How much of popular psychology holds up?

We rated all 751 concepts in the Shrink Atlas on one honest scale, from strong to mixed to teaching model. 64% rest on strong or well established evidence. 5% are contested, and some of them are ideas people quote as settled fact.

By Shariq Refai, MD, MBA, FAPA, board-certified psychiatrist · Data updated September 29, 2026 · Version 2026-Q3 · Download the data · Cite this report

Key findings

• 64% hold up well. 235 concepts rate strong and 248 rate well established.
• 39 are contested. They've been studied, but the findings don't agree or didn't replicate cleanly. That list includes Decision Fatigue, Growth Mindset, Attention Restoration, Self-Actualization.
• 146 are still emerging. Real research, promising, and early.
• 83 are teaching models, and each of those pages says so.
• Shrink Thinking has the most solid footing (78% strong or well established). Shrink Becoming has the most ideas still in motion (48%).

The whole atlas at a glance

• Strong
• Well established
• Emerging
• Mixed
• Educational model

By realm

The atlas has six realms. Here's how solid the ground is in each one.

Popular ideas that are contested

These are the ones worth a second look before you repeat them. Each has real research behind it, and each has findings that pull in different directions.

• Decision Fatigue Decision fatigue has been studied across cognitive psychology and behavioral science.
• Growth Mindset The core idea is well studied, but effect sizes on grades and performance are modest and contested across large reviews. Brief classroom interventions can help some students, especially those who are struggling, though results vary. Treat it as a useful stance, not a guaranteed lever.
• Attention Restoration Attention restoration theory has moderate support, with studies linking natural settings to improved focus. Effect sizes and mechanisms are still debated. The broad idea that gentle rest helps focus is reasonable and useful.
• Self-Actualization Self-actualization is an influential humanistic idea, but the need hierarchy it sits within hasn't held up well in empirical tests.
• Loss Aversion It's foundational and widely cited, but recent work debates its size and generality, especially for small or routine stakes.
• Mental Toughness Mental toughness is popular in sport psychology but suffers from inconsistent definitions and heavy overlap with resilience and personality traits. Some components relate to performance, yet the construct as a whole is contested. Treat it as a useful umbrella idea rather than a precise measured trait.
• Core Affect The two-dimensional view of feeling is well supported in emotion research and shows up across many studies. Whether it's the true bedrock of all emotion is still debated, so treat core affect as a strong working model rather than a proven mechanism.
• Grit Grit predicts some outcomes such as persistence and achievement, but effects are modest and it overlaps strongly with conscientiousness. Critics question how much it adds beyond established traits. Treat it as a helpful lens with honestly limited and debated evidence.
• Compassion Fatigue Compassion fatigue is studied mainly in caregiving and helping professions with moderate support. It's a descriptive concept rather than a formal diagnosis. Terminology and boundaries with related ideas like burnout are still debated.
• Priming Semantic priming is robust, but many social and behavioral priming effects haven't held up in large replication attempts.
• Willpower Self control matters for outcomes, but the specific willpower as depletable fuel model has struggled to replicate consistently. The safer reading is that structure and habit reduce the need for willpower. Treat the strength model as contested and the prevention approach as sound.
• Post-Traumatic Growth Post-traumatic growth is widely reported, but self-report and retrospective measures may not match actual measured change.
• Present Bias Intertemporal choice studies consistently show people favor smaller sooner rewards over larger later ones in a present-weighted way. The pattern is robust, though how best to model it's debated. It's a well-supported feature of human choice.
• Embarrassment Embarrassment is recognized as a distinct self-conscious emotion with characteristic displays and a role in smoothing social slips. Its boundaries with shame are somewhat debated, so the distinction is well supported in outline while some details remain open.
• Envy Envy is a recognized emotion rooted in social comparison, and research distinguishes benign and malicious forms with different effects. The two-form model has support but is still debated, so the distinction is moderately, not firmly, settled.
• Altruism Altruism is well studied, and empathy driven helping has solid support, though the existence of purely selfless motivation remains philosophically and empirically debated. The behaviors are real and measurable. Interpretations of the underlying motives vary across researchers.
• Ego Depletion Ego depletion was influential but has faced serious replication problems, with large multi lab studies finding small or null effects. Belief and motivation appear to shape any results seen. Present it honestly as a contested model, not a reliable fact, and a case study in scientific self correction.
• Nature and Restoration A large body of research associates nature exposure with attention restoration and reduced stress, supported by attention restoration theory. Effects are typically modest and vary by context and person. The overall pattern is fairly consistent even if mechanisms are still debated.
• Productive Failure Productive failure has supportive meta-analytic evidence, but it's actively debated against direct instruction and results depend on design.
• Reconsolidation Reconsolidation is well supported in animal studies, but human findings and clinical applications have had replication problems.
• Yerkes-Dodson Law The original work was limited animal research, and the tidy inverted U curve hasn't held up consistently across human studies.
• Dunning-Kruger Effect The original studies are frequently cited, but the standard interpretation is contested, and some researchers argue parts of the pattern can arise from statistical artifacts and regression to the mean. It's best treated as a useful educational model and a caution about self-assessment rather than a settled law.
• Actor-Observer Bias It's a classic attribution idea, but meta-analytic work found the actor-observer asymmetry is much smaller and less consistent than originally claimed.
• Gottman Four Horsemen These patterns come from decades of observational research coding real couple conversations, where their frequency correlated with later divorce. The work is influential but based on specific samples and lab observation, and some prediction claims have been debated.
• Grit and Purpose Grit predicts some outcomes but studies show it overlaps heavily with conscientiousness and its unique effect is often small. The link between purpose and sustained motivation is supported across several lines of research. Combine the two carefully, since grit's independent power is debated.
• Zeigarnik Effect Early interrupted task findings haven't replicated consistently, and effect sizes in later work are modest.
• Cognitive Forcing Strategy Cognitive forcing strategies are well described, but studies testing them in clinicians have shown limited or inconsistent benefit.
• Basic Emotions The basic emotions view has strong historical influence and some cross-cultural support, especially for facial expressions. It's also seriously contested, with critics arguing culture and construction play a larger role, so the evidence is genuinely mixed.
• Discrete Emotions Whether emotions are distinct natural kinds is actively contested, with basic emotion and constructionist accounts disagreeing on the evidence.
• Constructed Emotion The theory of constructed emotion is influential and generates active research, and it fits some findings that trouble older views. It remains contested and far from settled, so treat it as a leading model under debate rather than established fact.
• Visualization Process focused imagery tends to help, while outcome focused positive fantasy can reduce effort, so results depend on the type of visualization.
• Fixed Mindset The distinction between fixed and growth beliefs is well described, but its measured impact on real outcomes is modest and debated. Fixed beliefs tend to correlate with avoidance and worse responses to setbacks, though people rarely hold one mindset across all domains. Treat it as a tendency, not a fixed diagnosis.
• Self-Actualization Needs Maslow's hierarchy is influential, but its strict ordering has little empirical support and findings conflict.
• Signature Strengths The VIA classification is widely used and studies link signature strength use to wellbeing and engagement. Effects are generally modest and some findings come from short interventions. The framework is respected but still debated on structure and measurement.
• Moral Licensing Experiments across consumer, ethical, and social domains show licensing effects, though some studies find them small or fail to replicate. The direction is fairly consistent while the magnitude is debated. Treat it as a real but modest and context-dependent effect.
• Somatic Marker Hypothesis An influential neuroscience theory with supportive evidence, though parts of it remain debated and not fully settled.
• Love Languages A popular framework with wide appeal but limited and mixed empirical support for its specific claims.
• Clinical Debiasing Debiasing strategies are widely discussed, but trials of their effect on clinical accuracy have produced inconsistent results.
• Self-Complexity The idea that self complexity buffers stress hasn't replicated consistently, and reviews report mixed and sometimes opposite effects.

The best supported ideas

The most connected concepts in the atlas with a strong evidence label. If you only learn a dozen ideas about how minds work, these are good ones.

Emerging, worth watching

Real constructs with growing evidence. Expect some of these to move as the research matures.

Vulnerability Self-Authorship Contentment Savoring Authenticity Micro-Recovery Repair Attempts Bids for Connection Active Recovery Character Strengths Identity Integration Awe Task Initiation Grounding Self-Transcendence Attentional Residue Self-Continuity Emotional Flooding Emotional Vocabulary Wisdom Healthy Dependence Social Baseline Theory Thought Labeling Intellectual Humility and 122 more in the dataset

Teaching models, labeled as such

These help explain real science, and they aren't research constructs on their own. We label them so nobody mistakes a metaphor for a finding.

Rest Boundaries Window of Tolerance Nervous System Regulation Emotional Safety Deliberate Rest Cognitive Humility Recovery Rituals Mental Forecasting Peak Performance Mental Clutter Habit Loop Self-Trust Feedback Loops in Skill Analysis Paralysis Cognitive Overload Adaptive Thinking Mental Availability Expected Value Prioritization Emotional Maturity Pacing Energy Management Self-Improvement Trap and 59 more in the dataset

How we rated

Every concept in the atlas carries exactly one label from the scale above, and the label shows on the concept page itself. Ratings are set against a written rubric and approved by Shariq Refai, MD, MBA, a board-certified psychiatrist. We rate how settled the research is for the idea as described, not for any treatment built on it. When a name is our own shorthand for a broader body of research, we label it an educational model even if the science underneath is solid.

We don't cite individual studies in the ratings themselves, and we never invent one. Where the evidence moves, the label moves, and the change shows up in the next quarterly snapshot. Our full approach is in our evidence standards.

What this report isn't

It isn't a systematic review or a meta analysis, and it isn't medical advice. One psychiatrist sets the final label, which keeps the scale consistent and also means reasonable experts might place a few ideas one step higher or lower. If you think we've got one wrong, tell us and we'll look at it.

How it stays current

This page reads the live data, so the numbers are always current. Each quarter we save a snapshot, recheck a quarter of the atlas against current research, and post what changed. Every concept is rechecked at least once a year, and each January brings a new edition.

Download the data

The full dataset is free under CC BY 4.0. Each row has the concept, its realm, its label, and a link to the concept page.

Download CSV Download JSON

Cite this report

Refai, S. (2026). How much of popular psychology holds up? The ShrinkDaily evidence report (2026 ed.). ShrinkDaily. https://shrinkdaily.com/research/evidence-report

Refai, S. (2026). ShrinkDaily evidence ratings [Data set, version 2026-Q3]. ShrinkDaily. https://shrinkdaily.com/research/evidence-ratings.csv

Journalists and researchers can reach Dr. Refai through our contact page.

Questions