The p factor is a proposed single dimension that captures a person’s overall tendency toward mental health problems, regardless of the specific diagnosis. Think of it as a general vulnerability score: the higher someone falls on this dimension, the more likely they are to experience a wide range of psychiatric symptoms, from depression and anxiety to substance use and psychotic experiences. The concept was formally introduced in 2014 by psychologists Avshalom Caspi and Terrie Moffitt, who named it the “p factor” as a deliberate parallel to the “g factor” of general intelligence.
Why Mental Health Disorders Cluster Together
The p factor grew out of a well-known puzzle in psychiatry: mental health conditions overlap far more than diagnostic manuals suggest. A person diagnosed with depression is more likely than average to also develop an anxiety disorder, a substance use problem, or even psychotic symptoms. This pattern isn’t limited to closely related conditions like anxiety and depression. The correlations between broad categories of mental illness, including internalizing problems (like depression and anxiety), externalizing problems (like conduct disorder and substance dependence), and psychotic experiences (like hallucinations and disorganized thinking), hover around 0.5. That’s a substantial overlap for categories that are supposed to be distinct.
Researchers had long organized psychiatric conditions into these three broad families. But the consistent overlap between them suggested something deeper was going on. If internalizing, externalizing, and psychotic experiences all correlate with each other, there may be one underlying dimension driving all three. That dimension is the p factor.
How the P Factor Works
The p factor represents the shared variance among all forms of psychopathology. In statistical terms, it’s the single thread running through hundreds of psychiatric symptoms, dozens of distinct diagnoses, and three broad domains of mental illness. A person with a high p factor score isn’t destined for any one disorder. Instead, they carry a general liability that can manifest as different conditions at different points in their life.
The original 2014 study, using data from 1,000 participants in a longitudinal birth cohort, tested several statistical models to see which best explained the structure of psychiatric disorders. A model that included one general psychopathology factor fit the data well, with symptom loadings on the general factor averaging 0.65. Notably, thought disorder symptoms (mania, OCD, and schizophrenia-related symptoms) loaded so heavily onto the general factor that they couldn’t form a separate dimension independent of it. Internalizing and externalizing symptoms loaded onto p as well, but also retained enough independence to remain as distinct subfactors.
The practical upshot: p represents low-to-high psychopathology severity, with thought disorder symptoms sitting at the extreme high end of the spectrum.
The Parallel to General Intelligence
The name “p factor” is intentional. For over a century, psychologists have known that different cognitive abilities, such as reasoning, memory, and processing speed, are all positively correlated. A person who scores well on one type of cognitive test tends to score well on others. The common thread among all these abilities is called the g factor, or general intelligence.
The p factor applies the same logic to psychopathology. Just as g doesn’t mean everyone has the same cognitive profile, a high p factor doesn’t mean a person will develop every mental disorder. It means their general vulnerability is elevated, and the specific form that vulnerability takes depends on other factors. One key difference has emerged from developmental research: the link between intelligence and psychopathology runs in both directions, but through different mechanisms. The influence of intelligence on later mental health problems appears to be largely genetic, while the influence of mental health problems on later cognitive decline is driven more by environmental factors, and increasingly so with age.
How It Changes Over Development
The p factor isn’t static. Research tracking girls from ages 14 to 21 found that the strength and proportion of variance explained by p steadily increased throughout adolescence and into early adulthood, reaching its peak at age 21. At that point, roughly 86% of the reliable variance in total psychopathology scores could be attributed to the general factor alone.
This trajectory is consistent with a theory called dynamic mutualism, which proposes that the p factor isn’t simply a fixed trait you’re born with. Instead, individual symptoms feed into and reinforce each other over time, gradually strengthening the connections between them. Internalizing symptoms were especially powerful predictors of future p factor scores, with their influence growing at each successive age. Conditions like ADHD, conduct disorder, and major depression showed increasingly strong connections to the general factor over time, while oppositional defiant disorder showed a slightly weaker link by age 21.
What a High P Factor Looks Like
People who score high on the p factor dimension tend to share a recognizable profile that goes beyond any single diagnosis. In the original study, higher p factor scores were associated with more life impairment, greater familial risk for mental illness, worse developmental histories, and more compromised brain function in early life.
Neuroimaging research has added detail to this picture. A meta-analysis of brain structure studies found that differences between people with psychiatric diagnoses and healthy controls concentrated in the same brain regions across six different disorders, with very few diagnosis-specific patterns. People with higher general psychopathology tend to have pervasively thinner cortical tissue across the brain’s outer surface, along with structural changes in cerebellar circuitry. In children around age 10, a single whole-brain measure of white matter integrity was associated with the p factor. Resting-state brain connectivity studies have identified a single dimension combining functional connectivity with environmental and behavioral variables that explained up to 17% of the variance in the data.
The picture that emerges is not of a brain with one broken circuit, but of a brain with broadly compromised structure and connectivity.
How Researchers Measure It
The p factor isn’t something you can measure with a single questionnaire or blood test. It’s a statistical construct, extracted from patterns in data using a family of techniques called factor analysis. Researchers collect symptom ratings across many different mental health domains, then use statistical models to identify the shared variation among them.
There are several competing approaches. A higher-order model treats the p factor as a second-order factor that arises from the correlations between internalizing, externalizing, and thought disorder dimensions. A bifactor model takes a different approach: every individual symptom loads simultaneously onto both the general p factor and onto a domain-specific factor (like internalizing) that is statistically independent of p. Each model produces somewhat different p factor scores and carries different assumptions about what the factor actually represents.
Why Some Researchers Are Skeptical
The p factor is one of the most influential ideas in modern psychopathology research, but it has drawn serious criticism. The most pointed critique targets the statistical models used to extract it. Symmetrical bifactor models, the most popular method, frequently produce anomalous results: negative variance estimates, factor loadings that flip from positive to negative, and specific factors that lose statistical significance. These problems aren’t rare edge cases. They appear regularly in published p factor studies.
Critics argue that these anomalies aren’t just technical nuisances. They fundamentally alter what the p factor is measuring. When factor loadings behave unpredictably, the general factor may not reflect a true overarching liability to psychopathology. Instead, its meaning can shift from study to study depending on which symptoms happen to dominate the model. Some researchers have recommended moving to alternative modeling approaches that anchor the general factor to a specific reference group, making its interpretation more stable and theoretically grounded.
There’s also a deeper conceptual question: does the p factor represent a real biological or psychological mechanism, or is it simply a statistical summary of the well-known fact that mental health problems tend to co-occur? The original authors proposed it reflects a genuine common cause, but the dynamic mutualism perspective suggests it could instead be an emergent property, something that arises from individual symptoms reinforcing each other over time rather than from a single underlying vulnerability.
What It Could Mean for Treatment
If the p factor reflects a real general liability, it has significant implications for how mental health treatment is designed. Rather than developing highly specific treatments for each individual diagnosis, a p-informed framework points toward transdiagnostic approaches that target the processes common to many disorders, things like emotion regulation, cognitive flexibility, and stress reactivity. For risk prediction, a general psychopathology score could help identify people who are broadly vulnerable before they develop any specific condition, allowing for earlier and more flexible intervention.
The field is still working out how to connect this statistical construct to specific treatment-relevant pathways in the brain and body. For now, the p factor remains primarily a research tool rather than something used in clinical practice. But it has already reshaped how researchers think about the boundaries between mental health conditions, pushing the field away from rigid diagnostic categories and toward a more dimensional view of psychological distress.