The NSS curriculum was officially launched in 2009. As a bold step forward for secondary education in Hong Kong, the NSS reform adopted several significant changes to the high school education system. The NSS curriculum reduced the total number of secondary school years (from 7 to 6 years) and increased the number of tertiary education curriculum years (from 3 to 4 years). As for public examination, a single Hong Kong Diploma of Secondary Education Examination (HKDSE) at the end of Secondary 6 replaced the previous two examinations: the Hong Kong Certificate of Education Examination and the Hong Kong Advanced Level examination. Guided by the framework and related learning goals, significant changes under the NSS curriculum consisted of designating Liberal Studies as one core subject and introducing an elective subject system and Other Learning Experience (OLE) as key elements of the curriculum (6). The core subject, Liberal Studies, creates opportunities for students to study and reflect on contemporary social and cultural issues concerning personal and social development, which would broaden students' knowledge base and increase their citizenship (6). The elective subjects include practical courses such as Engineering and Production, Business, Accounting and Financial Studies to prepare students for future career development. Additionally, courses such as Ethics and Religious Studies, Moral and Civic Education, and Community Service are provided to improve students' lifelong skills and to inspire their full potential and whole-person development.

The flexible and practical structure of the NSS curriculum allows schools to organize activities to develop students' noncognitive skills (4). The core subject, Liberal Studies, covers topics under the themes “self and personal development,” “society and culture,” and “science, technology, and the environment” for students to understand themselves as well as their relationships with society and the environment (7). Liberal Studies is expected to improve students' self-understanding, social awareness, and responsibility through investigating a variety of social issues (8–10). The introduction of OLE explicitly reflects the learning goals of the NSS curriculum, which stress the development of students' generic skills, values, and attitudes. The OLE provides various vocational development education and related life planning training activities for students to gain understanding of their potential and find purpose in life (11). Under the OLE, each student needs to spend 405 hours of extra-curricular activities in five learning areas (Moral and Civic Education, Community Service, Career-related Experiences, Aesthetic Development, and Physical Development) and complete Student Learning Profile (SLP). In practice, some schools organized a variety of aesthetic and physical activities, such as dance classes, to strengthen students' confidence, perseverance, resilience, and positive coping strategies (4). Along with the subject offer, the NSS curriculum also encourages a shift in teaching philosophies and methods to promote students' noncognitive development. Teachers are expected to be the key enactors to ensure students have “opportunities for developing diverse ways of learning according to their interests, needs and abilities to achieve the aims of education” (3). For example, reflexivity is a central pedagogical principle that is particularly emphasized in inquiry-based learning activities (8). Teachers applying reflexivity are aware of the influence of personal perspectives in constructing “knowledge” and thus more likely to support students to reflect in-depth and develop personal opinions (12).

At the initial stage of the reform, researchers identified various challenges schools and teachers face in translating government policy into school-based practices and implementing innovative curricula in real classrooms, including teachers' heavy workloads, limited understanding of the reform, insufficient professional training, and students' learning diversity (13, 14). Research has also demonstrated that effective school leadership and management are vital in gauging successful curriculum reform. For example, Lai and Cheung (15) interviewed school principals and found that shared instructional leadership and decision-making contributed to effective school-based curriculum implementation.

Following the first round of completion (i.e., 2009–2015), researchers have a more comprehensive picture to examine the NSS curriculum implementation. Empirical studies have shown that effective and examination-oriented practices coexist in curriculum implementation (9, 10). Studies on effective practices further identified effective school practices, such as collaborative inquiry, instructional strategy, and catering to learner diversity (16). However, some gaps between the ambitious objectives and the actual implementation exist. For example, although teachers concur with the reform objectives to promote students' critical thinking, they face pedagogical challenges, such as limited updated resources, which could hinder their support for students (9, 10). Additionally, there is limited empirical evidence supporting the hypothesis that the NSS reform promotes students' noncognitive development. Existing research has mainly focused on skills closely related to academic learning, including positive attitudes of learning and self-concepts regarding academic work (4). Among the limited studies on this topic, Cheung et al. (17) examined some aspects of students' noncognitive development, such as communication skills and problem-solving. They revealed that students perceived themselves as performing well in establishing good interpersonal relationships but as weak in self-regulated and reflective learning. Chan et al. (18) examined the influence of the NSS curriculum on students' transition from school to the work market and found an improvement in students' generic competencies. Yuen et al.'s (4) study based on 8,122 secondary school students indicated that moderate progress had been made in students' overall performance in generic skills, positive values, and attitudes over the key learning areas. Thus, further examination of different aspects of noncognitive skills against the backdrop of the NSS reform is needed.

The first research gap involves a lack of validated measurements of students' perceptions of the NSS curriculum in the existing literature. Primarily, most measures of academic or school satisfaction have been developed in the West. Moreover, imported measures have rarely been validated in the Chinese context (19). Obviously, a lack of validated assessment tools on perceptions of the NSS curriculum would hinder meaningful analyses of the effectiveness of the educational reform. Only Cheung et al. (17) developed and validated a scale to examine different stakeholders' perceptions of the NSS curriculum. As the items of Cheung et al.'s scale were related to the policy recommendations and reform objectives, this scale mainly focused on the quality of student learning under the NSS curriculum but did not cover attributes including self-understanding, resilience, emotional competence, and purpose in life, which are profoundly critical for nurturing students' positive development in the long run (19). Against the above background, this study attempted to examine the psychometric properties of the “Perceptions of the New Secondary School Curriculum” (PNSC). This scale was developed to measure four dimensions of Hong Kong students' perceptions of the NSS curriculum, including their fondness for and interest in the junior/senior secondary curriculum and perceived benefits of the junior/senior secondary curriculum in promoting positive and holistic development. Shek and Chai (20) conducted exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) to examine the structure of the scale. In the present study, we further examined the convergent, discriminant, and factorial validities (i.e., factorial invariance) of the scale.

The second research gap involves a lack of investigation of students' voices over the educational reform in the literature. Most existing studies have mainly adopted the perspectives of school heads and teachers as they are the frontline reformers at the school level (i.e., expert views). Due to the top-down nature of educational reforms, research has often focused on identifying on-site difficulties; the discrepancies between the understandings of the public, school, and policymakers; and characteristics of effective schooling (4, 17, 21). Students' perceptions of the curriculum and school programs reflect their academic satisfaction, which is an essential indicator of educational transformation success (20, 22). Marchiondo et al. (23) defined academic satisfaction as “the attraction or positive feelings that a student associates with the college or program in question” (p. 610). Scholars also found a positive relationship between academic satisfaction and students' developmental outcomes. For example, Rowe, Stewart, and Patterson (24) found that positive learning experiences in schools could strengthen students' health and well-being. On the contrary, students who do not like school or feel disconnected from school tend to have lower academic performance or even drop out (25, 26). However, the voice of secondary students, who are the direct recipients of educational services and major stakeholders, has received insufficient attention. Gaps might also exist between students' and other stakeholders' perceptions (4, 20). Among the limited studies examining students' experiences, Cheung et al. (17) used a large and representative sample to examine students' perceptions. They asked students to indicate how well the students in their schools performed in some aspects of noncognitive development (e.g., growth mindset and communication skills). However, how students individually perceive their own development was not examined. Additionally, the survey did not provide sufficient information about how students perceive the benefits of the NSS curriculum, as it mainly focused on the “objective” outcomes instead of perceived evaluation.

The third research gap is that it is not clear whether students perceive the junior and senior curricula differently. Some studies revealed a decreasing trend in school satisfaction reported by students (27). One possible reason is that students' academic stress might gradually increase when they reach the senior secondary level. In particular, senior high school students in Hong Kong face heavy academic stress due to the shortened school year and the competitive final examination under the NSS curriculum. Research on Chinese students revealed higher levels of perceived academic stress reported by students in higher grades (e.g., senior secondary school years) than junior secondary students (28). Zhu and Shek (29) examined the perceived effectiveness of a school-based curriculum promoting positive youth development in the Chinese context. They found that students in junior grades showed significantly more positive evaluations than did students in senior grades. However, other studies have not found a close correlation between grade and school satisfaction (30). Nevertheless, there is a need to understand the potential difference between students' perceptions of the junior and senior secondary curricula.

The fourth research gap concerns a lack of investigation of gender differences in students' perceptions of the NSS curriculum. Previous studies on school and academic satisfaction revealed that girls tended to report higher engagement and satisfaction with school than did boys (31). However, gender differences were not observed in other studies. For example, Løhre et al.'s study conducted with 149 boys and 119 girls revealed no gender differences in school well-being (32). Similarly, some studies based on Chinese students did not find any gender differences in children's perceived school satisfaction (33). Because this issue is not well addressed under the NSS curriculum, further examination is needed.

The fifth research gap is that, despite the good intention for the NSS curriculum to change the exam-oriented system and promote all students' noncognitive skills, we have little knowledge of whether high- and low-performing students differ in their perceptions of the benefits of the NSS curriculum in promoting their noncognitive development. Theoretically speaking, the NSS curriculum should apply the “no loser” principle in Hong Kong's education and promote whole-person development of all students despite their academic performance levels. As the new system introduces more selective and practical subjects, low-achieving students should have more opportunities to develop their potential in different aspects and could benefit more from the system than high-performing students. However, empirical studies have shown inconclusive findings concerning the relationship between students' performance and their satisfaction with school. Previous studies revealed that high-performing students often received higher levels of teacher expectation and support, reported more positive school experiences, and benefitted more from the school curriculum, which might contribute to more positive perceptions of the school curriculum (34, 35). For example, Zhou et al.'s (36) study, conducted with Chinese high school students, revealed that high-achieving students perceived higher levels of need satisfaction at schools, which in turn improved their future academic achievement. Similar results were found in the context of Hong Kong that gifted students with high performance reported higher levels of satisfaction than average students did (37). On the other hand, high-performing students in Chinese contexts often exhibit perfectionism (37) and experience heavier social expectations (38), which could negatively influence their perceptions of the school curriculum. Nevertheless, empirical evidence examining the relationship between students' performance and their perceptions of the NSS curriculum is scarce, thus calling for further studies addressing this research gap.

This study attempted to answer the following research questions to address the above research gaps:

Research Question 1: What are the psychometric properties of the PNSC as a measure of the students' perceptions of the NSS curriculum?

Research Question 2: What are high school students' perceptions of the NSS curriculum indexed by the PNSC?

Research Question 3: Do students perceive the junior and senior secondary school curricula differently? As the school work and examination pressure are often heavier in senior secondary studies than in junior secondary studies (39), we hypothesized that perceptions of the junior secondary curriculum would be more positive than those of the senior secondary curriculum (Hypothesis 1).

Research Question 4: Do boys and girls differ in their perceptions of the secondary school experience? Based on previous findings (31, 40), we expected that girls would have more favorable perceptions of the secondary school curriculum than boys would (Hypothesis 2).

Research Question 5: Do the perceptions of the NSS curriculum differ among students with different academic and school performance? Based on previous findings of studies conducted in the Chinese contexts (36, 37), we hypothesized that students with better academic and school performance would have more favorable perceptions of the NSS curriculum. If the NSS curriculum really emphasizes both academic and nonacademic skills, students with different academic performance levels would likely not differ appreciably in either direction in their perceptions. It is also possible that low-achieving students would perceive more benefits from the NSS curriculum because it seeks to change the role of examination scores as the only criterion of student performance. Therefore, we explored the relationships between students' academic performance and their perceptions of the NSS curriculum without making hypotheses.

This study is a part of a 6-year longitudinal project entitled Positive Adolescent Training through Holistic Social Programs (P.A.T.H.S.) initially launched in the 2009/2010 school year. The data used in the present study were collected every year between 2009/2010 and 2015/2016. This project aimed to examine adolescent development of high school students in Hong Kong, including their positive youth development, family functioning, academic performance, and perceptions of the NSS curriculum. Participants were from 28 schools in different areas of Hong Kong, including Kowloon, New Territories, and Hong Kong Island. All students in Grade 7 of the 28 schools were invited to participate in this study. A questionnaire was administered to participants annually for six consecutive years across secondary school education. Students responded to the questionnaire in classrooms during school hours. Written informed consent was obtained from students, parents, and school administrators before the data collection. Researchers provided information about the purpose, confidentiality, and anonymity principles of the research to all involved parties. The project was reviewed and approved by the authors' university.

In the present study, we used a sample of 3,498 Secondary 6 students who participated in the sixth wave of data collection in 2016 (i.e., the last year of the NSS curriculum). Among them, 1,669 were girls (47.7%), 1,821 were boys (52.1%), and 9 did not report their gender (0.2%). The mean age of the participants was 17.33 years.

We first examined the skewness and kurtosis of all items in the PNSC to test the normality. According to Curran et al. (44), the absolute values of skewness and kurtosis should be lower than 2 and 7, respectively, so that the requirement of normality can be met and maximum likelihood estimation would be appropriate for CFA. Results showed that for items in the PNSC subscales, absolute values of skewness ranged between 0.277 and 0.691, which were all below 2. In addition, absolute values of kurtosis ranged from 0.009 to 0.613, which were lower than 7. Therefore, all items can be considered to demonstrate a normal distribution (44).

The means, SD, and correlations of variables are summarized in Table 1. We found significant correlations between the four factors of the PNSC (rs ranged between 0.552 and 0.772, p < 0.01, see Table 1). We used AVE and CR to assess the convergent validity (45). As shown in Table 2, the AVE values for all four factors were greater than 0.5, and the CR values were above 0.70, supporting the good convergent validity of the scale (45). Following Fornell and Larcker's recommendations (45), we assessed the discriminant validity of the scale by comparing the square root of the AVE for each factor and the correlation involving the factor. The results showed that the values of the square root of the AVE for each factor ranged between 0.889 and 0.914, which were larger than the correlations under investigation ranging from 0.552 to 0.772. Thus, the discriminant validity of the scale was supported. Results of the reliability test showed good internal consistency of the PNSC (Cronbach's alpha = 0.967, McDonald's omega = 0.967, see Table 2) and all four factors (Cronbach's alpha values ranged between 0.946 and 0.953, McDonald's omega values ranged between 0.947 and 0.952, see Table 2).

Based on the work of Shek and Chai (20), we performed CFA on the 4-factor PNSC model with the entire sample. Results revealed that the baseline model (Model 0) demonstrated a good fit {χ² ₍₁₂₁₎ = 1,852.567; CFI = 0.977; NNFI = 0.971; RMSEA = 0.065 [90% CI: 0.063 to 0.068]; SRMR = 0.024}. The satisfactory fit provided basis for performing multigroup CFA to examine measurement invariance across genders.

Multigroup CFA were conducted on a subsample of boys (N = 1,821) and a subsample of girls (N = 1,669). The baseline model (Model 0) was tested separately by gender in both subsamples to gauge its factorial stability (48). The baseline model demonstrated a good fit to the data in both the boy subsample {χ² ₍₁₂₁₎ = 1,038.85; CFI = 0.978; NNFI = 0.972; RMSEA = 0.066 [90% CI: 0.062–0.07]; SRMR = 0.022} and the girl subsample {χ² ₍₁₂₁₎ = 1,162.212; CFI = 0.968; NNFI = 0.96; RMSEA = 0.073 [90% CI: 0.069–0.077]; SRMR = 0.031}, illustrating the factorial stability in both subsamples.

Next, we tested the configural invariance model (Model 1), metric invariance model (Model 2), scalar invariance model (Model 3), and error variance invariance model (Model 4). As shown in Table 3, Model 1 showed adequate fit to the data {χ² ₍₂₄₂₎ = 2,201.062; CFI = 0.974; NNFI = 0.967; RMSEA = 0.07 [90% CI: 0.067–0.072]; SRMR = 0.025}, supporting the invariance of the factorial structure across the boy and girl subsamples. In Model 2, factor loadings were constrained to be the same across genders {χ² ₍₂₅₆₎ = 2,223.99; CFI = 0.973; NNFI = 0.968; RMSEA = 0.068 [90% CI: 0.065–0.07]; SRMR = 0.025}. As shown in Table 3, results of the χ² tests revealed significant differences (Δχ² = 22.93, Δdf = 14, p < 0.01) between Models 1 and 2, between Models 2 and 3 (Δχ² = 55.83, Δdf = 14, p < 0.001), and between Models 3 and 4 (Δχ² = 150.07, Δdf = 18, p < 0.001). As suggested by Cheung and Rensvold (49), we referred to the value of difference-in-CFI (ΔCFI ≤ 0.01) instead of changes in χ² due to the large sample size. The comparison between Model 2 and Model 1 yielded a ΔCFI (<0.001) below 0.01, suggesting invariance in factor loadings across genders. In Model 3, both factor loadings and measurement intercepts were assumed to be equal across genders {χ² ₍₂₇₀₎ = 2,279.824; CFI = 0.973; NNFI = 0.969; RMSEA = 0.067 [90% CI: 0.064–0.069]; SRMR = 0.026}. The value of ΔCFI was below 0.01, suggesting invariance in measurement intercepts across the two subsamples (see Table 3). Lastly, Model 4 constrained factor loading, measurement intercept, and the error variance of each variable to be equal across the boy and girl subsamples {χ² ₍₂₈₈₎ = 2,429.895; CFI = 0.971; NNFI = 0.969; RMSEA = 0.067 [90% CI: 0.064–0.069]; SRMR = 0.026}. The value of ΔCFI (0.002, see Table 3) was below 0.01, denoting the invariance of measurement error for each item across genders.

The percentage responses of students to the items on the PNSC are summarized in Tables 4, 5 for the junior and senior secondary curricula, respectively. Concerning the junior secondary curriculum, the majority of students liked the curriculum (67.38%) and agreed that the curriculum was interesting (66.04%), encouraged them to reflect (59.15%), and enhanced learning interests (63.46%). Students also acknowledged the benefits of the curriculum in promoting their positive and holistic development. For example, more than two-thirds of the participants agreed that the junior curriculum helped them establish their values and deepen their self-understanding.

As for the senior secondary curriculum, students demonstrated neutral to positive attitudes toward the curriculum. Slightly more than half of the students liked the curriculum (51.43%) and agreed that the curriculum was interesting (55.15%) and enhanced learning interests (55.55%). Around 60% of students believed the curriculum encouraged self-reflection. As for the perceived benefits of the senior secondary curriculum, students perceived that the curriculum helped them establish values (63.18%), deepen self-understanding (66.52%), improve the ability to deal with adverse situations (63.61%), improve emotional competence (59.83%), and find their purpose in life (63.92%).

On the other hand, although a significant proportion of the respondents acknowledged the benefits of the curriculum in promoting their noncognitive development in both junior and senior secondary education, substantial proportions reported relatively negative perceptions. Concerning the junior secondary curriculum, around one third of students stated negative perceptions (see Table 4). For example, 40.11% of students disagreed that the curriculum encouraged them to reflect. More than 41% of students generally disagreed that the curriculum helped them find their life's purpose. As for the senior secondary curriculum, about 40% of students gave negative responses on items under fondness for the curriculum, and one third generally disagreed on items under perceived benefits of the curriculum. For example, more than 48% of students gave negative responses to the item “I like the senior secondary curriculum.” Around 40% of students generally disagreed that the senior curriculum improved their emotional competence.

As displayed in Table 6, results of paired t-tests showed that students had more positive perceptions of the junior secondary curriculum (mean = 3.682) than of the senior secondary curriculum (mean = 3.493; t = 10.779, p < 0.001, Cohen's d = 0.183). However, no significant difference was observed in perceived benefits between junior and senior secondary education (t = 0.235, p = 0.814). Generally, students liked the junior secondary curriculum more than the senior secondary curriculum, supporting Hypothesis 1.

We conducted a MANOVA to examine gender differences in subscales of the PNSC. We first tested the homogeneity of covariance matrices with Box's M test. The result of Box's M test was 145.266 with a p-value <0.001, which was interpreted as significant according to the cutoff suggested by Huberty and Petoskey (50). Thus, this assumption was violated. As researchers have argued, Box's M test is sensitive to large sample sizes and could detect very small departures from homogeneity (51). Following the recommendation of Tabachnick et al. (51), we adopted Pillai's trace criterion as it is the most robust statistic for protection against departures from the assumptions. MANOVA results showed that girls generally possessed more positive perceptions of the curriculum than did boys [Pillai's trace = 0.011, F_{(4, 3643)} = 10.068, p < 0.001, η² = 0.011, see Table 7], which generally supported Hypothesis 2. Girls showed higher levels of fondness for and interest in the junior secondary curriculum (F = 13.58, p < 0.001, η² = 0.004). However, we found no gender difference in fondness for and interest in the senior secondary curriculum. With regard to the perceived benefits of the curriculum, a significant gender difference was observed for both junior and senior secondary curricula. As compared with boys, girls perceived more benefits of the junior secondary curriculum (F = 15.37, p < 0.001, η² = 0.004) and the senior secondary curriculum (F = 10.52, p < 0.01, η² = 0.003).

We performed a MANOVA to test potential mean differences between academic performance and PNSC scores. Similarly, Box's M test was performed to check the homogeneity of covariance matrices. Based on Huberty and Petoskey's guideline (50), we reported Pillai's trace in the MANOVA as a significant value of Box's M test was observed. We found a significant multivariate effect for the four factors of perceptions of the NSS curriculum based on students' perceived academic performance [Pillai's trace = 0.058, F_{(8, 4, 010)} = 14.973, p < 0.001, η² = 0.029]. As shown in Table 8, MANOVA results showed that students' academic performance levels were positively related to all subscales of perceptions of the NSS curriculum. Additionally, results of post-hoc comparisons revealed significant differences between all pairs in comparisons (p < 0.001). That is, as compared with the other two groups, high-performing students liked the NSS curriculum the most, whereas low-performing showed the lowest levels of fondness for/interest in the curriculum. We found the same patterns when comparing the perceived benefits of the NSS curriculum by academic performance. Notably, the mean score of low-performing students' fondness for/interest in the senior school curriculum was 2.82 (SD = 1.18), suggesting that they slightly disliked the senior secondary curriculum.