Child Behavior Checklist (CBCL): What It Measures

The Child Behavior Checklist, commonly called the CBCL, is a standardized parent-report questionnaire used to identify emotional and behavioral problems in children and adolescents. Developed by Thomas Achenbach, it has become one of the most widely used instruments in child psychology and psychiatry, translated into dozens of languages and applied in both clinical and research settings worldwide. The checklist asks parents to rate over a hundred specific behaviors, from “argues a lot” to “withdrawn,” and organizes them into scales that help clinicians spot patterns that might otherwise be dismissed as normal childhood difficulty.

What the CBCL Actually Measures

The CBCL asks a parent or caregiver to rate about 120 problem items on a three-point scale: not true, somewhat or sometimes true, and very true or often true. Responses are scored into two broad groupings. Internalizing problems capture difficulties directed inward, such as anxiety, depression, withdrawal, and physical complaints without a clear medical cause. Externalizing problems capture difficulties directed outward, like aggression, rule-breaking, and defiance. A Total Problems score sums everything up. These broad-band scores have shown solid reliability across studies and across cultures, and the subscales beneath them successfully distinguish between children with different levels and types of difficulty.

Beneath the broad bands sit narrower syndrome scales. These include clusters like Anxious/Depressed, Withdrawn/Depressed, Somatic Complaints, Social Problems, Thought Problems, Attention Problems, Rule-Breaking Behavior, and Aggressive Behavior. A later addition introduced DSM-oriented scales, designed to map more directly onto diagnostic categories used by clinicians. These include Anxiety Problems, Affective Problems, Attention Deficit/Hyperactivity Problems, Oppositional Defiant Problems, and Conduct Problems, among others.

A natural question is whether these DSM-oriented scales are better than the older syndrome scales at flagging children who actually meet criteria for a diagnosis. Research suggests the answer is mixed. In one study of clinic-referred children, the DSM-oriented scales corresponded with clinical diagnoses derived from structured interviews, but they did not consistently outperform the original syndrome scales. The Anxiety Problems scale was the only one that showed a clear advantage over its syndrome counterpart.¹ Another study found the DSM-oriented scales were more specific but sometimes had weaker overall associations with diagnoses than the syndrome scales, and recommended using both sets together for the most complete picture.²

Age Versions and Who Can Fill It Out

The CBCL comes in two main versions. The CBCL/1.5–5 covers toddlers and preschoolers, and the CBCL/6–18 covers school-age children and teenagers. The preschool version has been validated internationally, with confirmatory analyses showing its factor structure holds up well outside the United States. A study applying it to Chinese preschoolers found that the U.S.-derived model fit the Chinese data, and overall rates of total behavior problems were largely similar between the two countries, though Chinese preschoolers scored higher on internalizing problems and American preschoolers scored higher on externalizing problems.³

The parent-report CBCL is part of a broader family of instruments. Teachers fill out a parallel form called the Teacher Report Form (TRF), and older children can complete a Youth Self-Report (YSR). Having multiple perspectives is useful, but it also introduces disagreement. Agreement among parents, teachers, and youth tends to range from low to moderate. Informants agree more about externalizing behaviors, which are visible and disruptive, than about internalizing problems like worry or sadness, which are internal and easier to miss.⁴ This is not necessarily a flaw. A child who acts out at school but not at home is giving real information through that inconsistency, and clinicians often treat the gaps between informants as clinically meaningful rather than just noise.

Sex and Age Patterns in CBCL Scores

A consistent finding across cultures is that boys and girls tend to show different behavioral profiles, and these differences shift with age. Boys typically score higher on externalizing problems during early and middle childhood. Girls begin to score higher on internalizing problems around age ten. One study that examined how sex and age interact found that boys displayed higher externalizing scores between ages six and roughly ten, while girls showed higher internalizing scores from about age ten onward.⁵ These effects were modest in size, but they matter for interpretation. A girl scoring at the 90th percentile on internalizing symptoms at age twelve may represent a different level of concern than a boy at the same percentile at the same age, because the norms already account for these patterns.

Cross-cultural data from a nine-culture comparison found remarkable consistency in these sex differences. Girls scored higher on somatic complaints and anxious/depressed symptoms, while boys scored higher on attention problems, delinquent behavior, and aggressive behavior, with this pattern holding across all cultures studied. Cultural variation existed, but it was generally small. Puerto Rican subjects tended to score highest and Swedish subjects lowest on most syndromes.⁶

Screening for ADHD

The CBCL was never designed as a diagnostic tool for any single disorder, but clinicians and researchers have long used its subscales as screening instruments. For ADHD, the evidence is fairly encouraging. The Attention Problems scale has consistently shown the highest discriminating power for identifying children who meet ADHD criteria, and adding other scales on top does not meaningfully improve accuracy.⁷ In a community mental health setting, the caregiver-reported Attention Problems scale produced an area under the curve of 0.74 for children and 0.73 for adolescents, outperforming all other subscales from both caregiver and teacher forms.⁸ The DSM-oriented Attention Problem scale has also been tested and found to be slightly better than the original Attention Problems scale at separating ADHD from non-ADHD cases.⁹

These numbers are respectable for a broadband screening tool but fall well short of what you would want for a standalone diagnostic test. An area under the curve in the low-to-mid 0.70s means the CBCL catches most kids with ADHD, but it also flags some who do not have it. The practical takeaway is that an elevated Attention Problems score is a useful signal to pursue a more thorough evaluation, not a diagnosis in itself.

The Autism Screening Question

Researchers have repeatedly tried to use the CBCL to screen for autism spectrum disorder, and the results are less encouraging than for ADHD. Children with autism do tend to show a characteristic pattern of elevated scores on certain CBCL scales, and one study found elevated scores on four scales in children with ASD compared to typically developing children.¹⁰ However, when tested against clinical diagnoses, the CBCL profiles showed low discriminative accuracy for ASD, with areas under the curve between 0.59 and 0.70. Meeting proposed ASD cutoffs on the CBCL turned out to be associated with general emotional and behavioral problems rather than autism specifically, leading to high false-positive rates.¹¹

Attempts to build a better ASD subscale within the CBCL have continued, using both statistical methods and input from clinical experts. A recent effort created and cross-validated a data-driven subscale across multiple independent samples. Despite having relatively low discriminative power overall, the data-driven subscale showed the best potential among existing CBCL-based approaches to screen for ASD.¹² The bottom line for clinicians is that the CBCL can hint at ASD, especially when other behavioral and emotional problems are not already at clinical levels, but it should not be relied upon as a primary autism screener.

Predicting Anxiety Disorders

For anxiety, the CBCL offers a somewhat different story. The Anxiety Problems scale has shown significant predictive ability across multiple anxiety diagnoses, including separation anxiety disorder, generalized anxiety disorder, and social phobia. It outperformed the older Anxious/Depressed syndrome scale for this purpose, as mentioned in the DSM-oriented scale research above. A study specifically examining the Anxiety Problems scale found it predicted separation anxiety, generalized anxiety, and the presence of any anxiety disorder with moderate accuracy, though it was not useful for predicting specific phobia on its own.¹³ Again, the instrument works best as a flag, pointing clinicians toward children who need deeper assessment rather than replacing that assessment.

The Dysregulation Profile and Future Risk

One of the more clinically interesting developments in CBCL research involves the so-called Dysregulation Profile, sometimes abbreviated CBCL-DP. This is not a standalone scale but a pattern of simultaneously elevated scores on the Anxious/Depressed, Attention Problems, and Aggressive Behavior syndrome scales. Children showing this combined profile seem to have broad difficulty regulating emotions, attention, and behavior all at once, and research suggests this pattern carries real prognostic weight.

A longitudinal study of children with disruptive behavior disorders found that those with higher dysregulation profile scores in childhood were at increased risk for ADHD and mood disorders in adolescence. The dysregulation profile score was the only significant predictor of a mood disorder at ages fourteen to fifteen, even after accounting for childhood diagnoses. It was not, however, associated with conduct disorder, substance use, or hospitalization during adolescence.¹⁴ A separate longitudinal study followed children into young adulthood and found that higher childhood dysregulation scores predicted substance use disorders, suicidality, and poorer overall functioning at age nineteen, even after adjusting for family income, parental education, and baseline psychiatric problems. The dysregulation profile was not, interestingly, a precursor of bipolar disorder specifically, despite some early speculation that it might be.¹⁵

How Well Do Early Scores Predict Later Diagnoses

Beyond the dysregulation profile, the broad-band Internalizing and Externalizing scores themselves carry predictive information. In a five-year follow-up study of children whose parents had panic disorder or depression, baseline Internalizing scores predicted later diagnoses of agoraphobia, generalized anxiety disorder, separation anxiety disorder, and social phobia. Externalizing scores predicted later disruptive behavior disorders and, somewhat surprisingly, major depression.¹⁶ Even at very young ages, the CBCL shows promise as an early indicator. One community study found that parent ratings on the checklist at ages five to six corresponded well with diagnoses determined by structured interviews a year and a half later.¹⁷

The practical implication here is that a high CBCL score in early childhood is not something to ignore or assume a child will grow out of. Early scores, especially in the clinical range, increase the probability of diagnosable problems later on. They do not guarantee them, but they justify closer monitoring and, in many cases, earlier intervention.

Detecting Traces of Trauma

One of the more striking applications of the CBCL is its potential to flag children who have experienced maltreatment, even when nobody has asked the child or caregiver directly about abuse or neglect. Children with histories of maltreatment score significantly higher on nearly all CBCL subscales. A recent study built prediction models using combinations of CBCL subscale scores alongside age, sex, and IQ, and achieved sensitivity above 0.90 and specificity above 0.96 for identifying children with maltreatment histories. Even a model using only CBCL scores, without demographic variables, achieved sensitivity of 0.91 and specificity of 0.86.¹⁸

The same study identified sensitive periods for different types of behavioral impact. Exposure to maltreatment at age five was particularly predictive of elevated scores on the Withdrawn and Thought Problems scales. Maltreatment between ages five and seven predicted elevated Somatic Complaints. Physical abuse specifically predicted somatic complaints and delinquent behavior, while emotional abuse predicted anxious/depressed symptoms and thought problems. These findings are valuable because direct questioning about trauma is not always possible or appropriate, especially in young children or situations where the caregiver may be involved in the maltreatment.

Research on maltreated preschoolers has further shown that trauma symptoms fully mediate the link between maltreatment and both internalizing and externalizing behaviors on the CBCL.¹⁹ Derived PTSD and dissociation scales pulled from CBCL items have also shown convergent validity in samples of neglected children.²⁰

The Maternal Depression Problem

The CBCL’s greatest strength is also its most persistent vulnerability: it depends entirely on the informant’s perception. And the informant’s own mental health affects what they report. Research has consistently found that mothers who are anxious or depressed tend to report more behavioral problems in their children than do non-depressed mothers or the children themselves.²¹ This raises a legitimate question: are depressed mothers seeing problems that are really there, or are they over-reporting because their own distress colors their perception?

The answer, based on the available evidence, is probably a bit of both. One study found that while maternal depression and child gender were both significantly associated with how mothers rated their children on the CBCL, the mothers’ ratings still successfully distinguished between children with and without psychiatric problems even after statistically removing the influence of maternal depression.²² So a depressed mother’s CBCL scores may be somewhat inflated, but they still carry real diagnostic information. The practical lesson is that clinicians should consider the parent’s own mental health when interpreting CBCL results, and combining parent report with teacher report or child self-report helps counterbalance any single informant’s bias.²³

How the CBCL Compares to Other Instruments

The CBCL is not the only behavioral rating system available. The Behavior Assessment System for Children (BASC) and the Youth Outcome Questionnaire (Y-OQ) are both widely used alternatives. When researchers compared how sensitive these three instruments were to changes during treatment, the Y-OQ turned out to be the most sensitive to change for parent-report measures. The CBCL and BASC performed comparably to each other, and for youth self-report, no significant differences in change sensitivity were found among any of the instruments.²⁴

This matters for how the CBCL gets used in practice. If the goal is measuring whether treatment is working over time, the CBCL is adequate but not the most sensitive option for parent-report tracking. Where the CBCL excels is in its enormous research base, cross-cultural validation, and the ability to compare a child’s scores to extensive normative data. Clinicians often choose their instrument based on what they are trying to accomplish: initial screening, ongoing monitoring, or detailed diagnostic workup.

Genetic Research and the CBCL

Because the CBCL produces standardized, quantitative scores on behavioral dimensions, it has become a workhorse in behavioral genetics research. Twin studies have long shown that behavior problems run in families, but more recent genomic work has used CBCL scores to estimate how much of the variation in childhood behavior can be attributed to common genetic variants. One genome-wide analysis found that common genetic variants accounted for a substantial proportion of variation in attention-deficit/hyperactivity symptoms, with estimates ranging from 37 to 71 percent depending on the informant. Externalizing problems and total problems also showed significant genetic influence.²⁵ For internalizing problems, a separate large meta-analysis of preschool children used CBCL scores and found that internalizing problems are heritable and moderately genetically stable from childhood into adulthood.²⁶

These findings do not mean specific behaviors are hardwired. They mean that a child’s temperamental vulnerabilities, the thresholds at which they react with anxiety, defiance, or inattention, have a biological substrate that the CBCL happens to be good at measuring. The checklist was designed as a clinical tool, but its standardized structure has made it unexpectedly valuable for studies that need to measure behavior precisely across thousands of participants.

When Physical Illness Complicates the Picture

One underappreciated complication arises when the CBCL is used with children who have chronic medical conditions. The Somatic Complaints scale, which is grouped under Internalizing problems, includes items about aches, pains, and physical symptoms. For a healthy child, endorsing many of these items is a meaningful signal of internalizing distress. For a child with a condition like spina bifida, some of those somatic complaints reflect the medical reality of their condition rather than psychological distress. Research has confirmed this concern: in children with spina bifida, significantly more of the variation in Somatic Complaints scores was unrelated to internalizing problems compared to able-bodied children. When fathers completed the checklist, the internalizing factor explained little variance in somatic complaints for either group. Clinicians working with medically complex children need to interpret the Somatic Complaints scale cautiously and consider whether physical symptoms are being double-counted as both medical and psychological problems.

Psychometric Properties Across Settings

A practical question for any clinician or researcher choosing to use the CBCL is whether the instrument works reliably across different populations and settings. A recent psychometric evaluation of the CBCL/6–18 using a rigorous item-response-theory approach found that the Internalizing, Externalizing, and Total Problem subscales were each measuring a single underlying construct and demonstrated adequate reliability. The scale successfully distinguished between children with different levels of behavioral problems.²⁷ Combined with the cross-cultural evidence showing that the CBCL’s factor structure holds up across diverse populations, this gives clinicians reasonable confidence that scores mean roughly the same thing regardless of where in the world they are collected, though local norms should still be used when available to account for culturally influenced variation in what parents consider problematic behavior.