2 Psychology IIc

Psychology — Intermediate

Psychology’s Limitations: A Deeper Look

The Replication Crisis

The Bare Essentials caveat introduced the replication crisis briefly. Here we go deeper — not to undermine everything psychology has produced, but to give you the tools to evaluate psychological claims with appropriate precision.

What Happened

In 2011, a large-scale collaborative project called the Reproducibility Project attempted to replicate 100 published psychology studies. The results were striking: only 36% of the studies produced results consistent with the original findings when repeated. A field that had been producing confident, peer-reviewed conclusions was revealed to have a substantial proportion of findings that didn’t hold up under independent scrutiny.

This wasn’t a scandal unique to psychology — similar problems have been found in medicine, nutrition science, and economics. But psychology’s replication crisis was particularly visible and has driven significant methodological reform within the field since then.

Why It Happened

Several factors contributed, most of them structural rather than the result of individual dishonesty:

Publication bias is probably the most significant. Academic journals have historically been far more likely to publish studies with positive results — we found an effect — than null results — we found nothing. This creates a systematic distortion: the published literature overrepresents findings that turned out to be flukes, because the studies that failed to find the same effect never made it into print. Researchers knew this, which created pressure — often unconscious — to find positive results.

Small sample sizes amplify statistical noise. Many classic psychology studies were conducted on surprisingly few participants — sometimes fewer than 30. Small samples produce unreliable estimates, and results that appear statistically significant in a small sample frequently disappear when the study is repeated with more people.

P-hacking refers to the practice of analyzing data in multiple ways until a statistically significant result appears, then reporting only that analysis. This isn’t necessarily deliberate fraud — it can happen through ordinary researcher decisions made without awareness of the cumulative effect. But the result is findings that look more robust than they are.

WEIRD populations compound all of the above. A small, statistically noisy study conducted on undergraduate psychology students at a Western university is being asked to produce conclusions about human psychology universally. The gap between the sample and the claim is enormous.

What Has Changed

The replication crisis prompted genuine reform efforts within psychology, and it’s worth acknowledging them:

Pre-registration — registering a study’s hypotheses and methods before collecting data, making post-hoc adjustment harder — has become increasingly standard.
Open data and open materials requirements at many journals make independent verification more possible.
Larger, multi-site studies with more statistical power are increasingly favored.
Registered replication reports — where multiple labs simultaneously attempt to replicate a finding — provide much stronger evidence than single studies.

These reforms are meaningful. Psychology in 2025 is methodologically more rigorous than psychology in 1985. But the reforms are uneven across the field, and a large body of older research has not been revisited. The textbooks, popular science books, and media coverage that shaped public understanding of psychology are still largely built on the pre-reform literature.

What This Means for You as a Learner

The replication crisis doesn’t mean psychology is useless or that nothing it has found is reliable. It means applying a calibrated skepticism:

Single studies prove little. A finding that has been replicated across multiple independent studies, in diverse populations, using different methodologies, is much more trustworthy than a single compelling experiment — however elegantly designed.
Effect sizes matter as much as statistical significance. A finding can be statistically significant — unlikely to be a fluke — while being so small in practical terms as to be nearly irrelevant to real life. Both pieces of information matter.
Be especially skeptical of surprising, counterintuitive findings that generated a lot of media coverage. These are precisely the findings most likely to have been the product of statistical noise, because surprising results are more publishable and more shareable.
Concepts with strong replication records — like the Big Five personality model, attachment theory, the stress response system, and the basic mechanisms of cognitive bias — can be held with considerably more confidence than findings from single studies.

This isn’t a reason for paralysis or wholesale dismissal. It’s a reason for the same calibrated, evidence-sensitive thinking that Level 2: Critical Thinking develops throughout — applied now to the field you’re currently studying.

Cultural and Demographic Biases in Research

The WEIRD problem was introduced in the Bare Essentials caveat and mentioned again in the Big Five section. Here we examine it properly — because it’s not just a methodological footnote. It shapes which questions get asked, which experiences get studied, and whose psychology gets treated as the default human psychology.

The Shape of the Bias

WEIRD — Western, Educated, Industrialized, Rich, Democratic — describes the population that has historically provided the vast majority of research participants in psychology. A 2008 analysis by Henrich, Heine, and Norenzayan found that Americans alone represented 68% of participants in studies published in top psychology journals, despite being roughly 5% of the world’s population. When the analysis was broadened to WEIRD populations generally, the figure rose to 96%.

This would be less problematic if human psychology were largely uniform across cultures. It isn’t.

Research has found substantial cross-cultural variation in:

Perception: The Müller-Lyer illusion — two lines of equal length appearing unequal due to arrow-like end caps — is one of the most famous demonstrations in psychology. WEIRD populations are among the most susceptible to it. Some non-Western populations show little to no susceptibility at all, suggesting that visual perception itself is shaped by environmental and cultural experience more than previously assumed.
Reasoning styles: WEIRD populations tend toward more analytic reasoning — breaking problems into components, focusing on focal objects, applying rules consistently regardless of context. Many East Asian populations tend toward more holistic reasoning — attending to relationships and context, accepting apparent contradictions, prioritizing the whole over the parts. Neither is superior; they’re different tools.
Self-concept: The distinction between independent and interdependent self-construal — whether you primarily understand yourself as an individual with personal attributes or as a node in a network of relationships — varies substantially across cultures and affects everything from motivation to emotional experience to moral reasoning.
Emotional expression and experience: Which emotions are considered appropriate to express, how emotions are labeled, and even which emotional states are culturally recognized varies across cultures in ways that affect both research findings and clinical practice.

Demographic Biases Within WEIRD Populations

Even within Western psychology, research has systematically underrepresented and misrepresented specific groups:

Women and girls were historically excluded from many research studies — including medical and psychological research — partly out of concern that hormonal variation would complicate results. The consequence is a body of findings assumed to apply universally that was actually generated predominantly from male participants. This has had direct clinical consequences: heart attack symptoms present differently in women than in men, but the male presentation became the default, leading to systematic misdiagnosis.

In psychology specifically, women and girls have been underdiagnosed with conditions like ADHD and autism because diagnostic criteria were developed primarily from research on boys and men, whose presentations tend to be more externally visible. This is the same pattern noted in the Neurodiversity section.

Racial and ethnic minorities have been both underrepresented as research participants and misrepresented in research conclusions. Intelligence research in particular has a deeply problematic history of conflating culturally specific test performance with innate cognitive capacity — a confusion with serious real-world consequences. Diagnostic categories in psychiatry have been applied unevenly across racial lines, with Black patients in particular historically more likely to receive severe diagnoses like schizophrenia and less likely to receive diagnoses like depression that lead to more supportive treatment.

Socioeconomic class shapes psychology in ways that research has only partially examined. Cognitive load research — work on how poverty affects decision-making by consuming mental bandwidth — is one area where class has been taken seriously as a variable, but it remains underexplored relative to its importance.

LGBTQ+ people were pathologized by mainstream psychology well within living memory — homosexuality was listed as a mental disorder in the DSM until 1973. The shift away from this position came partly from research and partly from sustained advocacy by affected communities. It’s a clear example of professional consensus being wrong in ways that caused direct harm, and of that consensus eventually being corrected — partly through exactly the kind of community pushback the program has emphasized throughout.

Who Gets to Ask the Questions

Behind the demographic biases in research participants is a related problem: who conducts the research and what questions they think to ask.

Research priorities are shaped by researcher backgrounds, funding sources, and institutional incentives. A research community that is itself predominantly WEIRD, predominantly white, and historically predominantly male will naturally generate different research questions than a more diverse one — not through malice, but through the ordinary operation of perspective. Questions that are invisible from a position of privilege tend not to get asked.

This is why the “nothing about us without us” principle discussed in the Neurodiversity section matters beyond neurodiversity. It applies to psychological research broadly: the people whose psychology is being studied have knowledge about their own experience that outside observation cannot fully access, and research conducted without their meaningful involvement will have systematic blind spots.

Applying This Critically

As with the replication crisis, the goal here is calibrated evaluation rather than wholesale dismissal:

Ask who was studied. When a psychological finding is presented as universal, check whether the research was conducted on a representative sample or a narrow population. Many findings that hold reliably within WEIRD populations may not generalize.
Pay particular attention to clinical applications. Diagnostic criteria, treatment protocols, and therapeutic models developed on narrow populations and applied universally can cause real harm to people whose experience doesn’t fit the assumed template.
Take seriously the knowledge produced by communities about their own experience. Lived experience is not infallible, but it is a form of evidence — and in areas where formal research has systematic blind spots, community knowledge often fills gaps that the literature hasn’t caught up with yet.
Notice when “universal” actually means “default.” The assumption that WEIRD psychology is human psychology is itself a cultural assumption, and recognizing it as such is the first step toward a more accurate picture.

Medicalisation of Normal Human Experience

The Bare Essentials caveat asked a question worth returning to now: who gets to decide what’s “normal” or what needs “fixing”? Medicalization is what happens when the answer to that question gets captured by professional, institutional, or commercial interests in ways that serve those interests more than the people being classified.

What Medicalization Is

Medicalization is the process by which human experiences, behaviors, and conditions that were previously understood as part of ordinary life — or as social, spiritual, or personal matters — become redefined as medical problems requiring diagnosis and treatment. This isn’t always wrong. Some experiences that were historically dismissed or moralized actually do benefit from a medical understanding: addiction, for instance, is better understood as a condition with neurological and psychological components than as a moral failing.

But medicalization has a consistent tendency to overshoot, and the consequences of overshooting are real: unnecessary treatment, iatrogenic harm (harm caused by the treatment itself), stigma, reduced self-understanding, and the substitution of a medical label for genuine engagement with the social or personal context driving the distress.

The DSM and Diagnostic Creep

The Diagnostic and Statistical Manual of Mental Disorders (DSM) is the primary classification system for psychiatric diagnosis in the United States, and its influence extends globally. It is worth knowing something about how it has evolved.

The first DSM, published in 1952, contained 106 diagnostic categories. The DSM-5, published in 2013, contains approximately 300. This expansion is not straightforwardly a sign of improved knowledge. Some of it reflects genuine scientific progress — conditions that were previously unrecognized or lumped together. But a significant portion reflects diagnostic creep: the gradual expansion of diagnostic categories to include experiences that sit within the normal range of human variation.

Several examples illustrate the pattern:

Grief is perhaps the clearest recent case. The DSM-IV included a “bereavement exclusion” that prevented someone from being diagnosed with major depressive disorder within two months of losing a loved one — acknowledging that the symptoms of grief and depression overlap, and that grieving is not a disorder. The DSM-5 removed this exclusion. A person experiencing profound sadness, loss of appetite, difficulty sleeping, and reduced ability to function two weeks after the death of someone they loved can now technically be diagnosed with major depressive disorder and prescribed antidepressants. The line between a normal human response to loss and a condition requiring medical treatment has been moved — not because of strong scientific evidence that early medicalization of grief improves outcomes, but through a committee process involving people with significant pharmaceutical industry ties.

Shyness and social discomfort were once understood as personality traits and social experiences. The category of social anxiety disorder — which has genuine clinical validity for severe cases — has gradually expanded in both formal criteria and popular understanding to encompass a much wider range of ordinary social nervousness. Marketing campaigns for SSRIs in the late 1990s explicitly targeted social shyness as a treatable condition before the clinical category had meaningfully expanded — a case where pharmaceutical marketing shaped diagnostic culture.

Childhood behavior has seen significant diagnostic expansion. Attention difficulties, restlessness, emotional intensity, and defiance in children can all be signs of genuine conditions that benefit from support. They can also be signs of developmentally normal behavior in environments poorly suited to children, responses to stress, or reflections of individual temperament. The rate at which children — particularly boys — are diagnosed with ADHD and medicated has grown substantially in recent decades, with significant geographic variation that is difficult to explain purely by genuine prevalence differences. This doesn’t mean ADHD isn’t real; it is, and it is underdiagnosed in some populations. It means the clinical category is being applied unevenly in ways that reflect something other than clinical precision.

Oppositional Defiant Disorder — a diagnosis applied to children who persistently argue with authority figures, refuse to comply with rules, and blame others for their mistakes — illustrates the political dimension of medicalization most starkly. The behaviors it describes are real and can indicate genuine distress. They can also be a reasonable response to genuinely unreasonable authority. The diagnosis locates the problem entirely in the child.

The Pharmaceutical Industry’s Role

It would be incomplete to discuss medicalization without naming the commercial interests that accelerate it. The pharmaceutical industry profits from the expansion of diagnostic categories — more diagnoses mean larger markets for treatments. This creates structural incentives to fund research that supports new or expanded categories, to support advocacy organizations that normalize the categories, to fund continuing medical education for prescribers, and to market directly to consumers in countries where that is permitted.

This is not a conspiracy theory. It is a documented pattern that has been the subject of serious academic and journalistic investigation. Former editors of major medical journals — including the New England Journal of Medicine and the British Medical Journal — have written critically about pharmaceutical industry influence on medical research and practice. The problem is structural: the same companies that manufacture treatments have significant influence over the evidence base used to justify those treatments.

None of this means pharmaceutical treatment is never useful. Psychiatric medication genuinely helps many people, and dismissing it wholesale would be as inaccurate as uncritical acceptance. It means applying the same critical evaluation the program advocates throughout: who funded this research, what are the incentive structures, and what does the independent evidence actually show?

The Line Worth Drawing

Medicalization becomes harmful when it:

Substitutes a label for understanding. Telling a child their distress is a disorder can close inquiry into what in their environment, relationships, or history might actually be driving it.
Creates treatment where the real need is accommodation or change. If a person’s distress is primarily driven by circumstances — poverty, an abusive relationship, an inaccessible work environment — treating their symptoms medically while leaving the circumstances unchanged is at best incomplete and at worst a way of making the person more functional within conditions they shouldn’t have to tolerate.
Applies to the normal range of a genuine spectrum. Many real conditions exist on a spectrum, and there is a meaningful difference between the experiences of people near the high-support end of that spectrum and people at the other end for whom the label primarily serves to normalize distress that has other causes.
Removes agency and self-understanding. A medical label, even an accurate one, can replace a person’s own narrative about their experience with a professional narrative — sometimes usefully, sometimes at the cost of their ability to make sense of their own life on their own terms.

The parallel to the social model discussed in the Neurodiversity section is direct: medicalization consistently locates the problem in the person rather than in the interaction between the person and their environment, social context, and circumstances. This framing is sometimes accurate and sometimes a profound evasion.

Holding Both Things True

Real conditions exist. People genuinely suffer from depression, anxiety, psychosis, personality disorders, and the full range of what the DSM attempts to catalogue. Dismissing psychiatric diagnosis entirely would leave many people without access to understanding and support that genuinely helps them.

The position this program takes — consistent with the caveat at the start of this topic — is that professional consensus is a useful input, not an authority. It is produced by human beings with interests, within institutions with incentives, using methods with limitations. Evaluating it with the same critical tools you’d apply to anything else is not cynicism. It’s appropriate epistemic hygiene.

How to Evaluate Psychological Claims Critically

The three previous subsections have described the landscape: a field with genuine value and real methodological problems, systematic biases in who gets studied and who asks the questions, and commercial pressures that shape what gets classified and treated. This final subsection turns that landscape into a practical toolkit — not a reason for paralysis, but a set of questions that help you navigate psychological claims with appropriate precision.

This is, at its core, an application of Level 2: Critical Thinking to a specific domain. If you’ve worked through that topic, many of these moves will be familiar. The goal here is to make them concrete and specific to psychology.

Questions Worth Asking About Any Psychological Claim

Who was studied, and how many of them? A finding from a study of 24 undergraduate psychology students at an American university tells you something about 24 undergraduate psychology students at an American university. Before accepting a generalization, ask how representative the sample was and how large. Small, unrepresentative samples produce unreliable estimates — and, as covered in the replication crisis section, many of psychology’s most famous findings came from exactly these kinds of studies.

Has it been replicated? A single study, however well-designed and peer-reviewed, is a weak basis for a strong claim. A finding that has been independently replicated across multiple studies, in different populations, using different methodologies is considerably more trustworthy. When a psychological claim is being presented as established fact, asking “replicated how many times, by whom, on what populations?” is entirely reasonable.

What’s the effect size? Statistical significance tells you that a finding is unlikely to be a fluke. It does not tell you that the effect is large enough to matter in practice. A study might find a statistically significant relationship between two variables that accounts for 1% of the variation in outcomes — real, but negligible for practical purposes. Effect size is the measure you actually want, and it’s often not what gets reported in popular coverage of psychological research.

Who funded this, and who benefits if the claim is true? Research funded by pharmaceutical companies, professional licensing bodies, or organizations with commercial interests in a particular finding should be evaluated with awareness of those interests. This doesn’t automatically invalidate the research — it’s a flag for additional scrutiny, not automatic dismissal. Ask whether independent research has produced consistent findings, and notice when independent replication is conspicuously absent.

What are the affected communities saying? As discussed in both the Neurodiversity and Medicalisation sections, the people whose psychology is being studied or whose behavior is being classified often have knowledge that outside observation cannot fully access. When there is a significant gap between professional consensus and the lived experience of the people the consensus is about, that gap is itself evidence worth taking seriously. It may mean the professional consensus is wrong, or it may mean the community’s experience is being misinterpreted — but it warrants genuine engagement rather than dismissal.

Is this a claim about mechanism or a claim about practical utility? These are separable. A therapeutic approach might have genuine practical utility even if the theoretical explanation for why it works is wrong or unproven — the NLP discussion earlier in this topic is a clear example. Conversely, a theoretically compelling mechanism doesn’t guarantee practical benefit. Keeping these two types of claims distinct helps you evaluate each on its own terms rather than accepting or rejecting a package wholesale.

Is this objective or subjective? The SOS principle from Level 2: Critical Thinking applies here. Some psychological claims are empirical questions with better and worse answers — does this treatment reduce symptom severity, does this assessment reliably predict outcomes, does this intervention improve wellbeing as measured by agreed criteria? These should be evaluated as objective claims, with the tools described above. Other questions — whether a particular life path constitutes flourishing, whether a person’s emotional response is appropriate, whether someone’s preferences need adjusting — are substantially subjective, and professional authority is not a trump card. A psychologist’s opinion about how you should live your life is an opinion, not a clinical finding.

Holding Uncertainty Without Paralysis

A risk in developing critical evaluation skills is overcorrection: if everything in psychology is contested, if replication is uncertain, if biases are pervasive, why trust any of it? This is the wrong conclusion, and it’s worth being explicit about why.

Uncertainty exists on a scale. The stress response system is not a controversial finding — it is supported by convergent evidence across neuroscience, physiology, evolutionary biology, and decades of clinical observation. The Big Five personality model is not perfectly validated, but it has been replicated cross-culturally and has genuine predictive power. Attachment theory has a large and consistent evidence base. The specific claims made in a 2003 study with 30 participants deserve more scrutiny than these foundational frameworks.

Calibrated evaluation means holding different claims with different degrees of confidence — not treating all psychological claims as equally uncertain or equally reliable. You can act on well-supported concepts while remaining appropriately skeptical of more contested ones. You don’t need certainty to make good decisions; you need honest assessment of the probability and the stakes.

It also means being willing to update. A critical approach to psychological claims is not about arriving at a fixed position of skepticism and defending it — it’s about remaining genuinely open to evidence, including evidence that contradicts your current view. This is the intellectual humility discussed in Critical Thinking Intermediate, applied to a domain where it’s particularly needed.

A Final Note on This Section’s Purpose

Psychology’s limitations are not a reason to dismiss psychology. They are a reason to engage with it the way this program has tried to engage with everything: with genuine curiosity, honest assessment of evidence, attention to who is speaking and from what position, and respect for your own experience and judgment as a legitimate source of knowledge.

The caveat at the start of this topic asked you to hold psychology’s value and its limitations in the same hand. This section has tried to give you the tools to do that well. The concepts throughout this topic — in personality, neurodiversity, trauma, and memetics — are genuinely useful. They become more useful, not less, when you hold them with appropriate precision rather than uncritical acceptance.

How It Connects

Critical Thinking: This section is, more than any other in this topic, a direct application of Critical Thinking to a specific domain. The replication crisis maps onto evidence evaluation and source reliability; the WEIRD problem maps onto recognizing sampling bias and overgeneralization; medicalization maps onto identifying commercial incentives and the appeal to authority fallacy; the evaluation toolkit draws directly on the intellectual virtues — particularly intellectual humility, fairmindedness, and intellectual honesty — developed in Critical Thinking Intermediate. If there is one section in this topic that rewards having done Critical Thinking thoroughly first, it is this one.

Science Intermediate: The replication crisis is not unique to psychology — it exists across several sciences, and Science Intermediate addresses the broader question of how science self-corrects and why its findings should be trusted with calibration rather than either blind faith or wholesale dismissal. The two sections together give a more complete picture of how scientific knowledge is produced, where it is reliable, and where it warrants ongoing scrutiny.

Neurodiversity: A Deeper Look and Trauma: Mechanisms and Healing: Both earlier sections in this topic applied the critical lens developed here before the lens itself was fully articulated — the ABA discussion, the evaluation of therapeutic options including NLP, and the “nothing about us without us” principle all anticipate what this section makes explicit. Reading this section after those two means you can now name more precisely what you were already doing.

Level 1 — External Barriers Intermediate: The demographic biases in psychological research — who gets studied, whose experience gets classified as normal, whose distress gets medicalized rather than structurally addressed — are a specific instance of the systemic barriers covered there. The intersectionality framework from that section applies directly to understanding why women, Black patients, and LGBTQ+ people have been systematically misserved by mainstream psychology.

Community & Cooperation: The principle that affected communities hold knowledge about their own experience that outside observation cannot fully access has implications beyond research methodology. It shapes how communities should be engaged — in research, in program design, in therapy — and connects to the broader theme of distributed knowledge and collective intelligence in that topic.

Level 3 — Systemic/Institutional Change: The pharmaceutical industry’s influence on diagnostic categories, the publication bias built into academic journals, and the demographic homogeneity of research communities are all examples of how institutional incentive structures shape what knowledge gets produced and what doesn’t. Level 3 addresses how to understand and work with exactly these kinds of systemic dynamics.

Practice Exercises

Comprehension

Explain the replication crisis in your own words. What structural features of academic publishing contributed to it, and what reforms have been introduced in response?
What does WEIRD stand for, and why does it matter that the vast majority of psychological research has been conducted on WEIRD populations? Give one example of a finding that doesn’t hold consistently across cultures.
What is medicalization? Describe one example where applying a medical framework to a human experience is genuinely helpful, and one example where it may overshoot. What distinguishes the two cases?

Reflection

Think of a psychological claim you’ve accepted — from a book, an article, a course, a therapist, or popular culture. Apply the evaluation questions from the final subsection: Who was studied? Has it been replicated? Who benefits if the claim is true? Does the affected community’s experience align with the professional consensus? What’s your updated level of confidence in the claim?
Think of a time when professional or institutional authority conflicted with your own experience or the experience of someone you know. How was that conflict resolved — or was it? Looking back with the framework from this section, how would you evaluate it now?
Before this section, what was your default attitude toward psychological claims — broadly trusting, broadly skeptical, or somewhere in between? Has working through this section shifted that in any way? If so, how?

Application

Find a psychological claim in popular media — a news article, a self-help book summary, a social media post — and apply the evaluation framework from this section systematically. What can you determine about the source, the sample, the replication status, and the incentive structures involved? What’s your overall assessment of how much confidence the claim warrants?
Return to the Pathways to Healing subsection in Trauma and apply the evaluation questions to two of the therapeutic approaches listed — one well-established, one more contested. What does the evaluation process reveal about each? Does it change which you’d consider pursuing or recommending?

Discussion

(Partner or group) Each person brings one psychological claim they’ve encountered recently. Evaluate them together using the framework from this section. Notice where the group agrees and where it doesn’t — and what drives the disagreement.
(Partner or group) Discuss the tension between the need to make decisions with incomplete or uncertain information and the desire for greater certainty before acting. How do you currently navigate that tension in practice? Does the calibrated uncertainty framework this section proposes feel workable, or does it raise difficulties?

Key Sources & Further Reading

On the replication crisis:

Open Science Collaboration — “Estimating the Reproducibility of Psychological Science” (2015): The original Reproducibility Project paper, published in Science. Freely available online and relatively accessible for a scientific paper. The primary source for the 36% figure.
Stuart Ritchie — Science Fictions (2020): A readable and thorough account of fraud, bias, negligence, and hype in science broadly, with psychology featuring prominently. More accessible than the primary literature and genuinely illuminating.

On WEIRD populations:

Joseph Henrich — The WEIRDest People in the World (2020): A substantial but accessible book making the case that Western psychology has systematically mistaken culturally specific traits for universal human nature. Henrich is one of the researchers who identified and named the WEIRD problem.

On medicalization:

Allen Frances — Saving Normal (2013): Written by the psychiatrist who chaired the DSM-IV task force and became one of the DSM-5’s most prominent critics. An insider account of how diagnostic categories are made and what goes wrong in the process.
Gary Greenberg — The Book of Woe (2013): A journalist’s account of the DSM-5’s development, covering the political and commercial dynamics behind diagnostic decisions. Readable and well-researched.

On pharmaceutical industry influence:

Ben Goldacre — Bad Pharma (2012): Thoroughly documented and accessibly written account of how pharmaceutical companies manipulate the evidence base. Goldacre is a physician and epidemiologist; his analysis is rigorous rather than conspiratorial.
Marcia Angell — The Truth About the Drug Companies (2004): A former editor of the New England Journal of Medicine making a similar case from an insider perspective.

On applying critical thinking to psychology:

Scott Lilienfeld, Steven Jay Lynn, John Ruscio & Barry Beyerstein — 50 Great Myths of Popular Psychology (2010): Systematically evaluates widely believed psychological claims against the actual evidence. A useful practical reference and a model for the kind of evaluation this section advocates.

Continue to Psychology Intermediate Part 4 →

Return to the main page.