Introduction

Since the first case of coronavirus disease-2019 (COVID-19) was detected, mankind has waged a protracted war against the virus. As of 18 May 2022, this war without gunpowder has lasted more than two years, plundering the lives of over six million people and the freedom of survivors (WHO, 2022). Although the outbreak of COVID-19 cannot halt individual growth, it indeed has multifactorial effects on human development, especially for young people. Previous studies have well documented that adolescents suffer from more mental disorders (e.g., depression and anxiety), more physical problems (e.g., neuroinflammation and neuroimmunoendocrine changes), as well as more behavioral problems (e.g., hyperactivity-inattention and emotional problems) during the lockdown period (De Miranda et al., 2020; De Figueiredo et al., 2021; Liu et al., 2011). A high level of school resources means that the school provides a caring and encouraging climate, clear rules and harmonious relationships, and adolescents actively participate in learning activities (Benson et al., 2011). According to the developmental assets theory, adolescents who obtain more school resources will develop more flourishing achievements and fewer harmful problems in the future (Scales et al., 2000). In the current issue, adolescents who experience more school resources may develop fewer problematic behaviors. Empirical studies have provided numerous indirect evidence on the association between school resources and problematic behaviors. For instance, a meta-analysis of longitudinal studies suggested that school climate was negatively associated with problem behaviors (Reaves et al., 2018). As for the teacher-student relationship, students who had less conflict and were more intimate with teachers reported less involvement in problematic behaviors (Pakarinen et al., 2018). Moreover, the higher level of school engagement could predict fewer problem behaviors among Chinese adolescents (Hu et al., 2019).

More specifically, school resources may be an effective predictor of IGD, school bullying, and victimization among adolescents. There are two theories that could contribute to the understanding of the complex linkage. First, Bandura (1978) claims in the social learning theory that individual behavioral outcomes are shaped by the environment. Adolescents may learn bullying behaviors through observation of aggressive behaviors at school. However, schools with diverse resources, such as an encouraging atmosphere, clear rules, and harmonious relations, will guide students to use non-violent approaches to cope with conflicts and spare no effects to prevent adolescents from exposure to aggressive behaviors. In this aspect, school resources are potential protectors for adolescents’ bullying behaviors, which has been repeatedly confirmed among Chinese adolescents in previous longitudinal studies. Teng and his colleagues (2020) revealed that students with more positive perceptions of school climate reported lower levels of bullying perpetration over time. As for victimization, Chen et al. (2021) suggested that adolescents who reported more engagement in school would be less exposed to violence in school and cyberspace. Moreover, Nie et al. (2021) demonstrated that perceptions of school climate mitigated the positive association between bullying and victimization.

Second, the general strain theory highlights that individuals may adopt maladaptive behaviors to alleviate negative emotions caused by environmental stress (Agnew, 1992). From this perspective, students in schools with tense atmospheres and discordant relationships will feel unhappy, unsupported, and anxious, and then they will develop maladaptive behaviors such as IGD to alleviate these negative feelings. On the contrary, if schools provide a comfortable environment and abundant resources, students may positively involve themselves in academic activities and achieve more. To a certain degree, schools providing more positive resources might play an important role in the prevention of IGD among students. Previous studies have indicated that adolescents who experience more school engagement, positive perceptions of school climate, and support from teachers and classmates are less likely to develop IGD (Yu et al., 2015; Ma et al., 2017; Zou et al., 2022). Furthermore, ** problem behaviors by avoiding or reducing these risks. Existing studies have ignored the fact that young people indeed have positive traits to protect themselves, and the environment also provides many positive factors promoting individual growth (Scales et al., 2000). Moreover, little attention has been paid to the cumulative effects of diverse positive factors. To address these knowledge gaps, the present study, based on the problem behavior theory (Jessor, 1987), aims to evaluate the protective roles of school resources and self-control ability in adolescents’ problematic behaviors in the post-pandemic era. Considering the theoretical and empirical evidence on the associations between school resources, self-control, and problematic behaviors, the present study hypothesizes that (Fig. 1):

Fig. 1
figure 1

The conceptual model of the relationship between school resources and problematic behaviors. Note: SR = School resources; SC = Self-control; SB = School bullying; SV = School victimization; IGD = Internet gaming disorder; T1 = Time 1, T2 = Time 2

Full size image

  • Hypothesis 1: School resources will positively predict self-control ability concurrently and longitudinally.

  • Hypothesis 2: School resources will negatively predict problematic behaviors concurrently and longitudinally.

  • Hypothesis 3: Self-control ability will mediate the relationship between school resources and problematic behaviors.

  • Hypothesis 4: One problematic behavior will mediate the relationship between school resources and other problematic behaviors.

  • Hypothesis 5: Self-control ability and one problematic behavior will sequentially mediate the relationship between school resources and other problematic behaviors.

Method

Participants and procedure

Using the random cluster sampling method, participants were recruited from one public middle school and one public high school in Hubei province, Central China. A sample of 789 students (Mage = 14.00 years, SD = 2.05, 418 boys) participated in the first data collection (T1: September, 2021) and completed self-reported questionnaires. 51 dropped out of the follow-up data collection (T2: January, 2022) for various reasons, such as absence. The final sample consisted of 738 adolescents and 70.05% of them were middle school students. Among the current sample, 50.65% were the only children in their families, and 95.44% were from families at or above the average economic level in the local cities. Moreover, Chi-square test and t-test reported that there were insignificant differences in gender (χ 2 (1) = 0.748, p = 0.387), T1 school resources (t (787) = -0.853, p = 0.394), T1 school bullying (t (52) = 1.261, p = 0.213), T1 school victimization (t (52) = 1.687, p = 0.098), and T1 IGD (t (53) = 1.833, p = 0.072) between the final sample and the lost sample, which indicated that the current findings would not be biased due to attrition.

Data was collected in the post-pandemic era. The Research Ethics Committee of the College of Education and Sports Sciences at Yangtze University provided ethical approval for the present study. Before the first formal survey, written consent was obtained from the school leaders, parents, and adolescents through the explanation of detailed information about the study. Well-trained researchers with negative test results for COVID-19 performed the survey within 50 min during school hours. Honest responses were encouraged by informing the anonymity of the survey. Participants did not receive any kind of gift for their participation. The procedures of the two surveys were the same.

Measures

School resources

The School Assets Subscale (SAS) from the Chinese version of the Developmental Assets Profile was adopted to assess adolescents’ school resources (Scales, 2011). The SAS includes 10 items, and one example item is “I have a school that cares about kids and encourages them”. Response options range from “1 = not at all or rarely” to “4 = extremely or almost always” on a four-point Likert scale. A higher total score indicates more quantity and higher quality of school resources. In previous empirical studies, this subscale suggested good reliability and validity in Chinese adolescents (Chang et al., 2020). In this study, the Cronbach’s α of this scale were 0.869 at T1 and 0.909 at T2.

Self-control

The Chinese version of the Brief Self-Control Scale (BSCS) was employed to measure adolescents’ self-control ability (Tangney et al., 2004; Unger et al., 2016). It consists of 13 items and rates on a five-point Likert scale ranging from “1 = not like me at all” to “5 = very much like me”. One sample item is “Sometimes I can’t stop myself from doing something, even if I know it is wrong”. A higher total score suggests a higher level of self-control ability. This scale had high internal consistency in the Chinese adolescent sample (Gu, 2020). In this study, its Cronbach’s α were 0.752 at T1 and 0.788 at T2.

School bullying and victimization

The Chinese version of the School Bullying and Victimization Scale (SBVS) was adopted to evaluate adolescents’ school bullying and victimization in the past six months (Deng et al., 2018). This scale composes two subscales, each of which includes six items. It rates on a five-point Likert scale ranging from “1 = never” to “5 = several times a week”. One example is “Others call me bad nicknames, scold me, make fun of me and ridicule me”. A higher total score on a subscale reflects a higher level of school bullying or victimization. This measure showed good reliability and validity among Chinese adolescents (Deng et al., 2018). In this study, the Cronbach’s α of the Bullying Subscale were 0.705 at T1 and 0.628 at T2, and the Cronbach’s α of the Victimization Subscale were 0.763 at T1 and 0.624 at T2. The Cronbach’s α of the total SBVS were 0.791 at T1 and 0.676 at T2.

Internet gaming disorder

The Chinese version of the Internet Gaming Disorder Questionnaire was used to measure adolescents’ IGD (Zhang et al., 2017). This scale has 11 items and responses on a three-point Likert scale (“1 = never” to “3 = often”). One example is “Would you skip homework to have more time to play online games”. A higher total score represents a higher level of IGD. This scale demonstrated good reliability and validity in a Chinese adolescent sample (Zhang et al., 2017). In this study, the Cronbach’s α were 0.811 at T1 and 0.827 at T2.

Demographic covariates

Several demographic factors were assessed at T1 in this study, including adolescent age, gender (1 = boys, 2 = girls), grade, only-child family (1 = yes, 2 = no), and family economic level (1 = above the average level, 2 = at the average level, 3 = below the average level).

Analytic plan

SPSS 26.0 and the PROCESS macro were employed to analyze the data. First, the Harman’s single-factor test was performed to evaluate the common method bias. Second, descriptive analysis and Pearson correlational analysis were performed to describe the basic associations among the main variables. Third, hierarchical regression analyses were used to assess the concurrent and longitudinal predictive effects among the main variables. Fourth, the PROCESS macro (model 81) was adopted to test the longitudinal mediating effects in the relationship between school resources, self-control and problem behaviors. Bootstrapped 95% confidence intervals (CIs) were calculated with 2,000 re-samplings to examine the multiple mediating effects.

Results

Common method bias

Given that the data in the current study was collected by self-reported questionnaires, Harmen’s single factor test was adopted to test the common method bias. At T1, there were eight factors with characteristic root values greater than one, and the first factor only explained 16.71% of the variance. At T2, there were nine factors with characteristic root values greater than one, and the first factor only explained 18.50% of the variance. Overall, the common method bias appeared no threat in the current results.

Descriptive and correlational analyses

Table 1 displays the descriptive and correlational results of the main variables. School resources were positively associated with self-control ability at both waves in medium effect sizes (rs > 0, ps < 0.01). Adolescents who experienced more school resources reported better self-control ability. School resources were negatively related to school bullying and victimization at T2 and IGD at both waves in small-to-medium effect sizes (rs < 0, ps < 0.05). Adolescents obtained more school resources showed less engagement in three problem behaviors. Self-control ability was also negatively correlated to three problem behaviors in small-to-medium effect sizes (rs < 0, ps < 0.01). Adolescents who developed a greater self-control ability tended to engage less in problem behaviors. These three problem behaviors were significantly and positively associated with each other (rs > 0, ps < 0.05). Moreover, adolescent gender and grade were significantly related to self-control ability and three problem behaviors, so they were included as control variables in the following analyses.

Table 1 The means, standard deviations and correlations among the variables
Full size table

Test of the concurrent and longitudinal predictive effects

As shown in Tables 2 and 3, school resources could predict self-control (β > 0, ps < 0.001), school victimization (β < 0, ps < 0.05) and IGD (β < 0, ps < 0.05) concurrently and longitudinally, supporting the first and second hypothesis. Adolescents who obtained more school resources tended to develop a higher level of self-control ability and less involvement in school victimization and IGD. Self-control ability had predictive effects on school bullying (β < 0, ps < 0.01) and IGD (β < 0, ps < 0.001) at the same time and five months later. A good self-control ability could significantly prevent adolescents from develo** these two problematic behaviors. School bullying and victimization could concurrently predict each other in small-to-medium effect sizes (β > 0, ps < 0.001). Furthermore, school victimization at T1 predicted bullying at T2. IGD and school bullying concurrently predicted each other in small effect sizes (β > 0, ps < 0.01). IGD and school victimization at T1 could also predict each other (β > 0, ps < 0.05). Adolescents involved in one problem behavior might also develop other more problem behaviors.

Table 2 Cross-sectional regression analyses
Full size table
Table 3 Longitudinal regression analyses
Full size table

Test of the longitudinal mediating effects

Three longitudinal mediating models were constructed in the PROCESS macro (model 81) to examine the multiple mediating effects in the current study (Fig. 2). Table 4 reports the summary of the multiple longitudinal mediating effects. Self-control ability at T1 significantly mediated the relationships between T1 school resources and T2 problematic behaviors, which supports the third hypothesis. School victimization at T2 could mediate the relationships between school resources at T1, IGD at T2 (β = -0.004, 95%CI: -0.008, -0.001) and school bullying at T2 (β = -0.008, 95%CI: -0.014, -0.003). T2 IGD could mediate the relationships between school resources at T1, school bullying at T2 (β = -0.007, 95%CI: -0.014, -0.002) and school victimization at T2 (β = -0.006, 95%CI: -0.013, -0.001). These findings jointly supported the fourth hypothesis. Moreover, T1 self-control and T2 school bullying had sequential mediating effects on the relationships between T1 school resources, T2 school victimization (β = -0.006, 95%CI: -0.009, -0.003) and T2 IGD (β = -0.004, 95%CI: -0.007, -0.002). T1 self-control and T2 IGD played chain mediating roles in the associations between T1 school resources, T2 school victimization (β = -0.004, 95%CI: -0.008, -0.001) and T2 school bullying (β = -0.005, 95%CI: -0.008, -0.002). These findings supported the fifth hypothesis. Overall, these three problem behaviors might coexist in youth development, but fortunately, school resources could prevent or reduce them directly or indirectly via develo** a higher level of self-control ability.

Fig. 2
figure 2

a, b, c The relationship between school resources and problematic behaviors. Note: SR = School resources; SC = Self-control; SB = School bullying; SV = School victimization; IGD = Internet gaming disorder; T1 = Time 1, T2 = Time 2. Gender and grade are covariates; Dashed line indicates a non-significant coefficient. * p < 0.05, ** p < 0.01, ***p < 0.001

Full size image
Table 4 Summary of the indirect effects
Full size table

Discussion

The outbreak of COVID-19 has brought many challenges to youth development. During this specific period, adolescents have suffered from more and more behavioral problems (Christner et al., 2021; Liu et al., 1990). Overall, self-control also plays a protective role in youth development during the post-pandemic period.

With regard to the interactions between problematic behaviors, the fourth hypothesis was also revealed. School bullying and victimization could predict each other concurrently in small-to-medium effect sizes, school victimization could predict bullying over time. IGD and school bullying concurrently predicted each other in small effect sizes. IGD and school victimization could also predict each other. Additionally, victimization could mediate the relationships between school resources, IGD, and bullying. Similarly, IGD mediated the relationships between school resources, bullying, and victimization. That is, once adolescents develop a problem behavior, they are more likely to engage in more problem behaviors if not timely and effectively intervened with (Jessor, 1987). These findings support several previous studies on the associations among problem behaviors (Kim et al., 2017; Huang et al., 2021; Li et al., 1990; Masten & Cicchetti, 2010; **ang et al., 2022b; Li et al., 2022). Moreover, these serial mediating effects also provide empirical support for the problem behavior theory, which claims that problem behaviors are determined by the interactions between the environmental system, personal system, and behavioral system (Jessor, 1987). Overall, it is necessary to jointly intervene in problem behaviors among adolescents.

Contributions and limitations

Theoretically, the current study has made several contributions to the related field. Initially, this study illustrates the importance of positive factors from individuals and schools in youth development during the recovery period of a public emergency. In combination with the previous evidence during the COVID-19 outbreak (Shek et al., 2021; **ang et al., 2022a), the current findings confirm the joint protective effects of positive factors on adolescent development during ordinary periods and emergencies. It encourages scholars to pay more attention to positive factors when researching developmental problems. Besides, the present study suggests the developmental cascading effects in the process of development. In a cross-level perspective, it reveals that the school environment will shape the development of self-control ability while this personal trait has significant effects on behavioral problems. From different aspects of the same level, the current finding confirms the chain reactions in different types of problem behaviors. This once again reminds scholars to focus on the mutual influence of various problems when exploring the causes and consequences, as individual development is a dynamic process that involves items affecting each other.

Additionally, the current study also provides some implications for related prevention and intervention programs. First, inferred from the current findings from school resources to developmental outcomes, schools may take on more responsibilities in promoting positive youth development. Specifically, schools have to create a free, equal, and harmonious environment and atmosphere for students and provide appropriate rules to monitor students’ performance at school. Teachers have to build good relationships with students, encouraging and supporting them when needed. In addition, schools need to actively engage with families to help young people deal with developmental problems. Second, apart from these external resources, schools should also guide and assist students to develop internal qualities and abilities, such as self-control capacity, through teaching activities or special programs. When a public emergency happens, such as the outbreak of COVID-19, it is often impossible for individuals to obtain external resources in time, but internal attributes function more effectively to protect individuals. Developmental psychologists have developed numerous programs to enhance individuals’ positive attributes. For instance, Shek and his team conducted the “Tin Ka ** P.A.T.H.S. Project” based on school activity in mainland China and significantly enhanced adolescents’ positive attributes (Zhu & Shek, 2020). Third, practitioners concerned with youth developmental issues are encouraged to conduct combined intervention programs to reduce or prevent adolescents’ problem behaviors, which is challenging to implement but may be more effective.

Despite these contributions to literature and practice, this study also has several limitations, pointing directions for future research. At first, the current study only recruited an adolescent sample from Central China, which limited the generalization of the current findings to young people from other cultures. Future studies may adopt a cross-cultural design to recruit diverse samples to obtain more general results. As for the data collection method, the self-reported questionnaires cannot avoid the unreal but socially expected responses. Future research is encouraged to use more objective approaches to collect data, such as other-reported questionnaires or experimental data. Another important limitation is the study design. The current study conducted a two-wave longitudinal design with a five-month time span, which is too short to examine the stability of the current findings. Scholars may adopt a longer follow-up study to obtain more stable results. Moreover, this study has attempted to reveal the interactions among different problem behaviors, but the current analytical approach may limit the examination of the chain effects among several problem behaviors due to the restriction of PROCESS Macro. Future research is encouraged to evaluate this developmental phenomenon through more advanced methods such as structural equation model techniques. Furthermore, the present study only chose school resources as environmental factors to investigate their positive effects on problematic behaviors. Actually, there are a large number of positive factors from family, school, community, and the whole society that could promote positive youth development and avoid negative outcomes. Scholars may pay more attention to other subsystems to explore more positive factors. And the comparative study of the protective factors and risk factors on adolescents’ problematic behaviors may also provide more critical guidance for practice. Last but also important, the current study only focused on the challenges that adolescents have faced during this pandemic. But recent research has also suggested that adolescents obtain positive experiences in this specific period, such as discovering oneself and connecting with family (Fioretti et al., 2020). Future studies may take a more integrative perspective to compare individuals’ positive and negative changes during this pandemic, which may provide more advice to promote adolescents’ psychological recovery in the post-pandemic era.

Conclusion

In summary, the results confirmed the assumptions about the multiple protective effects of school resources and self-control on adolescents’ problem behaviors. Specifically, school resources could negatively predict IGD and victimization, and self-control mediated these associations. Moreover, one problematic behavior could also mediate the associations between self-control and another problematic behavior. These findings contribute to the previous literature and several theories. Practically, these findings also provide some implications for practitioners to prevent or intervene in adolescents’ problem behaviors during the post-pandemic period.