed firms and their founders' backgrounds. Eesley and Roberts (2012) tracked all graduates and faculty of MIT and Stanford for decades, examining firm founding and success rates.

Key finding detail: Serial entrepreneurs (those who have succeeded before) have higher success rates — but this is a self-selection effect, not evidence that repeat attempts cause success. Controlling for experience and founding team quality substantially reduces apparent advantages of specific demographic characteristics.

Replication status: The systematic evidence consistently shows formal education is not disadvantageous (and often helpful) for entrepreneurial success; the famous dropouts are dramatic outliers in a large population of unsuccessful dropouts.

Relevance to luck science: Directly supports Chapter 9's debunking of survivorship bias in entrepreneurship advice. The dropout narrative is constructed entirely from visible successes and ignores the invisible population of failed dropout entrepreneurs.

Study 19: Sports Illustrated Jinx Analysis

Researchers: John Leonard and colleagues (various informal analyses); formal statistical analysis by multiple sports researchers

Year: Ongoing Published in: Various sports analytics publications and popular accounts

Core finding: Athletes featured on the cover of Sports Illustrated magazine appear to perform worse in the season following their cover appearance — the "Sports Illustrated Jinx." Statistical analysis reveals this is a regression to the mean phenomenon with no causal mechanism.

Methodology: Comparison of athletes' performance statistics before cover appearance (the season that earned them the cover) vs. after cover appearance. Multiple sports and decades analyzed.

Key finding detail: Athletes appear on the cover because they just had an exceptional performance season — an outlier season by definition. The next season is statistically likely to be more average simply because the random variation that contributed to the exceptional season is unlikely to repeat at the same extreme level. No causal mechanism (psychological pressure from fame, jinx, etc.) is required.

Replication status: The statistical explanation (regression to the mean) is mathematically confirmed. The absence of a causal jinx is supported by the fact that the same regression pattern appears for any selection criterion that uses extreme recent performance.

Relevance to luck science: One of the textbook's primary applied examples of regression to the mean (Chapter 8). Demonstrates how a statistical artifact can be misinterpreted as a meaningful phenomenon.

Theme 6: Labor Markets, Discrimination, and Structural Luck

Study 20: Are Emily and Greg More Employable? (Bertrand & Mullainathan)

Researchers: Marianne Bertrand, Sendhil Mullainathan Year: 2004 Published in: American Economic Review

Core finding: Identical resumes sent to job postings, differing only in whether the applicant name was distinctively White-sounding (Emily Walsh, Greg Baker) or distinctively African American-sounding (Lakisha Washington, Jamal Jones), received 50% more callbacks for the White-sounding names (9.65% vs. 6.45% callback rate).

Methodology: 5,000 resumes submitted to 1,300 real job postings in Boston and Chicago newspapers across a range of occupations. Resumes were constructed to be otherwise equivalent. Names were selected based on frequency data in birth certificates to ensure clear racial signal. Low-quality and high-quality resume versions tested.

Key finding detail: The callback gap was present across occupations, industries, and employer sizes. Higher-quality resumes improved callback rates for White-sounding names more than for African American-sounding names — suggesting that resume quality investments yield smaller returns for African American applicants. Callback rates were unrelated to actual applicant qualifications.

Replication status: One of the most-replicated findings in discrimination research. Multiple follow-up audit studies across countries, contexts, and demographic characteristics (gender, age, disability) have found consistent evidence of name-based discrimination, though effect sizes vary.

Relevance to luck science: Definitive evidence that surname (a constitutive luck factor — one's name is assigned at birth based on family and culture) substantially affects labor market outcomes. Directly supports Chapter 18's structural luck analysis and Chapter 39's ethics discussion.

Theme 7: Timing, Technology, and Macro Luck

Study 21: Bill Gross's Timing Analysis (Idealab)

Researcher: Bill Gross (practitioner analysis) Year: 2015 Published in: TED Talk (no formal academic publication)

Core finding: In an analysis of 200+ Idealab companies and 100+ external startups, timing accounted for 42% of the difference between startup success and failure — more than team (32%), idea (28%), business model (24%), or funding (14%).

Methodology: Post-hoc analysis of startups that Gross knew well. Rated each on five factors (timing, team, idea, business model, funding) and correlated ratings with outcomes (success vs. failure). The methodology is informal and subject to hindsight bias, but the results are directionally consistent with academic research on timing.

Key finding detail: Examples: Airbnb succeeded partly because the 2008 financial crisis made people willing to rent their homes for the first time; earlier house-sharing startups with the same idea failed. Z.com failed before YouTube; YouTube succeeded in the same space partly because broadband penetration had reached a tipping point.

Replication status: This is a practitioner analysis, not a peer-reviewed study. Academic research on startup timing is consistent with the directional finding that market timing matters substantially (e.g., Agarwal & Gort, 2002). The specific percentages should be treated as illustrative rather than precise estimates.

Relevance to luck science: Directly cited in Chapter 31 (Timing and Luck). Provides the most accessible empirical anchor for the claim that macro timing is a dominant success factor.

Study 22: The S-Curve and Technology Adoption

Researchers: Everett M. Rogers (foundational); widespread subsequent documentation

Year: 1962 (Rogers' Diffusion of Innovations); ongoing subsequent research Published in: Diffusion of Innovations (Rogers, 1962, multiple editions)

Core finding: New technologies and innovations spread through populations in a characteristic S-shaped curve: slow initial adoption by innovators and early adopters, rapid growth through the early and late majority, and eventual saturation. The specific stages create time-sensitive opportunity windows.

Methodology: Rogers synthesized hundreds of diffusion studies from agriculture, medicine, education, and other domains. The S-curve pattern emerged consistently across technologies, time periods, and cultures.

Key finding detail: Each segment of the S-curve represents different types of participants with different risk tolerance, information needs, and social roles. Early adopters benefit from first-mover advantages but face high failure rates. Early majority participants enter during the high-growth phase when opportunities are abundant.

Replication status: The qualitative S-curve pattern is extremely robust across technology contexts. The specific timing and shape vary considerably by technology type.

Relevance to luck science: Foundational for Chapter 31's treatment of timing luck and Chapter 33's technology luck chapter. The S-curve explains why entering a market or platform at different stages produces dramatically different opportunities.

Theme 8: Opportunity Recognition and Alertness

Study 23: Kirznerian Alertness in Experimental Settings

Researchers: Multiple, including Gaglio & Katz (2001), Ardichvili et al. (2003)

Year: 2001–2003 Published in: Entrepreneurship Theory and Practice; Journal of Business Venturing

Core finding: Entrepreneurs differ from non-entrepreneurs in systematic patterns of information processing — they are more likely to notice "signals" in their environment that suggest unmet needs or potential opportunities. This alertness appears to be trained and teachable, not purely dispositional.

Methodology: Cognitive interview studies comparing expert entrepreneurs to novices and non-entrepreneurs in how they process market information. Think-aloud protocols, information search studies.

Key finding detail: Experienced entrepreneurs search differently: they are more likely to notice anomalies, to make connections between disparate pieces of information, and to ask "why isn't this done differently?" rather than accepting existing arrangements. This pattern is partially trainable through experience.

Replication status: Mixed; the general finding that entrepreneurs process market information differently is well supported, but the mechanisms and trainability are less well-established.

Relevance to luck science: Directly supports Chapter 30's treatment of opportunity recognition as a trainable perceptual skill, not just an innate disposition.

Study 24: Curiosity and Information Foraging

Researchers: Multiple, including Pirolli & Card (1999); Berlyne (1960); Kashdan, Rose & Fincham (2004)

Year: Various Published in: Psychological Review; Psychological Science

Core finding: Curiosity increases the depth and breadth of information search. Curious individuals explore more widely, persist longer in information-gathering, and make more unexpected cross-domain connections. Trait curiosity correlates with broad knowledge, creativity, and subjective well-being.

Methodology: Experimental studies measuring information seeking behavior under varying curiosity conditions; trait curiosity scales correlated with outcome measures; foraging studies tracking information search patterns online.

Key finding detail: Berlyne (1960) distinguished between "perceptual curiosity" (driven by novel/unusual stimuli) and "epistemic curiosity" (drive to learn). Epistemic curiosity is the more powerful predictor of knowledge breadth and discovery. Pirolli and Card (1999) developed "information foraging theory" showing that information seeking follows optimal foraging mathematics.

Replication status: The general relationship between curiosity and information breadth is robust. Specific mechanisms are areas of active research.

Relevance to luck science: Directly supports Chapter 26's claim that curiosity is a structural luck strategy. Curious individuals inhabit wider information environments and are more likely to encounter serendipitous discoveries.

Theme 9: Ethics, Fairness, and Luck Distribution

Study 25: Moser's Patent Record Study

Researcher: Petra Moser Year: 2005 Published in: American Economic Review

Core finding: Analysis of 19th-century world's fair exhibits and patent data reveals that the types of inventions produced in a country are strongly shaped by the country's existing resource endowments and institutional structures — not just individual inventor talent. Countries with good cotton agriculture produced textile innovations; countries with coal produced mining innovations.

Methodology: Systematic coding of exhibits at 19th-century world's fairs plus patent records from multiple countries, correlated with geographic and institutional characteristics.

Relevance to luck science: An often-overlooked macro example of structural luck: even the domains in which individual talent can express itself are substantially determined by where one happens to be born and the resources available there. Relevant to Chapter 18 and Chapter 39.

Study 26: Corak's Intergenerational Mobility — Beyond the Curve

Researcher: Miles Corak Year: 2013 and subsequent Published in: Journal of Economic Perspectives; NBER Working Papers

Core finding: Beyond the cross-national Great Gatsby Curve, Corak documents that intergenerational mobility in the United States is lower among children born to parents in the lowest quintile than commonly assumed, and that specific geographic regions within the US have mobility rates similar to Denmark while others have mobility rates similar to developing countries.

Methodology: Analysis of tax records and survey data to track income of parents and children across multiple decades. Chetty et al.'s Opportunity Atlas (2018) extends this work with neighborhood-level granularity.

Key finding detail: The "Opportunity Atlas" work by Chetty and colleagues reveals that growing up in certain zip codes within the same city produces dramatically different lifetime earnings for otherwise similar children. Geography (a pure constitutive luck factor) has measurable, large effects on life outcomes.

Replication status: Findings robust and replicated using multiple data sources by multiple research teams.

Relevance to luck science: Reinforces Chapter 18's structural luck analysis with geographic granularity. "Where you were born" is not just a sociological observation but a measurable predictor of lifetime income with documented mechanisms.

Study 27: Audit Study on Gender and Hiring (Goldin & Rouse)

Researchers: Claudia Goldin, Cecilia Rouse Year: 2000 Published in: American Economic Review

Core finding: When major symphony orchestras adopted blind auditions (auditioning behind screens so evaluators couldn't see who was playing), the fraction of women advancing to the final round and being hired increased significantly.

Methodology: Compared outcomes of orchestra auditions with and without screens across multiple orchestras over multiple decades. Used a natural experiment: different orchestras adopted screens at different times, allowing before-after comparisons.

Key finding detail: Blind auditions increased the probability of a woman being advanced by about 50%. The effect was larger in final rounds (where decisions about hiring are made) than in preliminary rounds. This suggests that implicit gender bias, not ability differences, was driving the gender gap.

Replication status: Highly influential; partially replicated and extended by subsequent audit studies in other domains. Some subsequent reanalyses of the Goldin-Rouse data have raised questions about specific effect size estimates, but the general finding that blind evaluation reduces gender bias has broad support.

Relevance to luck science: Definitive demonstration that gender (a constitutive luck factor) affects professional outcomes through bias mechanisms — and that structural redesign (blind auditions) can partially counteract this luck effect. Directly relevant to Chapter 18 and Chapter 39.

For full citations, consult the Bibliography (Appendix 3). For methods context — including how to evaluate replication status, effect sizes, and the limits of any of these studies — see Appendix B: Research Methods Primer.