A total of 76 children and adolescents aged 9–15 years participated in the study, including 40 with a clinical diagnosis of ASD and 36 TD controls. Participants in the ASD group were recruited from patients receiving care at the Department of Psychiatry, Nara Medical University Hospital, between March 2021 and January 2024. Participants with ASD were recruited from an outpatient psychiatric service during the study period, with eligible patients enrolled upon providing informed consent. TD participants were recruited employing convenience sampling from the local community. Diagnoses of ASD were established by at least two experienced psychiatrists or clinical psychologists according to the criteria of the Diagnostic and Statistical Manual of Mental Disorders, 5th edition (DSM-5) (1). To ensure diagnostic accuracy, only individuals who also met ASD classification criteria on the Autism Diagnostic Observation Schedule, Second Edition (ADOS-2; Modules 3–4) were included. The ADOS-2 is widely recognized as the gold-standard observational assessment for ASD, with strong validity in differentiating autism from other developmental and psychiatric conditions (23).

Cognitive functioning was assessed using the Das–Naglieri Cognitive Assessment System (DN-CAS), a theory-driven instrument that evaluates four cognitive processing domains: planning, attention, simultaneous processing, and successive processing (24). Participants with a total DN-CAS score below 70 were excluded to minimize the influence of intellectual disability, consistent with previous studies that regard a score below 70 as indicative of significant cognitive impairment (24). Individuals with other neurological or psychiatric disorders were also excluded from the study.

Demographic information, including age and sex, was obtained from caregiver questionnaires and medical records. Written informed consent was obtained from all parents or legal guardians, and assent was obtained from the children when appropriate. The study protocol was approved by the Ethics Committee of Nara Medical University (Approval Nos. 1319, 2535).

The Autism Spectrum Quotient–Japanese version (AQ-J) is a 50-item self-report questionnaire that evaluates autistic traits across five core domains: social skills, attention switching, attention to detail, communication, and imagination. Each item is rated on a 4-point Likert scale, with higher scores indicating greater levels of autistic traits. Example items include “I find it difficult to make new friends” (social skills), “I frequently get so absorbed in one thing that I lose sight of other things” (attention switching), and “I find it difficult to read between the lines in conversations” (communication). The AQ-J has been validated for use in Japanese populations and demonstrates good internal consistency (Cronbach’s α = 0.77–0.86) (25). Because all participants were under 16 years of age, assessments were completed by a parent or caregiver.

The Rosenberg Self-Esteem Scale (RSES) is a widely used 10-item measure designed to estimate global self-worth and self-esteem (26). It includes positively and negatively worded statements, such as “I take a positive attitude toward myself” and “I feel I do not have much to be proud of.” Each item is rated on a four-point Likert scale, ranging from “strongly disagree” to “strongly agree,” with higher scores showing greater self-esteem. The Japanese version, translated and validated by Yamamoto et al., has shown high internal consistency (Cronbach’s α = 0.81) and good construct validity (27, 28). Individuals with ASD often present with low self-esteem, making this measure especially useful for capturing clinically meaningful variation in self-worth within this population. The RSES, originally developed for adolescents and widely used in late childhood and adolescence, was considered appropriate for participants aged 9–15 years (26, 29). The scale was completed by the children themselves, with brief explanations provided to younger participants as needed; caregivers did not influence responses.

Parenting style was assessed using the Parental Nurturance Profile Scale (PNPS), a validated instrument that evaluates parental attitudes and behaviors across supportive and maladaptive domains. The PNPS consists of multiple subscales, including involvement, positive responsivity, respect for will, overprotection, inconsistency, harsh discipline, positive parenting, and negative parenting. Higher scores on positive domains (e.g., involvement, responsivity, respect for will) indicate greater parental warmth and support, whereas higher scores on negative domains (e.g., harsh discipline, overprotection, inconsistency) reflect more controlling or maladaptive parenting practices. The PNPS has been widely applied in Japanese child and adolescent research, where it has illustrated strong internal consistency, construct validity, and clinical utility in distinguishing parenting behaviors associated with children’s socio-emotional adjustment (30, 31). Although the PNPS was not developed specifically for ASD, it has been applied in clinical and developmental research including children with neurodevelopmental and emotional–behavioral difficulties. In this study, the PNPS was used to capture general dimensions of parenting attitudes relevant across diagnostic groups.

Analyses were conducted in the combined TD + ASD sample and separately in the ASD and TD groups to examine dimensional and group-specific associations among autistic traits, parenting attitudes, and self-esteem. Within the ASD group, parallel regression analyses using parent-reported autistic traits (AQ-J) and clinician-rated autism severity (ADOS-2) were performed to distinguish subjective trait expression from clinician-observed symptom severity.

All analyses were two-tailed with a significance level set at α = 0.05. Descriptive data are presented as mean ± standard deviation (SD). Group differences between individuals with ASD and TD individuals in continuous variables were examined using independent-samples t-tests, whereas sex distribution was compared using chi-squared (χ²) tests. Effect sizes are reported as Cohen’s d for continuous variables and Cramer’s V for categorical variables. The 95% confidence intervals refer to the mean differences between groups.

To examine associations between autistic traits, parenting attitudes, and self-esteem, linear regression analyses were first conducted. Self-esteem, assessed by the RSES total score, was entered as the dependent variable. Parent-reported autistic traits (AQ-J total score) and parenting attitudes (PNPS assessed by the parent) were entered as independent variables. In the pooled TD + ASD sample, age was included as a covariate in Model 1, and age and diagnostic status (ASD vs. TD) were included in Model 2. Additional regression analyses were conducted separately in the TD group and the ASD group, with age included as a covariate. In analyses restricted to the ASD group, clinician-rated autism symptom severity assessed by the ADOS-2 total score was examined alongside parenting attitudes to compare the relative associations of clinician-rated severity and parent-reported autistic traits with self-esteem.

Next, path analyses were conducted to examine the direct and indirect associations among autistic traits, parenting attitudes, and self-esteem using observed variables. Autistic traits (AQ-J total score), parenting attitudes, and self-esteem were modeled simultaneously, with parenting attitudes hypothesized to mediate the association between autistic traits and self-esteem. Parenting attitudes were modeled separately as positive and negative domains based on PNPS subscales. Analyses were performed for the pooled TD + ASD sample as well as separately for the TD and ASD subgroups. Given the cross-sectional design, these paths represent statistical associations rather than causal relationships. Two-tailed p values and 95% confidence intervals were calculated, and age was included as a covariate in all models.

Finally, partial correlation analyses were conducted as supplementary analyses to further characterize the associations among autistic traits, parenting attitudes, and self-esteem while controlling for potential confounding variables. In the pooled TD + ASD sample, partial correlations were calculated controlling for age and diagnostic status. In analyses restricted to the ASD group, partial correlations were calculated controlling for age only and included associations with ADOS-2 total scores. Partial correlation coefficients (r) and corresponding significance levels are reported in Supplementary Tables 5, 6.

A total of 76 participants were included in the analysis (ASD group: n = 40; TD comparison group: n = 36). There were no significant group differences in age (p = 0.76, d = -0.07) or sex distribution (p = 0.71). Group comparisons of demographic variables, autism-related traits, self-esteem, and parenting style are summarized in Table 1. Regarding autism-related traits assessed by the AQ-J, the ASD group had significantly higher scores than the TD group in social skills (p < 0.01, d = -1.43), attention switching (p < 0.01, d = -1.27), local details (p = 0.02, d = -0.68), communication (p < 0.01, d = -1.62), imagination (p < 0.01, d = -0.94), and total AQ-J score (27.58 ± 8.08 vs. 14.30 ± 6.42; p < 0.01, d = -1.83). On the RSES, the ASD group reported significantly lower self-esteem than the TD group (p < 0.01, d = 0.94). For parenting style, a significant between-group difference was observed only for the PNPS subscale of involvement (p = 0.02, d = 0.74). No significant differences were observed on other PNPS subscales, including positive responsivity (p > 0.99, d = -0.05), respect for will (p > 0.99, d = 0.17), overprotection (p = 0.20, d = -0.53), inconsistency (p > 0.99, d = 0.07), harsh discipline (p = 1.00, d = -0.17), positive parenting (p > 0.99, d = 0.33), and negative parenting (p > 0.99, d = -0.32). Within the ASD group, symptom severity was evaluated using the ADOS-2. The mean scores were 3.5 ± 1.0 for the language and communication domain, 11.6 ± 2.9 for reciprocal social interaction, and 0.4 ± 0.8 for restricted and repetitive behaviors, with a mean total score of 12.1 ± 3.1. These findings confirm the presence of core autism-related impairments in communication and social reciprocity within the ASD group.

In the pooled sample including TD and ASD participants, linear regression analyses were conducted to examine the associations of parent-reported autistic traits and parenting attitudes with self-esteem (Table 2). In Model 1, which adjusted for age, higher levels of parent-reported autistic traits (AQ-J total score) were significantly associated with lower self-esteem (β = −0.471, p < 0.001). Neither positive parenting (β = 0.186, p = 0.104) nor negative parenting (β = −0.021, p = 0.858) showed a significant association with self-esteem. In Model 2, which additionally adjusted for diagnostic status (ASD vs. TD), the negative association between autistic traits and self-esteem remained significant (β = −0.371, p = 0.011). Parenting attitudes again did not show significant independent associations with self-esteem (positive parenting: β = 0.180, p = 0.116; negative parenting: β = −0.033, p = 0.776). These findings indicate that autistic traits were independently associated with self-esteem across the pooled sample, even after accounting for age and diagnostic group.

To examine whether these associations differed by diagnostic group, additional regression analyses were conducted separately within the TD group (Supplementary Table 1). After adjusting for age, parent-reported autistic traits were not significantly associated with self-esteem in TD participants (β = −0.087, p = 0.616). Similarly, neither positive parenting (β = 0.283, p = 0.122) nor negative parenting (β = 0.063, p = 0.733) showed a significant association with self-esteem in this group. Parallel analyses were conducted within the ASD group (Supplementary Table 2). In contrast to the TD group, higher levels of parent-reported autistic traits were significantly associated with lower self-esteem after adjusting for age (β = −0.442, p = 0.008). Positive parenting (β = 0.115, p = 0.519) and negative parenting (β = −0.171, p = 0.345) were not significantly associated with self-esteem.

In the pooled TD + ASD sample, parent-reported autistic traits were associated with parenting attitudes (Table 3). Specifically, higher autistic trait levels were associated with higher levels of negative parenting at the coefficient level after adjustment for age (Model 1: β = 0.32, p = 0.005). This association remained statistically significant after further adjustment for diagnostic status (ASD vs. TD) (Model 2: β = 0.40, p = 0.011), although the overall model fit did not reach statistical significance. In contrast, autistic traits were not significantly associated with positive parenting in either model, although a trend toward lower positive parenting was observed when controlling for age alone (Model 1: β = −0.23, p = 0.051).

When analyses were stratified by diagnostic group, no statistically significant associations were observed between autistic traits and parenting attitudes in either group. In the TD group (Supplementary Table 3), higher autistic traits showed a trend-level association with increased negative parenting (β = 0.31, p = 0.058), whereas the association with positive parenting was weak and non-significant (β = −0.20, p = 0.225). Similarly, in the ASD group (Supplementary Table 4), autistic traits were not significantly associated with parenting attitudes; however, a non-significant tendency toward greater negative parenting was observed (β = 0.29, p = 0.092), while no association was evident for positive parenting (β = −0.13, p = 0.457).

Within the ASD group, linear regression analyses were conducted to examine the associations of parenting attitudes and clinician-rated autism symptom severity with self-esteem (Table 4). Self-esteem (RSES total score) was entered as the dependent variable, and positive parenting, negative parenting, and ADOS-2 total scores were entered as independent variables, with age included as a covariate. Clinician-rated autism symptom severity, as assessed by the ADOS-2 total score, was not significantly associated with self-esteem (β = 0.057, p = 0.732). Similarly, neither positive parenting (β = 0.107, p = 0.591) nor negative parenting (β = −0.305, p = 0.123) showed a significant association with self-esteem in this model.

Partial correlation analyses in the pooled TD + ASD sample (Supplementary Table 5) provided complementary support for the regression findings. After controlling for age and diagnostic status, higher parent-reported autistic traits were moderately associated with lower self-esteem (r = −0.52, Bonferroni-adjusted p < 0.001). Autistic traits were also positively associated with negative parenting (r = 0.32, p < 0.05) and inversely associated with positive parenting, although the latter did not reach statistical significance after correction. In addition, positive parenting was positively correlated with self-esteem (r = 0.30, p < 0.05), whereas negative parenting showed a weak and non-significant inverse association with self-esteem after correction. These correlational patterns were consistent with the direction of effects observed in the regression analyses.

Within the ASD group (Supplementary Table 6), partial correlation analyses controlling for age indicated that higher parent-reported autistic traits were significantly associated with lower self-esteem (r = −0.50, Bonferroni-adjusted p < 0.05). In contrast, clinician-rated autism symptom severity (ADOS-2 total score) was not significantly correlated with self-esteem. Negative parenting showed a moderate, non-significant inverse association with self-esteem, whereas positive parenting was weakly and positively correlated with self-esteem; however, neither association remained statistically significant after Bonferroni correction. Autistic traits were moderately correlated with negative parenting, but this association did not reach adjusted significance.

Children and adolescents with ASD exhibited significantly lower self-esteem than their TD (8, 32). Low self-esteem in autism is well documented and often co-occurs with depression and anxiety (9, 17). In both groups, self-esteem has been shown to be inversely associated with depressive symptoms (7), reinforcing its significance as a risk factor for mental health difficulties. Beyond group comparisons, higher levels of autistic traits were associated with lower self-esteem in the combined sample, consistent with evidence that even subclinical traits may undermine self-concept (33). In contrast, autism symptom severity as assessed by the ADOS-2 was not directly associated with self-esteem, suggesting that clinician-rated symptom severity alone may be less closely related to self-evaluative outcomes. Previous studies have indicated that factors such as alexithymia, rather than autism severity per se, are associated with low self-esteem in autistic adolescents (34). Taken together, these findings suggest that vulnerabilities in self-esteem may reflect not only autistic traits but also broader social-emotional characteristics and contextual responses. Given that poor self-esteem in ASD is associated with depression, anxiety, and suicidality (35, 36), identifying risk and protective factors remains an important clinical and research priority.

Our findings showed no broad differences in overall parenting style between the ASD and TD groups, consistent with meta-analytic evidence indicating comparable levels of parental warmth and support across diagnostic groups (37). However, higher levels of autistic traits were associated with more negative parenting attitudes and showed modest associations with lower positive parenting, echoing prior reports that parents of children with ASD may exhibit slightly greater control or negativity (38, 39). These associations may reflect the increased stress and challenges associated with managing communication difficulties or behavioral characteristics (19, 40). Importantly, substantial variability was observed within each group, and many parents of children with ASD reported highly supportive parenting attitudes. Taken together, these findings suggest that differences in parenting attitudes may be less strongly related to diagnostic status per se than to individual stress levels, coping strategies, and contextual supports (37–39).

Parenting style was related to self-esteem outcomes at the correlational level. Tendencies toward negative parenting—such as harshness, inconsistency, or overprotection—were associated with poorer self-esteem, with a more pronounced pattern observed in youth with ASD (42). Previous research has shown that harsh parenting is linked to greater emotional and behavioral difficulties in children (43), which may in turn undermine self-evaluative processes. In contrast, positive parenting characterized by warmth, acceptance, and support for autonomy showed modest associations with higher self-esteem, consistent with findings in neurotypical and clinical populations (41, 42). Positive parenting has also been associated with reduced vulnerability to stress and adversity (22), suggesting that supportive parenting contexts may be relevant to self-esteem development in autistic youth. Although subgroup analyses did not consistently yield statistically significant results, the overall pattern of associations across groups was directionally consistent.

Parenting processes are often conceptualized within transactional frameworks, in which child characteristics and parenting behaviors are thought to influence one another over time. Although cross-sectional data cannot establish causality, such models highlight the potential for bidirectional associations between child difficulties and parental responses (44). In this context, our findings are broadly consistent with a transactional perspective, in that higher autistic traits tended to co-occur with elevated negative parenting tendencies, which were in turn associated with lower self-esteem (43). Conversely, positive parenting attitudes were modestly associated with higher self-esteem, suggesting that supportive relational contexts may be relevant to children’s self-evaluative processes. The absence of a direct association between clinician-rated autism severity (ADOS-2) and self-esteem further suggests that relational and contextual factors may be more closely linked to self-esteem than symptom severity alone. Other factors, such as social support and self-regulation, may also play important roles in shaping self-esteem (17).

Promoting self-esteem in children with ASD may be an important clinical consideration, given its broad associations with mental health and overall well-being (6). Prior intervention studies suggest that parent-focused programs aimed at enhancing warmth, praise, and autonomy support can improve parent–child interactions (45), and meta-analytic evidence indicates benefits for a range of ASD-related outcomes (46). Because parenting practices are modifiable, supportive parenting contexts may be relevant targets for psychosocial support. In addition, parental well-being warrants attention, as higher parental stress has been associated with harsher parenting practices (40, 47). Mindfulness-based programs have been shown to reduce parental stress and reactivity (47), highlighting the potential value of approaches that address child and parent factors.

This study is limited by its cross-sectional design, reliance on parent- and self-reported measures, and modest sample size. Longitudinal research is needed to clarify the temporal ordering and potential causal directions among autistic traits, parenting attitudes, and self-esteem (48). Accordingly, all associations identified in the present analyses should be interpreted as correlational rather than causal. Future studies would benefit from incorporating observational and multi-informant approaches, as well as larger and more diverse samples. In addition, peer relationships and experiences such as bullying warrant further consideration, given their established associations with self-esteem (15, 49). Cross-cultural replication will also be important to evaluate the generalizability of these findings.