Ten-Dimensional Measure Links Personality to Mental Health and Substance Use

The world’s most widely used measure of childhood adversity may be telling scientists only half the story. A new study published in the Journal of Child & Adolescent Trauma suggests that the ACE-10 questionnaire, the ten-item screening tool that has shaped decades of research into adverse childhood experiences, is not the single unified scale that clinicians and researchers have long assumed it to be. Instead, the instrument appears to capture two statistically distinct dimensions of childhood adversity, and, critically, those two dimensions point toward different adult outcomes. Abuse and neglect reported on the scale were significantly associated with depression and anxiety in adulthood, while household dysfunction together with sexual abuse tracked more closely with adult drug use. The findings, drawn from a U.S. general population sample of 667 women, arrive amid a growing international debate about whether the familiar practice of summing adverse childhood experiences into a single score has outlived its scientific usefulness.

The concept of adverse childhood experiences, or ACEs, traces its origins to a landmark 1998 investigation by Vincent Felitti and colleagues, who documented striking relationships between childhood abuse and household dysfunction and the leading causes of death in adults. That original study organized ten categories of adversity: emotional, physical, and sexual abuse; emotional and physical neglect; and five forms of household dysfunction, including domestic violence, substance abuse, mental illness in the home, parental separation, and incarceration of a household member. In the decades since, the cumulative ACE score has become one of the most recognizable numbers in public health, cited in policy documents, incorporated into pediatric screening programs, and used to justify trauma-informed care initiatives across school districts, courts, and health systems. The underlying premise is deceptively simple: more adversity in childhood produces greater risk in adulthood, in a roughly dose-response fashion. But that simplicity, critics have argued, conceals important complexities about how different forms of adversity relate to one another and to later disease.

The new research, conducted by Matthias Woeckener of the University of Nebraska at Kearney and Ethan A. Marshall of the University of Houston-Downtown, tackles a question that has dogged the field for years: does the ACE-10 actually measure one thing, or several? Psychometricians refer to this as the dimensionality of an instrument, and it matters enormously for interpretation. If the ten items reflect a single underlying latent factor, then a total score is a legitimate and efficient summary of a person’s adversity burden. If, however, the items cluster into separate factors, then collapsing them into one number risks averaging away meaningful differences, much as a physician might obscure distinct diseases by reporting a single generic “illness score.” Previous factor-analytic work has been inconsistent, with some studies supporting a one-factor structure, others finding two or three factors, and still others recommending domain-specific subscales. The inconsistencies have fueled calls, from researchers including McLennan, MacMillan, and Afifi, to question whether ACE questionnaires should be used as diagnostic-style instruments at all.

To resolve the question, the researchers recruited a U.S. general population sample of 667 women who completed the ACE-10 along with well-validated measures of mental health and substance use: the Patient Health Questionnaire-9 for depression, the Generalized Anxiety Disorder-7 scale for anxiety, and the Drug Abuse Screening Test for drug use. The analytical strategy unfolded in two stages. First, the team employed both exploratory factor analysis and confirmatory factor analysis to determine the latent structure of the ten adversity items. Exploratory factor analysis allows the data themselves to suggest how many underlying dimensions are needed to account for the pattern of correlations among items, while confirmatory factor analysis then tests pre-specified models against the data, comparing how well a single-factor model fits relative to a multi-factor alternative. Because the ACE items are binary, coded simply as present or absent, the authors used estimation approaches appropriate for ordinal data, drawing on methodological literature comparing robust maximum likelihood with diagonally weighted least squares estimators.

The results were unambiguous on the central question. While both the one-factor and two-factor models demonstrated statistically acceptable fit to the data, the two-factor solution proved more appropriate, providing a better account of the correlations among the ten items. The composition of the two factors was revealing. The first factor grouped together the items reflecting abuse and neglect: emotional abuse, physical abuse, emotional neglect, and physical neglect. The second factor captured household dysfunction, including living with someone with mental illness or substance problems, witnessing domestic violence, parental separation, and household incarceration, along with sexual abuse. The pairing of sexual abuse with household dysfunction rather than with the other maltreatment items is one of the study’s most intriguing findings, and it converges with earlier work by Ford and colleagues, who also found that sexual abuse behaved anomalously in factor analyses of ACE items, as well as with prior evidence linking childhood sexual abuse to distinct long-term trajectories, including heightened vulnerability to later sexual victimization.

Establishing the structure of the scale was only half the task. The researchers then examined criterion validity, asking whether the two empirically derived factors predicted different adult outcomes. Using a series of structural equation models, a framework that allows researchers to model relationships among latent variables while accounting for measurement error, the team found a striking dissociation. The abuse and neglect factor was significantly associated with adult depression and anxiety, consistent with the extensive literature tying direct maltreatment to internalizing psychopathology across the life course. The household dysfunction and sexual abuse factor, by contrast, was significantly associated with adult drug use. In other words, the two dimensions of the ACE-10 were not interchangeable indicators of a single generic risk; they carried distinct predictive information about the kinds of problems individuals would face decades later. Had the study relied on the traditional total score alone, this differential patterning would have been invisible, absorbed into an undifferentiated adversity index.

The authors situate these results within a broader rethinking of how childhood adversity should be conceptualized and measured. A growing body of work, including latent class analyses by Baldwin, Danese, and Coid, has shown that children tend to experience adversity in patterned configurations rather than as random assortments of isolated events, and that these configurations carry differential risks for mental health outcomes. Studies by Negriff and by Giano and colleagues have similarly demonstrated that specific ACE domains are not equal in their effects, with maltreatment sometimes exerting stronger influences on adolescent and adult mental health than household dysfunction. At the same time, methodological critiques have accumulated: retrospective self-reports of childhood adversity show imperfect agreement with prospective records; the binary scoring of items discards information about severity, timing, and chronicity; and the original ten categories omit many adversities now recognized as consequential, such as poverty, community violence, and discrimination. The new findings add a psychometric dimension to these critiques, suggesting that even within the existing instrument, the arithmetic of a simple sum conceals a multidimensional reality.

For clinicians and public health practitioners, the implications are consequential. Screening programs that use total ACE scores to trigger interventions may be flagging the right people for the wrong reasons, or missing the mechanistic pathways most relevant to each individual. If abuse and neglect drive internalizing disorders like depression and anxiety, while household dysfunction and sexual abuse are more tightly linked to substance use, then prevention and treatment resources might be allocated more effectively by attending to the specific composition of a child’s adversity rather than its raw quantity. The findings also carry implications for research design: studies that regress outcomes on total ACE scores may be imposing a constraint, namely a single latent factor, that the data themselves reject, potentially biasing effect estimates and obscuring genuine associations. The authors caution, however, that the two-factor structure should not be reified; factor solutions are models, not discoveries of natural kinds, and replication across diverse samples, including men, adolescents, and non-Western populations, remains essential.

The study is not without limitations that temper generalization. The sample consisted exclusively of women recruited from the U.S. general population, and the authors note that prior work has raised questions about measurement invariance across gender, race, and ethnicity, meaning the two-factor structure may not hold identically in other groups. The data were collected online, a methodology whose data quality has been scrutinized, though the authors cite evidence supporting the reliability of reputable research platforms. The use of retrospective self-report introduces the possibility of recall bias, a well-documented issue in adversity research. The authors also report a technical note: because their estimation software did not generate the standardized root mean square residual for the confirmatory models, substituting a weighted variant of limited recommended utility, they omitted that fit index from their reporting, relying instead on the remaining fit statistics. None of these caveats undermines the central result, but together they define the agenda for replication.

What the study ultimately delivers is a sharper picture of an instrument that has become a fixture of modern public health. The ACE-10, on this evidence, is best understood not as a single gauge of childhood adversity but as a composite of two related yet distinguishable constructs, each with its own downstream signature in adult life. That realization does not diminish the profound importance of the original ACE research, which transformed scientific and public understanding of how early experiences shape lifelong health. Rather, it refines it, pointing toward a future in which adversity is measured, analyzed, and ultimately treated with the specificity it deserves. As the authors conclude, the ten items of the ACE-10 represent a multidimensional construct, and those dimensions are associated with distinct negative outcomes in adulthood, a conclusion that should prompt researchers and practitioners alike to think twice before reducing a childhood to a single number.