Title Perceived parental expectations and their role in academic and psychosocial functioning Author(s) Gregory Arief D. Liem, Zi Yang Wong, Melvin Chan, Woon Chia Liu, Nur Izzati Zainudin, Ser Hong Tan, Kenneth K. Poon, Trivina Kang and Siow Chin Ng This is the author’s accepted manuscript of the following article: Liem, G. A. D., Wong, Z. Y., Chan, M., Liu, W. C., Nur Izzati Zainudin, Tan, S. H., Poon, K. K., Kang, T., & Ng, S. C. (2025). Perceived parental expectations and their role in academic and psychosocial functioning. International Journal of Behavioral Development . https://doi.org/10.1177/01650254251321444 SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 1 Perceived Parental Expectations and Their Role in Academic and Psychosocial Functioning International Journal of Behavioral Development Gregory Arief D. Liem*, Zi Yang Wong, Melvin Chan, Woon Chia Liu, Nur Izzati Zainudin, Ser Hong Tan, Kenneth K. Poon, Trivina Kang National institute of Education Nanyang Technological University, Singapore Siow Chin Ng Research and Management Information Division Ministry of Education, Singapore NOTE: * Corresponding author information: Gregory Arief D. Liem, National Institute of Education, Nanyang Technological University, 1 Nanyang Walk, 637616, SINGAPORE (e-mail: gregory.liem@nie.edu.sg ). **The views and opinions expressed in this paper are solely those of the authors and do not necessarily reflect the official policies or positions of the Ministry of Education, Singapore, or any affiliated institutions. SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 2 Abstract Parental and teacher expectations for students’ academic performance, expressed as confidence in students’ academic potential, may play a critical role in fostering holistic development. However, the specific mechanisms through which these perceived expectations influence academic motivation and psychosocial well-being remain underexplored. This research tests a hypothesized model of the role of mathematics self-concepts in linking perceived parental and teacher expectations to self- determined motivation and achievement goals in the subject and various aspects of psychosocial well-being. Two studies were conducted with Singaporean students (Study 1: N = 2,632, M age = 14.05 years; Study 2: N = 450, M age = 13.98 years), with surveys administered at two points during the academic year. Results consistently showed: (a) perceived parental expectations had a more salient role in mathematics self-concepts compared to perceived teacher expectations, (b) mathematics self-concepts partially mediated the relationships between perceived parental expectations and motivation, goals, and well-being (except connectedness), (c) these relationships were relatively stable over time, and (d) perceived teacher expectations complemented perceived parental expectations in predicting mathematics self-concepts, connectedness, and happiness. These findings highlight the distinct yet complementary roles of parents and teachers in shaping developmental outcomes, emphasizing the importance of considering cultural context and offering insights for educational practices to support students’ academic and psychosocial functioning. Keywords: parental academic expectations, self-determined motivation, achievement goals, psychosocial well-being SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 3 Perceived Parental Expectations and Their Role in Academic and Psychosocial Functioning: Parents play a critical role in fostering adolescents’ positive development (Bornstein, 2019). Research on parental influence has explored its various dimensions, including autonomy support (Vasquez et al., 2016), control (Grolnick & Pomerantz, 2009), and expectations (Wentzel et al., 2016). Studies on parental expectations, however, often focus on the highest educational attainments parents envision for their children (e.g., Aceves et al., 2020), overlooking expectations based on confidence in a child’s academic potential (but see e.g., Ma et al., 2018), and few studies have juxtaposed parental with teacher expectations (but see e.g., Wentzel et al., 2016). Moreover, certain parenting practices, such as control, may not mean and manifest uniformly across cultures (Grolnick & Pomerantz, 2009; Ryan & Deci, 2020). Therefore, examining how perceived parental expectations relate to students’ school and psychosocial functioning, particularly in academically driven societies like Singapore, is both theoretically and practically important. Like in most East Asian societies, parents in Singapore hold high academic expectations for their children, mainly rooted in the societal emphasis on education as a pathway to success and cultural values like family honor, interdependence, and hard work (Chao, 1994; Hwang, 2012; Tan, 2012). These expectations are often explicitly communicated, with parents actively managing their children’s education, including study schedules, enrichment activities, and performance monitoring (Chan et al., 2022). While parental expectations also play a role in children’s academic development in Western contexts, they are often conveyed in ways that emphasize the children’s personal choice and interest, reflecting broader societal norms that prioritize self-realization and self-determination over familial reputation and collective achievement (Markus, 2016). Alongside parents, teachers play a crucial role in shaping students’ self-beliefs, motivation, and well-being particularly in the context of a subject as important as mathematics. Parents often set expectations grounded in long-term aspirations in their child’s future, directly influencing the child’s sense of competence, motivation, and well-being (Chao, 1994; Hwang, 2012). Teachers, on SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 4 the other hand, provide expectations based on students’ observed performance and classroom behavior, offering immediate and performance-driven feedback that helps reinforce students’ academic confidence and motivation (Rubie-Davies et al., 2010). Mathematics, being a subject often regarded as foundational for educational and professional success, tends to attract considerable attention and high expectations from both parents and teachers. While mathematics may be less directly shaped by familial practices, such as reading or storytelling, compared to language-based subjects (Fan & Chen, 2001), parents contribute to students’ perceived competence and motivation in the subject through broader messages about diligence, perseverance, and role obligation. Teachers complement this by creating opportunities for students to experience success and develop competence through structured feedback and classroom support. Thus, these distinct yet complementary roles underscore the crucial role of both parents and teachers in fostering students’ mathematics self-belief and motivation and promoting their psychosocial well-being. This paper reports two studies that collectively examined Singaporean secondary school students’ perceptions of parental and teacher academic expectations and their links to academic and non-academic outcomes. Academic outcomes included academic self-concepts (Muenks et al., 2018), self-determined motivation (Howard & Gagné, 2017), and mastery and performance goals (Liem & Senko, 2022) in mathematics. The second study also addressed domain-general psychosocial well-being outcomes, including engagement, perseverance, optimism, connectedness, and happiness (Kern et al., 2016). Both studies assessed the independent and joint roles of perceived parental and teacher expectations and explored academic self-concepts as a mediator. As Figure 1 illustrates, these hypothesized relationships were analyzed at two time points to assess their relative stability across the academic year. Together, this research has the potential to advance our understanding of how perceived social contexts shape the academic and psychosocial development of Singaporean students and provide insights for effective partnerships between teachers and parents in this development. SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 5 A Self-Determination Theory Perspective on Holistic Development Self-Determination Theory (SDT) asserts that individuals possess inherent psychological needs for autonomy, competence, and relatedness, and the fulfillment of these needs is essential for not only volitional functioning but also psychosocial development (Ryan & Deci, 2000, 2017, 2020). That is, the satisfaction of these needs influences individuals’ emotional health, social development, and resilience, forming the foundation for their self-determined motivation and flourishing. In this regard, SDT emphasizes the role of social contexts in facilitating need satisfaction: social environments that nurture needs enable individuals to build a strong sense of self-worth, positive interpersonal connections, and adaptive coping strategies, while environments that thwart needs can lead to disengagement, anxiety, and diminished well-being. The recent addition of Relationship Motivation Theory (Reeve, 2022; Ryan & Deci, 2017) as a mini-theory within SDT further solidifies the pivotal role of high-quality social relationships, characterized by authenticity and emotional support, in individuals’ need satisfaction and personal growth, social development, and psychological well-being. Parents and teachers, as significant socialization agents, play a crucial role in satisfying students’ needs by communicating expectations in ways that respect their autonomy, foster their competence, and nurture their sense of relatedness (Ryan & Deci, 2017, 2020). When students experience autonomy (feeling volitional and authentic), competence (feeling capable and effective), and relatedness (feeling connected and valued), these needs form the foundation for their self- determined motivation, a type of motivation based on internalized values and personal interests (Howard & Gagné, 2017), and various dimensions of their well-being, including a sense of engagement, positive emotions, optimism, connectedness, and happiness (Kern et al., 2016). These mechanisms highlight the critical role of parental and teacher expectations within a broader socioecological framework that supports students’ holistic development. Within SDT, academic expectations are closely tied to the concept of academic structure, which involves parents and teachers setting clear, consistent, and realistic standards of performance SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 6 (Grolnick & Pomerantz, 2009; Jang et al., 2010). This structure often includes providing resources or guidance to help students accomplish tasks effectively (Grolnick et al., 2014). While academic structure is primarily associated with satisfying competence needs (Jang et al., 2010), it also has the potential to support relatedness and autonomy. When parents and teachers communicate academic expectations that convey confidence in the student’s abilities, acknowledge their prior performance, and consider the student’s perspective, such expectations can contribute to the satisfaction of relatedness and autonomy needs (Cheon et al., 2020; Grolnick & Pomerantz, 2009). Conceptualized this way, perceived parental and teacher expectations not only satisfy students’ competence needs and strengthen academic self-concepts (Benner et al., 2021; Chen et al., 2022; Muenks et al., 2018) but also fulfill their needs for relatedness and autonomy, which in turn foster self-determined motivation (Griffith & Grolnick, 2014; Luo et al., 2016). As such, students’ perceived parental and teacher expectations were hypothesized to positively predict their mathematics self-concepts and self-determined motivation. The relationships between perceived parental and teacher expectations and achievement goals are less explored in the literature. Achievement goal theory posits that the goals students adopt drive engagement and shape learning experiences (Elliot & Hulleman, 2017). The theory broadly distinguishes between mastery goals, which orient students towards learning and task- or self- referenced competence and performance goals, which focus students on grades and normatively referenced competence (Liem & Senko, 2022). 1 Existing research suggests that the relations between students’ perceived parental expectations and adoption of mastery and performance goals are mediated by their behavioral (e.g., efforts to organize) and cognitive (e.g., setting personal standards of attainment) strategies (Madjar et al., 2015). Beyond this mediation, students might construe parental and teacher expectations of academic excellence as encouragement to both learn 1 The achievement goal literature (Elliot & Hulleman, 2017) identifies these goals as mastery-approach and performance-approach. Each has a corresponding avoidance-oriented counterpart: mastery-avoidant goals aim to prevent a decline in competence, while performance-avoidant goals seek to avoid being outperformed or appearing incompetent. Our study focuses solely on the two approach goals, which we refer to simply as mastery and performance goals for the sake of clarity. SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 7 and develop skills (mastery goals) and outperform their peers (performance goals), providing a basis for the hypothesized positive direct paths from perceived parental and teacher expectations to the two achievement goals. Research exploring the role of parental expectations in child well-being is expanding. Traditionally, this body of work has focused on the negative effects of parental expectations, particularly the stress and pressure they can impose on children. These expectations often involve unreasonably high goals (Ma et al., 2018) and are sometimes communicated in controlling and conditional ways (Grolnick & Pomerantz, 2009; Haines & Schutte, 2023). Such forms of expectation have been associated with decreased well-being and quality of life (Assor et al., 2004; Cassidy & Boulos, 2023). Conversely, when parental expectations are communicated with confidence in the child’s abilities (e.g., My parents think I can do well in school) or posed neutrally (e.g., How well do you think your parents expect you to do in math?), they tend to correlate with more adaptive outcomes, such as grit (Won & Lee, 2023) and self-efficacy and valuing (Luo et al., 2016). Additionally, such expectations can act as a protective factor against burnout (Sorkkila et al., 2017). The dimensions of well-being examined in this study include engagement, perseverance, optimism, connectedness, and happiness—collectively referred to as EPOCH (Kern et al., 2016). These dimensions are critical for understanding adolescent well-being as they are developmentally relevant psychosocial characteristics. The EPOCH framework aligns with both hedonic and eudaimonic perspectives on well-being and the concept of flourishing, characterized by both feeling good and functioning well (Huppert & So, 2013). Considering the nature of perceived parental and teacher expectations in this study, perceived parental and teacher expectations were expected to be positively associated with the EPOCH well-being dimensions. Mediation of Academic Self-Concepts Academic self-concepts represent students’ perceptions of their academic abilities. They play a crucial role in how parental (Yamamoto & Holloway, 2010) and teacher (Liem, 2024) SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 8 expectations exert their influence on student motivation and well-being. The role of academic self- concepts in linking perceived parental and teacher expectations to self-determined motivation, achievement goals, and well-being is supported both theoretically and empirically. For example, Marsh and Craven (2006) spotlight the importance of positive self-concepts as “a hot variable that makes good things happen, facilitating the realization of full human potential” (p. 134), suggesting that self-concepts are valuable as an outcome and also as a mediator through which other desirable outcomes are realized. From the SDT perspective, the communication of academic expectations by parents or teachers represents a form of competence support that enhances self-determined motivation and well-being by satisfying students’ needs and, in turn, boosting their sense of competence or academic self-concepts (Bureau et al., 2022; Grolnick & Pomerantz, 2009). These various process models, emphasizing the central role of competence-related self- beliefs, have been supported by multiple studies (e.g., Benner et al., 2021; Chen et al., 2022). Chen et al., for instance, found that the relation between perceived parental expectations and STEM career aspirations was fully mediated by students’ mathematics and science self-efficacy. Bureau et al.’s (2022) recent meta-analysis also identified a sense of competence as the strongest predictor of self-determined motivation, surpassing autonomy and relatedness. From the achievement goal standpoint, perceived ability is theoretically considered an antecedent of achievement goals, where a high perceived ability leads to the adoption of approach-oriented goals, whereas a low perceived ability triggers the pursuit of avoidance-oriented goals (Elliot & Hulleman, 2017). Finally, emerging evidence underscores the significant role of academic self-concepts in non- academic well-being, with Torppa et al. (2023) highlighting their importance due to the central role of schoolwork in adolescents’ lives, especially in crucial subjects such as mathematics. This suggests that students’ academic self-concepts not only influence their academic functioning but also extends to psychosocial well-being, particularly those dimensions most related to perceived competence, such as engagement, perseverance, optimism, and happiness. Therefore, mathematics SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 9 self-concepts were hypothesized as a partial mediator that link perceived parental and teacher expectations to self-determined motivation, mastery and performance goals, and well-being. The Present Study This research was conducted in Singapore, an Asian meritocratic society where education is highly valued as a pathway to a secure future, including a stable job and comfortable life. Consequently, it is common for Singaporean parents to place a strong emphasis on academic achievement, expecting their children to excel in school (Tan, 2012). With a strong emphasis on filial piety, children in Singapore and most East Asian societies view academic success as a way to fulfill their social roles and obligations. Thus, studying becomes a means not only for personal development and securing a future but also for upholding the family’s reputation (Chao, 1994; Hwang, 2012). These cultural characteristics naturally lead to heightened competitiveness within the school environment, and can increase students’ fear of failure, anxiety, and stress (Davie, 2017). Beyond this cultural backdrop, our research focused on perceived parental and teacher expectations in mathematics, a subject widely regarded as foundational for academic progression, higher education opportunities, and career pathways. Mathematics holds a uniquely high value for both parents and teachers, who often emphasize its importance due to its instrumental role in shaping future educational and professional success. It is also frequently perceived as a challenging subject, making it a critical area for understanding how students’ perceived expectations and self- concepts are connected with their motivation and well-being. Furthermore, given the documented decline in adolescents’ motivational beliefs (Eccles & Wigfield, 2020) and psychosocial well-being (Marquez & Long, 2021), this research highlights the importance of fostering motivation and well- being through healthy academic expectations by parents and teachers as key socialization agents. In sum, this research investigated the role of perceived parental and teacher expectations in students’ mathematics self-concept, mathematics motivation, and psychosocial well-being. Additionally, we examined how mathematics self-concepts partially channel the paths linking perceived expectations to motivation and well-being. As illustrated in Figure 1, Study 1 primarily SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 10 addressed self-determined motivation and achievement goals as academic outcomes. Building on this, Study 2 revisited these relationships and expanded the scope to include psychosocial well- being outcomes. While acknowledging that both perceived parental and teacher expectations are important for academic outcomes (e.g., Wentzel et al., 2016) and given the lack of similar studies in Singapore, we refrained from making predictions about the relative salience of these perceived expectations on mathematics self-concepts, mathematics motivation, and psychosocial well-being. We assessed our hypothesized model at two time points and connected the Time 1 (T1) and Time 2 (T2) models via auto-regressive paths (Cole & Maxwell, 2003). While we expected the hypothesized relationships to remain relatively stable across time points, we anticipated smaller effect sizes for the T2 relationships (i.e., standardized regression weights) due to the typically large auto-regressive (stability) paths. By accounting for the effects of corresponding T1 factors, the T2 relationships illuminate how changes in perceived parental and teacher expectations relate to changes in mathematics self-concept, which then relate to changes in motivation and well-being. Additionally, the hypothesized model allowed the examination of the delayed effects of T1 perceived expectations on changes in T2 motivation and well-being through the various linking variables specified in the model (MacKinnon et al., 2007). This design promises insights that could be valuable for educational practices that consider the timing of their implementation. Methods Participants and Procedure Study 1 involved 2,632 students (44.8% female; M age = 14.05, SD age = 0.90), from eight secondary schools in Singapore. The students were proportionately distributed across grades: 33.3% in Secondary 1, 34% in Secondary 2, and 32.7% in Secondary 3. In terms of mathematics proficiency, 12.2% of the students were in Band 1, 30% in Band 2, and 57.8% in Band 3, with higher bands indicating higher proficiency levels. Study 2, comprising an independent sample of 450 students (47.6% female; M age = 13.98, SD age = 0.94) from a different secondary school, showed similar demographic characteristics. In terms of grade, 29.6% of the students were in Secondary 1, SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 11 36% in Secondary 2, and 34.4% in Secondary 3. These students took their mathematics course in Band 1 (10%), Band 2 (36%), and Band 3 (54%). The demographic profiles of the samples from both studies are notably similar, providing a basis for making a broad comparison of their findings. The studies reported in this paper are parts of the larger projects (OER 09/22 GADL and OER 34/22 PKLK) funded by the Ministry of Education, Singapore. The survey, conducted in English as a medium of instruction in Singapore schools, took approximately 30 minutes and was administered in classroom settings by teachers at the respective schools. Students were encouraged to respond truthfully and were assured of the confidentiality of their responses. Prior to administering the survey, necessary ethical and data collection approvals were obtained from the Ministry of Education and Nanyang Technological University’s Institutional Review Board (IRB-2021-268). Both student assent and parental consent were secured. The timing of the survey administrations was strategically planned. The T1 survey occurred at the beginning of Term 2, about 10 weeks into the school year, after students received their Term-1 results. This timing was chosen as it likely influenced the formation of students’ perceived parental and teacher expectations. There was a slight between-study difference in the timing of the T2 survey. It was conducted in Term 4—just as students were preparing for their year-end exams in Study 1 and after the year-end exams in Study 2. The interval between the two survey administrations was approximately 24 weeks for Study 1 and 28 weeks for Study 2. No intervention was conducted between the two survey administrations. Measures Studies 1 and 2 were two independent studies. They administered similar, but unidentical, surveys. All the measures described below were rated on a 7-point Likert scale, ranging from 1 ( Strongly Disagree ) to 7 ( Strongly Agree ). The psychometric properties of subscales in both studies are detailed in Table 1. Academic Expectations. The Parental Expectation subscale, validated by Liu and Wang (2007) for use in Singapore, was adapted to mathematics (4 items; e.g., My parents believe I can do well in math). This subscale was administered in both Studies 1 and 2. However, the Teacher SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 12 Expectation subscale, also by the same authors, was administered in the two studies with slightly different wording. In Study 1, participants rated their perceived teacher’s expectations of their individual performance (4 items; e.g., My math teacher is confident that I can get good results in the subject). In Study 2, participants rated their perceived teacher’s expectations for their class as a whole (4 items; e.g., Our math teacher is confident that we can get good results). Mathematics Self-Concepts. Students’ perceptions of their mathematics abilities were measured by a five-item subscale (e.g., I am good at math) of Marsh’s (1993) Self-Description Questionnaire-II. This subscale, previously used in Singapore with sound psychometric properties (e.g., Liem et al., 2015), was consistently administered across both studies. Mathematics Motivation. Self-determined motivation and mastery and performance goals represented the three academic motivation outcomes. In both studies, self-determined motivation was measured by 14 items drawn from the Self-Regulation Questionnaire (Ryan & Connell, 1989). These items are organized into four subscales reflecting different motivational regulations: external regulation (4 items; I will get into trouble if I don’t), introjected regulation (4 items; I want my parents to see me as a good child), identified regulation (3 items; I want to learn new skills in math), and intrinsic regulation (3 items; I think math is fun). These items were preceded by the stem sentence “I do my math work because...” to standardize the response context. The composite score for self-determined motivation was calculated using the formula (2 × Intrinsic + Identified – Introjected – 2 × External) provided by the authors to strategically weight different types of regulation to reflect their varying impacts on self-determination. In Study 1, the Achievement Goal Questionnaire-Revised (Elliot & Murayama, 2008) was used to assess performance goals (3 items; My goal in my math class is to perform better than the other students) and mastery goals (3 items; My goal in my math class is to learn as much as possible). In Study 2, the 3 × 2 Achievement Goal Questionnaire (Elliot et al., 2011) was used, with performance goals measured by the other-approach goal subscale (3 items; My goal in math class is to do well compared to others), whereas mastery goals measured using the task-approach (3 SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 13 items; My goal in math class is to answer a lot of questions correctly) and self-approach (3 items; My goal in math class is do better than I typically do) goal subscales. As discussed by Elliot et al. (2011) performance(-approach) and other-approach goals are conceptually aligned, so are mastery(- approach) and the combined task-approach/self-approach goals. Well-Being. In addition to the motivation measures, Study 2—but not Study 1—administered the EPOCH Measure of Adolescent Well-Being (Kern et al., 2015) to assess adolescents’ developmentally relevant psychosocial well-being. It consisted of five 4-item subscales: engagement (e.g., I get completely absorbed in what I am doing), perseverance (e.g., Once I make a plan to get something done, I stick to it), optimism (e.g., I think that good things are going to happen to me), connectedness (e.g., There are people in my life who really care about me), and happiness (e.g., I love life). Background Factors (Covariates) . In both studies, gender was coded 0 for males and 1 for females, and school grades ranged from 1 (Secondary 1) to 3 (Secondary 3). Study 1 used mathematics course bands (1-3), determined by students’ Primary School Leaving Examination (PSLE) mathematics scores, as a proxy for prior mathematics achievement, with higher bands indicating greater proficiency. The PSLE is a national standardized assessment taken by all Primary-6 students in Singapore at the end of their primary education. It evaluates proficiency in four subjects: mathematics, English, mother tongue language, and science. Schools assign students to one of three mathematics course bands based on their PSLE mathematics scores. In Study 2, student-reported PSLE alphabetical grades in mathematics were converted to numerical values as an indicator of their mathematics ability. These grades were derived from students’ raw mathematics scores (ranging from 0 to 100) which were standardized into T-scores with a M of 50 and a SD of 10. This standardization ensured comparability of scores across cohorts. Data Analysis Confirmatory factor analysis (CFA) was conducted in both studies to examine the multidimensionality of the focal substantive measures, with the resulting item factor loadings and SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 14 error variances used to calculate omega coefficients for subscale reliability. To test the hypothesized model, Study 1 adopted full structural equation modeling (SEM) and Study 2 used path analysis with error adjustment due to a smaller ratio between sample size and estimated parameters. To account for measurement errors without estimating a measurement model, error- adjusted means σh 2 *(1- ωh), were applied in Study 2, where σh is the estimated variance of a variable and ωh is the reliability estimate of the same variable (see Brown, 2015). To reduce the shared method variance, cross-wave correlations between pairs of the residuals of observed variables were freely estimated (Cole & Maxwell, 2003; Marsh, 1993). Both studies accounted for hierarchical data (students nested in classes) using Mplus 8.8’s (Muthén & Muthén, 1998-2023) TYPE = COMPLEX option with the mathematics class set as the cluster variable and the robust maximum likelihood (MLR) estimator was used in all analyses. Missing data were 16.05% at T1 and 8% at T2 in Study 1, and 14.4% at T1 and 17.6% at T2 in Study 2. Little’s MCAR test indicated that data were missing completely at random for Study 2 (χ 2 [115] = 122.59, p > .05) but not for Study 1 (χ 2 [4295] = 5101.45, p < .001). As one of the recommended strategies (Allison, 2010), we conducted binary logistic regression to test if Study 1’s T2 missingness could be predicted by T1 variables (e.g., math self-concepts) and covariates. The results suggested that T2 missingness was mainly predicted by math course band, with the odds of T2 missingness being 3.32 times more likely for students from the lowest band than highest band. This suggests that the data for Study 1 were missing at random. To address the issue, missing data were handled using the Mplus default Full Information Maximum Likelihood (FIML) and all analyses were estimated using the robust maximum likelihood (MLR) estimator. Model fit was evaluated using the Comparative Fit Index (CFI) and the Root Mean Square Error of Approximation (RMSEA). CFIs of > .90 and > .95 and RMSEAs of < .08 and < .05 indicate good and excellent fit. When interpreting findings, we wanted to guard against Type-1 error (false positive) due to our large sample size in Study 1 (Keith, 2006), or Type-2 error (false negative) because of our relatively smaller sample size in Study 2 could have inflated standard SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 15 errors of regression weights (MacKinnon et al., 2007). Thus, we adopted Keith’s (2006) guidelines to determine the nature of effects with interpretability and applied significance of standardized effects in educational research ( .05 ≤ β s < .10 are considered small, .10 ≤ βs < .25 moderate, β s ≥ .25 large). We considered standardized betas if they met or exceeded Keith’s (2006) benchmark of a medium effect size (β s > .10). Results Measurement Analysis Table 1 shows the descriptive statistics, distributional characteristics, and psychometric properties for all the measured variables in both studies. Because of its relatively large sample size, a preliminary CFA in Study 1 tested a measurement model comprising all the 18 latent factors at both time points (9 at each time point: 2 expectations, 1 math self-concepts, 4 behavioral regulations, 2 achievement goals). The model yielded an excellent fit (χ 2 = [1866] = 6247.22, p < .001, CFI = .959, RMSEA = .030). Because of the smaller sample size and more variables in Study 2, the CFAs were conducted on T1-T2 responses for each category of constructs: perceived expectations (χ 2 = [141] = 308.83, p < .001, CFI = .923, RMSEA = .051); mathematics self- concepts (χ 2 = [127] = 70.89, p < .001, CFI = .981, RMSEA = .060); behavioral regulations (χ 2 = [297] = 703.27, p < .001, CFI = .906, RMSEA = .055); achievement goals ( χ 2 = [112] = 162.25, p < .001, CFI = .984, RMSEA = .032), and well- being (χ 2 = [673] = 1137.17, p < .001, CFI = .926, RMSEA = .039). All item factor loadings were significant at p < .001. Table 1 shows that each substantive factor in both studies had strong within-construct validity as reflected in the substantial means of factor loadings. Table 1 also shows that participants’ responses were normally distributed, as values were within acceptable ranges for skewness ( + 2) and kurtosis ( + 7). Omega coefficients in Study 1 ranged between .83 and .95 at T1 and between .86 and .95 at T2, whereas those in Study 2 ranged between .69 and .89 at T1 and between .72 and .91 at T2, showing acceptable reliability of all the subscales. We also computed an intra-class correlation (ICC) for each variable. The bulk of SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 16 variance (over 95%) resided at the student level rather than at the class level (ICC M s = .037 at T1 and .046 at T2 for Study 1 and .035 at T1 and .042 at T2 for Study 2), supporting the appropriateness of the individual-level analysis in this study. Correlation Analysis Table 2 reports within and between-wave correlations among factors. It shows a generally consistent trend across studies and time points that (a) mathematics self-concepts were correlated more strongly with perceived parental expectations than perceived teacher expectations, (b) mathematics self-concepts had slightly larger correlations with self-determined motivation than with achievement goals, and (c) motivation variables were more strongly correlated with perceived parental expectations than perceived teacher expectations. As Table 2 also shows, in Study 2, well- being factors correlated more strongly with perceived parental expectations than perceived teacher expectations. Lastly, mathematics self-concepts were moderately correlated with well-being factors. Together, these correlations demonstrated the relative stability of inter-variable relationships across the two time points, despite using different measures to assess perceived teacher expectations and achievement goals in the studies. They also provided a foundation for further multivariate analyses using SEM (Study 1) and path analytic (Study 2) models, controlling for shared variance among variables, including covariates. Main Analysis The main analysis was conducted to test the hypothesized model in Figure 1. Background factors (gender, school grade, prior achievement) were included as covariates in the model. Study 1. SEM yielded excellent fit (χ 2 = [819] = 4066.91, p < .001, CFI = .946, RMSEA = .039). Figure 2 shows the significant betas while Table S1 reports all betas in the model. Findings demonstrated that, at T1, perceived parental and teacher expectations were associated with mathematics self-concepts, which were then associated with self-determined motivation, mastery goals, and performance goals. The difference of the paths linking perceived parental and teacher expectations to mathematics self-concepts was significant ( Δχ 2 = [1] = 57.88, p < .001). A SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 17 significant interaction effect pointed to the benefit of having high levels of both perceived parental and teacher expectations for mathematics self-concepts (see Figure 4a). Replicating the T1 relationships, at T2, perceived parental and teacher expectations were associated with mathematics self-concepts, which were then associated with self-determined motivation, mastery goals, and performance goals. Similar to T1, the difference between the paths from perceived parental and teacher expectations to mathematics self-concepts was significant albeit marginally ( Δχ 2 = [1] = 3.78, p < .10). These observed T2 relationships were significant even after controlling for autoregressive paths, showing that changes in perceived parental and teacher expectations were associated with changes in mathematics self-concepts which were then associated with changes in self-determined motivation and the two achievement goals. Additionally, perceived teacher expectations had direct associations with self-determined motivation, mastery goals, and performance goals, whereas perceived parental expectations had direct associations only with the two achievement goals. These direct links were consistent at both time points, attesting to the relative stationarity of these relationships. Study 2. Path analysis, with correction for unreliability via error-adjusted scores, yielded excellent fit ( χ 2 [123] = 181.02, p < .001, CFI = .986, RMSEA=.032). Figure 3 shows the significant betas whereas Table S2 reports all beta in the model. Results demonstrated that, at T1, perceived parental and teacher expectations were associated with mathematics self-concepts. Although the path from perceived parental expectations to mathematics self-concepts was doubled the path from perceived teacher expectations to mathematics self-concepts, they were not significantly different ( Δχ 2 = [1] = 1.66, p > .05), likely due to the lower statistical power associated with a smaller sample size. Mathematics self-concepts, in turn, were associated with self-determined motivation, mastery goals, performance goals, engagement, perseverance, optimism, and happiness. The link between mathematics self-concepts and connectedness was not significant. Consistent with Study 1, perceived parental and teacher expectations had a synergistic effect on mathematics self-concepts (see Figure 4b), so did they on connectedness and happiness (see Figures 4c and 4d). SINGAPOREAN STUDENTS’ PERCEIVED PARENTAL EXPECTATIONS 18 Beyond the auto-regressive paths, at T2, perceived parental expectations were associated with mathematics self-concepts, which were then associated with self-determined motivation, mastery goals, performance goals, engagement, perseverance, optimism, and happiness. Similar to T1, the path from mathematics self-concepts to connectedness at T2 was not significant. The path from perceived teacher expectations to mathematics self-concepts was also not significan