Narcissism Moderates the Association Between Autonomy-Supportive Parenting and Adolescents’ Prosocial Behavior

During regular school hours, adolescents whose parents or legal guardians provided consent were instructed to complete an anonymous, questionnaire-based survey under the supervision of trained graduate students. The study procedures were approved by the Institutional Review Board at the Northwest Minzu University and school authorities in China before data collection. In Study 1, conducting an a priori power analysis was challenging because the expected effect size of the interaction relied on the “shape” of the interaction, which was hypothesized to be exploratory in this research project. The researchers thus adopted a more flexible sample size plan.² Specifically, a priori power analysis using G*power (Faul et al., 2007) was conducted for the main effect only (N _minimum = 103 with 80% statistical power) to have the lowest boundary of the required sample size. The small-to-medium effect size was employed in this power analysis based on prior meta-analytical findings (Thielmann et al., 2020; Vasquez et al., 2016). Subsequently, the researchers consulted prior research investigating similar constructs using linear regression to guide the sample size plan (Li et al., 2023).

In total, a convenience sample of 318 secondary school students (M_age = 12.91) volunteered to participate in Study 1.³ The sample had slightly more girls (52.5%) than boys. More details regarding participants’ family backgrounds can be viewed in the supplementary materials (Table S1).

All self-reported, Chinese-administrated measurements were carefully selected based on existing psychometric properties. The brevity and simplicity of the measurements were prioritized to decrease participation burdens, given that completing relevant tasks with long formats might be difficult for young adolescents (Gogol et al., 2014).

Prosocial behavior was assessed using a subscale from the Strengths and Difficulties Questionnaire (Goodman et al., 1997; Liu et al., 2013). This subscale has five items (e.g., “I am helpful if someone is hurt, upset, or feeling ill”), and each was assessed on a 3-point scale varying from 0 (not true) to 2 (certainly true). The summed scores were computed, and higher scores indicated higher prosocial behavior. In Study 1, this subscale had borderline acceptable internal consistency (Cronbach’s alpha = 0.65 and McDonald’s omega = 0.65) but was consistent with prior research on Chinese adolescents (Teuber et al., 2022).

Autonomy-supportive parenting was assessed using the Perceived Parental Autonomy Support Scale, a measure culturally adapted by Wang et al. (2007). This questionnaire includes 16 items, separated by the father’s and mother’s dimensions (e.g., “My father/mother is willing to consider issues from my perspective”). All items were scored on a 5-point scale running from 1 (strongly disagree) to 5 (strongly agree). The average scores were calculated, with higher scores indicating a greater perception of autonomy-supportive situations provided by both parents. This scale had good internal consistency in Study 1 (Cronbach’s alpha = 0.89 and McDonald’s omega = 0.89).

Narcissism was measured using the Childhood Narcissism Scale (Thomaes et al., 2008), a brief assessment demonstrating good psychometric properties in Chinese adolescents (Xu et al., 2020). This scale contains ten items (e.g., “I am a very special person”), rated on a 4-point scale varying from 1 (strongly disagree) to 4 (strongly agree). Average scores were calculated, with higher scores indicating greater narcissism. In Study 1, this scale had adequate internal consistency (Cronbach’s alpha = 0.79 and McDonald’s omega = 0.79).

Study 1 controlled for adolescents’ age, sex, highest education levels of their parents, and family wealth. Sex was dummy-coded, with 0 representing girls and 1 representing boys. Regarding the measurement of parental education levels, adolescents responded to two items, one for the father and another for the mother, represented by three categories (1-middle school or lower, 2-high school, and 3-undergraduate education or higher). The scores across those items were combined into a composite score, with a higher score indicating a higher parental education level. Since young adolescents often have difficulties accurately reporting family income, resulting in a high non-response rate, a four-item family affluence scale was employed (Boyce et al., 2006) as a proxy reflecting the common features of family wealth. A summed score was created for this scale, with higher scores indicating higher family wealth.

Leveraging against R software (R Core Team, 2022), data analyses were first conducted by summary statistics and followed by bivariate correlations using the R package corrplot (Wei & Simko, 2017). Summary statistics present the means, standard deviations, skewness and kurtosis values, and internal consistency for the variables. Since only a small number of missing values (less than 1%) appeared in Study 1, the researchers replaced the missing values with an expectation-maximization algorithm after confirming Little’s (1988) missing completely at random test.

To examine research hypotheses, a hierarchical multiple regression was performed, in which covariates were entered in Step 1, main effects in Step 2, and the interaction in Step 3. Notably, this analysis was conducted using the Ordinary Least Squares method. Before conducting this regression, multicollinearity was examined using the variance inflation factor. The results showed that all predictors were lower than 2. The significant interaction term was further assessed with simple slope analyses, visualized for the moderator between −2 and 2 standard deviations using the R package InterActive (McCabe et al., 2018). This range was selected because it generally represents the variable’s observed range. To provide more comprehensive information for the interaction, the Johnson-Neyman technique was also performed using the R package interactions (Long, 2022) to inspect the regions of significance. This technique complemented simple slope analyses by providing a full range of the moderator affecting the study association from statistically non-significant to significant (Lin, 2020). Following the guidelines outlined by Green (2010), the interaction’s effect size was interpreted as small, medium, and large, corresponding to R² values of 0.008, 0.07, and 0.19.⁴

A series of additional analyses were conducted to examine the robustness of the research findings. First, analyses were conducted by separating father and mother autonomy support because they might play a differential role in adolescents’ prosocial behavior (Vrolijk et al., 2020). Second, analyses were conducted using Poisson regression to evaluate whether the main and interaction effects were replicable. This type of regression was performed considering that a self-reported ordinal scale measured prosocial behavior in Study 1 and that the summed scores of this scale generated ordinal (instead of continuous) data. Prior research has indicated that analyzing an ordinal outcome while assuming its continuity might distort estimates of effect sizes and inflate false positive rates (Rohrer & Arslan, 2021). Third, missing data were handled with regression-based multiple imputations (Enders, 2022). One hundred imputed datasets were generated, and the pooled parameter estimates were obtained according to Rubin’s (1987) recommendations. This extended analysis was done to ensure the robustness of the findings under different imputation methods. Finally, a post hoc power analysis would be implemented to estimate the statistical power associated with the current sample size, considering the specific shape of the interaction and the magnitude of the effect size. When performing those additional analyses, the previously mentioned covariates were also controlled.

Summary statistics and bivariate correlations are all presented in Table S2 for the sake of space limitation. As shown in Fig. 2, autonomy-supportive parenting, narcissism, and prosocial behavior were all positively correlated.

Table 1 presents the results of the regression analysis in Study 1. In total, the model explained an 8% variance in prosocial behavior. The main effects examined in the second step showed that narcissism was positively related to prosocial behavior, whereas autonomy-supportive parenting was not. The first hypothesis was thus not supported. In Step 3, the two-way interaction was significant, further explaining the 2% variance. The effect size of this interaction was small to medium.

Figure 3 illustrates the patterns of this significant interaction, showing that high narcissism amplified the positive association between autonomy-supportive parenting and adolescents’ prosocial behavior. Specifically, autonomy-supportive parenting showed a positive correlation with prosocial behavior at one and two standard deviations above the mean of narcissism. Conversely, at the mean level of narcissism, as well as one and two standard deviations below the mean, this association became non-significant, exhibiting a relatively flat slope. The examination of regions of significance, as illustrated in Fig. S1, showed that autonomy-supportive parenting was positively associated with prosocial behavior in adolescents with 0.05 standard deviations above the mean in narcissism. Additionally, 41.82% of observations in narcissism fell within this region of significance.

Additional steps on this interaction pattern should be taken because a crossover point is shown in Fig. 3. From a descriptive point of view, in the presence of high autonomy-supportive parenting, adolescents presenting with higher (versus lower) narcissism reported a higher intercept for prosocial behavior. In contrast, in the presence of low autonomy-supportive parenting, adolescents with higher narcissism reported a lower intercept for prosocial behavior. This interaction pattern seemingly supported the differential susceptibility hypothesis (Hypothesis 2c). Confirmatory analysis was conducted following point and interval estimates of the crossover point (Widaman et al., 2012). The results showed that the crossover point (C = 2.76, SE = 0.46) and the interval estimate (95% CI = [1.65, 3.67]) fell within the observed range of autonomy-supportive parenting in Study 1. The differential susceptibility hypothesis of narcissism was thus confirmed.

First, the results separating father and mother autonomy support, as reported in Tables S3 and S4 and Figs. S2 and S3, still substantially resembled those with a combined single-score analysis. Second, the results based on Poisson regression converged on the same conclusions as the linear regression, as shown in Table S5. Thus, the original analyses with the combined score of autonomy-supportive parenting and linear regression were retained, and the same analyses were discontinued in the subsequent studies. Third, as detailed in Table S6, the results using multiple imputations were largely consistent with those obtained using the single imputation method. While multiple imputations offer notable advantages over single imputation, the data exhibited a missing data pattern that adhered to the missing completely at random assumption, with a low percentage of missingness. Considering the computational convenience of single imputation, as emphasized by Javanbakht et al. (2022), the original choice of using the single imputation method for the analyses was maintained. Finally, according to the generated interaction shape and identified effect size, a post hoc power analysis showed that the current sample size could yield approximately 80% statistical power with a two-tailed test (Sommet et al., 2023).

In Study 2, a priori power analysis was not performed because data collection was conducted based on a large school collaboration project in which adolescents from (almost) entirely public schools voluntarily participated. A large-scale sample size in this regard would ensure sufficient statistical power to perform the subsequent analyses (Sommet et al., 2023). Participants in Study 2 were convenience-based and totaled 2098 adolescents, with an average age of 15.70 years.⁵ The sample contained slightly fewer girls (44.1%) than boys. Participants’ family backgrounds can be viewed in Table S1.

Summary statistics and correlations between all variables appear in Table S7. As illustrated in Fig. 4, the more autonomy-supportive parenting and narcissism the participants reported, the more prosocial behavior they tended to show. Notably, social desirability was also positively associated with prosocial behavior and autonomy-supportive parenting.

Table 2 shows the results of the regression analysis in Study 2. In total, the model explained an 18% variance in prosocial behavior. Regarding the main effects tested in the second step, both autonomy-supportive parenting and narcissism were positively related to prosocial behavior, accounting for an additional 9% variance in prosocial behavior. The first hypothesis was hence supported. In addition to the main effects, the interaction effect examined in the third step reached significance and helped explain a 1% variance. Consistent with the first study, the effect size of this interaction was small to medium. The second hypothesis was again supported.

Similar to those reported in the first study, in Study 2, high narcissism enhanced this positive association (see Fig. 5). Specifically, a positive correlation was observed between autonomy-supportive parenting and prosocial behavior at the mean, as well as at one and two standard deviations above the mean level of narcissism. However, this association was not significant at one and two standard deviations below the mean level of narcissism. As shown in Fig. S4, autonomy-supportive parenting demonstrated a positive association with prosocial behavior in the region at −0.85 standard deviations above the mean in narcissism, including 78.79% of the observations.

However, the inspection of the crossover point did not support the differential susceptibility pattern identified in the first study. Statistically, the point estimate (C = 1.67, SE = 0.48) was within the observed range of autonomy-supportive parenting in Study 2, but the lowest boundary of this point (95% CI = [0.72, 2.59]) fell outside the observed range. Instead, the interaction pattern seemed to support “vantage sensitivity” (Pluess & Belsky, 2013), in which high narcissism amplified the positive association of autonomy-supportive parenting on adolescents’ prosocial behavior, particularly in the context of high autonomy-supportive parenting, but not that of low autonomy-supportive parenting.

First, the results exhibited, as shown in Table S8, teacher autonomy support was more strongly linked to adolescents’ prosocial behavior than autonomy-supportive parenting. However, the interaction effect between teacher autonomy support and narcissism was not significant, whereas the interaction between autonomy-supportive parenting and narcissism remained significant. The results suggested that teacher autonomy support might facilitate adolescents’ prosocial behavior as well, but the moderating role of narcissism seemed contextually specific to autonomy-supportive parenting. Second, as reported in Tables S9–S12 and the associated Figs. S5–S8, when taking specific dimensions of prosocial behavior into account, the subscales of altruistic behavior, public good, and trait prosociality exhibited similar findings to those with a unidimensional score of prosocial behavior. The inspection of the point and interval estimates showed that the lowest boundaries of 95% CI fell outside of the observed range of autonomy-supportive parenting (C = 0.86, SE = 0.91, 95% [−0.92, 2.64] for altruistic behavior; C = 1.89, SE = 0.46, 95% [0.98, 2.80] for public good; C = 1.61, SE = 0.64, 95% [0.35, 2.87] for trait prosociality). However, with regard to relational behavior subscale, adolescents manifesting high narcissism in Study 2 seemed more susceptible than others to being affected by autonomy-supportive parenting, for better and for worse. The point and interval estimate also supported the differential susceptibility pattern (C = 2.11, SE = 0.32, 95% [1.48, 2.73]), as identified in the first study.

The investigation in Study 3 was conducted in the middle of two consecutive academic years, sampled one year apart. In the first assessment, adolescents reported all previously mentioned control and study variables except for narcissism. One year apart, adolescents again completed those questionnaires and additionally reported their narcissistic traits. ⁷One year follow-up was chosen because autonomy-supportive parenting and prosocial behavior are relatively stable in the short run (Bülow et al., 2022; Te Brinke et al., 2023).

With the support of the principals of two public schools, the researchers recruited 650 adolescents who completed both assessments. Data from 21 adolescents were eliminated from the final sample due to substantial missing information relevant to the third study, resulting in a final sample of 629 adolescents. The average age of the sample was 12.86 years. Within the sample, 48.5% were girls. Participants’ family backgrounds can be viewed in Table S1.

Summary statistics and the matrix of bivariate correlations between the study variables are presented in Table S13. As shown in Fig. 6, prosocial behavior (Time 2) was positively related to autonomy-supportive parenting assessed at both time points. Autonomy-supportive parenting and narcissism were only concurrently associated, and narcissism (Time 2) was not significantly related to prosocial behavior assessed at both time points.

Table 3 presents the regression analysis in Study 3. The overall model explained an 8% variance in prosocial behavior. In the second step, examination of the main effects accounted for an additional 3% variance. Change in autonomy-supportive parenting was positively related to prosocial behavior (Time 2). Beyond the main effects, a significant interaction effect between change in autonomy-supportive parenting and narcissism was found, explaining an additional 1% variance. Similar to those reported in Studies 1 and 2, the effect size of this interaction was small to medium. In Study 3, the first and second hypotheses were both supported.

The inspection of the two-way interaction exhibited similar patterns as those indicated in Studies 1 and 2 (see Fig. 7). Specifically, Study 3 revealed a positive correlation between changes in autonomy-supportive parenting and adolescents’ prosocial behavior at the mean level of narcissism, as well as at one and two standard deviations above this mean. However, this relationship was not significant at one and two standard deviations below the mean level of narcissism. As illustrated in Fig. S9, changes in autonomy-supportive parenting were positively associated with prosocial behavior in adolescents who scored −0.75 standard deviations above the mean in narcissism. Furthermore, 81.72% of the observations in narcissism were within this region of significance. In Study 3, the point estimate (C = −0.40, SE = 0.53) and the associated confidence boundaries (95% CI [−1.44, 0.64]) fell within the observed range of autonomy-supportive parenting. The differential susceptibility hypothesis of narcissism was again supported.

First, as reported in Table S14, neither changes in teacher autonomy support nor changes in peer autonomy support exhibited significant main and interaction effects. Those supplementary analyses again suggested that the moderating role of narcissism was situationally specific to autonomy-supportive parenting. Second, a post hoc power analysis, indicated that the current sample size could yield approximately 95% statistical power (Sommet et al., 2023).

With the permission of parents or legal guardians, adolescents were informed that they were signed up for an anonymous, voluntary, non-harm experimental study investigating memory recall and prosocial behavior. Data collection was conducted within a secondary school setting, with the permission and cooperation of the school administration. This study took place during regular school hours in designated classrooms, where the adolescents were first instructed to complete the manipulation task and then answer a few well-validated questionnaires. Upon completing the questionnaires, adolescents were allowed to enter a lottery, with the chance to receive 1–20 Chinese yuan as compensation.

Participants in Study 4 were 118 adolescents (M _age = 12.42; 53.4% girls) randomly assigned into two manipulation conditions: autonomy-supportive context (n = 59) and autonomy-suppressive context (n = 59). No significant differences in participants’ sociodemographic characteristics were revealed between the two conditions. Although obtaining a large and well-powered sample size is ideal for probing this hypothesized interaction effect, researchers must balance the resource limits and the desired statistical power, particularly considering that collecting quasi-experimental data at school become increasingly costly and challenging (Mayeux & Kraft, 2017). However, it should be noted that similarly sized adolescent samples have been employed in prior experimental studies (e.g., Li et al., 2023). Participants’ family backgrounds in Study 4 can be viewed in Table S1.

Experimental study often operates under strict time constraints, necessitating more efficient data collection methods. Using the brief version of the scales allows researchers to focus on the core dimensions of the variable of interest and help prevent participant fatigue, ensuring that participants remain attentive and provide accurate responses throughout this investigation.

Prosocial behavior was measured by a brief version of the scale employed in Study 2. This 15-item scale has been validated among Chinese adolescents, showing similar psychometric properties to the longer version (Zhang & Kou, 2011). In Study 4, the internal consistency improved compared to the second study (Cronbach’s alpha = 0.94 and McDonald’s omega = 0.94).

Autonomy-supportive parenting was manipulated by instructing adolescents to vividly recall an experience related to interacting with their parents. This type of recall manipulation task has been demonstrated to be effective when studying social relationships (Dang & Liu, 2023). Specifically, adolescents were randomly assigned to one of the manipulation conditions (autonomy-supportive versus autonomy-suppressive). In the autonomy-supportive condition, adolescents were given written instructions, including several practices documented as autonomy-supportive interpersonal styles (e.g., acknowledging adolescents’ perspectives, providing choices, and meaningful rationales). In contrast, in the autonomy-suppressive condition, adolescents were given the meaningfully opposite instructions, containing controlling language by using verbs (e.g., “should”). Those instructions, described in detail in the supplementary materials, were crafted based on the conceptualization of autonomy support and prior research (Benita et al., 2014; Jungert et al., 2021). After reading those instructions, participants were first asked to recall relevant experiences and write down the real experiences in full detail and subsequently completed a four-item manipulation check questionnaire (e.g., “At this moment, I feel that my parents are willing to consider issues from my perspective.”; Cronbach’s alpha = 0.84 and McDonald’s omega = 0.84) rated on a 5-point scale.

Narcissism was measured using the same scale employed in Study 2. In Study 4, the internal consistency was good (Cronbach’s alpha = 0.72 and McDonald’s omega = 0.74).

Due to the quantity constraints of this survey, an abridged five-item version of the social desirability scale (Schuessler et al., 1978) was administered in Study 4. The researchers intentionally limited this scale to the most representative items while including enough items to capture this construct effectively. The internal consistency of this scale was adequate (Cronbach’s alpha = 0.75 and McDonald’s omega = 0.74).

The inherent challenge presented by a small sample size might increase the risk of generating biased model estimates. Monte Carlo cross-validation (with 200 repetitions and holding out 20% of the sample in each repetition) was thus conducted to provide a more robust and reliable estimate of model performance (Song et al., 2021). Specifically, Monte Carlo cross-validation randomly splits the data into training and validation sets over multiple iterations. This flexibility is particularly advantageous with a small data set, as it allows for a wide variety of data combinations, enhancing the robustness of the model estimate.

Adolescents in the autonomy-supportive condition reported significantly higher levels of autonomy-supportive parenting (M = 3.95, SD = 0.64) than those in the autonomy-suppressive condition (M = 2.60, SD = 0.36), F (1, 116) = 3.16, p < 0.001. The manipulation was therefore deemed effective in creating conditions that led to autonomous orientations.

Summary statistics and intercorrelations for the study variables are provided in Table S15. Figure 8 visualizes the correlation matrix for the variables.

Table 4 summarizes the results of the regression analysis in Study 4. The mode totally explained a 36% variance in prosocial behavior. The main effects tested in the second step revealed no significant association between manipulation conditions and adolescents’ prosocial behavior. Yet the two-way interaction examined in the final step approached significant levels, additionally explaining up to a 7% variance. The effect size of this interaction was medium. The second hypothesis was again supported.

As shown in Fig. 9, at one and two standard deviations above the mean level of narcissism, adolescents in the autonomy-supportive condition exhibited significantly higher intercepts for prosocial behavior compared to those in the autonomy-suppressive condition. In contrast, at two standard deviations below the mean level of narcissism, adolescents in the autonomy-supportive condition displayed significantly lower intercepts for prosocial behavior than their counterparts in the suppressive condition. Furthermore, at the mean and one standard deviation from the mean level of narcissism, the intercept difference in prosocial behavior between the two conditions was not statistically significant. As depicted in Fig. S10, the experimental conditions showed a positive association with prosocial behavior in adolescents who scored 0.15 standard deviations above the mean in narcissism. Notably, 41.53% of the observations were within this region of significance.

The interaction patterns echoed those identified in the previous studies. Again, adolescents presenting high narcissism showed heightened susceptibility to both the beneficial and adverse effects of autonomy-supportive parenting. The inspection of point and interval estimates showed that the crossover point (C = 2.44, SE = 0.19) and 95% CI [2.06, 2.82] fell within the observed range of narcissism in Study 4.

The results of cross-validation exhibited that when the regression model is generalized to another sample, 30% of the variance in prosocial behavior will likely be accounted for the variables under investigation. Additionally, the cross-validated mean square error of 0.68 suggested that, on average, the model predicting prosocial behavior scores will likely deviate from the observed scores in the new sample by 0.79 points on a 7-point scale. Given those prediction accuracy indices, the findings of Study 4 were validated internally and might have good generalization ability.

Along with those findings, the current findings must also be evaluated within the context of several limitations. First, prosocial behavior in the present studies was predominantly constructed as a global and homogeneous variable. As indicated by the findings in Study 2, the interaction between autonomy-supportive parenting and narcissism might be distinctively linked to different types of prosocial behavior. Future research should develop a comprehensive evaluation of the prosocial behavior spectrum based on diverse motives, situations, and target to relate them to autonomy-supportive parenting and narcissism (Carlo & Padilla-Walker, 2020). Similarly, more recent theoretical movements have proposed differentiating several underlying components of narcissism (Crowe et al., 2019; Miller et al., 2021), although its multidimensional assessment in youth is still in infancy. Future research might therefore also consider unpacking each dimension of narcissism. Second, delving into the moderating role of narcissism holds significant theoretical and practical relevance, but the conditional process underlying the positive association between autonomy-supportive parenting and prosocial behavior is far more complex than what is investigated. One agenda for future initiatives should thus elaborate on this positive association by exploring additional dispositional moderators (e.g., sympathy; Xu & Zhang, 2023; grit; Lan et al., 2019) to discuss the unexplained variance found in the present studies. A third limitation that needs to be mentioned is that genetic factors might present a potential confound in the current findings because autonomy-supportive parents might transfer genetic dispositions associated with prosocial behavior to their children (Kretschmer, 2023). Future studies might use a genetically informed design to disentangle potential genetic and environmental processes that explain these study associations. Fourth, examining the crossover two-way interaction, by theory, requires a large sample size (Sommet et al., 2023), but unfortunately, due to time or financial constraints, two of the four studies contained limited sample sizes, which might generate biased estimates and inflated Type I errors. Thus, readers should be cautious when interpreting narcissism as a differential susceptibility trait, and future studies should replicate this finding using well-powered surveys/experiments with large samples. Finally, this research included only Chinese adolescents. Although that cultural context is well-suited to addressing study associations, one caveat is that the generalizability of the current findings might be restricted by certain cultural boundaries. Future studies should consider recruiting samples from multiple cultural contexts to replicate the present findings.

Those limitations notwithstanding, the present studies demonstrate important implications at both the theoretical and practical levels. Regarding theoretical implications, this research contributes to enriching the universality of the self-determination theory in an East Asian society. The findings also add to the growing bodies of work by adopting the comprehensive self-determination theory framework to gain a deep understanding of the complex interaction between socialization experiences and dispositional characteristics related to adolescents’ prosocial behavior. Further, exploring the association between narcissism and prosocial behavior contributes to discussing the general Dark Triad framework in relation to the association between narcissism and the remaining two Dark traits. The current findings challenge the predominantly negative connotation associated with narcissism, suggesting that its nuanced role in adolescents’ prosocial behavior warrants a more differentiated consideration. Especially under the influence of autonomy-supportive parenting, certain aspects of narcissism may paradoxically enhance adolescents’ prosocial behavior. Additionally, the present studies are of theoretical relevance to the differential susceptibility theory because narcissistic adolescents might exhibit pronounced responses to both negative and positive situational influences.

Through an examination of those associations with diverse research designs, this research also provides important insights into developing practical activities. First, the research suggests that facilitating autonomy-supportive parenting practices is beneficial to adolescents’ prosocial behavior. Educators and practitioners can, for instance, organize structured presentations via parent meetings online or at school, highlighting the critical roles of autonomy-supportive practices in adolescents’ prosocial behavior. During such meetings, educators and practitioners can also exemplify specific autonomy-supportive practices for parents, such as considering the adolescent’s point of view and providing meaningful rationales for their guidance or decisions. Importantly, the findings also indicate that autonomy-supportive parenting might not benefit all adolescents equally, arguing against the one-size-fits-all approach. The moderating role of narcissism played contributes to developing personalized initiatives that hold considerable promise for educators and practitioners. For instance, according to the assessment of autonomy-supportive parenting and narcissism, educators or practitioners can better locate adolescents in need of improved intervention or prevention efficacy.