Open Access Article
This Open Access Article is licensed under a Creative Commons Attribution 3.0 Unported Licence
Profiles of chemistry identity among Chinese high school students: antecedents and outcomes
Mutong Niu
ab,
Haoran Sunab,
Qianfeng Zhangab and
Yurong Liu*ab
aSchool of Chemistry and Chemical Engineering, Henan Normal University, Xinxiang, Henan 453007, China. E-mail: liuyur66@163.com
bFaculty of Education, Henan Normal University, Xinxiang 453007, China
First published on 1st June 2026
Abstract
Chemistry identity is closely associated with students’ engagement in chemistry learning, academic performance, and chemistry-related career development. However, chemistry identity is not a unitary construct; rather, it comprises multiple dimensions, including interest, competence, performance, and recognition, and students may exhibit different configurations of these identity dimensions. To examine heterogeneity in high school students’ chemistry identity patterns, this study investigated 1431 high school students and employed factor mixture modeling, grounded in the four-dimensional framework of science identity, to identify latent profiles of chemistry identity. In addition, the one-step approach and the BCH method were used to examine the relationships of antecedent and outcome variables with profile membership. The FMM results suggested three probabilistic profile patterns of chemistry identity: the academic competence-career alienation profile (37.6%), the robust identity profile (51.0%), and the interest-career driven profile (11.4%). These profiles should be interpreted as latent response patterns reflecting both overall level differences in chemistry identity and dimension-specific tendencies, rather than as sharply discrete student categories. Gender, perceived teacher support, and chemistry self-efficacy significantly predicted profile membership. Specifically, higher levels of teacher support were associated with a greater likelihood of membership in more adaptive chemistry identity profile patterns, whereas higher chemistry self-efficacy was more likely to predict membership in positive profiles such as the interest-career driven profile. In contrast, grade level, parental occupation, and class leadership role did not significantly predict profile membership. Further analyses revealed significant differences among the three profiles in chemistry academic achievement and chemistry university aspirations, with the interest-career driven profile showing the highest levels, followed by the robust identity profile, and the academic competence-career alienation profile showing the lowest levels. These findings indicate substantial heterogeneity in high school students’ chemistry identity and highlight perceived teacher support and chemistry self-efficacy as important variables associated with more adaptive chemistry identity profiles.
Introduction
In the PISA 2025 Science Framework, the affective factors influencing competencies were broadened from the concept of scientific attitudes in earlier frameworks to the more comprehensive construct of science identity (OECD, 2023). In China, the Ministry of Education issued the Opinions on Strengthening Science Education in Primary and Secondary Schools in the New Era, which states that efforts should be made to “comprehensively improve students’ scientific literacy, cultivate adolescents with the potential to become scientists and the willingness to devote themselves to scientific research, plant the seeds of science in children's minds, and guide them to dream of becoming scientists” (Ministry of Education of the People's Republic of China, 2023). Although this policy statement does not explicitly use the academic term science identity, its emphasis on emotional identification with science, career aspirations, and value commitment is highly consistent with the core connotations of science identity. The introduction of science identity provides a new perspective for fostering scientific literacy and highlights the practical and long-term significance of cultivating students’ science identity.Identity has a strong influence on how students interact, participate, perform, and learn (Lave and Wenger, 1991; Bishop, 2012). Previous research has also shown that science identity is significantly associated with students’ science-related career aspirations (Carlone, 2017). However, primary and secondary science education has traditionally emphasized students’ understanding of scientific knowledge and the nature of science, while students’ science identity and related aspirations have received comparatively less attention. Given that science identity is considered a key factor in enabling students to think, make decisions, and act like scientists, it deserves greater consideration in science education (Li, 2025). Among Chinese high school students, science learning is organized around relatively independent school subjects, such as physics, chemistry, biology, and geography. Therefore, students’ science identity may be better understood by examining how they identify with specific science subjects. In this context, chemistry identity can be regarded as an important subject-specific form of science identity. Empirical studies have shown that levels of science identity vary across subjects (Vincent-Ruz and Schunn, 2018; Chen and Wei, 2022). In the Chinese educational context, however, limited attention has been paid to the heterogeneity of high school students’ chemistry identity. In particular, existing research has yet to fully reveal the differential combinations and latent patterns of students’ multidimensional chemistry identity. Therefore, it is necessary to examine the current status of high school students’ science identity from the perspective of chemistry as a school subject. Such an effort is important for helping educators better understand students’ motivation for learning chemistry and their future career intentions (Pfeifer et al., 2024).
Science identity theory
Current research on science identity is mainly grounded in two influential models. First, Carlone and Johnson proposed a three-dimensional model of science identity consisting of competence, performance, and recognition (Carlone and Johnson, 2007). Within this framework, individuals are considered to possess a strong science identity when they demonstrate scientific competence, perform effectively in scientific activities, and receive recognition from both others and themselves as being capable in science. Subsequently, Hazari et al. extended and revised this framework through large-scale quantitative research on physics identity using data from the PRISE project. They argued that, for high school students and early undergraduate students, science identity is still in the process of development and has not yet fully stabilized. In this context, interest – defined as an individual's willingness to learn about and engage in science – serves as an important driving force in identity development (Hazari et al., 2010). Accordingly, the framework was expanded into a four-dimensional model of science identity, including interest, competence, performance, and recognition. Building on this four-dimensional identity framework, recent chemistry education research has further clarified how science identity perspectives can be operationalized in chemistry-specific contexts. Hosbein and Barbera argued that identity measures in chemistry education should be theoretically grounded and used the physics identity framework, including performance/competence, recognition, and interest, as the starting point for defining science and chemistry identity (Barbera, 2020). Their qualitative study showed that students’ perceptions of science and chemistry identity could be aligned with broader affective and social-cognitive constructs, including mastery experiences, verbal persuasion, vicarious experiences, situational interest, and mindset. This work is particularly relevant to the present study because it provides a theoretical bridge between the components of chemistry identity and students’ beliefs about their chemistry-related competence, the recognition they receive from others, and the interest they develop in chemistry learning.In a subsequent study, Hosbein and Barbera (Hosbein and Barbera, 2020) developed and evaluated novel science and chemistry identity measures in university-level chemistry courses. Their work provided validity and reliability evidence for the Measure of Chemistry Identity (MoChI) and further examined the relations among mastery experiences, verbal persuasion, situational interest, and science or chemistry identity. Although MoChI was developed in university-level chemistry contexts, this work is important for the present study because it clarifies how chemistry identity can be measured as a theoretically grounded, multidimensional construct. Importantly, their findings suggested that mastery experiences alone were not directly related to chemistry identity; rather, the more theoretically supported model indicated an indirect relation through verbal persuasion and situational interest. These findings suggest that chemistry identity is shaped not only by students’ successful performance in chemistry, but also by interest-supporting experiences and recognition-related feedback. In the context of the present study, perceived teacher support can be understood as an important source of recognition-related feedback and verbal persuasion, whereas chemistry self-efficacy reflects students’ competence-related beliefs about their ability to succeed in chemistry learning. Therefore, examining perceived teacher support and chemistry self-efficacy together with multidimensional chemistry identity profiles is theoretically warranted.
These chemistry-specific findings are consistent with the broader view that identity is not formed statically; rather, it is gradually constructed through continuous interactions with others and the surrounding environment, which gives it a distinctly social character. Accordingly, experiences in science education contexts, such as interactions with peers, teachers, and instructional materials, shape individuals’ understandings of themselves. It is therefore necessary to pay attention to scientific participation and its relationship with science identity development in order to help students develop positive science identities (Kim and Sinatra, 2018). Compared with the three-dimensional model, the four-dimensional framework is particularly relevant for studying adolescent students because it incorporates interest as a key component of identity development, which is important for understanding young students’ science-related self-perceptions and future choices (Hazari et al., 2010). Therefore, the present study adopts the four-dimensional science identity framework as its theoretical foundation to examine the characteristics of chemistry identity among Chinese high school students.
From a broader perspective, identity is a complex and multidimensional construct that refers to the way individuals develop multiple identities across different social contexts and integrate them into an overall understanding of the central question, “Who am I?” (Vincent-Ruz and Schunn, 2021). In general, identity can be divided into three levels: personal identity, social identity, and science identity (Hazari et al., 2010). Among these, Aschbacher et al. defined science identity as students’ perceptions of who they are, what they believe they are capable of doing, and what kind of person they hope to become (Aschbacher et al., 2010).
Developmental pathways of science identity
Previous studies have shown that the development of science identity is influenced by multiple factors. Previous studies have summarized three major characteristics of science identity and pointed out, first, that students’ science identity is jointly shaped by external social structures and the development of the individual self (Wenger, 1998; Zhao et al., 2024). On this basis, related research has further identified three key pathways through which students’ science identity develops (Vincent-Ruz and Schunn, 2018; Jiang and Wei, 2025): a sense of belonging and connectedness to the community, the alignment of internal and external attitudinal factors, and the degree of correspondence between school science and authentic science.The first pathway, belonging and connectedness, emphasizes whether students experience acceptance and recognition during science learning. It focuses on how interactions with and support from significant others, such as teachers, peers, and parents, influence the formation of science identity. This pathway reflects a social constructionist perspective, according to which science identity originates not only from individuals’ self-perceptions but is also shaped by others’ recognition and broader social contexts (Moon et al., 2025). The second pathway posits that science identity is shaped by relatively stable attitudinal factors, encompassing both intrinsic and extrinsic aspects (Aschbacher et al., 2010). From this perspective, students' science interest, perceived value of science, and self-efficacy or competence beliefs are regarded as important factors closely associated with science identity development (Maltese and Tai, 2010; Oettingen and Gollwitzer, 2015). However, there remains some debate in the current literature as to whether these attitudinal variables are antecedents of science identity or constituent components thereof (Trujillo and Tanner, 2014). The third pathway, the correspondence between school science and authentic science, highlights the situated and dynamic nature of science identity. It concerns whether classroom instruction can provide learning experiences that approximate authentic scientific practice. For example, students may construct their chemistry identity when they are given opportunities to think and inquire like chemists in classroom settings that simulate authentic scientific contexts (Carlone and Johnson, 2007; Archer et al., 2010; Aschbacher et al., 2010).
Taken together, students’ science identity should not be understood as the product of a single factor. Rather, it is gradually constructed and developed through lived experiences and social interactions within families, schools, and broader social contexts (Aschbacher et al., 2010). Based on this theoretical framework, it is necessary to situate the formation of science identity within specific disciplinary contexts and to further examine how it is manifested in chemistry learning, as well as how it relates to learning outcomes and career development.
Variables related to chemistry identity
Existing research has shown that chemistry identity is closely related not only to students’ perceptions of their roles within the discipline, but also to their academic performance and future developmental trajectories. Specifically, students with higher levels of chemistry identity tend to demonstrate better chemistry achievement (Ardura and Pérez-Bitrián, 2018) and are more likely to pursue chemistry-related majors or career pathways in the future (Stets et al., 2017). These findings suggest that chemistry identity plays an important mediating and integrative role between individuals’ motivational systems and their career decision-making.Building further on the three developmental pathways of science identity discussed above, the variables examined in this study can be situated within specific theoretical dimensions. Perceived teacher support and parental occupation reflect influences at the levels of belonging and social structure, capturing how the support and resources students obtain in family and school contexts may shape their identity development. Research has shown that perceived teacher support is an important factor associated with students’ chemistry learning and enhancing their chemistry academic performance (Shin and Chang, 2022; Ryan and Deci, 2023; Liu et al., 2025). From the perspective of self-determination theory, teacher support can satisfy students’ basic psychological needs for competence, autonomy, and relatedness, thereby enhancing their learning motivation and learning efficiency (Ryan and Deci, 2017). Positive teacher–student interactions, especially those that are responsive to students’ needs, are linked to more positive chemistry learning experiences (Gan and Peng, 2024). Studies have further shown that such support is associated with students’ learning motivation, cognitive processes, learning behaviors, and emotional experiences, and is also linked to deeper and more active engagement in chemistry learning through stronger emotional investment and more positive learning attitudes (Li et al., 2021; Chen et al., 2023a, 2023b). Ultimately, these supportive experiences may contribute to higher chemistry achievement and help students develop a stronger chemistry identity.
In addition, chemistry self-efficacy corresponds to the pathway concerning the alignment of internal and external attitudinal factors, reflecting how students’ cognitive evaluations of their own chemistry-related capabilities may be linked to the internalization and stabilization of identity. Chemistry self-efficacy generally refers to students’ confidence in their ability to successfully complete chemistry tasks, master chemistry skills, and meet the demands of chemistry coursework (Chang, 2015). Previous studies have shown that self-efficacy influences students’ engagement, effort, and performance, as well as their course selection and future career choices (Bandura, 1986; Lau and Roeser, 2002; Taasoobshirazi and Glynn, 2009; Ramnarain and Ramaila, 2018). Students with high self-efficacy are more likely to complete tasks seriously, persist for longer periods, and invest greater effort, whereas those with low self-efficacy are more likely to respond superficially and give up easily. Research has also indicated that stronger self-efficacy is linked to stronger science identity (Flowers III and Banda, 2016; Cwik and Singh, 2022).
The present study
Against the background of the new college entrance examination reform in Henan Province, the province has adopted the “3 + 1 + 2” subject-selection model for the National College Entrance Examination. Under this model, Chinese, mathematics, and a foreign language are compulsory national examination subjects; students select one subject from physics and history, and two subjects from ideological and political education, geography, chemistry, and biology (Henan Provincial Department of Education, 2022). Because most high schools in China do not offer an integrated science curriculum as an independent subject, students are instead required to select several relatively independent disciplines according to their personal interests and developmental plans. As a result, students’ science learning experiences during high school vary considerably. Differences in the types and combinations of subjects that students study may lead to substantial differences in their science learning experiences and disciplinary participation pathways. In this context, Chinese students’ levels of identity across different science subjects may not be consistent (Vincent-Ruz and Schunn, 2021), which may also create difficulties in their interpretation of and response to items intended to measure a “general science identity” (Chen and Wei, 2022). Therefore, it is necessary to distinguish among subject-specific identities and examine them from the perspective of specific disciplines. Previous studies have shown that disciplinary identity is highly domain-specific. For example, some students may develop a strong identity in chemistry while displaying a relatively weak identity in language-related subjects (Hazari et al., 2020; Guo et al., 2022). On this basis, the present study focuses on chemistry as a specific disciplinary context and examines the current status of chemistry identity among Chinese high school students.At the same time, chemistry identity is inherently multidimensional, with different dimensions jointly shaping whether students see themselves as “chemistry people” (Hazari et al., 2010). This multidimensional structure implies that students may vary across dimensions and, in turn, exhibit different configurations of identity characteristics. For example, some students may show relatively strong performance in understanding chemical knowledge and engaging in chemistry-related practices, while remaining weak in self-recognition or recognition by others, resulting in an overall lower level of chemistry identity (Guo et al., 2022). This suggests that relying solely on variable-centered analyses of average effects may obscure the heterogeneity in how identity dimensions combine within the population (Ferguson et al., 2020). Moreover, person-centered approaches are particularly suitable for exploring the development and cultivation of identity (e.g., science identity), as they can identify multiple meaningful latent groups within a sample, where individuals within these groups exhibit a high degree of similarity in their characteristic patterns (Jung and Wickrama, 2008; Lockhart and Rambo-Hernandez, 2024). Previous person-centered studies have classified science identity among rural high school students and identified an optimal four-profile solution reflecting four traditional identity statuses: achievement, foreclosure, moratorium, and diffusion (Lockhart et al., 2024). Because identity is also shaped by cultural background, researchers in the Chinese context have used cluster analysis to examine the group characteristics of secondary school students’ science identity under a four-dimensional framework including scientific learning competence, interest and performance in science courses, and science career interest, and identified four groups: compliant identity, potential identity, committed identity, and balanced identity (Wang and Yao, 2021).
However, research on the group characteristics of chemistry identity among Chinese high school students remains limited. To address this gap, the present study employs factor mixture modeling from a person-centered perspective to identify latent subgroups of chemistry identity and, on this basis, to provide more targeted implications for differentiated instruction.
H1: high school students exhibit distinct latent profiles of chemistry identity across chemistry learning competence, chemistry classroom interest and performance, chemistry career interest, and chemistry recognition.
H2: perceived teacher support and chemistry self-efficacy are significantly associated with students’ chemistry identity profile membership.
H3: gender, grade level, parental occupation, and class leadership role are significantly associated with students’ chemistry identity profile membership.
H4: chemistry identity profiles differ significantly in chemistry academic achievement and chemistry university aspirations.
Methods
Participants
At the school level, this study used a convenience sampling strategy. Four senior high schools in one city were invited to participate. To ensure that the sample covered students with different levels of prior academic achievement, the participating schools were selected across the local admission-score batches of general senior high schools. In this city, general senior high schools are grouped, from higher to lower admission selectivity, into first-, second-, and third-batch schools according to the cut-off scores used in senior high school admission. The number of questionnaires distributed across the different admission-score batches was kept approximately comparable so as to include students from different academic achievement ranges. Within each participating school, intact classes were randomly selected as the sampling units. The survey was administered only to Grade 10 and Grade 11 students, as Grade 12 students were preparing for the National College Entrance Examination and had a heavy academic workload. The on-site administration and collection of the questionnaires lasted one week, during which a total of 2028 paper questionnaires were distributed. After collection, all paper questionnaires were manually entered into an Excel database by the research team. Prior to formal data analysis, the questionnaires were screened according to predefined exclusion criteria to identify invalid responses. Specifically, a questionnaire was considered invalid and excluded if it met any of the following criteria: first, more than one-third of the items were missing; second, the response pattern contained 12 or more consecutive identical answers. Because the main scale used in this study contained 23 items, and because the long-string index criterion suggests that a string of identical responses exceeding half of the total scale length may indicate careless responding (Huang et al., 2012; Curran, 2016), a cut-off value of 12 was adopted in the present study. Third, questionnaires were also excluded if they showed clearly invalid or uninterpretable response patterns, such as repeatedly selecting the same response option across conceptually unrelated items or providing obviously random or meaningless responses (Sjöström et al., 1999). An Excel macro was written to implement these screening rules and identify invalid responses. After data cleaning, 1431 valid questionnaires were retained for analysis, including 697 boys and 734 girls, yielding a valid response rate of 70.56%.Survey administration
This study was conducted in January 2026 across four schools, and the distribution, collection, and data entry of the paper questionnaires were completed over a two-week period. Ethical approval for the study was obtained from the Academic Ethics Committee of Henan Normal University and other relevant institutions, and informed consent was obtained from the parents or legal guardians of all student participants. Prior to the formal administration of the questionnaire survey, the research team explained to all participants the purpose of the study and the intended use of the collected data. All personal information was kept confidential in strict accordance with relevant national and provincial laws and regulations. Participation was entirely voluntary, and no additional compensation was provided to the students.Instruments
The self-report instruments used in this study were adapted from previously validated questionnaires or published measures. Because the survey was administered to Chinese high school students, the research team first translated the relevant English items into Chinese and, where necessary, adapted the wording to the high school chemistry context. For example, when a construct was measured in the chemistry domain, disciplinary terms such as “science” were replaced with “chemistry.” Cognitive interviews were then conducted to examine whether students understood the translated and context-adapted items as intended. In this study, Tourangeau's four-stage cognitive interviewing method was employed (Tourangeau, 1984; Deng et al., 2021). The purpose of the cognitive interviews was not to redevelop the constructs, but to identify potential problems in students’ comprehension of item wording, disciplinary context, response options, and distinctions between similar items after translation and adaptation. High school students were asked to think aloud while responding to the questionnaire items, so that potential sources of bias in the four cognitive stages of comprehension, information retrieval, judgment, and response could be directly identified.In the first round of cognitive interviews, two items from the chemistry self-efficacy questionnaire – “I am confident that I can do well on chemistry tests” and “I believe I can earn a grade of A in chemistry” – were perceived by most students as having essentially the same meaning. In addition, students expressed uncertainty about how to interpret the grade “A.” In the source item used for translation, the first statement explicitly referred to students’ confidence in doing well on chemistry tests. Therefore, in the present adaptation, we treated it as reflecting students’ self-perceived capability in a test-related chemistry context. By contrast, the fourth item was designed to reflect an objective standard of performance and thus represented an externally referenced evaluation. Based on the interview findings, the first item was revised to: “I am confident that I can perform well on chemistry tests and achieve results that satisfy me.” The fourth item was revised to: “I believe that I can achieve excellent results or rank among the top students in chemistry tests.” In addition, for the fifth item, originally phrased as “I am certain that I can understand chemistry,” students felt that the term “chemistry” alone was too vague and did not refer to specific content or school level. Accordingly, this item was revised to: “I am certain that I can master the knowledge and skills required in high school chemistry courses.” To enhance transparency and replicability, the full list of self-report questionnaire items used in this study, including the Chinese wording administered to students and the corresponding English translations, is provided in Table S3 in the SI.
Statistical analysis
All statistical analyses were conducted using Mplus (Version 8.3) and R (Version 4.3.3). First, descriptive statistics, including means, skewness, kurtosis, and correlation coefficients, were calculated to provide a preliminary examination of data distribution characteristics and relationships among variables. According to Curran et al. (Curran et al., 1996), data can be regarded as approximately normally distributed when the absolute value of skewness is less than 2 and the absolute value of kurtosis is less than 7.Because data were collected through self-report questionnaires from the same respondents, the results might be affected by common method bias (CMB). To assess potential CMB in the data, the unmeasured latent method factor approach was adopted (Podsakoff et al., 2003). Specifically, two confirmatory factor analysis (CFA) models were compared. First, a baseline model (M1) including all core constructs – perceived teacher support, chemistry self-efficacy, and chemistry identity – was specified. Second, on the basis of the baseline model, a global method factor orthogonal to all substantive latent variables was added to construct Model 2 (M2). The severity of common method bias was then evaluated by comparing changes in model fit indices (ΔCFI, ΔTLI, ΔRMSEA, and ΔSRMR) between the two models. Following Wen et al., if the inclusion of the method factor substantially improves model fit (i.e., ΔCFI ≥ 0.10, ΔTLI ≥ 0.10, and ΔRMSEA ≤ −0.05, ΔSRMR ≤ −0.05), serious common method bias is considered to be present (Wen et al., 2018).
To examine the structural validity of the measurement instruments, confirmatory factor analyses were conducted separately for the perceived teacher support, chemistry self-efficacy, and chemistry identity scales. Given that educational measurement data may exhibit non-normality and heteroscedasticity, robust maximum likelihood estimation (MLR) was used to improve the robustness of parameter estimates and standard errors (Kline, 2023). Model fit was considered acceptable when the comparative fit index (CFI) and Tucker–Lewis index (TLI) were at or above 0.90, and the root mean square error of approximation (RMSEA) and standardized root mean square residual (SRMR) were at or below 0.08 (Hu and Bentler, 1999). Because chemistry identity was conceptualized as a multidimensional construct, in addition to testing the overall multidimensional CFA model, separate unidimensionality tests were performed for each subdimension to confirm the basic unidimensionality of the corresponding subscales. Internal consistency reliability was assessed using McDonald's ω, with values above 0.70 indicating good reliability (McDonald, 2013).
Measurement invariance is a prerequisite for meaningful cross-group comparisons (Riordan and Vandenberg, 1994). Therefore, gender-based measurement invariance was tested for the perceived teacher support, chemistry self-efficacy, and chemistry identity scales using R. First, single-group CFA models were estimated for each scale using robust maximum likelihood (MLR) estimation (Muthen and Muthen, 2017), and the model fit indices RMSEA (<0.080), SRMR (<0.090), CFI (>0.900), and TLI (>0.900) were reported. Subsequently, measurement invariance was tested by fitting a series of increasingly restrictive models, including configural invariance, weak factorial invariance (metric invariance), strong factorial invariance (scalar invariance), and strict invariance (residual invariance or invariant uniqueness) (Widaman and Reise, 1997). RMSEA, SRMR, CFI, and TLI were reported for each model. Invariance was supported when the changes in fit indices satisfied ΔCFI ≤ 0.010 and ΔRMSEA ≤ 0.015 (Chen et al., 2008).
Factor mixture modeling was then conducted in Mplus, with Excel used as an auxiliary tool for visualization, to identify latent subgroups of chemistry identity and to explore differences in students’ academic adaptation across these subgroups. Models with an increasing number of latent profiles were estimated step by step, and model comparison and selection were based on fit indices. In determining the optimal number of latent classes, multiple indicators were considered jointly. Entropy and the average latent class probabilities were used to evaluate classification accuracy, with values closer to 1 indicating more reliable classification; both indices should ideally reach at least 0.80 (Jung and Wickrama, 2008). In addition, relative fit indices such as the Akaike information criterion (AIC), Bayesian information criterion (BIC), and adjusted Bayesian information criterion (aBIC) were used to assess model fit, with lower values indicating better fit. The Lo–Mendell–Rubin likelihood ratio test (LMRT) and bootstrap likelihood ratio test (BLRT) were used to compare adjacent class solutions, and significant LMRT and BLRT results indicate that a model with k classes fits better than a model with k − 1 classes. Furthermore, when selecting the number of classes, it is also necessary to ensure that the number of participants in each class is not too small; each class should include at least 1% of the total sample (Jung and Wickrama, 2008). Although relative fit indices served as the primary basis for determining the number of classes, the interpretability and substantive meaning of the classes were also taken into account.
Finally, the one-step approach and the BCH method (Ferguson et al., 2020) were applied, based on the optimal model and the selected number of latent classes, to examine the relationships between latent profile membership and other relevant variables.
Suitability for measurement scales
| Item1 | Item2 | |
|---|---|---|
| Original text |  |
 |
| Translation (in English) | I am confident I will do well on chemistry tests | I believe I can earn a grade of ‘A’ in chemistry |
| Revised text |  |
 |
| Translation (in English) | I am confident that I can perform well on chemistry tests and achieve results that satisfy me | I believe that I can achieve excellent results or rank among the top students in chemistry tests |
| χ2 (df) | p-Value | CFI | TLI | RMSEA [90% CI] | SRMR | Omega | |
|---|---|---|---|---|---|---|---|
| Note: italicized values indicate that they meet the model fit criteria (CFI and TLI ≥ 0.90, RMSEA ≤ 0.08, SRMR ≤ 0.10 and omega ≥ 0.70). | |||||||
| Complete model | |||||||
| Perceived teacher support | 172.024(5) | <0.001 | 0.974 | 0.948 | 0.153 [0.134–0.173] | 0.022 | 0.941 |
| Chemistry self-efficacy | 104.226(5) | <0.001 | 0.984 | 0.968 | 0.118 [0.099–0.138] | 0.017 | 0.936 |
| Chemistry identity four-factor | 1820.328(224) | <0.001 | 0.918 | 0.907 | 0.071 [0.068–0.074] | 0.051 | — |
| Individual factor of the chemistry identity | |||||||
| Learning competence | 319.885(20) | <0.001 | 0.942 | 0.919 | 0.102 [0.093–0.112] | 0.039 | 0.873 |
| Classroom performance and interest | 142.820(14) | <0.001 | 0.956 | 0.934 | 0.080 [0.069–0.092] | 0.034 | 0.826 |
| Career interest | 26.835(2) | <0.001 | 0.989 | 0.967 | 0.093 [0.064–0.126] | 0.019 | 0.842 |
| Recognition | 15.270(2) | <0.001 | 0.997 | 0.992 | 0.068 [0.039–0.102] | 0.008 | 0.933 |
Results
To assess the potential impact of common method bias, the unmeasured latent method factor approach was employed (Podsakoff et al., 2003). The baseline model (M1) demonstrated a good fit to the data (CFI = 0.929, TLI = 0.922, RMSEA = 0.060, SRMR = 0.045). After introducing the latent method factor into the baseline model, the resulting changes in model fit were as follows: ΔCFI = −0.001, ΔTLI = −0.004, ΔRMSEA = 0.001, and ΔSRMR = 0.004. All of these changes were far below the cut off values recommended by Wen et al. for indicating serious common method bias (Wen et al., 2018). In other words, model fit did not improve substantially after the inclusion of the method factor, and the error indices even showed slight increases. These results suggest that common method bias was not a major concern in the present study.Descriptive statistics and correlational analysis
The means, standard deviations, and correlation coefficients of all variables are presented in Table 3. The skewness values ranged from −0.373 to 0.050, and the kurtosis values ranged from −0.173 to 0.264, all of which fell within the acceptable thresholds for normal distribution (Curran et al., 1996). In addition, all variables were significantly and positively correlated with one another. Specifically, teacher support, self-efficacy, and the four dimensions of chemistry identity – chemistry learning competence, chemistry classroom interest and performance, chemistry career interest, and chemistry recognition – were all significantly positively correlated with chemistry academic achievement (Table 3).| Variables | M | SD | Skew | Kurt | 1 | 2 | 3 | 4 | 5 | 6 | 7 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Note: *p < 0.050, **p < 0.010, ***p < 0.001, PTS = perceived teacher support; CSE = chemistry self-efficacy; CLC = chemistry learning competence; CCIP = chemistry classroom interest and performance; CCI = chemistry career interest; CR = chemistry recognition; CAA = chemistry academic achievement. | |||||||||||
| 1. PTS | 3.797 | 0.892 | −0.373 | −0.115 | 1 | ||||||
| 2. CSE | 3.369 | 0.942 | −0.171 | −0.105 | 0.361** | 1 | |||||
| 3. CLC | 3.212 | 0.719 | −0.051 | −0.173 | 0.329** | 0.803** | 1 | ||||
| 4. CCIP | 3.615 | 0.655 | −0.203 | 0.264 | 0.453** | 0.609** | 0.689** | 1 | |||
| 5. CCI | 3.092 | 0.801 | −0.191 | 0.252 | 0.352** | 0.544** | 0.648** | 0.679** | 1 | ||
| 6. CR | 2.666 | 0.899 | 0.05 | −0.022 | 0.286** | 0.593** | 0.718** | 0.547** | 0.683** | 1 | |
| 7. CAA | 0.064* | 0.337** | 0.380** | 0.269** | 0.306** | 0.359** | 1 | ||||
Measurement invariance testing
The results showed that, for the perceived teacher support scale, the configural invariance model (M1), metric invariance model (M2), and scalar invariance model (M3) across gender all demonstrated good model fit. Further comparisons of changes in fit indices between models indicated that the changes in ΔCFI and ΔRMSEA were all within acceptable ranges, suggesting that the scale exhibited a stable measurement structure across gender groups (Putnick and Bornstein, 2016).For the chemistry self-efficacy scale, the cross-group analyses likewise showed a high degree of stability. From the configural invariance model to the metric invariance model, ΔCFI was −0.001 and ΔRMSEA was −0.016; from the metric invariance model to the scalar invariance model, ΔCFI was −0.002 and ΔRMSEA was −0.009. The overall magnitude of change was small and met the criteria for measurement invariance.
Similarly, the chemistry identity scale also demonstrated good measurement consistency across gender groups. From the configural invariance model to the metric invariance model, ΔCFI was −0.001 and ΔRMSEA was −0.001; from the metric invariance model to the scalar invariance model, ΔCFI was −0.005 and ΔRMSEA was −0.001. None of these changes exceeded the recommended cut off values.
Taken together, all three core scales passed the test of measurement invariance across gender. These findings indicate that there was no systematic bias in how male and female students understood and responded to the scale items, and that the scale scores could validly reflect differences in the underlying latent constructs. This provides a reliable measurement basis for subsequent comparative analyses across groups. Detailed results for each questionnaire are presented in the SI (S1. Measurement Invariance Test Results).
Testing H1: latent profiles of chemistry identity
Preliminary latent profile analysis indicated that the profile means across the four dimensions displayed a largely parallel pattern, suggesting that the differences among the candidate profiles were primarily reflected in overall level differences rather than sharply distinct dimensional configurations (Morin et al., 2016). Previous research has pointed out that, in person-centered analyses, when multiple indicators share a strong overall level effect, shape differences among profiles may be masked by level differences, and profiles that are primarily distinguished by level differences may also reflect a higher-order construct underlying the observed dimensions (Morin and Marsh, 2015). Because chemistry identity comprises multiple interrelated dimensions, individual differences among students may simultaneously manifest as shared overall level differences across dimensions and latent subgroup differences (Morin et al., 2016). In such cases, purely categorical models may need to extract additional classes to absorb the unexplained common variation within classes, whereas factor mixture modeling can further capture such within-class variation through continuous latent factors (Clark et al., 2013; Asparouhov and Muthén, 2014). Therefore, the present study further employed factor mixture modeling (FMM). FMM integrates factor analysis and latent class analysis within a single framework, enabling the simultaneous representation of continuous latent variation and categorical latent subgroups (Şen and Cohen, 2024). Moreover, comparing factor analysis, latent class analysis, and FMM is a commonly recommended analytical strategy in latent structure research (Chen et al., 2015). Thus, FMM was used not to demonstrate that students fall into sharply separated empirical categories, but to better represent the possibility that chemistry identity profiles may contain both an overall level component and dimension-specific patterning across the four identity dimensions.Because the log-likelihood function of mixture models may contain multiple local maxima, robust maximum likelihood estimation (MLR) in Mplus was used (Muthen and Muthen, 2017). In addition, a multiple random-start strategy was adopted by increasing the number of random sets of starting values during parameter estimation. Specifically, 500 random starts were specified in the initial stage and 100 final-stage optimizations were used (STARTS = 500 100), in order to ensure replication of the best log-likelihood value and the stability of the results (Asparouhov and Muthén, 2019; Morin et al., 2020). The results showed that the best log-likelihood value was successfully replicated across multiple runs, indicating that the model had converged to a global optimum.
Based on a comprehensive evaluation of multiple model fit indices, the three-profile solution was ultimately selected as the optimal model. Although AIC, BIC, and aBIC continued to decrease as the number of classes increased, the four-profile model did not show a significant improvement over the three-profile model in the LMRT test (p > 0.05), and its smallest class proportion dropped to 10.9%, suggesting possible overextraction. By contrast, the three-profile model not only showed substantially better information criteria than the two-profile model, but also yielded the highest entropy value (0.752). This value indicates acceptable classification quality, although profile membership should still be interpreted with some caution. In addition, the class proportions were relatively balanced, with the smallest class accounting for 11.4% of the sample, which meets recommended standards for class stability in latent profile analysis. Therefore, considering the statistical fit indices, classification quality, and the substantive interpretability of the model, the three-profile solution was identified as the best latent structure of chemistry identity among high school students.
Profile 1 (37.6% of the sample) was characterized by relatively high scores on chemistry learning competence, but markedly low scores on chemistry career interest and chemistry recognition. This pattern suggests that these students possess academic competence in chemistry but lack a clear career vision related to the subject. Accordingly, this profile was labeled the academic competence-career alienation profile.
Profile 2 (51.0% of the sample) represented the largest group in the sample. Its pattern was the most balanced across all dimensions, and it showed the highest score among the three groups on chemistry recognition. This profile reflects a relatively stable sense of belonging to and adaptation within the chemistry domain. It was therefore labeled the robust identity profile.
Profile 3 (11.4% of the sample), although showing relatively lower scores in self-evaluated competence and current chemistry identity, demonstrated very high levels of chemistry classroom interest and performance as well as chemistry career interest. This pattern indicates strong intrinsic motivation and a clear future-oriented career focus. Accordingly, this profile was labeled the interest-career driven profile. To improve readability, the three profiles can be summarized in plain language as follows: Profile 1 reflects relatively strong chemistry learning competence but limited chemistry career interest and chemistry recognition; Profile 2 reflects a relatively balanced and stable chemistry identity pattern; and Profile 3 reflects interest- and career-driven engagement with chemistry, accompanied by comparatively limited recognition.
Overall, these labels are interpretive shorthand used to summarize characteristic response patterns of the latent profiles, rather than directly observed empirical categories. Given the largely parallel pattern observed in the preliminary LPA and the moderate entropy of the selected FMM solution, the three profiles should be interpreted cautiously as probabilistic chemistry identity patterns. These patterns reflect both overall level differences in chemistry identity and relative dimension-specific tendencies, particularly in chemistry career interest and chemistry recognition. In this qualified sense, the results support H1 (Fig. 1 and Table 4).
 | ||
| Fig. 1 Results of the Factor mixture modeling fit for chemistry identity. | ||
| Profile | AIC | BIC | aBIC | Entropy | LMRT | BLRT | Proportions |
|---|---|---|---|---|---|---|---|
| Profile-2 | 75![[thin space (1/6-em)]](https://www.rsc.org/images/entities/char_2009.gif) 340.672 |
75830.292 |
75534.863 |
0.743 | 0.000 | 0.000 | 0.346/0.654 |
| Profile-3 | 74815.165 |
75431.139 |
75059.470 |
0.752 | 0.635 | 0.000 | 0.376/0.510/0.114 |
| Profile-4 | 74356.560 |
75098.887 |
74650.979 |
0.743 | 0.259 | 0.000 | 0.234/0.109/0.472/0.185 |
Testing H2 and H3: variables associated with chemistry identity profile membership
Among the demographic variables, gender was the only significant predictor of profile membership. Relative to Profile 3, male students had much lower odds of belonging to Profile 1 (OR < 0.001) and Profile 2 (OR = 0.040), and were also less likely to belong to Profile 1 than to Profile 2 (OR = 0.004). Substantively, this indicates that male students were most concentrated in Profile 3, whereas female students were more likely to be classified into Profiles 1 and 2, especially Profile 1.Perceived teacher support showed a clear gradient pattern across profiles. For each one-unit increase in teacher support, the odds of belonging to Profile 1 rather than Profile 3 were reduced to 0.151 (an 84.9% reduction), and the odds of belonging to Profile 2 rather than Profile 3 were reduced to 0.342 (a 65.8% reduction). Teacher support also significantly differentiated Profile 1 from Profile 2 (OR = 0.441, a 55.9% reduction). These results indicate a clear ordering, with teacher support highest in Profile 3, intermediate in Profile 2, and lowest in Profile 1.
Chemistry self-efficacy primarily distinguished Profile 3 from the other two profiles. For each one-unit increase in self-efficacy, the odds of belonging to Profile 1 rather than Profile 3 were reduced to 0.008 (a 99.2% reduction), and the odds of belonging to Profile 2 rather than Profile 3 were reduced to 0.084 (a 91.6% reduction). However, self-efficacy did not significantly distinguish Profile 1 from Profile 2. This suggests that high self-efficacy was a defining characteristic of Profile 3, whereas Profiles 1 and 2 were relatively similar on this variable.
By contrast, grade level, parental occupation, and class leadership role were not significant predictors of profile membership, as their confidence intervals all included 1. Thus, chemistry identity profile membership in the present sample appeared to be more strongly related to gender and contextual supports than to grade, parental occupational background, or class leadership status.
Overall, the results supported H2, as perceived teacher support and chemistry self-efficacy were significantly associated with chemistry identity profile membership. H3 was partially supported: gender was significantly associated with profile membership, whereas grade level, parental occupation, and class leadership role were not (Table 5).
| Class1 vs. Class3 | Class2 vs. Class3 | Class1 vs. Class2 | |
|---|---|---|---|
| Note: gender was coded as female = 0 and male = 1. Grade was coded as 10th grade = 1 and 11th grade = 2. Parents' career was a dichotomous variable indicating whether at least one parent was employed in a chemistry- or engineering-related occupation (yes = 0, no = 1). Teacher support and self-efficacy were measured on a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree). Leader was a dichotomous variable indicating whether the student held a class leadership position (yes = 0, no = 1). 95% CI for OR. *p < 0.050, **p < 0.010, ***p < 0.001. | |||
| Variable | B(OR) [lower, upper] | B(OR) [lower, upper] | B(OR) [lower, upper] |
| Gender | −8.70(<0.001) [0.000, 0.000] | −3.21(0.040) ***[0.017, 0.064] | −5.49(0.004) *[0.001, 0.008] |
| Grade | −0.10(0.902) [0.639, 1.274] | −0.02(0.984) [0.729, 1.264] | −0.09(0.917) [0.570, 1.474] |
| Parents' career | 0.08(1.079) [0.598, 1.560] | −0.05(0.948) [0.562, 1.334] | 0.13(1.138) [0.336, 1.941] |
| Teacher support | −1.89(0.151) ***[0.091, 0.250] | −1.07(0.342) ***[0.240, 0.488] | −0.82(0.441) ***[0.327, 0.595] |
| Self-efficacy | −4.83(0.008) ***[0.000, 0.274] | −2.48(0.084) ***[0.051, 0.138] | −2.36(0.094) [0.003, 3.106] |
| Leader | −0.05(0.952) [0.662, 1.241] | −0.03(0.967) [0.649, 1.284] | −0.02(0.984) [0.811, 1.158] |
Testing H4: differences in chemistry academic achievement and chemistry university aspirations across chemistry identity profiles
The BCH analyses revealed highly significant differences among the three latent profiles in both chemistry academic achievement (χ2 = 191.542, p < 0.001) and chemistry university aspirations (χ2 = 224.850, p < 0.001). Moreover, both outcome variables showed a consistent stepwise distribution across the three profiles. With regard to chemistry academic achievement, Profile 3 (interest-career driven profile) had the highest score (M = 0.528), followed by Profile 2 (robust identity profile) (M = 0.083), whereas Profile 1 (academic competence-career alienation profile) had the lowest score (M = −0.505). Pairwise comparisons indicated that all differences between profiles were statistically significant. A similar descending pattern was observed for chemistry university aspirations, with scores decreasing from Profile 3 (M = 3.388) to Profile 2 (M = 3.104) and then to Profile 1 (M = 2.281), and all pairwise differences were statistically significant. Notably, the university aspiration score of Profile 1 was clearly below the scale midpoint, indicating a tendency toward alienation from chemistry-related careers. By contrast, Profile 3 showed the most favorable pattern on both academic achievement and university aspirations, indicating that the interest-career driven profile was associated with higher chemistry academic achievement and stronger intentions to pursue chemistry university aspirations.These results supported H4, indicating that chemistry identity profiles differed significantly in both chemistry academic achievement and chemistry university aspirations (Table 6).
| Profile | Chemistry academic achievement M (SE) | Chemistry university aspirations M (SE) |
|---|---|---|
| Note: *p < 0.050, **p < 0.010, ***p < 0.001. | ||
| Academic competence-career alienation profile | −0.505 (0.052) | 2.281 (0.053) |
| Robust identity profile | 0.083 (0.035) | 3.104 (0.034) |
| Interest-career driven profile | 0.528 (0.054) | 3.388 (0.063) |
| χ2(df) | 191.542***(2) | 224.850***(2) |
| Pairwise comparisons | 1 < 2, 1 < 3, 2 < 3 | 1 < 2, 1 < 3, 2 < 3 |
Discussion
The present study suggests that chemistry identity among Chinese high school students is characterized by heterogeneity that is partly level-based and partly configurational. Students exhibited different response patterns across the dimensions of chemistry learning competence, chemistry classroom performance and interest, chemistry career interest, and chemistry recognition. However, these profiles should not be interpreted as sharply discrete student types. Rather, consistent with the methodological distinction between level and shape effects in person-centered analyses, the profile solution appears to reflect a substantial overall chemistry identity component together with several dimension-specific tendencies. Moreover, the three chemistry identity profiles differed significantly in both chemistry academic achievement and chemistry university aspirations, indicating that chemistry identity is closely associated not only with students' current academic performance but also with their future developmental orientations. These findings further suggest that examining chemistry identity solely at the level of overall mean scores may be insufficient to capture the complex differences among students in the process of identity development. This point is particularly relevant given that students' science career interest and science attitudes may decline as they grow older (Barmby et al., 2008; Potvin and Hasni, 2014; Tröbst et al., 2016), which underscores the practical significance of investigating the mechanisms underlying such heterogeneity. Accordingly, the present study adopted a person-centered perspective to identify distinct chemistry identity profiles among Chinese high school students and further examined their associations with learning-related variables. This approach not only enriches the existing research on chemistry identity in the Chinese educational context but also provides new evidence for understanding patterns of chemistry identity development across different cultural settings.Profiles of chemistry identity among Chinese high school students
The factor mixture analysis in the present study suggested three probabilistic profile patterns of chemistry identity among Chinese high school students. These patterns should be understood as summaries of latent response tendencies rather than as fixed or naturally occurring categories. Importantly, the profiles were not entirely independent of overall chemistry identity level; instead, they appeared to combine level-based differences with relative differences across specific dimensions. This finding is consistent with the general conclusion of existing person-centered identity research, namely that students' disciplinary identity is not a single continuum but rather comprises different latent subgroup structures (Lockhart et al., 2024). Notably, chemistry recognition scores were consistently the lowest among the four dimensions across all three profiles. Previous research has suggested that identity possesses a distinctly social-constructive nature: its formation involves not only individuals' self-understanding and meaning-making but also depends on sociocultural participation and recognition from others through interaction (Le et al., 2019). Accordingly, compared with dimensions such as competence and interest, which can be developed through individual autonomous learning, the formation of chemistry recognition places greater demands on the external feedback environment. This may be an important reason for the consistently lower scores on this dimension across all three profiles. This cross-profile commonality suggests that, regardless of students' levels of competence and interest, creating more opportunities for students to receive chemistry recognition may be a shared priority for current chemistry education.Profile 1, the academic competence-career alienation profile, accounted for 37.66% of the total sample. As shown in the profile plot, students in this group maintained relatively high levels of chemistry learning competence and chemistry classroom interest and performance, with both scores around 3.6, but showed markedly lower levels of chemistry career interest and chemistry recognition. Compared with traditional perspectives that explain students’ science-related career choices primarily in terms of academic achievement or ability, science identity theory emphasizes that strong learning competence and good classroom performance do not necessarily translate into stronger disciplinary identity or career orientation. A well-developed science identity also requires the joint development of dimensions such as recognition and sense of belonging (Lee and Mun, 2023). Therefore, the “high competence–low recognition” separation pattern observed in Profile 1 provides support for this theoretical perspective based on a sample of Chinese high school students. It also indicates that, in the present sample, chemistry learning competence and chemistry recognition did not necessarily show parallel patterns.
Existing research also supports this finding. Studies have shown that even students who perform well in chemistry learning may not necessarily view themselves as “chemistry people” (Castano et al., 2025). Similarly, research in physics education has found that achievement or ability alone is insufficient to fully explain whether students develop a strong disciplinary identity and career orientation (Hazari et al., 2010). In addition, Profile 1 was also characterized by relatively high classroom interest but low career interest, suggesting that positive classroom-level experiences do not necessarily correspond to a higher level of future career orientation. Previous research has likewise indicated that although students’ science competence and perceived classroom value are related to university aspirations, the relationship is not a simple one-to-one correspondence. Compared with students who merely believe that they “can do well in science,” those who also perceive personal meaning and future utility in science learning are more likely to report stronger STEM career interest (Aschbacher et al., 2010). This is consistent with the pattern observed in Profile 1: these students scored relatively high on chemistry learning competence and chemistry classroom interest and performance, but did not show a corresponding level of chemistry career interest.
Overall, Profile 1 suggests that these students’ strengths were mainly reflected in their current academic performance and classroom participation, but these strengths did not develop in parallel with corresponding levels of career interest and chemistry recognition. Previous studies suggest that future-oriented career interest and its connection with identity may also be related to perceived value, science-related experiences, contextual resources, and social support (Archer et al., 2012; Jones et al., 2021). However, because the present study did not directly examine these factors, they are discussed here only as contextual references for interpreting the findings, rather than as direct explanations for the formation of profile differences. The fact that this profile represented a substantial proportion of the sample, 37.6%, suggests that improving academic achievement alone may be insufficient to promote the development of chemistry identity among these students. Educators may also need to attend to how students can establish meaningful connections between chemistry learning and their personal development.
Students in Profile 2 displayed moderate levels across all four dimensions – chemistry learning competence, chemistry classroom performance and interest, chemistry career interest, and chemistry recognition – exhibiting a relatively balanced overall structure but lacking prominent strengths or distinctive features. Drawing on identity development research, this pattern can be understood as a relatively stable but not yet fully developed identity state (Meeus et al., 1999), in which students maintained a basic level of engagement with and recognition of chemistry learning but had not yet developed a stronger career orientation or a more clearly defined sense of disciplinary belonging. This profile accounted for the largest proportion of the sample (51.0%), suggesting that for a substantial portion of Chinese high school students, chemistry represents a “stable but not prominent” learning domain. Similarly, previous person-centered studies have also reported large intermediate identity subgroups. For example, in a science identity study based on the exploration–commitment framework, Lockhart et al. identified a large Foreclosed class (94/156, approximately 60.3%) in which the Z-scores on both dimensions were close to zero (Exploration = −0.035, Commitment = −0.027), reflecting a relatively intermediate positioning (Lockhart et al., 2024). However, because that study was based on a two-dimensional identity framework, its results cannot be directly mapped onto the four-dimensional profiles identified in the present study. Nevertheless, the fact that both studies identified large intermediate subgroups suggests, to some extent, that a stable but not prominent identity state may be a relatively common pattern in disciplinary identity development.
Profile 3, the interest–career driven profile, accounted for 11.4% of the total sample, representing the smallest proportion among the three profiles. In terms of the profile pattern, students in this group scored the highest on chemistry classroom performance and interest, while also maintaining a relatively high level of chemistry career interest. Their chemistry learning competence was at a moderately high level but was not the highest among the three profiles, and their chemistry recognition did not reach the highest level either. Previous research has demonstrated that interest and future-related utility value play important roles in promoting students' sustained engagement in disciplinary learning, enhancing academic performance, and strengthening career orientation (Hidi and Renninger, 2006; Hulleman and Harackiewicz, 2009; Hulleman et al., 2010). From this perspective, the elevated levels of classroom interest and career interest in Profile 3 constitute the key psychological characteristics that distinguish this group from the other profiles. At the same time, the relatively small proportion of this profile is noteworthy. This proportion indicates that, in the present sample, students who simultaneously exhibited high levels of both classroom interest and career interest were relatively few. This finding can be situated within existing research suggesting that classroom-based science interest and science career interest do not share a straightforward correspondence; the link between the two may be influenced by factors such as career awareness and supportive learning experiences (Archer et al., 2012; DeWitt and Archer, 2015; Kang et al., 2023). However, the present study did not directly examine these mechanisms, and this interpretation therefore requires further investigation in future research. Accordingly, Profile 3 can be understood as a student group characterized by relatively well-developed interest and a clearer future orientation; yet its small proportion also suggests that this favorable combination was not prevalent in the present sample. This finding also resonates with the emphasis in identity research on the recognition dimension – namely, that interest, competence, or performance do not necessarily translate into a correspondingly high level of identity (Potvin and Hazari, 2014), which helps explain why Profile 3, despite having high classroom interest and career interest, did not show the highest level of chemistry recognition.
The influence of antecedent variables on students’ membership in chemistry identity profiles
The findings revealed that gender, perceived teacher support, and chemistry self-efficacy significantly predicted students' membership in the chemistry identity latent profiles. The following discussion addresses the significant effects of these three variables, drawing on theoretical analysis and existing empirical research.Perceived teacher support
Perceived teacher support significantly predicted students' chemistry identity profile membership and exhibited a gradient pattern: higher levels of perceived teacher support were associated with a greater likelihood of belonging to Profile 3 and Profile 2, and a lower likelihood of belonging to Profile 1. In light of the profile characteristics, higher perceived teacher support was more likely to correspond to profiles in which classroom interest and career interest were more prominent, rather than to a uniform increase across all identity dimensions. In other words, differences in perceived teacher support effectively differentiated students' profile membership. Previous research has shown that student–teacher positioning in classroom interactions (Hazari et al., 2015), as well as teacher feedback and teacher recognition, are closely associated with students' science identity development (Kim, 2018; Chiu, 2024). Extending these findings, the present study, through latent profile analysis, further revealed that perceived teacher support was not only associated with students' chemistry identity but also differentiated students' likelihood of belonging to different identity profiles. Specifically, teacher support was related not only to the overall level of chemistry identity but also to the way students' scores were configured across the dimensions of chemistry learning competence, chemistry classroom performance and interest, chemistry career interest, and chemistry recognition. This finding provides more nuanced person-centered evidence for understanding the relationship between teacher support and chemistry identity, suggesting that the role of teacher support is reflected not only in elevating overall identity levels but also in differentiating membership across distinct identity profiles.Students' grade level, class leadership position, and parental occupation type did not significantly predict chemistry identity profile membership. First, regarding grade level, this finding is consistent with the conclusion of Vincent-Ruz and Schunn, who, after conducting measurement invariance testing and group-based analyses, reported that the mean scores on science identity did not differ significantly between seventh- and ninth-grade students, and emphasized the comparability of science learning arrangements across the two grade levels (Vincent-Ruz and Schunn, 2018). From a person-centered perspective, the present study further corroborates this conclusion, indicating that grade level was not only unrelated to the overall level of chemistry identity but also unable to effectively differentiate students' membership across different chemistry identity profiles. According to Carlone and Johnson's science identity framework (Carlone and Johnson, 2007), the formation of chemistry identity depends on individuals' practical engagement in specific learning contexts and recognition from the learning community, rather than being directly determined by the natural progression of grade level. However, the present study did not directly examine the specific relationship between grade level and classroom experiences, and this interpretation therefore awaits further verification in future research. Second, regarding class leadership position, identity theory suggests that identities from different sources need to be verified and reinforced within their corresponding contexts (Stets and Burke, 2000). The experience of serving as a class leader primarily strengthens general role identities such as “leader” or “responsibility bearer,” and its influence tends to operate through general pathways such as learning autonomy, learning engagement, and time allocation. Previous research has also indicated that its effect on specific academic outcomes is relatively limited (Deng et al., 2020). The present finding is consistent with this view, further suggesting that the general role identity conferred by class leadership may not effectively differentiate students' profile membership across the dimensions of chemistry identity. Finally, regarding parental occupation, this finding diverges from some previous research. Prior studies have shown that students with science university aspirations tend to come from families with science-supportive environments, and that parents, as important sources of science knowledge, have a significant influence on students' science interest and career choices (Dabney et al., 2013; Rüschenpöhler and Markic, 2020; Stahl et al., 2021). However, the present results suggest that having at least one parent employed in a chemistry- or engineering-related occupation does not necessarily equate to the presence of a chemistry-supportive environment at home. One possible explanation is that, in the context of the increasingly fast pace of contemporary Chinese society, opportunities for family members to interact and communicate around science-related topics may be relatively limited, making it difficult for parents’ chemistry-or engineering-related occupational backgrounds to be effectively translated into a direct influence on students' chemistry identity (Chen et al., 2023). From a person-centered perspective, the present study further indicates that parental occupation was not only unrelated to the overall level of students' chemistry identity but also unable to effectively differentiate profile membership.
Relationships between latent profiles and outcome variables
The BCH analyses showed significant differences among the three chemistry identity profiles in chemistry academic achievement, following the pattern Profile 3 > Profile 2 > Profile 1. Notably, Profile 1 had the highest self-rated score on the chemistry learning competence dimension among the three profiles, yet showed the lowest chemistry academic achievement. By contrast, Profile 3 had a relatively lower self-rated score on the chemistry learning competence dimension, but showed the highest chemistry academic achievement. This pattern indicates that students’ self-rated chemistry learning competence does not directly correspond to their chemistry academic achievement.This finding is consistent with previous discussions of the relative contributions of cognitive ability and non-cognitive factors to academic achievement. Prior research has suggested that only approximately 25% of the variance in students’ academic achievement can be attributed to cognitive ability, whereas as much as 75% may stem from non-cognitive factors (Kuncel et al., 2004). As a core component of non-cognitive factors, motivation may predict academic achievement even more strongly than cognitive ability itself (Meyer et al., 2019). Moreover, even after controlling for prior achievement, motivation remains a key predictor of subsequent academic achievement (Steinmayr and Spinath, 2009).
In relation to the profile characteristics identified in the present study, Profile 3 had the highest scores among the three profiles on both chemistry classroom interest and performance and chemistry career interest. As interest is an important component of learning motivation (Hidi, 2006), it may have played an important role in the higher chemistry academic achievement observed in this profile. By contrast, although Profile 1 reported relatively high self-rated chemistry learning competence, it scored lower on interest-related dimensions, which corresponded with its lower chemistry academic achievement. From a person-centered perspective, the present study further shows that high self-rated competence and high academic achievement do not necessarily co-occur within the same profile, a pattern that may be difficult to reveal through previous variable-centered research. However, given the cross-sectional design of the present study, the causal relationship between interest and academic achievement requires further examination in future longitudinal research.
In addition to chemistry academic achievement, the BCH analyses also showed significant differences among the three profiles in chemistry university aspirations, with Profile 3, the interest-career driven profile, showing the highest level and Profile 1, the academic competence-career alienation profile, showing the lowest level. It should be noted that chemistry university aspirations are conceptually related to, but not identical with, the chemistry career interest dimension of chemistry identity. Chemistry career interest is one component of the multidimensional chemistry identity scale and was used to identify students’ latent profiles; it mainly reflects students’ interest and value orientation toward future chemistry-related development within their identity structure. By contrast, chemistry university aspirations were measured as an outcome variable and refer specifically to students’ self-reported likelihood of choosing chemistry or a chemistry-related university major after high school. Therefore, the higher level of chemistry university aspirations observed in Profile 3 should be interpreted in light of its higher chemistry career interest, rather than as an outcome entirely independent of the profile-defining indicators.
Lockhart et al. similarly found that science identity classifications were significantly associated with students’ science career interest (Lockhart et al., 2024). Notably, Profile 3 showed the highest chemistry university aspirations but the lowest score on chemistry recognition among the three profiles. This suggests that chemistry university aspirations are not simply determined by the overall level of chemistry identity, but are more closely related to the specific configuration of students’ scores across different identity dimensions, especially the prominent levels of chemistry classroom interest and performance and chemistry career interest. From a person-centered perspective, the present study reveals this differentiated pattern, which may be difficult to capture through previous variable-centered research.
Implication for practice
The practical implications of the present study should be understood as exploratory and interpretive rather than prescriptive. Because the identified profiles are probabilistic latent patterns rather than directly observed student types, they should not be used as a fixed basis for assigning students to instructional groups. First, the present study found that scores on the chemistry recognition dimension were relatively low across all three profiles, a trend that was particularly pronounced in Profiles 1 and 3. It should be noted that, as discussed in the preceding section, differences in item phrasing and response anchors across dimensions may affect the direct comparability of mean scores, and therefore this result should not be simply interpreted as indicating that chemistry recognition is necessarily the weakest aspect of students' identity development; rather, it should be regarded as a trend worthy of further attention. With this caveat in mind, chemistry recognition, as a core component of chemistry identity, nonetheless suggests that teachers should attend to students' experiences of both “self-recognition” and “recognition from others” in chemistry learning (Wang and Yao, 2021).Second, the present study found that Profile 3 (the interest-career driven profile) showed the highest levels of both chemistry academic achievement and university aspirations, whereas Profile 1 (the academic competence-career alienation profile), despite having relatively high self-rated learning competence, showed the lowest levels of university aspirations and academic achievement. This contrast suggests that relying solely on strengthening knowledge mastery may be insufficient to help students develop a stable identity. Guiding students to actively reflect on the connections between chemistry learning and their own future development – that is, helping students perceive the utility value of learning chemistry (Fryer and Ainley, 2019; Walton and Crum, 2020) – may be more critical. In the context of China's New Gaokao reform, previous research has indicated that students' interest in chemistry learning is the primary factor influencing their decision to select and continue studying chemistry (Qiang and Yan, 2021). Teachers can create authentic scientific contexts that connect chemistry knowledge with real-world applications and strengthen chemistry career education (Qiang and Yan, 2021), helping students construct meaningful connections among “learning content, real-world applications, and future planning,” which may help students perceive the value of learning chemistry more clearly (McWhirter et al., 2000).
Third, the present study found that gender significantly predicted profile membership, with female students being more likely to be classified into Profiles 1 and 2 rather than Profile 3. As key agents in students' gender socialization processes (Rayaprol et al., 2023), teachers should be mindful of the influence of gender stereotypes on instructional expectations (Perander et al., 2020) and provide all students with equal opportunities to “think and act like scientists” through inquiry-based learning approaches (Zhai et al., 2014), thereby reducing the impact of gender differences on chemistry identity development.
Limitations and future directions
Although the present study adopted a person-centered approach to reveal the latent heterogeneity in high school students' chemistry identity and further examined the roles of relevant antecedent and outcome variables, several limitations remain that warrant attention in future research.First, the present study employed a cross-sectional research design in which all variables were measured at a single time point. Consequently, the results primarily reflect associations among variables rather than supporting stronger causal inferences. Chemistry identity is inherently developmental and contextual; its formation and change typically constitute a dynamic process in which students gradually construct their identity through sustained classroom participation, social interactions, and disciplinary learning experiences. Although the present study identified distinct latent profiles and found that perceived teacher support and chemistry self-efficacy were significantly associated with profile membership, bidirectional effects or stage-specific changes may exist among these variables and chemistry identity that cannot be fully captured through cross-sectional data. Future research could employ longitudinal designs to further examine the developmental trajectories of students' chemistry identity across different grade levels, as well as the sustained mechanisms through which teacher support, self-efficacy, and other variables are related to the identity development process.
Second, the interpretation of the latent profiles should also be treated with caution. Although the three-profile FMM solution was selected based on statistical fit, classification quality, and substantive interpretability, the preliminary LPA showed a largely parallel pattern across the four chemistry identity dimensions. This indicates that the profiles were influenced by an overall level component of chemistry identity. Therefore, the profiles identified in this study should not be interpreted as sharply separated latent classes or fixed empirical student types. Rather, they should be understood as probabilistic chemistry identity patterns that combine overall level differences with dimension-specific tendencies. Future research could further examine the stability and validity of these patterns using longitudinal data, independent samples, and alternative modeling approaches such as bifactor or bifactor-ESEM models to more directly evaluate the global and specific components of chemistry identity. In addition, because the latent profiles identified in the present study are based on cross-sectional data, their operationalizability in instructional practice requires further verification. Although the present study reveals differentiated characteristics across chemistry identity profiles, these findings should be regarded as providing directions for future research rather than as direct evidence for immediate instructional intervention. Future studies could further investigate whether utility-value interventions can support chemistry identity development among students who show low chemistry career interest but relatively high academic competence, and which classroom feedback strategies may be effective in enhancing students’ chemistry recognition experiences. Answers to these questions will require further examination through longitudinal tracking studies and instructional intervention experiments.
Third, the present study has certain limitations regarding sample selection. Participants were primarily drawn from 10th- and 11th-grade students at four schools of varying academic levels within a single city, using convenience sampling with class-level administration. More specifically, the sample was not intended to be statistically representative of all Chinese high school students, but rather to provide coverage of students from different school levels within the local context. 12th-grade students were not included due to the practical constraints of academic pressure they face. As a result, the present findings should be interpreted as applying primarily to 10th- and 11th-grade students in the sampled city and should not be generalized uncritically to students in other regions, school systems, or grade levels across China. Although this approach was operationally feasible and facilitated coverage of students from schools of different levels, the geographic scope of the sample was relatively limited and did not encompass 12th-grade students. In particular, students' chemistry learning experiences and identity construction processes may differ across regions, school types, and contexts characterized by varying levels of examination pressure in China. In addition, because Grade 12 students are at a distinctive stage of examination preparation, subject-selection decision-making, and future planning, their chemistry identity patterns and related motivational characteristics may differ from those observed in Grades 10 and 11. Future research could conduct comparative studies across multiple regions and school types on a larger scale and include 12th-grade students in the analysis to enhance the representativeness and generalizability of the findings.
Fourth, the present study has potential for further expansion in its examination of antecedent variables. Grounded in relevant theoretical frameworks of science identity and informed by key antecedent variables that have been repeatedly validated in prior research, the present study selected variables with strong theoretical foundations, representativeness, and explanatory power – including perceived teacher support, parental occupation, and chemistry self-efficacy – from both the external sociocultural context and internal psychological mechanism levels. These variables reflect, to a certain extent, the critical sources of influence on students' chemistry identity formation, including school support, family background, and individual competence beliefs, and thus possess a solid theoretical basis and practical significance. However, the development of chemistry identity is not determined by a single factor or a limited number of factors; rather, it is a dynamic construction process shaped by the joint operation of multiple factors across family, school, and broader sociocultural contexts. Accordingly, the current model has not fully incorporated other potentially important factors, such as peer support, family science interactions, classroom instructional approaches, perceived disciplinary value, sense of belonging, opportunities for scientific practice, and the degree of alignment between school science and authentic science. Future research could, within a more comprehensive theoretical framework, incorporate additional contextual and psychological variables relevant to chemistry identity formation and investigate potential mediation, moderation, and sequential mechanisms among different factors, thereby providing a more holistic explanation of the developmental patterns of Chinese high school students' chemistry identity.
Fifth, the present study also has a limitation in its measurement of chemistry university aspirations. In this study, chemistry university aspirations were assessed using a single self-report item. Although this approach was practical and relevant in the context of a large school-based survey, it is less robust than a multi-item scale, provides less comprehensive coverage of the construct, and does not allow internal consistency to be evaluated in the same way. Therefore, findings related to chemistry university aspirations should be interpreted with caution. Future research could employ multi-item measures to assess this construct more comprehensively and further validate the present findings.
Conclusions
Grounded in a four-dimensional framework of science identity, the present study employed factor mixture modeling to examine patterns of high school students' chemistry identity: the academic competence-career alienation profile, the robust identity profile, and the interest-career driven profile, suggesting meaningful heterogeneity in high school students' chemistry identity patterns. Further analysis revealed that gender, perceived teacher support, and chemistry self-efficacy significantly predicted students' profile membership, and significant differences were observed across profiles in both chemistry academic achievement and chemistry university aspirations, with the “interest-career-driven” profile exhibiting the most favorable outcomes. These findings indicate that teacher support and chemistry self-efficacy are important variables associated with students’ chemistry identity patterns, and that interest and career orientation play critical roles in students' chemistry learning development. The present study provides empirical evidence for understanding the differentiated characteristics of Chinese high school students' chemistry identity and offers exploratory implications for supporting students' chemistry identity development and career awareness education in high school chemistry teaching.Conflicts of interest
There are no conflicts to declare.Data availability
Due to ethical considerations and privacy protection requirements, the raw data supporting the conclusions of this manuscript will not be made publicly available. These data contain sensitive information, and public sharing may compromise the privacy of the research participants. However, the anonymized data supporting the main findings of this study are available from the corresponding author, Yurong Liu, upon reasonable request. Researchers who wish to access the data will be required to sign a data access agreement specifying how the data may be used and ensuring that participant confidentiality is maintained. For further information regarding the data and conditions of access, please contact liuyur66@163.com.The supplementary information (SI) contains the measurement invariance test results for the perceived teacher support, chemistry self-efficacy, and chemistry identity scales, standardized factor loadings for the questionnaires, and the Chinese questionnaire items with corresponding English translations. See DOI: https://doi.org/10.1039/d6rp00114a.
Acknowledgements
The paper was supported by the project “Model Construction and Practice of a School-Locality ‘U-G-S-T-P’ Professional Learning Community for Promoting High-Quality Teacher Development” (Project No. ZXSC26006), under the Teacher Education Collaborative Quality Improvement Program (South China Normal University Cluster), which is supervised by the Department of Teacher Education, Ministry of Education of China, funded by the Jack Ma Foundation, and established by the China Teacher Development Foundation.References
- Archer L., DeWitt J., Osborne J., Dillon J., Willis B., and Wong B., (2010), “Doing” science versus “being” a scientist: Examining 10/11-year-old schoolchildren's constructions of science through the lens of identity, Sci. Educ., 94(4), 617–639 DOI:10.1002/sce.20399.
- Archer L., DeWitt J., Osborne J., Dillon J., Willis B. and Wong B., (2012), Science Aspirations, Capital, and Family Habitus: How Families Shape Children's Engagement and Identification With Science, Am. Educ. Res. J., 49(5), 881–908 DOI:10.3102/0002831211433290.
- Ardura D. and Pérez-Bitrián A., (2018), The effect of motivation on the choice of chemistry in secondary schools: adaptation and validation of the Science Motivation Questionnaire II to Spanish students, Chem. Educ. Res. Pract., 19(3), 905–918 10.1039/C8RP00098K.
- Aschbacher P. R., Li E. and Roth E. J., (2010), Is science me? High school students’ identities, participation and aspirations in science, engineering, and medicine, J. Res. Sci. Teach., 47(5), 564–582 DOI:10.1002/tea.20353.
- Asparouhov T. and Muthén B., (2014), Auxiliary Variables in Mixture Modeling: Three-Step Approaches Using Mplus, Struct. Eq. Model.: Multidiscipl. J., 21(3), 329–341 DOI:10.1080/10705511.2014.915181.
- Asparouhov T. and Muthén B., (2019), Random Starting Values and Multistage Optimization.
- Bandura A., (1986), The Explanatory and Predictive Scope of Self-Efficacy Theory, J. Soc. Clin. Psychol., 4(3), 359–373 DOI:10.1521/jscp.1986.4.3.359.
- Barbera J., (2020), Alignment of theoretically grounded constructs for the measurement of science and chemistry identity. Chem. Educ. Res. Pract., 21(1), 371–386 10.1039/C9RP00193J.
- Barmby P., Kind P. M., and Jones K., (2008), Examining Changing Attitudes in Secondary School Science, Int. J. Sci. Educ., 30(8), 1075–1093 DOI:10.1080/09500690701344966.
- Betz N. E. and Hackett G., (1981), The relationship of career-related self-efficacy expectations to perceived career options in college women and men, J. Counsel. Psychol., 28(5), 399–410 DOI:10.1037/0022-0167.28.5.399.
- Bishop J. P., (2012), “She's Always Been the Smart One. I’ve Always Been the Dumb One”: Identities in the Mathematics Classroom, J. Res. Math. Educ., 43(1), 34–74 DOI:10.5951/jresematheduc.43.1.0034.
- Carlone H. B., (2017), Disciplinary Identity as Analytic Construct and Design Goal: Making Learning Sciences Matter, J. Learn. Sci., 26(3), 525–531 DOI:10.1080/10508406.2017.1336026.
- Carlone H. B. and Johnson A., (2007), Understanding the science experiences of successful women of color: Science identity as an analytic lens, J. Res. Sci. Teach., 44(8), 1187–1218 DOI:10.1002/tea.20237.
- Castano G., Dou R., Linh N., Mohammed N., Lopez A. and Underwood S. M., (2025), Identity Development in Chemistry: The Social Functionality and Moral Significance of Being (Considered) a “Real” Chemist, JACS Au, 5(9), 4408–4426 DOI:10.1021/jacsau.5c00769.
- Chang B. L., (2015), Effects of racialized tracking on racial gaps in science self-efficacy, identity, engagement, and aspirations: Connection to science and school segregation.
- Chen J. A. and Usher E. L., (2013), Profiles of the sources of science self-efficacy, Learn. Individual Differ., 24, 11–21 DOI:10.1016/j.lindif.2012.11.002.
- Chen S. and Wei B., (2022), Development and Validation of an Instrument to Measure High School Students’ Science Identity in Science Learning, Res. Sci. Educ., 52(1), 111–126 DOI:10.1007/s11165-020-09932-y.
- Chen F., Curran P. J., Bollen K. A., Kirby J., and Paxton P., (2008), An Empirical Evaluation of the Use of Fixed Cutoff Points in RMSEA Test Statistic in Structural Equation Models, Soc. Methods Res., 36(4), 462–494 DOI:10.1177/0049124108314720.
- Chen Y., Wen Z. and Gu H., (2015), Factor mixture modeling: an integration of latent class analysis and factor analysis, Adv. Psychol. Sci., 23(3), 529–538.
- Chen J., Lin C.-H., and Chen G., (2023a), Adolescents’ self-regulated and affective learning, teacher support and digital reading literacy: A multilevel latent profile approach, Comput. Educ., 205, 104883 DOI:10.1016/j.compedu.2023.104883.
- Chen S., Wei B. and Zhang H., (2023b), Exploring high school students’ disciplinary science identities and their differences, Int. J. Sci. Math. Educ., 21(2), 377–394 DOI:10.1007/s10763-022-10257-7.
- Chiu T. K. F., (2024), Using self-determination theory (SDT) to explain student STEM interest and identity development, Instr. Sci., 52(1), 89–107 DOI:10.1007/s11251-023-09642-8.
- Clark S., Muthén B., Kaprio J., D’Onofrio B., Viken R. and Rose R., (2013), Models and Strategies for Factor Mixture Analysis: An Example Concerning the Structure Underlying Psychological Disorders, Struct. Equ. Model.: Multidisc. J., 20 DOI:10.1080/10705511.2013.824786.
- Cribbs J., Huang X. and Piatek-Jimenez K., (2021), Relations of mathematics mindset, mathematics anxiety, mathematics identity, and mathematics self-efficacy to STEM career choice: A structural equation modeling approach, Sch. Sci. Math., 121(5), 275–287 DOI:10.1111/ssm.12470.
- Curran P. G., (2016), Methods for the detection of carelessly invalid responses in survey data, J. Exp. Soc. Psychol., 66, 4–19 DOI:10.1016/j.jesp.2015.07.006.
- Curran P. J., West S. G., and Finch J. F., (1996), The robustness of test statistics to nonnormality and specification error in confirmatory factor analysis, Psychol. Methods, 1(1), 16–29 DOI:10.1037/1082-989X.1.1.16.
- Cwik S. and Singh C., (2022), Self-efficacy and perceived recognition by peers, instructors, and teaching assistants in physics predict bioscience majors’ science identity, PLoS One, 17(9), e0273621 DOI:10.1371/journal.pone.0273621.
- Dabney K. P., Chakraverty D., and Tai R. H., (2013), The Association of Family Influence and Initial Interest in Science, Sci. Educ., 97(3), 395–409 DOI:10.1002/sce.21060.
- Deng W., Li X., Wu H., and Xu G., (2020), Student leadership and academic performance, China Econ. Rev., 60, 101389 DOI:10.1016/j.chieco.2019.101389.
- Deng J. M., Streja N., and Flynn A. B., (2021), Response Process Validity Evidence in Chemistry Education Research, J. Chem. Educ., 98(12), 3656–3666 DOI:10.1021/acs.jchemed.1c00749.
- DeWitt J. and Archer L., (2015), Who Aspires to a Science Career? A comparison of survey responses from primary and secondary school students, Int. J. Sci. Educ., 37(13), 2170–2192 DOI:10.1080/09500693.2015.1071899.
- Donnay D. A. C. and Borgen F. H., (1999), The incremental validity of vocational self-efficacy: An examination of interest, self-efficacy, and occupation, J. Counsel. Psychol., 46(4), 432–447 DOI:10.1037/0022-0167.46.4.432.
- Ferguson S. L., Moore E. W. G., and Hull D. M., (2020), Finding latent groups in observed data: A primer on latent profile analysis in Mplus for applied researchers, Int. J. Behav. Dev., 44(5), 458–468 DOI:10.1177/0165025419881721.
- Flowers III A. M. and Banda R., (2016), Cultivating science identity through sources of self-efficacy, J. Multicult. Educ., 10(3), 405–417 DOI:10.1108/JME-01-2016-0014.
- Fryer L. K. and Ainley M., (2019), Supporting interest in a study domain: A longitudinal test of the interplay between interest, utility-value, and competence beliefs, Learn. Instruct., 60, 252–262 DOI:10.1016/j.learninstruc.2017.11.002.
- Gan Y. and Peng J., (2024), Effects of teacher support on math engagement among Chinese college students: A mediated moderation model of math self-efficacy and intrinsic value, Child. Youth Serv. Rev., 156, 107369 DOI:10.1016/j.childyouth.2023.107369.
- Guo X., Deng W., Hu K., Lei W., Xiang S. and Hu W., (2022), The effect of metacognition on students’ chemistry identity: the chain mediating role of chemistry learning burnout and chemistry learning flow, Chem. Educ. Res. Pract., 23(2), 408–421 10.1039/D1RP00342A.
- Hazari Z., Sonnert G., Sadler P. M. and Shanahan M., (2010), Connecting high school physics experiences, outcome expectations, physics identity, and physics career choice: A gender study, J. Res. Sci. Teach., 47(8), 978–1003 DOI:10.1002/tea.20363.
- Hazari Z., Cass C. and Beattie C., (2015), Obscuring power structures in the physics classroom: Linking teacher positioning, student engagement, and physics identity development: TEACHER POSITIONING AND STUDENT PHYSICS IDENTITY, J. Res. Sci. Teach., 52(6), 735–762 DOI:10.1002/tea.21214.
- Hazari Z., Chari D., Potvin G. and Brewe E., (2020), The context dependence of physics identity: Examining the role of performance/competence, recognition, interest, and sense of belonging for lower and upper female physics undergraduates, J. Res. Sci. Teach., 57(10), 1583–1607 DOI:10.1002/tea.21644.
- Henan Provincial Department of Education, (2022), Henan's comprehensive reform plan for the college entrance examination released: The “3 + 1 + 2” model will be implemented and the division between arts and sciences will be abolished, Henan Provincial People's Government.
- Hidi S., (2006), Interest: A unique motivational variable, Educ. Res. Rev., 1(2), 69–82 DOI:10.1016/j.edurev.2006.09.001.
- Hidi S. and Renninger K. A., (2006), The Four-Phase Model of Interest Development, Educ. Psychol., 41(2), 111–127 DOI:10.1207/s15326985ep4102_4.
- Hosbein K. N. and Barbera J., (2020), Development and evaluation of novel science and chemistry identity measures, Chem. Educ. Res. Pract., 21(3), 852–877 10.1039/C9RP00223E.
- Hu L. and Bentler P. M., (1999), Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives, Struct. Eq. Model.: Multidisc. J., 6(1), 1–55 DOI:10.1080/10705519909540118.
- Huang J. L., Curran P. G., Keeney J., Poposki E. M. and DeShon R. P., (2012), Detecting and Deterring Insufficient Effort Responding to Surveys, J Bus Psychol, 27(1), 99–114 DOI:10.1007/s10869-011-9231-8.
- Huangfu Q., Wei N., Zhang R., Tang Y. and Luo G., (2023), Social support and continuing motivation in chemistry: the mediating roles of interest in chemistry and chemistry self-efficacy, Chem. Educ. Res. Pract., 24(2), 478–493 10.1039/D2RP00165A.
- Hulleman C. S., Godes O., Hendricks B. L. and Harackiewicz J. M., (2010), Enhancing interest and performance with a utility value intervention, J. Educ. Psychol., 102(4), 880–895 DOI:10.1037/a0019506.
- Hulleman C. S. and Harackiewicz J. M., (2009), Promoting Interest and Performance in High School Science Classes, Science, 326(5958), 1410–1412 DOI:10.1126/science.1177067.
- Jiang Z. and Wei B., (2025), Understanding Science Identity Development Among College Students, Sci. Educ., 34(3), 1797–1824 DOI:10.1007/s11191-023-00478-9.
- Jones M. G., Chesnutt K., Ennes M., Mulvey K. L. and Cayton E., (2021), Understanding science career aspirations: Factors predicting future science task value, J. Res. Sci. Teach., 58(7), 937–955 DOI:10.1002/tea.21687.
- Jung T. and Wickrama K. A. S., (2008), An Introduction to Latent Class Growth Analysis and Growth Mixture Modeling, Soc. Person. Psych., 2(1), 302–317 DOI:10.1111/j.1751-9004.2007.00054.x.
- Kang J., Salonen A., Tolppanen S., Scheersoi A., Hense J., Rannikmäe M., et al., (2023), Effect of Embedded Careers Education in Science Lessons on Students’ Interest, Awareness, and Aspirations, Int. J. Sci. Math. Educ., 21(1), 211–231 DOI:10.1007/s10763-021-10238-2.
- Kenny D. A., Kaniskan B. and McCoach D. B., (2015), The performance of RMSEA in models with small degrees of freedom, Soc. Methods Res., 44(3), 486–507 DOI:10.1177/0049124114543236.
- Kim M., (2018), Understanding children's science identity through classroom interactions, Int. J. Sci. Educ., 40(1), 24–45 DOI:10.1080/09500693.2017.1395925.
- Kim A. Y. and Sinatra G. M., (2018), Science identity development: an interactionist approach, IJ STEM Educ., 5(1), 51 DOI:10.1186/s40594-018-0149-9.
- Kline R. B., (2023), Principles and Practice of Structural Equation Modeling, Guilford Publications.
- Kuncel N. R., Hezlett S. A. and Ones D. S., (2004), Academic Performance, Career Potential, Creativity, and Job Performance: Can One Construct Predict Them All? J. Person. Soc. Psychol., 86(1), 148–161 DOI:10.1037/0022-3514.86.1.148.
- Lau S. and Roeser R. W., (2002), Cognitive Abilities and Motivational Processes in High School Students’ Situational Engagement and Achievement in Science, Educ. Assess., 8(2), 139–162 DOI:10.1207/S15326977EA0802_04.
- Lave J. and Wenger E., (1991), Situated Learning: Legitimate Peripheral Participation, Cambridge University Press.
- Le P. T., Doughty L., Thompson A. N. and Hartley L. M., (2019), Investigating Undergraduate Biology Students’ Science Identity Production, LSE, 18(4), ar50 DOI:10.1187/cbe.18-10-0204.
- Lee G.-G. and Mun S., (2023), From science motivation to science identity: The mediating effect of science achievement according to gender, EURASIA J. Math. Sci. Tech. Ed., 19(10), em2341 DOI:10.29333/ejmste/13633.
- Lent R. W., Brown S. D. and Larkin K. C., (1987), Comparison of three theoretically derived variables in predicting career and academic behavior: Self-efficacy, interest congruence, and consequence thinking, J. Counsel. Psychol., 34(3), 293–298 DOI:10.1037/0022-0167.34.3.293.
- Li G., (2025), Through the lens of the Guidelines for Science Education Work: Reflections on pathways to science identity among primary and secondary school students, Hubei Soc. Sci., (8), 87–95 DOI:10.13660/j.cnki.42-1112/c.016601.
- Li H., Zhang A., Zhang M., Huang B., Zhao X., Gao J. and Si J., (2021), Concurrent and longitudinal associations between parental educational involvement, teacher support, and math anxiety: The role of math learning involvement in elementary school children, Contemp. Educ. Psychol., 66, 101984 DOI:10.1016/j.cedpsych.2021.101984.
- Liu Y., Niu M. and Sun H., (2025), A Pattern Analysis of Chinese High School Students’ Chemistry Achievement Goals Based on Latent Profile Analysis, J. Chem. Educ. DOI:10.1021/acs.jchemed.5c00735.
- Lockhart M. E. and Rambo-Hernandez K., (2024), Investigating engineering identity development and stability amongst first-year engineering students: a person-centred approach, Eur. J. Eng. Educ., 49(3), 411–433 DOI:10.1080/03043797.2023.2262412.
- Lockhart M. E., Kwok O.-M. and Yoon M., (2024), Investigating science identity classifications of rural high school students: a person-centered approach, Int. J. Sci. Educ., 46(16), 1671–1689 DOI:10.1080/09500693.2023.2293681.
- Maltese A. V. and Tai R. H., (2010), Eyeballs in the Fridge: Sources of early interest in science. Int. J. Sci. Educ., 32(5), 669–685 DOI:10.1080/09500690902792385.
- McDonald R. P., (2013), Test theory: a unified treatment, Routledge, Taylor & Francis Group.
- McWhirter E. H., Crothers M. and Rasheed S., (2000), The effects of high school career education on social–cognitive variables, J. Counsel. Psychol., 47(3), 330–341 DOI:10.1037/0022-0167.47.3.330.
- Meeus W., Iedema J., Helsen M. and Vollebergh W., (1999), Patterns of Adolescent Identity Development: Review of Literature and Longitudinal Analysis, Dev. Rev., 19(4), 419–461 DOI:10.1006/drev.1999.0483.
- Meyer J., Fleckenstein J. and Köller O., (2019), Expectancy value interactions and academic achievement: Differential relationships with achievement measures, Contemp. Educ. Psychol., 58, 58–74 DOI:10.1016/j.cedpsych.2019.01.006.
- Ministry of Education of the People's Republic of China, (2023), Opinions on Strengthening Science Education in Primary and Secondary Schools in the New Era.
- Moon S., Guan S.-S. A., Vargas J. H., Lin J. C. P., Kwan P., Saetermoe C. L., et al., (2025), Critical Mentorship in Undergraduate Research Experience BUILDs Science Identity and Self-Efficacy. Int. J. Sci. Math. Educ., 23(2), 321–341 DOI:10.1007/s10763-024-10476-0.
- Morin A. J. S. and Marsh H. W., (2015), Disentangling Shape from Level Effects in Person-Centered Analyses: An Illustration Based on University Teachers’ Multidimensional Profiles of Effectiveness, Struct. Eq. Model.: Multidisc. J., 22(1), 39–59 DOI:10.1080/10705511.2014.919825.
- Morin A. J. S., Boudrias J.-S., Marsh H. W., Madore I. and Desrumaux P., (2016), Further Reflections on Disentangling Shape and Level Effects in Person-Centered Analyses: An Illustration Exploring the Dimensionality of Psychological Health, Struct. Eq.Model.: Multidisc. J., 23(3), 438–454 DOI:10.1080/10705511.2015.1116077.
- Morin A. J. S., McLarnon M. J. W. and Litalien D., (2020), Mixture modeling for organizational behavior research, in Handbook on the Temporal Dynamics of Organizational Behavior, Griep Y. and Hansen S. D. (eds.), Edward Elgar Publishing.
- Muthen L. K. and Muthen B., (2017), Mplus Version 8 User's Guide, Muthen & Muthen.
- OECD, (2023), Agency in the Anthropocene: Supporting document to the PISA 2025 Science Framework.
- Oettingen G. and Gollwitzer P. M., (2015), Self-Regulation in Adolescence, Cambridge University Press.
- Perander K., Londen M. and Holm G., (2020), Anxious girls and laid-back boys: teachers’ and study counsellors’ gendered perceptions of students, Cambridge J. Educ., 50(2), 185–199 DOI:10.1080/0305764X.2019.1653825.
- Pfeifer M. A., Zajic C. J., Isaacs J. M., Erickson O. A. and Dolan E. L., (2024), Beyond performance, competence, and recognition: forging a science researcher identity in the context of research training, IJ STEM Educ., 11(1), 19 DOI:10.1186/s40594-024-00479-2.
- Podsakoff P. M., MacKenzie S. B., Lee J.-Y. and Podsakoff N. P., (2003), Common method biases in behavioral research: A critical review of the literature and recommended remedies, J. Appl. Psychol., 88(5), 879–903 DOI:10.1037/0021-9010.88.5.879.
- Potvin P. and Hasni A., (2014), Analysis of the Decline in Interest Towards School Science and Technology from Grades 5 Through 11, J. Sci. Educ. Technol., 23(6), 784–802 DOI:10.1007/s10956-014-9512-x.
- Potvin G. and Hazari Z., (2014), The Development and Measurement of Identity across the Physical Sciences, in 2013 Physics Education Research Conference Proceedings, Engelhardt P. V., Churukian A. D. and Jones D. L. (eds.), American Association of Physics Teachers, pp. 281–284.
- Putnick D. L. and Bornstein M. H., (2016), Measurement invariance conventions and reporting: The state of the art and future directions for psychological research, Dev. Rev., 41, 71–90 DOI:10.1016/j.dr.2016.06.004.
- Qiang W. and Yan Y., (2021), A Research on the Factors of Chemistry Subject Selection against the Background of New College Entrance Examination: Based on the Survey after the Implementation of Chongqing's “3 + 1 + 2” Program, J. Teach. Educ., 8(5), 70–77 DOI:10.13718/j.cnki.jsjy.2021.05.010.
- Ramnarain U. and Ramaila S., (2018), The relationship between chemistry self-efficacy of South African first year university students and their academic performance, Chem. Educ. Res. Pract., 19(1), 60–67 10.1039/C7RP00110J.
- Rayaprol A., Raman U., Farooqui S. A. and Joshi A., (2023), Gender Sensitisation for School Teachers: Agents of Change in the Classroom, Sociol. Bull., 72(4), 534–545 DOI:10.1177/00380229231196769.
- Riordan C. M. and Vandenberg R. J., (1994), A central question in cross-cultural research: Do employees of different cultures interpret work-related measures in an equivalent manner? J. Manage., 20(3), 643–671 DOI:10.1177/014920639402000307.
- Rüschenpöhler L. and Markic S., (2020), Secondary school students’ acquisition of science capital in the field of chemistry. Chem. Educ. Res. Pract., 21(1), 220–236 10.1039/C9RP00127A.
- Ryan R. M. and Deci E. L., (2017), Self-Determination Theory: Basic Psychological Needs in Motivation, Development, and Wellness, Guilford Press.
- Ryan R. M. and Deci E. L., (2023), Self-Determination Theory, in Encyclopedia of Quality of Life and Well-Being Research, Springer, pp. 6229–6235.
- Sadler P. M., Sonnert G., Hazari Z. and Tai R., (2012), Stability and volatility of STEM career interest in high school: A gender study, Sci. Educ., 96(3), 411–427 DOI:10.1002/sce.21007.
- Şen S. and Cohen A., (2024), A Systematic Review of Factor Mixture Model Applications, JMEEP, 15(2), 79–93 DOI:10.21031/epod.1423427.
- Shin H. and Chang Y., (2022), Relational support from teachers and peers matters: Links with different profiles of relational support and academic engagement, J. Sch. Psychol., 92, 209–226 DOI:10.1016/j.jsp.2022.03.006.
- Sjöström O., Holst D. and Lind S. O., (1999), Validity of a questionnaire survey: the role of non-response and incorrect answers, Acta Odontol. Scand., 57(5), 242–246 DOI:10.1080/000163599428643.
- Stahl G., Scholes L., McDonald S. and Lunn J., (2021), Middle years students’ engagement with science in rural and urban communities in Australia: exploring science capital, place-based knowledges and familial relationships, Pedag., Cult. Soc., 29(1), 43–60 DOI:10.1080/14681366.2019.1684351.
- Steegh A., Höffler T., Höft L. and Parchmann I., (2021), Exploring science competition participants’ expectancy-value perceptions and identification: A latent profile analysis, Learn. Instruct., 74, 101455 DOI:10.1016/j.learninstruc.2021.101455.
- Steinmayr R. and Spinath B., (2009), The importance of motivation as a predictor of school achievement, Learn. Individ. Differ., 19(1), 80–90 DOI:10.1016/j.lindif.2008.05.004.
- Stets J. E. and Burke P. J., (2000), Identity Theory and Social Identity Theory, Soc. Psychol. Q., 63(3), 224 DOI:10.2307/2695870.
- Stets J. E., Brenner P. S., Burke P. J. and Serpe R. T., (2017), The science identity and entering a science occupation, Soc. Sci. Res., 64, 1–14 DOI:10.1016/j.ssresearch.2016.10.016.
- Taasoobshirazi G. and Glynn S. M., (2009), College students solving chemistry problems: A theoretical model of expertise, J. Res. Sci. Teach., 46(10), 1070–1089 DOI:10.1002/tea.20301.
- Tai R. H., Liu C. Q., Maltese A. V. and Fan X., (2006), Planning Early for Careers in Science, Science, 312(5777), 1143–1144 DOI:10.1126/science.1128690.
- Tourangeau R., (1984), Cognitive sciences and survey methods, in Cognitive aspects of survey methodology: Building a bridge between disciplines, Jabine T. B., Straf M. L., Tanur J. M. and Tourangeau R. (eds.), National Academy Press, pp. 73–100.
- Tröbst S., Kleickmann T., Lange-Schubert K., Rothkopf A. and Möller K., (2016), Instruction and Students’ Declining Interest in Science: An Analysis of German Fourth- and Sixth-Grade Classrooms, Am. Educ. Res. J., 53(1), 162–193 DOI:10.3102/0002831215618662.
- Trujillo G. and Tanner K. D., (2014), Considering the Role of Affect in Learning: Monitoring Students’ Self-Efficacy, Sense of Belonging, and Science Identity, LSE, 13(1), 6–15 DOI:10.1187/cbe.13-12-0241.
- Vincent-Ruz P. and Schunn C. D., (2018), The nature of science identity and its role as the driver of student choices, IJ STEM Educ., 5(1), 48 DOI:10.1186/s40594-018-0140-5.
- Vincent-Ruz P. and Schunn C. D., (2021), Identity Complexes and Science Identity in Early Secondary: Mono-topical or in Combination with Other Topical Identities, Res. Sci. Educ., 51(S1), 369–390 DOI:10.1007/s11165-019-09882-0.
- Walton G. M. and Crum A. J., (2020), Handbook of Wise Interventions, Guilford Publications.
- Wang Q. and Yao B., (2021), Current Status and Group Characteristics of Secondary School Students’ Science Identity, Educ. Meas. Eval., (9), 38–47 DOI:10.16518/j.cnki.emae.2021.09.005.
- Wang N., Tan A.-L., Zhou X., Liu K., Zeng F., and Xiang J., (2023), Gender differences in high school students’ interest in STEM careers: a multi-group comparison based on structural equation model, IJ STEM Educ., 10(1), 59 DOI:10.1186/s40594-023-00443-6.
- Wen Z., Huang B. and Tang D., (2018), Preliminary Work for Modeling Questionnaire Data, J. Psychol. Sci., 41(1), 204–210 DOI:10.16719/j.cnki.1671-6981.20180130.
- Wenger E., (1998), Communities of practice: Learning, meaning, and identity, Cambridge University Press.
- Widaman K. F. and Reise S. P., (1997), Exploring the measurement invariance of psychological instruments: Applications in the substance use domain, in The science of prevention: Methodological advances from alcohol and substance abuse research, Bryant K. J., Windle M. and West S. G. (eds.), American Psychological Association, pp. 281–324 DOI:10.1037/10222-009.
- Zhai J., Jocz J. A. and Tan A.-L., (2014), ‘Am I Like a Scientist?’: Primary children's images of doing science in school, Int. J. Sci. Educ., 36(4), 553–576 DOI:10.1080/09500693.2013.791958.
- Zhang J. and Zhou Q., (2022), Chinese chemistry motivation questionnaire II: adaptation and validation of the science motivation questionnaire II in high school students, Chem. Educ. Res. Pract., 24(1), 369–383 10.1039/D2RP00243D.
- Zhao M., Ozturk E., Law F., Joy A., Deutsch A. R., Marlow C. S., et al., (2024), Reciprocal Associations Between Science Efficacy, STEM Identity and Scientist Career Interest Among Adolescent Girls within the Context of Informal Science Learning, J. Youth Adolesc., 53(2), 472–484 DOI:10.1007/s10964-023-01868-6.
- Zimmerman B. J. and Kitsantas A., (1997), Developmental Phases in Self-Regulation: Shifting From Process Goals to Outcome Goals, J. Educ. Psychol., 89(1), 29–36 DOI:10.1037/0022-0663.89.1.29.
| This journal is © The Royal Society of Chemistry 2026 |