Student academic performance: the influence of peers’ problem behaviors and moderating effect of educational expectation Open Access

Friendship plays a critical role in children’s well-being and development, positively influencing social, emotional, and cognitive growth (Matheson et al., 2007; Ladd, 1999; Newcomb and Bagwell, 1995). During adolescence, peer influence intensifies as teenagers seek acceptance, often adopting behaviors without fully understanding the consequences (Moldes et al., 2019). At this stage, adolescents increasingly turn to their peers rather than families for guidance, particularly in decision-making and moral development (Uslu, 2013). Thus, fostering positive peer networks is vital. Although risk-taking is a common part of adolescence, research suggests that youth who avoid such behaviors generally show better adjustment (Willoughby et al., 2007). Recent studies also highlight the adverse academic effects of peers’ problematic behaviors. For example, Chen and Lin (2015) found that poor peer attendance negatively affects individual exam performance. Similarly, Goller et al. (2023) showed that re-enrolling dropouts can harm the academic outcomes of their classmates if placed in the same subject. Gao et al. (2019) further noted that students engaged in physical conflicts are more likely to drop out and experience academic decline.

Given the limited research on the relationship between student academic achievement and peers' problematic behaviors, this study seeks to address two key gaps in the existing literature. Responding to the call by Butler-Barnes et al. (2015), the first objective is to examine the influence of peers’ problematic behaviors—such as skipping classes, engaging in physical altercations, and dropping out—on individual academic outcomes, using data from the China Education Panel Survey (CEPS). Secondly, this study explores the potential moderating role of self-expectation in shaping the relationship between peer misconduct and academic performance. While prior research has primarily concentrated on direct associations, the moderating effect of self-expectation remains underexplored. To bridge the identified gaps, the following research questions have been formulated.

The remainder of this paper is structured as follows: Section 2 provides a comprehensive review of the relevant literature. Section 3 outlines the theoretical foundation and hypothesis development, while Section 4 details the research methodology. Section 5 presents the empirical findings, which are further analyzed in Section 6. Finally, the paper concludes with a discussion of the implications, limitations, and suggestions for future research.

Generally, our study is supported by two theories. Firstly, the linkage between peers’ problem behavior and student academic performance is supported by problem Behavior Theory (PBT).

PBT explains how psychosocial protective and risk factors influence adolescents' involvement in various problem behaviors. These behaviors include delinquency, tobacco use, alcohol abuse, illicit drug use, early sexual activity, aggression, and risky driving (Jessor, 1987). It suggests that these behaviors are interconnected and often stem from underlying psychosocial factors or influences, reflecting a broader pattern of risk-taking or problem behavior during adolescence (McAlister et al., 1984). This theory posits that an individual’s problematic behavior can be influenced by both their personal traits and external environmental factors (Jessor, 1987). In particular, personality traits, such as a tendency toward obedience to authority or rebelliousness, interact with perceived environmental factors to influence smoking and other high-risk behaviors in adolescents. Relevant environmental factors include the attitudes and behaviors of friends, including their actual behaviors and even just spending time with them.

In addition, the selection of moderator (educational expectation) in this study is supported by expectancy-value theory (EVT). It is a prominent framework in the study of academic motivation, extensively used to predict and explain students' task choices, learning persistence, and academic performance (Wigfield and Eccles, 2000). According to this theory, students' expectations for success and their subjective valuation of tasks directly influence their achievement-related choices, the effort they invest, their persistence in the face of challenges, and their overall performance (Wigfield and Eccles, 2000). EVT posits that students' decisions to engage in and sustain participation in any given activity are influenced by their success expectations and the personal values they associate with that activity. Consequently, EVT serves as a suitable theoretical and pedagogical framework for elucidating motivational behaviors among students (Ortlieb and Marinak, 2013). This theoretical framework is relevant to our context, as we suggest that students' educational expectations will influence the relationship between peers' problematic behavior and their academic performance.

In this study, supported by PBT, the number of peers participating in actions such as missing class, getting involved in fights, and leaving school is used as a measure of their behavioral inclinations. Problem behaviors are typically characterized as maladaptive actions that hinder an individual’s ability to adjust socially and often involve the violation of societal norms, leading to adverse effects on communities (Lin et al., 2004; Lassi et al., 2011). Such behaviors include aggression, disciplinary violations in school, and various forms of antisocial conduct, all of which negatively impact adolescents’ mental health and developmental trajectories (Gross et al., 2009). When established during childhood, these behaviors may persist into adulthood, potentially manifesting as personality disorders and increasing the risk of substance abuse and criminal activity (Narusyte et al., 2017). Within the existing literature, scholars have explored the detrimental effects of problem behaviors on academic performance. For example, Thombs (2009) identified a significant negative relationship between semester academic outcomes and several alcohol-related indicators, including the frequency and intensity of elevated breath alcohol concentrations. Likewise, peer-related behaviors such as bullying (Menken et al., 2022) and anti-school attitudes (Nelson and DeBacker, 2008; Véronneau et al., 2008) have been found to impair cognitive development. Other contributing risk factors include low socioeconomic status, lack of parental supervision, family instability, and experiences of bullying (Resnick et al., 2004).

The influence of peers has been extensively examined within educational research (Billings et al., 2014). Adolescence, a critical developmental stage, is characterized by a strong desire for social conformity (Hamm et al., 2014). During this period, youths often adopt the academic behaviors and values of their peers to avoid the stigma and social exclusion associated with nonconformity. As a result, adolescents tend to associate with peers who exhibit similar attitudes and behaviors (Parker et al., 2015; Rambaran et al., 2017). Empirical evidence has highlighted the substantial role peers play in shaping adolescents’ behavioral, cognitive, and emotional engagement (Wang et al., 2018), as well as their academic motivation and performance (Blansky et al., 2013). In particular, the quality of close friendship networks has been positively associated with academic outcomes (Wu and Wang, 2023). However, peer influence can also operate negatively; studies show that deviant behaviors often cluster within peer groups, and frequent interaction among such peers may amplify individual deviant actions (Denault and Poulin, 2012; Keijsers et al., 2012).

To the best of our knowledge, the first line of research on peer influence has focused on peer characteristics, including factors such as group home size interaction (Osei and Gorey, 2019), peer quality (Berthelon et al., 2019), relationships within learning communities (Brouwer et al., 2022), physical well-being (Huang et al., 2021), and the proportion of only-child peers in the classroom (Cai et al., 2022). These influences have been shown to persist over time (Lépine and Estevan, 2021). Negative peer effects may stem from factors such as the presence of non-local students (Hu, 2018) or experiences of peer victimization (Fite et al., 2014). Overall, peers exert a significant influence on students’ academic outcomes, encompassing both beneficial and detrimental impacts.

Additionally, several studies have demonstrated a link between peers' problem behaviors and student learning outcomes. Kassarnig et al. (2017) found a strong association between peer attendance patterns, highlighting a notable peer effect or homophily in both class attendance and academic performance—suggesting that students whose close friends frequently skip classes may experience declines in their own academic achievement. Similarly, Jiang (2023) reported a significant negative relationship between deviant peer behavior and academic performance, with the effect remaining robust even after controlling for selection bias in peer relationships. Overall, consistent with Zhang et al. (2021), positive peer behaviors are shown to enhance academic performance, whereas deviant peer behaviors have detrimental effects.

Beyond absenteeism, prior research has also examined the academic impact of peers’ decisions to drop out of school. Goller et al. (2023) found that re-enrolled dropouts negatively affected the academic performance of peers when placed in the same subject, whereas subject changes by dropouts had a positive effect. Additionally, Dong et al. (2023) demonstrated that low-achieving students are more influenced by the positive spillover effects of middle-achieving peers, particularly in terms of academic performance and learning attitudes.

In summary, the studies mentioned above indicate that peer behaviors play a crucial role in shaping student academic performance. In our context, peers' problem behaviors (skipping classes, fighting, dropping out of school) have an adverse impact on student academic outcomes. One of the foundational frameworks for this study is problem-behavior theory (Jessor, 1987). The core principle of this theory posits that all behavior results from interactions between the individual and their environment. Problematic behaviors are defined as those deemed undesirable by conventional societal standards. The theory particularly emphasizes the external context of adolescent and young adult life, focusing on the stressors and satisfactions that influence the manifestation of these problematic behaviors (De Leo and Wulfert, 2013). Relevant external factors, such as the behaviors of friends, significantly influence students' personal success. Thus, drawing from the discussions above and grounded in PBT, we formulate the following hypothesis:

In addition to the aforementioned studies, existing research has also examined the impact of personal expectations on students' academic performance. Educational expectation, as defined by Pinquart and Ebeling (2020), encompasses the projected academic achievements at school or university. This includes both short-term expectations for individual course grades and long-term aspirations for ultimate educational attainment, such as graduating from high school or university.

In response to the inquiries posited by Khalaf et al. (2024), Mansour et al. (2024), Saleh and Maigoshi (2024), and Saleh et al. (2024), this research primarily concentrates on the context within emerging nations. In the context of China, both parental educational expectations and student educational expectations have been demonstrated to be significant predictors of science achievement (Liu et al., 2006). According to Wang and Benner (2014), it is grounded in an individual’s cognitive capability, realistic circumstances, and the anticipations that parents have for their children’s or adolescents' future academic accomplishments. Furthermore, they have been found to have a significant correlation with academic performance in middle schools (Qian, 2008).

A growing body of research has examined how self-expectations can buffer the effects of problem behaviors and associated risks. For example, Chang (2002) found that optimism moderated the relationship between perceived stress and psychological symptoms, with highly optimistic individuals exhibiting fewer symptoms under stress. Similarly, Ouyang et al. (2023) revealed that adolescents’ self-educational expectations and associations with deviant peers mediated the link between parental expectations and problem behaviors, highlighting the protective role of both parental and adolescent expectations. McDade et al. (2011) reported that adolescents with higher expectations of attending college were more likely to engage in healthy behaviors, such as regular exercise and reduced smoking, in young adulthood. Harris et al. (2002) also observed that low educational expectations were associated with increased engagement in risky behaviors. Moreover, expectations have been shown to intensify exam-related anxiety (Burns, 2004), amplify the benefits of parental support (Gerard and Booth, 2015), and buffer the negative effects of teacher expectations on academic performance (Benner and Mistry, 2007). These findings suggest that self-expectations play a significant and enduring role in shaping adolescents’ behavioral and developmental outcomes.

Within the context of EVT, an individual’s behavior is linked to the expectations they maintain and the subjective values (or valences) attributed to various instrumental actions and their potential outcomes (Feather, 1992). Besides, expectations involve beliefs about the ability to perform an action to achieve a successful outcome, as well as beliefs about the positive and negative consequences that may follow (Feather, 1992). Applying this to our context, it can be seen that students with higher self-expectations experience fewer negative consequences from the influence of problem behaviors. Consequently, drawing upon the discussions above and supported by EVT, the following hypothesis is proposed.

Below, Figure 1 illustrates the conceptual framework utilized in this study.

As illustrated in Figure 1, the factors (peer’s problem behaviors) exerting a negative impact on student academic performance are presented on the left-hand side, and the link between peer’s problem behaviors and student academic performance is moderated adversely by students' educational expectations.

This study employs secondary data analysis, a method that utilizes existing datasets to validate prior findings and investigate new or complementary research questions emerging from the literature (Johnston, 2014). The data were drawn from the China Education Panel Survey (CEPS), conducted by Renmin University of China in 2017, which surveyed over 10,000 students from 112 secondary schools nationwide (CEPS, 2017). The dataset encompasses diverse information, including students’ mental health and study behaviors. For this analysis, relevant variables were selected, yielding a final sample of over 8,500 observations after excluding cases with missing data from the original 10,750 eighth-grade respondents.

The primary aim of this study is to examine the factors influencing student academic performance. To this end, as shown in Panel A, we use the number of peers engaged in behaviors such as skipping class (Absent), fighting (Fight), and dropping out of school (Quit_School) as measures of peer behavioral tendencies. Following the approach of Li et al. (2019, 2024), our variables are coded binary; for example, the variable ‘educational expectation’ is coded as 1 for ‘College and above’ and 0 for ‘High school and below.’

In addition, as outlined in Liu et al. (2020) and depicted in Panel B of Table 1, student academic performance is assessed using standardized tests in three fundamental subjects: Chinese, mathematics, and English. Furthermore, as depicted in Panel C, we employed participants' educational expectations (Expectation) as a moderator to explore potential moderating effects. As presented in Panel D of Table 1, five control variables are incorporated. In particular, based on previous studies, three categories of control variables related to our context were chosen for this study. In line with previous studies (Jiang et al., 2018), the first and second categories encompass socio-demographic and family characteristic variables. These include indicators such as whether the student is the only child in the family (One_Child), household registration category (Registration_Type), and proficiency in managing the local language (Dialect). The third category pertains to the attributes of social networks, as outlined by DeLay et al. (2016), encompassing variables such as the count of participants' close friends (Friend_No).

In this section, the regression models adopted in this study are listed as follows.

In the proposed models, three distinct problem behaviors displayed by the close friends of participants, namely skipping class (Absent), engaging in fights (Fight), and dropping out of school (Quit_School), were chosen as independent variables. The dependent variable comprises the student academic performance in the standardized tests for three key subjects. Additionally, the educational expectation (Expectation) of participants was selected as a moderating variable in Model 2, 4, and 6.

In this section, the study’s empirical findings are showcased.

Table 2 presents the descriptive statistics for student performance and their self-expectations in the pooled sample. The average Chinese score is 80.98, with the highest score being 142.50 (the maximum achievable score on the test). The standard deviation for student performance is 20.64. Notably, the self-expectations were rated on a scale of 1–10 in the questionnaire.

Below, the justifications of the correlation matrix are presented.

Table 3 shows significant positive effects of positive factors, such as having high self-expectations, on students' performance in Chinese, mathematics, and English. Conversely, when friends frequently engage in risky activities, such as fighting, skipping class, and dropping out of school, it negatively impacts student performance.

This section presents the empirical results of the statistical analysis.

The empirical results presented in Table 4 partially address our research questions. Specifically, Models 1, 3, and 5 reveal that peers' problematic behaviors—such as skipping classes, fighting, and dropping out—significantly and negatively affect students’ performance across all three subjects in the Chinese exams. These findings support our first research question concerning the impact of peers’ problematic behaviors on academic performance, thereby providing evidence in favor of Hypothesis 1.

Addressing the second research question—how students' educational expectations influence the effect of peers' problematic behaviors on academic performance—our results indicate that high educational expectations amplify the negative impact of peers skipping classes on Chinese subject performance. Conversely, elevated educational expectations mitigate the adverse effects of peers' involvement in physical fights on overall academic outcomes. However, no significant moderating effect of educational expectations was observed regarding the relationship between peers' dropout rates and academic performance across all subjects. Therefore, hypothesis 2 is only partially supported.

To address potential endogeneity, we utilized the copula correction method, following Gui et al. (2023) and Park and Gupta (2012), with an internal instrumental variable approach via the `REndo` package in R. Table 5 displays the copula-corrected regression estimates for six model specifications, accounting for endogeneity in discrete variables Absent, Fight, Quit_School, and Expectation. Consistently across all models, Expectation shows a strong, statistically significant positive effect on the outcome variable, while Absent, Quit_School, and Registration_Type exhibit robust negative effects. The copula correction terms (PStar variables) are mostly non-significant, indicating that residual endogeneity has been sufficiently addressed through the copula approach. Models 3 through 6 incorporate interaction terms, revealing significant moderation effects: Expectation mitigates the negative impact of Absent (positive interaction) but intensifies the negative effect of Fight (negative interaction). These results highlight the importance of considering both direct and moderated effects when modeling the influence of endogenous predictors, and confirm the robustness of the primary associations even after correcting for endogeneity.

Some researchers have recognized that children’s interactions with their peers form a vital developmental context for children and adolescents (Kindermann et al., 1996; Ryan, 2000). Research has shown that best friends tend to exhibit similar behaviors, such as the frequency of skipping classes (Kandel, 1978) and the amount of time devoted to homework (Cohen, 1977). Brown et al. (1993) posited that peers choose each other based on common traits, reinforcing rather than changing an adolescent’s characteristics. However, they noted that not all traits are shared among peers, leaving open the question of how differing characteristics are handled. Consistent with these studies, the regression statistics indicated that the number of close friends with problem behaviors (i.e. skipping classes, dropping out of school, and fighting with others) is negatively related to academic performance in all of the subjects. Indeed, both school environment and peer influence play significant roles in students' academic performance (Korir and Kipkemboi, 2014). The quality of peer friendships and the academic progress of friends can forecast adolescents' learning engagement and academic success (Lessard and Juvonen, 2018). Our findings corroborate this notion by indicating that students who have a higher number of close friends engaged in problem behaviors are less likely to achieve academic success.

Physical bullying, whether occurring alone or alongside verbal bullying, is generally associated with poorer academic performance (Mundy et al., 2017). Our findings indicate that students who have more close friends involved in physical fights exhibit significantly lower academic achievement. This relationship may be explained by the fact that poor academic performance and signs of fatigue at school can reflect youths’ exposure to violence (Lepore and Kliewer, 2013).

Victimization often results in poor academic outcomes due to its detrimental psychological effects, such as hindering classroom participation (Ladd et al., 2008). Studies found that perceived harassment can lead to feelings of depression and loneliness, negatively affecting grades and attendance (Juvonen et al., 2000). Frequent victimization has also been linked to lower grades and test scores, mediated by depressive symptoms (Schwartz et al., 2005). Moreover, students feeling safe at school tend to perform better academically (Milam et al., 2010). Exposure to school violence demands victims' time and energy to handle the aftermath and manage negative emotions (Côté-Lussier and Fitzpatrick, 2016). Trauma theory suggests that adolescents exposed to violence, such as through friends involved in fights, may experience stress-induced brain changes that can negatively affect skills like communication, memory, and concentration, leading to poorer academic performance (Cobos-Cali et al., 2018). Such issues may increase the uncertainty students experience while studying at schools. In such a school environment, high student turnover rates throughout the academic year can negatively impact the learning experience for all students (Raudenbush et al., 2011). From this perspective, if the peers of participants are involved in violent behaviors, it could potentially lead to lower academic performance, consequently impacting the participants' own performance negatively. In summary, our findings indicate that students with friends who frequently skip classes tend to perform poorly across all subjects in exams. Chen and Lin (2015) found a significant negative correlation between peer attendance and the exam performance of individual students. According to self-categorization and social influence theory, the norms related to absence within referent groups play a crucial role in explaining high levels of absence, even when the norms of the larger organizational units, like departments, are taken into account (Bamberger and Biron, 2007).

Exposure to peers whose values do not support strong academic performance can encourage similar behaviors. These peer groups may directly influence adolescent behavior through imitation, or indirectly by instilling and internalizing norms and attitudes that favor these behaviors (South et al., 2007). The characteristics of adolescents' peer networks, especially their centrality within these networks and their friends' academic performance, are crucial mediators of the link between mobility and dropout rates (South et al., 2007). Gao et al. (2019) observe a strong correlation between both push and pull factors and student dropout rates. It was discovered that students with dropout friends, those who maintain regular contact with such friends, and those actively encouraged to drop out by a dropout friend, are significantly more prone to dropping out.

Again, as adolescents shift from parental relationships to peer interactions, the influence of peers on their social actions and mental structures is predicted to increase (Brown et al., 1993), with peers also playing a vital role in academic achievement. Such impacts can be interpreted as the influences stemming from collective groups such as family and social networks. In the domain of cultural studies, Western countries commonly exhibit elevated individualism scores, while Asian nations like China tend to achieve higher scores in collectivism. Individualism emphasizes competition and self-reliance, whereas collectivism places importance on interdependence, family cohesion, and cooperation (Triandis, 1995). As per Tan et al. (2021), despite rising individualism among Chinese youth, collectivism remains prevalent in Chinese society and education. Essentially, students' values, behaviors, and learning outcomes are greatly influenced by their environment. Thus, students with many dropout friends may have poorer academic outcomes, as these friends could transmit certain detrimental values or attitudes. According to Tan et al. (2021), Chinese youth display both collectivist and individualist tendencies, with a stronger lean towards collectivism. This leads students to gauge themselves against peer preferences, impacting their value acquisition and academic advancement (Zhao and Zhao, 2022). In essence, our study reveals that greater interaction with peers involved in physical conflicts or school dropout negatively affects academic performance in subjects like Chinese, mathematics, and English.

Again, the EVT asserts that students' achievement-related choices, behaviors, and persistence are influenced by their ability beliefs, task-specific expectations, and subjective task values. Within this framework, educational expectations are conceptualized as a moderating factor in the relationship between peers' problematic behaviors and students' academic performance. The regression analysis identified a negative association between the number of close friends exhibiting problem behaviors and academic performance in Chinese, mathematics, and English. These findings differ from some previous studies, which emphasize the significant influence of adolescents’ academic expectations on educational success (Lucio et al., 2012). Individuals with varying educational expectations tend to show different levels of effort and determination to achieve their goals. Consistent with this literature, our results suggest that students’ educational expectations can amplify the negative impact of peers skipping classes on Chinese academic performance.

It has been highlighted that individuals with higher educational expectations are more likely to improve their achievements, as they tend to have more specific goals (Carroll et al., 2009). Theoretically, however, the expectancy-value construct is made up of two distinct yet interconnected dimensions: a student’s belief in their ability to succeed (expectancy beliefs) and the subjective value they assign to the task (Shang et al., 2023). Expectancy beliefs depend on perceptions of the difficulty of tasks and the individual’s belief in their ability to overcome these difficulties to achieve success (Eccles and Wigfield, 2002). it is theorized that the more challenging a student perceives a given task to be, the lower the expectancy belief the student would attach to it (Eccles and Wigfield, 2002). Consequently, the student’s motivation to overcome the challenge may decrease if their expectancy beliefs are low. In this sense, individuals might perform worse on certain tasks when they have high personal expectations.

Indeed, an individual’s anxiety response becomes more pronounced as their expectations intensify. Studies, such as the one by Brandmo et al. (2019), have demonstrated a positive correlation between self-expectation and test anxiety, indicating that higher levels of self-expectation are associated with increased test anxiety. In high-stress school environments, an elevation in academic expectations could potentially heighten school-related stress and hinder students' academic performance (Kaplan and Kaplan, 2005). In our context, elevated levels of students' educational expectations amplify the adverse association between peers' problem behaviors and academic performance in mathmetics. In certain scenarios, heightened future expectations could serve as a hindrance to enhancing academic performance. In summary, adolescents with high educational expectations often engage in beneficial academic behaviors and set enduring goals, which have a lasting impact on their academic development (Harris et al., 2002; Gu et al., 2024). However, these expectations can lead to irrational beliefs that exacerbate stress, especially when they can’t cope, increasing academic pressure. This stress is particularly impactful on Asian students' psychological well-being (Cao et al., 2022; Jailani et al., 2020). Furthermore, Higgins' self-discrepancy theory (1987) posits that high self-expectations can intensify test anxiety, as discrepancies between one’s actual, ideal, and ought selves can induce anxiety. This supports the self-discrepancy theory, suggesting that disparities between the real self, ideal self, and ought-to self can lead to dissatisfaction, potentially causing depression and impacting academic development (Higgins, 1987).

Once again, Tan et al. (2021) suggested that collectivism continues to play a major role in Chinese society, with education aiming to promote collectivistic values among students, even though there’s been a growing trend of individualistic orientation among Chinese youth. They also discovered that both collectivistic and individualistic orientations are present among Chinese youth, with collectivism being more prominent than individualism. In simpler terms, a mismatch between personal and group expectations can cause stress for individuals striving to achieve their goals. Students' educational expectations (an individualistic perspective) are often overshadowed by peer behavior (a collectivistic perspective), suggesting that adolescents are more influenced by group dynamics. In other words, supported by EVT, our findings indicate that the negative correlation between students' interactions with peers involved in physical fights and their performance in all subjects is negatively moderated by their educational expectations.

The primary aim of this study is to investigate the relationship between peers' problematic behaviors and students' academic performance, as well as the potential moderating role of students' educational expectations. The findings reveal that peers' behaviors—such as skipping classes, engaging in conflicts, and dropping out—negatively impact student academic outcomes, with these effects being amplified by students' own educational expectations. This study makes two key contributions: first, it provides a novel examination of how peers' skipping, dropout, and conflict behaviors influence academic performance; second, building on Moè (2016), it offers new insights into how students' educational expectations may moderate these peer effects.

Consistent with our results, various theoretical and practical contributions can be inferred. From a theoretical perspective, this study expands our understanding of the relationship between peers' problem behaviors and students' educational expectations. Specifically, it highlights that peers' problematic behaviors, such as skipping classes and engaging in fights, can negatively affect students' academic performance. This finding aligns with the principles of PBT.

Additionally, EVT offers valuable insights into the sources of motivation, activity choice, and persistence among students of various age groups. According to EVT, an individual’s choices, persistence, and performance in activities are largely determined by their beliefs about their potential performance and the extent to which they value the activity (Wigfield and Eccles, 2000). Drawing on this theory, our study makes a theoretical contribution by examining the moderating role of educational expectations in the relationship between peers' problematic behavior and academic performance. Additionally, we extend the application of EVT to the context of understanding how peers' problematic behavior impacts academic outcomes.

Furthermore, in terms of practical implication, it was found that peer influence on disruptive behavior was lower when students perceived their teacher’s instruction to be more supportive and interesting, but higher when the teacher used more ability differentiation (Müller et al., 2018). Thus, educators and school personnel are advised to offer assistance and guidance to individuals grappling with friendships involving problem behaviors. For example, implementing a buddy scheme within the school can offer students additional supportive resources for cultivating their social networks. Furthermore, education policymakers are recommended to establish informal guidelines that assist family members and educators in offering support to students in bridging expectation gaps.

However, this study is subject to certain limitations. Firstly, identity control theory (Burke, 2007) suggests that external individuals’ (e.g. parental) education expectations represent assessments of significant others, while self-education expectations function as individual benchmarks for present social roles; misalignment between the two can induce stress, influencing development and intrinsic motivation (Moè, 2016). Our study did not find evidence for the mediating role of stress, personal value, and intrinsic motivation in the relationship between peers' problem behavior and student academic performance. Future research could delve deeper into this topic by exploring the potential mediation of stress and intrinsic motivation. Secondly, the participants' ethnicity was not considered, suggesting the need for future research encompassing diverse ethnic backgrounds.

Besides, research conducted by McAlister et al. (1984) provides evidence that personality traits interact with environmental factors to predict problematic behavior in adolescents. However, within the framework of EVT, the present study did not examine the potential role of personality traits in moderating the relationship between peers' problematic behavior and academic performance. Consequently, future research could investigate how personality characteristics, such as those identified through the MBTI assessment, may influence this association. Similarly, the perceived environment, as outlined in PBT (Jessor, 1987), refers to the subjective interpretation of influences exerted by one’s parents and peers. Since this discussion has primarily focused on the impact of peers' problem behaviors on academic performance, future research could investigate the influence of parental factors in greater depth.

We extend our sincere gratitude to the anonymous reviewers for their valuable insights and comments on previous versions of this paper.