In 2009, Barbara Fredrickson — Kenan Distinguished Professor of Psychology at the University of North Carolina at Chapel Hill, one of the most cited researchers in positive psychology, principal investigator of a multimillion-dollar research lab — published a trade book titled Positivity: Top-Notch Research Reveals the 3 to 1 Ratio That Will Change Your Life (Crown, 2009). The book was built around a number she had co-derived four years earlier in the field’s flagship journal: 2.9013. According to the book, this was the minimum ratio of positive to negative emotions a person needed in order to “flourish” rather than “languish.” Fredrickson wrote, in a sentence that would later be quoted back at her with some frequency: “Just as zero degrees Celsius is a special number in thermodynamics, the 3-to-1 positivity ratio may well be a magic number in human psychology.”

The number had four decimal places. It came from a paper Fredrickson had co-authored with Marcial Losada, a Chilean business consultant, titled “Positive Affect and the Complex Dynamics of Human Flourishing,” published in American Psychologist in October 2005 (DOI: 10.1037/0003-066X.60.7.678). The paper’s centerpiece was a set of differential equations — the same Lorenz equations that produce the famous “butterfly” strange attractor in nonlinear dynamics. Fredrickson and Losada claimed these equations, applied to the dynamics of human emotion, generated a precise mathematical threshold separating flourishing from languishing. Above 2.9013, flourishing. Below, languishing. There was also an upper limit at 11.6346 beyond which positivity allegedly became destabilizing.

The paper was a sensation. By January 2014, it had accumulated more than 320 citations. The 3:1 ratio became a fixture of corporate leadership training, executive coaching, wellness programs, and the kind of TED-talk-and-airport-bookstore science that flows from psychology departments into the broader management consulting economy. The math made it sound rigorous. The decimal places made it sound precise. The Lorenz attractor — which any educated reader vaguely associated with chaos theory and “the butterfly effect” — made it sound like physics.

Then, in July 2013, American Psychologist published a paper titled “The Complex Dynamics of Wishful Thinking: The Critical Positivity Ratio” by Nicholas J. L. Brown, Alan D. Sokal, and Harris L. Friedman (DOI: 10.1037/a0032850). It was a 13-page demolition. It demonstrated, in detail, that the Lorenz equations had been misapplied; that the parameter values had been borrowed from fluid dynamics with no theoretical justification; that the precise 2.9013 number had no actual derivation from the model; that the data did not meet the basic requirements (continuous variables evolving smoothly over time) for the kind of differential-equation modeling the paper purported to perform; and that the entire mathematical apparatus was, in Brown, Sokal, and Friedman’s careful phrasing, “entirely unfounded.”

Two months later, American Psychologist issued a partial retraction. The journal formally withdrew the mathematical modeling component of the 2005 paper, including the specific ratios of 2.9 and 11.6. Fredrickson published a response defending her broader empirical research on positive emotions while conceding that she had “neither the expertise nor the insight” to defend Losada’s mathematical model. Losada himself, contacted by The Chronicle of Higher Education, declined to defend the paper and indicated he was too busy with consulting.

This article walks through what was claimed, what the math actually was, what happened in 2013, and — most importantly for strategists who routinely encounter “research-backed” precise numerical claims in soft-science contexts — what pattern of red flags would have flagged this one years earlier.

What Fredrickson & Losada 2005 Claimed

The 2005 American Psychologist paper proceeded in two registers, which is part of why it took so long to be challenged.

In its empirical register, the paper reported data from two samples — one of college students (N = 188) tracking daily emotional experience over 28 days, and a second sample of community adults (N = 86) doing the same. Each participant’s daily ratio of positive to negative emotions was computed. Participants were also categorized as “flourishing” or “non-flourishing” based on a separate psychological well-being measure (Keyes’ Mental Health Continuum). The empirical finding: participants categorized as flourishing had higher mean positivity ratios (around 3.2:1) than non-flourishing participants (around 2.3:1). This is the kind of correlation between two self-report measures that fills the social-psychology literature, and on its own would have been a perfectly ordinary, publishable, modestly interesting finding.

What made the paper extraordinary was its mathematical register. Fredrickson and Losada claimed that the boundary between flourishing and non-flourishing was not merely an empirical regularity but a mathematically derivable threshold — derived from a nonlinear dynamics model that Losada had developed for studying business team performance. The model, they wrote, was an application of the Lorenz equations:

dx/dt = σ(y − x) dy/dt = x(r − z) − y dz/dt = xy − bz

These are the equations Edward Lorenz derived in 1963 to model atmospheric convection — the equations whose strange attractor in three-dimensional phase space gave rise to the term “butterfly effect” and to the broader popular fascination with chaos theory. Fredrickson and Losada repurposed the three variables x, y, z to represent quantities of emotional and interpersonal experience, and claimed that running the model forward generated a phase-space trajectory that bifurcated — that is, qualitatively changed behavior — at a positivity ratio of exactly 2.9013. Below that ratio, the system collapsed into a fixed point (languishing). Above it, the system entered the chaotic butterfly attractor (flourishing). At a still higher value of 11.6346, the system became unstable and the trajectory diverged.

The paper’s verbal claim was that these equations described “the inner workings of the affective system that distinguishes flourishing from languishing.” The number 2.9013 was presented not as a fitted parameter but as a mathematical consequence of the equations themselves — analogous to deriving π from the geometry of a circle.

What The Math Was Supposed To Show

To make the claim plausible to a reader of American Psychologist, Fredrickson and Losada relied on a particular reader’s mathematical sophistication being precisely calibrated: high enough to be impressed by the differential equations, low enough not to actually work through them.

The reader was supposed to understand that:

  1. Complex human systems are non-linear, and so should be modeled with non-linear dynamics rather than linear regression.
  2. The Lorenz equations are the canonical example of a non-linear dynamical system that produces qualitatively different behavior in different parameter regimes.
  3. A phase transition in such a system can produce an exact mathematical threshold — a number with no error bars, derived from the equations themselves.
  4. Therefore, the 2.9013 ratio was not a statistical estimate but a mathematical fact about the system being modeled, more akin to the boiling point of water at standard pressure than to a regression coefficient.

This framing was rhetorically powerful. It is also the framing that, when subjected to actual mathematical scrutiny by people who knew what they were looking at, fell apart entirely.

The trouble was that almost no one with the relevant mathematical training was reading American Psychologist, and the people who were reading it were not, on the whole, equipped to evaluate whether Losada’s purported application of the Lorenz equations had any of the mathematical properties he claimed. The paper was reviewed by psychologists. It was cited by psychologists. It influenced corporate training programs designed by HR consultants. The number 2.9013, with its spurious four-decimal precision, propagated through eight years of citations, leadership development workshops, and the popular science press essentially unchallenged.

What The Math Actually Was: Brown, Sokal & Friedman 2013

Nick Brown was, in 2011, a 50-something British graduate student enrolled in a part-time master’s program in applied positive psychology at the University of East London. He came to the program with no background in psychology but a long career in IT and a sharp ear for claims that sounded too tidy. When he encountered the Fredrickson-Losada paper as required reading, the four-decimal precision of “2.9013” bothered him in the way it would bother any reader from a quantitative discipline who knew that real-world social science estimates do not usually have four significant figures of precision.

Brown began working through the actual mathematics. He found problems quickly and brought them to Harris L. Friedman, a psychologist at the University of Florida known for editorial independence in the positive-psychology subfield. Friedman in turn brought in Alan Sokal, the New York University physicist most famous for the 1996 “Sokal hoax” in which he published a deliberately nonsensical postmodernist parody, “Transgressing the Boundaries,” in the cultural studies journal Social Text. Sokal had spent the subsequent two decades arguing publicly against the misuse of mathematical and scientific terminology in soft-science writing. The Fredrickson-Losada paper was, for him, an unusually pure specimen of the genre he had been criticizing.

The Brown-Sokal-Friedman critique (DOI: 10.1037/a0032850) identified errors at multiple levels.

Misapplication of the Lorenz equations. The Lorenz equations describe a specific physical system: a fluid layer heated from below and cooled from above (Rayleigh-Bénard convection). The variables x, y, z and the parameters σ, r, b have specific physical meanings in that context — convective velocity, temperature differences, vertical temperature variation. When Losada appropriated the equations for emotional dynamics, he assigned new meanings to the variables (such as “inquiry,” “advocacy,” “positivity/negativity ratio”) without any derivation showing why these psychological quantities would obey the same differential relationships as fluid convection variables. Brown, Sokal, and Friedman pointed out that there is no general principle by which arbitrary differential equations transfer between physical and psychological domains; the mathematical structure of a model is only as meaningful as the physical or causal mechanism it represents. The equations had been borrowed wholesale, with the variables relabeled, and no argument provided for why the relabeled system should describe anything at all.

Arbitrary parameter values. The parameters σ = 10, b = 8/3, r = 28 that produce the famous butterfly attractor in the Lorenz system are derived from specific physical properties of atmospheric convection. Losada used precisely these values — the canonical “Lorenz” parameters from the 1963 paper — for his emotional dynamics model. Brown, Sokal, and Friedman noted that there was no justification given for why the parameter values appropriate for atmospheric fluid convection would be the correct values for human emotion. Different parameter values produce qualitatively different dynamics, and almost any threshold one liked could be generated by choosing parameters appropriately. The 2.9013 number was a consequence not of psychology but of Losada’s having borrowed Lorenz’s atmospheric parameters and run the equations forward.

No actual derivation of 2.9013. Brown, Sokal, and Friedman worked through the model and noted that the specific number 2.9013 was never actually derived in either the Losada (1999) paper or the Fredrickson-Losada (2005) paper. The papers asserted the threshold; they did not show the calculation. When Brown attempted to reconstruct the derivation, no consistent calculation produced the asserted number. The four-decimal precision was not the result of a calculation more precise than its inputs warranted; it was a number with no documented source, presented with a precision that obscured its absence of derivation.

The data did not satisfy the model’s requirements. Differential-equation modeling of the kind Losada attempted requires continuous variables evolving smoothly over time — i.e., a dynamical system whose state can be observed at arbitrarily fine time resolution. The actual data Losada had worked from in his 1999 business-teams paper consisted of categorical codings of speech acts (statements coded as “inquiry,” “advocacy,” “positive,” “negative,” “other”) in 60-minute meetings. These are discrete, categorical, and sparse — exactly the kind of data for which differential equations are an inappropriate modeling framework. The Fredrickson-Losada 2005 paper inherited this incongruity without acknowledging it.

The “butterfly” was a picture, not data. The Fredrickson-Losada paper included a figure showing the canonical Lorenz attractor — the famous butterfly shape — and presented it as a visualization of the dynamics of high-performing teams or flourishing individuals. Brown, Sokal, and Friedman noted that this figure was simply a computer simulation of the Lorenz equations with the standard atmospheric parameters. It was not a plot of any actual emotional or interpersonal data. The same picture would appear in any introductory chaos-theory textbook.

The Brown-Sokal-Friedman paper closed with a recommendation that future researchers exercise far greater caution when borrowing advanced mathematical apparatus from other fields, and that journal editors require reviewers competent in the mathematical methods being applied. Their conclusion on the specific claim: the proposition that a critical positivity ratio of 2.9013 exists is, as a mathematical fact about Losada’s stated model, “entirely unfounded.”

The 2013 American Psychologist Reckoning

The Brown-Sokal-Friedman paper went online at American Psychologist on July 15, 2013. Within days it had been written up in Discover magazine (by the blogger Neuroskeptic, under the headline “Death of a Theory”), in The Chronicle of Higher Education, in the British Psychologist magazine (where Adam Anderson’s piece “The Pernicious Positivity Ratio” appeared in August 2013), and across the science-blogging ecosystem. Sokal’s involvement gave the story unusual reach — the framing in much of the press was that this was, in spirit, a second Sokal affair: a physicist demonstrating that a celebrated piece of soft-science work was mathematically incoherent.

In September 2013, American Psychologist published a formal retraction notice. Its key language: “the modeling element of this article is formally withdrawn as invalid and, along with it, the model-based predictions about the particular positivity ratios of 2.9 and 11.6.” The notice preserved as “valid and unaffected” the empirical correlation between positivity ratios and well-being measures in the two samples.

The retraction was partial in a specific and consequential way. The journal did not retract the entire 2005 paper. It retracted only the mathematical modeling — the differential equations, the derivation of the precise thresholds, the appeal to nonlinear dynamics. The empirical descriptive statistics survived. This compromise was contested. Brown, Sokal, and Friedman pointed out in a 2014 follow-up that without the mathematical model, the empirical data alone provided no evidence whatsoever for a critical threshold — they showed only that, on average, people categorized as flourishing reported more positive emotions than people categorized as non-flourishing, which is approximately what “flourishing” means.

Fredrickson’s Response

In the same September 2013 issue of American Psychologist, Fredrickson published “Updated Thinking on Positivity Ratios” (DOI: 10.1037/a0033584). The response was unusual in tone. Fredrickson conceded the mathematical critique substantially:

  • She accepted that the Losada nonlinear-dynamics model could not be defended.
  • She acknowledged that she had “neither the expertise nor the insight” to evaluate the mathematical claims when she had originally co-authored the paper, and that she had relied on Losada’s representation of the model’s validity.
  • She removed the chapter on Losada’s mathematical model from subsequent editions of her trade book Positivity.

She defended, however, the broader empirical claim that higher positivity ratios are predictive of better mental health outcomes — pointing to a body of subsequent empirical work by herself and others. She did not endorse the specific 2.9013 or 11.6346 thresholds.

Marcial Losada, when contacted by The Chronicle of Higher Education, indicated that he was “very busy with [his] consulting work” and declined to mount a substantive defense of the model. He did not publish a response in American Psychologist.

In 2014, Brown, Sokal, and Friedman published a further response in the same journal titled “The Persistence of Wishful Thinking” (DOI: 10.1037/a0034962). Their position: Fredrickson’s defense of “a positivity ratio” without the precise threshold conflated two distinct claims. That higher ratios correlate with well-being measures is a correlation between two operationalizations of similar constructs, not evidence of a discoverable mathematical regularity. The continued use of the phrase “positivity ratio” in the post-retraction literature, they argued, traded on the apparent rigor of the now-discredited mathematical work.

What’s Honest To Say About Positive Psychology Now

It would be a mistake to extrapolate from the Fredrickson-Losada episode to the conclusion that all of positive psychology is bunk. The actual landscape, post-2013, is more textured.

Fredrickson’s broaden-and-build theory of positive emotions — the broader theoretical framework that long predates the 3:1 ratio and is independent of the Losada model — describes positive emotions as expanding momentary thought-action repertoires and building enduring psychological resources. The theory has accumulated independent empirical support, including some replicated effects. It does not require any precise mathematical threshold to be useful. Stripped of the Losada apparatus, it is a defensible mid-level theory about why positive affect matters.

Specific findings about positive emotions and outcomes — that gratitude interventions can modestly improve well-being, that positive affect predicts some health markers, that experiences of positive emotion correlate with broader well-being — survive in attenuated form. Effect sizes are typically smaller than the initial enthusiastic literature suggested. The Open Science Collaboration’s psychology replication project (Science, 2015) found that positive-psychology effects fared no better and no worse than the field as a whole, which is to say that roughly half the high-profile findings did not replicate at the original effect size.

The precise mathematical claims do not survive. There is no critical positivity ratio of 2.9013. There is no upper limit of 11.6346. There is no Lorenz-attractor dynamics underlying human emotional life. The 3:1 number that propagated through corporate training and self-help, presented as a scientifically derived threshold, is — as a mathematical claim — invalid.

Calibration matters. The honest claim is something like “positive emotions, in moderation and frequency, are associated with various good outcomes, with effect sizes that are modest and context-dependent.” This is true and useful. It is also a much weaker claim than “science has discovered the magic number 2.9013.”

What This Means For Strategists Evaluating Quantitative Behavioral Claims

The Fredrickson-Losada episode is unusually clean as a teaching case because the source of the error was so specific. The claim was false in a way that, in retrospect, was readable from the surface features of the original paper. The strategist or executive who routinely encounters “research-backed” numerical claims in management, marketing, leadership, and behavioral design contexts can extract several pattern-recognition heuristics from it.

Heuristic 1: precise non-round numbers in soft-science claims warrant scrutiny. “2.9013” is suspicious in a way that “about 3” is not. Four-decimal precision implies that something has been measured or derived to that precision. In social science, this is almost never warranted. Real estimates have error bars; theoretical thresholds derived from well-established mathematical models can be precise, but the underlying model needs to be well-established. When a precise non-round number appears in a behavioral claim — especially one being marketed in trade-book or consulting contexts — ask: where does the precision come from? If the answer is “from a mathematical model,” the next question is whether the model is itself defensible.

Heuristic 2: appeals to “complex dynamic models,” “nonlinear systems,” or “chaos theory” in psychology should be treated with elevated skepticism. These terms are technically meaningful in physics, applied mathematics, and engineering, where they refer to specific kinds of differential-equation systems with rigorously characterized behaviors. They are also enormously popular as rhetorical decoration in soft-science writing, where they typically signal little more than “this system is complicated.” The technical apparatus of nonlinear dynamics requires that you have a defensible model — derived from causal mechanism, not analogically borrowed from another discipline — whose variables and parameters have empirical content. When you encounter Lorenz equations, strange attractors, or phase-space language in a psychology or management context, ask: is this a derived model, or a rebranded one?

Heuristic 3: the gap between empirical correlation and mathematical theorem is wider than it looks. The empirical data in the Fredrickson-Losada 2005 paper showed a modest correlation between positivity ratios and a well-being categorization. This is a perfectly ordinary social-science finding. The mathematical apparatus was layered on top to upgrade the claim from “associated with” to “mathematically derived.” Any time you see this pattern — observed correlations being elevated to mathematical laws via an appended model — the relevant question is whether the model adds explanatory power or merely the appearance of rigor.

Heuristic 4: prestige institutional venues do not eliminate this class of error. The Fredrickson-Losada paper was published in American Psychologist, the flagship journal of the American Psychological Association. It was peer-reviewed. It was cited 320+ times before the critique appeared. The mathematical errors were present from publication; they simply went undetected for eight years because the reviewers, citers, and readers were not equipped to detect them. Institutional vetting reduces some classes of error but does not eliminate the class of error that arises when methodological sophistication required to evaluate a claim exceeds the methodological sophistication of the evaluators.

Heuristic 5: when a researcher concedes the math but defends “the broader research program,” ask what specifically survives. Fredrickson’s 2013 response retained the rhetorical structure of the original claim (positivity ratios matter) while abandoning the specific basis on which the claim had been distinguished from a routine correlation. This is a common pattern in post-retraction defenses across many fields. The substantive question is not whether the researcher continues to defend the work but what specific empirical or theoretical content survives the critique. In this case, the answer was: a modest correlation between positivity and well-being, of the kind that the field has known about for decades, with effect sizes that do not justify the marketing.

The Bigger Pattern: Corporate Training Built On “Science-y” Frameworks

The Fredrickson-Losada case sits in a larger ecosystem of corporate training and executive development frameworks that lean on the rhetorical authority of science without the substantive basis. The 3:1 ratio is one specimen of a recognizable type:

  • The Mehrabian 7-38-55 rule (covered separately in this hub) — the claim that communication is 7% words, 38% tone, 55% body language, derived by extrapolating from two narrow lab studies on emotional inconsistency, marketed as a universal communication principle.
  • Personality “types” (MBTI, DISC, Enneagram) — discrete-category frameworks with weak test-retest reliability and limited predictive validity, marketed as scientifically derived insight into individual differences.
  • Leadership “archetypes” — frameworks of 4, 6, 9, or 12 categorical leader types, typically derived from clustering analyses on self-report data that have not been independently validated, marketed as discovered structure rather than analytic convenience.
  • Precise productivity ratios — claims like “the 80/20 rule applies to your team’s output” or “the 70-20-10 model of learning,” presented as discoveries rather than as approximate heuristics with weak empirical basis.
  • Brain-based learning frameworks — content built on long-debunked claims about right-brained vs. left-brained thinking, “learning styles” (visual/auditory/kinesthetic), or “neuroplasticity” as a vague invocation rather than a specific mechanism.

What these share with the 3:1 positivity ratio is a particular rhetorical structure: a precise-looking numerical or categorical claim, attached to an appeal to scientific authority, marketed in a trade or consulting context where the underlying methodology is not subjected to the kind of scrutiny it would face within the academic discipline that produced it. None of these is necessarily worthless — most started from real, if modest, empirical observations. What is unreliable in each case is the precise quantitative or categorical structure that has been overlaid for marketability.

For a strategist allocating training budget, leadership development time, or assessment tools, the operational implication is not “ignore behavioral science.” It is “calibrate the strength of any specific claim to the strength of the underlying evidence.” A claim like “positive interactions tend to outweigh negative ones in healthy relationships and teams” is defensible. A claim like “the ratio must be at least 2.9013 to 1” requires the model that Brown, Sokal, and Friedman demonstrated does not exist. The first is a useful generalization; the second is a marketing fiction.

The general lesson — applied across the half-dozen frameworks above and the dozens more that proliferate in the corporate training market — is that the precision of a number is not, by itself, evidence of its truth. In domains as noisy as human behavior, a number that looks too precise to be true probably is.

Sources

  • Fredrickson, B. L., & Losada, M. F. (2005). Positive affect and the complex dynamics of human flourishing. American Psychologist, 60(7), 678–686. DOI: 10.1037/0003-066X.60.7.678
  • Brown, N. J. L., Sokal, A. D., & Friedman, H. L. (2013). The complex dynamics of wishful thinking: The critical positivity ratio. American Psychologist, 68(9), 801–813. DOI: 10.1037/a0032850. Open-access preprint: arXiv:1307.7006
  • Fredrickson, B. L. (2013). Updated thinking on positivity ratios. American Psychologist, 68(9), 814–822. DOI: 10.1037/a0033584
  • Brown, N. J. L., Sokal, A. D., & Friedman, H. L. (2014). The persistence of wishful thinking. American Psychologist, 69(6), 629–632. DOI: 10.1037/a0034962. Open-access preprint: arXiv:1409.4837
  • Anderson, A. (2013). The pernicious positivity ratio. The Psychologist (British Psychological Society), August 2013.
  • Losada, M. (1999). The complex dynamics of high performance teams. Mathematical and Computer Modelling, 30(9–10), 179–192. DOI: 10.1016/S0895-7177(99)00189-2
  • Fredrickson, B. L. (2009). Positivity: Top-Notch Research Reveals the 3 to 1 Ratio That Will Change Your Life. New York: Crown Publishers.
  • “Fredrickson-Losada ‘positivity ratio’ paper partially withdrawn.” Retraction Watch, September 19, 2013. retractionwatch.com
  • Neuroskeptic. “Positivity Ratio Criticized In New Sokal Affair.” Discover Magazine, July 16, 2013.
  • Correction to Fredrickson and Losada (2005). American Psychologist, 68(9), 822. (Formal retraction notice.)
  • Friedman, H. L., & Brown, N. J. L. (2018). Implications of debunking the “critical positivity ratio” for humanistic psychology: Introduction to special issue. Journal of Humanistic Psychology, 58(3), 239–261. DOI: 10.1177/0022167818762227

Frequently Asked Questions

Is positive psychology bunk?

No. Positive psychology as a research program has produced genuine, if modest, findings about the role of positive emotions, character strengths, gratitude practices, and meaning-making in well-being. The Fredrickson-Losada 3:1 ratio is a specific bad claim within a broader research program that includes both defensible findings and other questionable ones. Treating the whole field as bunk would be as much a calibration error as accepting the 3:1 ratio uncritically. The honest summary is that positive psychology is a normal research program with normal mixed results, presented in popular culture with unusual confidence and unusual marketing intensity.

What about Fredrickson’s other work?

Fredrickson’s broaden-and-build theory of positive emotions — distinct from the Losada model — has accumulated independent empirical support and remains a serious framework in emotion research. Some of her specific findings (e.g., on gratitude interventions, on the experience of distinct positive emotions) have replicated; others have not. Her status as a leading figure in positive psychology is not undermined by the Fredrickson-Losada retraction; what is undermined is the specific claim that there exists a precise mathematical threshold for human flourishing.

What about the broaden-and-build theory specifically?

Broaden-and-build proposes that positive emotions expand a person’s momentary thought-action repertoire (broaden) and, over time, build enduring psychological, social, and intellectual resources (build). This is a mid-level theoretical claim that has been tested in dozens of studies with mixed but generally supportive results. It does not require any precise mathematical threshold to be useful. It is independent of the Lorenz-equation model that was retracted. Stripped of the 3:1 marketing apparatus, it remains a serviceable framework for thinking about why positive affect matters in domains beyond mere subjective enjoyment.

Should I distrust any leadership framework that includes specific numbers?

Not automatically — but the question to ask is where the numbers come from. Numbers derived from large, well-conducted studies with replicated effects (e.g., effect sizes from meta-analyses) are reasonably trustworthy as approximate guides. Numbers presented as precise mathematical thresholds derived from theoretical models should be treated with elevated skepticism unless the model itself is well-established and the derivation is transparent. As a quick heuristic: if the number has more decimal places than the underlying data would warrant, and if the model’s “derivation” is not actually shown in the paper, the number is probably marketing rather than mathematics.

Did the partial retraction actually fix anything?

Partially. The 2013 retraction removed the specific 2.9013 and 11.6346 thresholds from the formal scientific record, and Fredrickson removed the Losada chapter from later editions of her book. However, the broader corporate training ecosystem that had built materials around the 3:1 ratio largely did not retract or update its content. As of the mid-2020s, the “Losada ratio” still appears in leadership development workshops, executive coaching materials, and wellness program content — typically with no mention of the retraction. The scientific correction was clean; the popular correction has been slow and incomplete.

Why did it take eight years for the math to be challenged?

Two reasons, both instructive. First, the audience that the paper was written for — academic psychologists publishing in American Psychologist — was, on the whole, not trained to evaluate differential-equation models from nonlinear dynamics. The mathematical apparatus was sophisticated enough to be impressive but specialized enough that the relevant experts (applied mathematicians, physicists) were not reading the paper. Second, the empirical register of the paper — the correlation between positivity ratios and well-being — was unobjectionable, and the mathematical register was easy to read past if you trusted the authors’ representation of it. Nick Brown’s eventual challenge required someone with mixed training (a part-time positive-psychology master’s student with a quantitative background) who happened to read the paper with the right mix of credulity toward the empirical claim and skepticism toward the four-decimal precision.

What’s the difference between this and the Mehrabian 7-38-55 case?

Mehrabian’s percentages came from two real but narrow experiments that were over-extrapolated to a universal communication principle by other people, often without Mehrabian’s consent or correction. The 3:1 positivity ratio was different in kind: the precise mathematical threshold was actively derived (or purported to be derived) by the original authors from a stated mathematical model, marketed by them in trade books and TED talks, and incorporated by them into a multimillion-dollar consulting and coaching ecosystem. The 7-38-55 case is primarily a story about misattribution; the 3:1 case is primarily a story about pseudo-mathematical apparatus being treated as a rigorous derivation when it was not.

What should I take away from this if I’m not a researcher?

The operational lesson, applied to any “science-backed” claim with a precise number attached: ask where the number comes from, and how. If the answer involves a model you cannot evaluate yourself, ask whether the people who could evaluate it have done so. If you can’t tell whether they have, assume some discount on the precision until you can. The four-decimal “2.9013” should always have been treated as suspicious; the lesson of the 2013 retraction is that, in soft-science marketing, precision is more often a rhetorical device than a measurement.

replication-crisis positive-psychology fredrickson-losada pseudoscience evidence-evaluation
Share this article
LinkedIn (opens in new tab) X / Twitter (opens in new tab)
Atticus Li

Experimentation and growth leader. CXL-certified CRO practitioner, Mindworx-certified behavioral economist (1 of ~1,000 worldwide). 200+ A/B tests across energy, SaaS, fintech, e-commerce, and marketplace verticals.

About LinkedIn Newsletter