The study team, led by pediatric researchers from Ningde Municipal Hospital and Pingnan County Hospital in Fujian Province, surveyed 3,050 children aged 6 to 12 years. Parents completed the Strengths and Difficulties Questionnaire, a widely validated screening instrument that measures five dimensions: emotional symptoms, conduct problems, hyperactivity and inattention, peer relationship problems, and prosocial behavior. Rather than simply adding these subscales into a total difficulties score, the researchers applied latent profile analysis, a person-centered method that asks whether the population contains subgroups of children who share similar response patterns across all five dimensions simultaneously.

The technical logic behind this approach matters. Variable-centered methods, such as ordinary correlation and regression on summary scores, assume that the sample is homogeneous and that associations apply uniformly to everyone. Latent profile analysis instead models the data as a mixture of unobserved subpopulations, estimating the probability that each child belongs to each profile. The researchers fit a series of competing models and selected the optimal solution using standard fit criteria, including the Akaike information criterion, the Bayesian information criterion, entropy as a measure of classification precision, and the requirement that each resulting class contain a practically meaningful proportion of children. The five-profile solution emerged as the best balance of statistical parsimony and interpretive clarity.

The largest group, comprising 56.8 percent of the children, was labeled the Normative profile, showing no marked elevations on any subscale. But nearly half the sample fell into one of four atypical profiles. The most common of these, at 18.1 percent, was Elevated Emotional Symptoms, in which children showed heightened anxiety and low mood without corresponding behavioral dysregulation. A further 12.3 percent belonged to a profile the authors called Low Hyperactivity and Low Prosocial Behavior, a quieter pattern in which children were neither restless nor particularly socially engaged. Two smaller but clinically concerning groups completed the picture: Elevated Conduct Problems and Low Prosocial Behavior, at 7.6 percent, and High Emotional and Multi-Domain Difficulties, at 5.2 percent, the latter representing children struggling across nearly every measured domain.

The most striking results came from the analysis of what distinguishes these profiles. Physical punishment showed a graded, dose-like cross-sectional association with profile membership. Compared with children who had never been physically punished, those with frequent exposure had substantially higher odds of belonging to the Elevated Conduct Problems and Low Prosocial Behavior profile and to the High Emotional and Multi-Domain Difficulties profile. Family conflict followed a strikingly similar pattern, with adjusted odds ratios of 3.42 for the conduct-problems profile and 3.55 for the multi-domain difficulties profile. In other words, children growing up amid frequent family conflict were roughly three and a half times more likely to land in the most troubled groups than their peers from low-conflict households, even after accounting for other measured characteristics.

Lifestyle factors told a complementary story. Excessive weekend screen time, defined in the study as more than two hours per day, was associated with higher odds of membership in the Elevated Emotional Symptoms profile and, with an adjusted odds ratio of 1.91, in the Elevated Conduct Problems and Low Prosocial Behavior profile. Short sleep duration, below the nine-hour threshold recommended for school-aged children, showed its most pronounced association with the conduct-problems profile, where it carried an adjusted odds ratio of 2.12. These associations were profile-specific: a given risk factor did not raise the odds of every atypical profile equally, which is precisely the kind of differentiated signal that a person-centered analysis is designed to detect and that a single total-difficulties score would obscure.

The differential patterning is what gives the study its practical bite. Emotional symptoms in middle childhood often fly under the radar because they do not disrupt classrooms the way hyperactivity or aggression do, yet the Elevated Emotional Symptoms profile captured nearly one in five children in this sample. The finding that weekend screen time, but not necessarily the family-conflict variables, was the factor most clearly tied to this profile suggests that different intervention levers may be needed for internally directed versus externally directed difficulties. Conversely, for children in the conduct-problems and multi-domain profiles, the dominant correlates were punitive discipline, household conflict and insufficient sleep, pointing toward family-system and sleep-hygiene interventions rather than screen-time limits alone.

The authors frame their conclusions as a call to prioritize the family parenting environment and healthy lifestyle habits in child health management. That framing is consistent with a substantial body of developmental science linking harsh discipline to externalizing behavior and linking sleep restriction to emotional and behavioral dysregulation, but the present study adds granularity by showing that these exposures cluster with particular symptom configurations. It is worth emphasizing the study’s cross-sectional design: the data capture associations at a single point in time, so they cannot establish whether physical punishment causes conduct problems, whether children with difficult behavior elicit harsher discipline, or whether both are shaped by unmeasured factors. Longitudinal follow-up would be needed to disentangle those directions of effect.

Methodologically, the study has notable strengths. The sample of 3,050 children is large enough to support stable profile estimation, and the use of multiple fit indices guards against over-extracting or under-extracting latent classes. The Strengths and Difficulties Questionnaire is one of the most extensively validated child mental health screens worldwide, and the anonymous administration of the survey, along with ethics approval from Pingnan County Hospital and informed consent from parents, addresses common concerns about reporting sensitive family variables. The research was supported by the Ningde Natural Science Foundation, and the authors report no competing interests.

For clinicians, educators and policymakers, the message is that childhood mental health screening may benefit from moving beyond aggregate scores toward pattern recognition. A child with moderate scores spread thinly across several subscales and a child with a sharp spike in emotional symptoms alone may both escape attention under a total-difficulties cutoff, yet they belong to different profiles with different correlates and potentially different intervention needs. As screen time and sleep debt continue to rise among school-aged children worldwide, and as debates over disciplinary practices persist, this study offers a data-driven map of how those everyday exposures align with the actual shapes of childhood distress, and a reminder that the household remains the most powerful lever available for shaping it.

Subject of Research: Latent profiles of emotional and behavioral problems and their environmental and lifestyle correlates among Chinese school-aged children

Article Title: Latent profiles of emotional and behavioral problems in Chinese school-aged children and their associations with environmental and lifestyle factors

Article References: Lu, Y., He, D., wang, T., Qian, F., Lin, Z., Zheng, Y., Zheng, M., Zhang, J., & Wang, J. (2026). Latent profiles of emotional and behavioral problems in Chinese school-aged children and their associations with environmental and lifestyle factors. BMC Psychology. https://doi.org/10.1186/s40359-026-05680-2

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05680-2

Keywords: latent profile analysis, emotional and behavioral problems, school-aged children, Strengths and Difficulties Questionnaire, physical punishment, family conflict, screen time, sleep duration, prosocial behavior, child mental health, BMC Psychology, China

The study was conducted between April and August 2023 by a team based at Shandong University’s School of Nursing and Rehabilitation, working with colleagues at the Second Qilu Hospital of Shandong University and the Qingdao Mental Health Center. Each participating family consisted of a child with autism spectrum disorder and one parent, forming what researchers call a dyad. Parents completed the Strengths and Difficulties Questionnaire, a brief parent-report instrument developed by child psychiatrist Robert Goodman that screens for emotional symptoms, conduct problems, hyperactivity, peer relationship difficulties, and prosocial behavior. Because parents observe children across everyday settings rather than in a single clinical session, parent-report measures of this kind are often used to capture the real-world expression of a child’s difficulties.

The headline finding was the sheer prevalence of difficulty: 81.3 percent of the children were flagged as having emotional or behavioral problems according to the questionnaire. That figure is consistent with a broader literature. A 2023 systematic review and meta-analysis published in Neuroscience and Biobehavioral Reviews documented high rates of co-occurring conditions in autistic children and adults, and longitudinal work following autistic people from childhood into early adulthood has shown that emotional and behavioral problems frequently persist across development. For families, this means the challenges associated with autism rarely stop at the core diagnostic features; they extend into anxiety, dysregulation, and conduct difficulties that shape daily life at home and at school.

What sets the new study apart is its analytical approach. Rather than averaging scores across the whole sample or treating problem severity as a single continuous scale, the researchers used latent profile analysis, a person-centered statistical method that searches for unobserved subgroups within a population. The technique asks a simple but powerful question: do children cluster into recognizable patterns of emotional and behavioral difficulty, or do their scores vary smoothly along one continuum? In this sample, the answer was that three distinct latent profiles best described the data, implying that autistic children do not form one uniform group but instead occupy meaningfully different positions with respect to their emotional and behavioral adjustment.

Identifying the profiles was only the first step. The team then wanted to know what distinguished families whose children fell into the more affected profiles from those in the less affected ones. To do this, they used the R3STEP procedure in the Mplus statistical software, an automatic three-step method that has become a standard tool in latent profile research. The procedure’s key virtue is that it accounts for classification uncertainty. Latent profile analysis never assigns a child to a profile with absolute certainty; instead, each child carries a probability of belonging to each group. Older approaches that simply assigned children to their most likely profile and then ran ordinary statistical tests treat those assignments as if they were error-free, which can bias estimates. The three-step method keeps the classification uncertainty in the model, producing more honest estimates of how observed variables relate to profile membership.

The covariates the researchers examined came from two domains of family life. The first was parenting stress, measured with an instrument whose Chinese-language version has been validated in prior psychometric work. Parenting stress in this context refers to the strain a parent experiences in the parenting role, encompassing perceptions of the child as demanding, a sense of incompetence or distress in the parental role, and restrictive or disrupted interactions between parent and child. The second domain was family management, assessed with a Chinese version of the Family Management Measure, a framework originally developed by nursing scientists to describe how families organize daily life around a child’s condition. Family management spans both difficult dimensions, such as difficulty managing the child’s condition and disruption of family routines, and positive dimensions, such as family life difficulty being managed with confidence, developmental competence in parents, and the family’s ability to maintain normalcy.

The associations that emerged were directionally coherent. Higher parenting stress was positively associated with membership in the profiles characterized by more emotional and behavioral problems. Negative family management, meaning families reporting greater difficulty and disruption in handling the child’s condition, was likewise positively associated with the high-problem profiles. Positive family management, by contrast, was inversely associated with membership in the most affected groups. In other words, after the model accounted for the probabilistic nature of profile assignment, families reporting more stress and more struggle tended to have children in the more difficult profiles, while families reporting confidence and competence in managing daily life tended to have children in the less affected ones.

The authors are careful about what this pattern does and does not establish. The study is cross-sectional, meaning stress, family management, and child outcomes were measured at the same point in time. That design cannot determine whether parenting stress drives children’s emotional and behavioral problems, whether challenging child behavior drives parental stress, or whether the two reinforce each other in a bidirectional loop. There is substantial evidence that such loops exist. A 2024 meta-analysis in Jornal de Pediatria synthesized studies linking parental stress to emotional and behavioral problems in offspring, and research on Chinese families of autistic children published in Autism Research has examined how child behavior problems and parental psychological distress relate to one another, with social support and cultural values as potential moderators. Other work has traced pathways from parental social support through family functioning to child outcomes, suggesting that the family system operates as an interconnected whole rather than as a set of independent variables.

Still, the practical implications of the study are relatively clear, and the authors frame them in terms of family-based support strategies. If parenting stress and negative family management are reliably associated with children landing in the most affected profiles, then interventions that reduce parental strain and strengthen families’ capacity to organize and manage daily life may be relevant to children’s emotional and behavioral adjustment. This aligns with a growing body of intervention research. Group counseling programs for parents of autistic children have been shown to improve psychological resilience, life satisfaction, and family functioning, and studies of family-school partnerships have linked them to lower parenting stress and better parent mental health. The family management framework, adapted over the past decade to include children with a range of conditions, gives clinicians a vocabulary for identifying families who feel overwhelmed by their child’s condition and for targeting support toward routine-building and confidence rather than only toward the child’s symptoms.

For the autism research community, the study’s contribution is methodological as much as substantive. By demonstrating that emotional and behavioral problems in autistic children fall into heterogeneous latent profiles, and that family variables predict profile membership even when classification uncertainty is modeled explicitly, the work adds to a shift away from treating autistic children as a single diagnostic block. Researchers have increasingly argued for plural understandings of autism, noting differences in developmental trajectories and symptom combinations, and the profile approach offers a quantitative way to capture that heterogeneity. The study was funded by the National Natural Science Foundation of China and the Natural Science Foundation of Shandong Province, and the authors report no competing financial interests. Its limitations, including the cross-sectional design and reliance on parent report, leave open the causal questions that longitudinal and multi-informant studies will need to address. But for families and clinicians, the message is concrete: the emotional lives of autistic children are entangled with the stress levels and management capacities of the households around them, and supporting the family may be one of the most direct routes to supporting the child.

Subject of Research: Latent profile analysis of parenting stress and family management in relation to emotional and behavioral problems in children with autism spectrum disorder

Article Title: Association of Parenting Stress and Family Management With Emotional and Behavioral Problems in Children With Autism Spectrum Disorders: A Latent Profile Analysis

Article References: Association of Parenting Stress and Family Management With Emotional and Behavioral Problems in Children With Autism Spectrum Disorders: A Latent Profile Analysis. (n.d.). https://doi.org/10.1007/s10803-026-07510-1

Image Credits: AI Generated

DOI: 10.1007/s10803-026-07510-1

Keywords: autism spectrum disorder, parenting stress, family management, emotional and behavioral problems, latent profile analysis, Strengths and Difficulties Questionnaire, R3STEP procedure, Mplus, parent-child dyads, family-based intervention, child mental health, cross-sectional study