




































 

 

 

 

 

Analysis of the Factors Influencing Jun-

ior High School Students’  

Academic Performance and the Con-

struction of a Prediction Model

Shunzhi Yan, Wenhui Li 

 
Shenyang Normal University, Shenyang 110136, Liaoning Province, China 

 

Abstract: Academic performance is a critical factor in determining 

the future academic path and the possibility of social class mobility 
of junior high school students. Exploring the influencing factors of 

academic performance and predicting them can help provide em-
pirical evidence for a comprehensive and objective evaluation of 

students’ academic performance. Using CEPS 2014- 2015 data, we 

analyze the influencing factors of junior high school students’ per-
formance based on the I-E-O model, use a nomogram to predict the 

likelihood of students’ achievement of excellent results, and visual-
ize and analyze the important factors. It was found that, in terms of 

psychological characteristics, students’ and their parents’ educa-

tional expectations had a much greater impact on academic per-
formance. Additionally, students’ motivation and persistence signif-

icantly influenced their academic performance. In terms of demo-

graphic characteristics, we found notable differences between boys 
and girls, with girls outperforming boys academically. In terms of 

student involvement, time spent on school homework and in cram 
schools has a significant impact on academic performance, but the 

impact of attending cram schools is negative. Finally, in terms of 

school characteristics, the level of school per student expenditure 
had a significant effect on achievement. 

Best Evidence in Chinese Education 2025; 21(2):2009-2037. 

DOI: 10.15354/bece.25.or026 

How to Cite: Yan, S. Z., & Li, W. H. (2025). Analysis of the factors influencing 

junior high school students’ academic performance and the construction of a 

prediction model. Best Evidence in Chinese Education, 2025, 21(2): 2009-

2037. 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2010 

Keywords: Junior High School Students, Academic Performance, Prediction Model, 

Educational Expectation 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

About the Authors: Shunzhi Yan, Shenyang Normal University, Shenyang 110136, Liaoning Province, China. E-

mail: xzsa13@126.com 

Wenhui Li, Shenyang Normal University, Shenyang 110136, Liaoning Province, China. E-mail: 
llwwhh1984@synu.edu.cn 

Correspondence to: Dr. Wenhui Li at Shenyang Normal University in China. 

Conflict of Interests: None 

Funding: No funding sources declared. 

AI Declaration: The author affirms that artificial intelligence did not contribute to the process of preparing the 

work. 

 
 

© 2025 Insights Publisher. All rights reserved. 

Creative Commons NonCommercial CC BY-NC: This article is distributed under the terms of the Crea-
tive Commons Attribution-NonCommercial 4.0 License (http://www.creativecommons.org/licenses/by-

nc/4.0/) which permits non-commercial use, reproduction and distribution of the work without further permission provided 

the original work is attributed by the Insights Publisher. 

mailto:xzsa13@126.com
mailto:llwwhh1984@synu.edu.cn
http://www.creativecommons.org/licenses/by-nc/4.0/
http://www.creativecommons.org/licenses/by-nc/4.0/


Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2011 

Introduction 

CADEMIC performance is a key indicator of students’ knowledge 

acquisition and learning outcomes, reflecting both individual pro-

gress and the effectiveness of educational systems (Huang et al., 

2016). Variations in academic performance among students from different 

backgrounds often reveal disparities in educational resource distribution and 

structural inequities, reflecting regional differences in education quality and 

access. Junior high school marks a critical stage in compulsory education, 

shaping adolescents’ self-identity and social development, while also deter-

mining future educational and career paths through the streaming mechanism 

based on high school entrance exam results. For junior high school students, 

academic performance is essential, serving as a core measure of individual 

developmental potential and educational efficacy. High academic perfor-

mance fosters confidence and motivation, encouraging deeper engagement 

with learning, which contributes to long-term educational success, career 

opportunities, and the development of high-quality human capital vital for 

societal sustainability. 

Literature Review 

Internal student factors can directly affect academic performance, and educa-

tional expectations are one of the important factors. Students’ educational 

expectations, encompassing both short-term (e.g., exam performance) and 

long-term academic goals, directly influence academic achievement through 

bidirectional relationships with performance outcomes (Pinquart & Ebeling, 

2020a; Pinquart & Ebeling, 2020b). Empirical evidence confirms significant 

positive correlations between students’ self-educational expectations and ac-

ademic performance, with cross-lagged analyses demonstrating reciprocal 

causation (Sanders et al., 2001; Pinquart & Ebeling, 2020b). Closely related 

to students’ self-educational expectations, parents’ educational expectations 

are also one of the most important factors affecting students’ performance. It 

substantially predicts academic performance through dual mechanisms: ac-

tive involvement in educational activities and psychological transmission of 

competence beliefs, which students internalize as personal behavioral stand-

ards (Yamamoto & Holloway, 2010). In addition to educational expectations, 

students’ learning motivation is an important influence on academic perfor-

mance. While both educational expectations and learning motivation influ-

ence academic behaviors, they constitute distinct psychological effects. Edu-

cational expectations reflect future-oriented beliefs about educational at-

tainment, whereas learning motivation drives goal-directed actions through 

cognitive activation (Yamamoto & Holloway, 2010; Moreira-Morales & 

García-Loor, 2024). Specifically, learning motivation comprises intrinsic 

A 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2012 

(task-inherent satisfaction) and extrinsic (external reward-driven) dimensions, 

both significantly predicting academic performance through behavioral en-

gagement (Entwistle et al., 1974). According to Hull’s formula, academic 

performance derives from the multiplicative interaction of motivational drive 

and study habits (Eysenck, 1957), with educators frequently attributing aca-

demic failures to deficiencies in these components (Entwistle et al., 1974). A 

SEM study showed that both internal and external motivation of students are 

significantly and positively correlated with academic performance and are 

important influences on academic performance (Nauzeer & Jaunky, 2019). 

In addition to internal student factors, school characteristics, as key 

external factors, significantly influence learning outcomes. Researchers have 

examined various aspects of the school environment, including climate, loca-

tion, facilities, division, and classroom conditions, to assess their impact on 

academic performance. Physically, classroom environment, location, and 

facilities remarkably affect student achievement (Iwuagwu, 2016); moreover, 

researchers have found that schools with more greenery show higher exam 

success rates, whereas barren landscapes negatively affect performance 

(Kweon et al., 2017). Psychologically, a positive and supportive school cli-

mate enhances academic performance, while a negative climate harms stu-

dents’ well-being and achievement (Podiya et al., 2025). Positive social rela-

tionships and a sense of belonging at school increase academic self-efficacy 

as well, which in turn improves performance (Zysberg & Schwabsky, 2020). 

The factors affecting academic performance certainly do not stop 

there, and a review of existing research on students’ academic performance 

reveals that there is a wealth of research in this area, covering a wide range 

of topics. Domestic studies have focused on individual student traits, school 

characteristics, and teacher attributes, offering insights into current perfor-

mance trends and their determinants. However, these studies often analyze 

influences through a single dimension, such as extracurricular tutoring (Guan, 

2022) or family-school cooperation (Zhang er al., 2020), after controlling for 

demographics, limiting findings to specific causal factors, and failing to re-

flect the complexity and diversity of academic performance. Additionally, 

most studies focus on university students, with insufficient attention to junior 

high school populations. 

In 1977, Astin proposed the Input-Environment-Outcome (I-E-O) 

model to analyze how higher education influences student development 

(Astin, 1977). The model categorizes educational processes into three com-

ponents: Input (students’ pre-enrollment background and social experiences), 

Environment (educational experiences and institutional characteristics), and 

Output (changes in students’ cognitive, skills, and value domains) (Zhang et 

al., 2020). It posits that student learning and competence development result 

from the interaction between pre-enrollment inputs and environmental fac-

tors during education. In 1984, Astin expanded the model by introducing the 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2013 

theory of Student Involvement, a key aspect of Environment, which 

measures the physical and mental energy students invest in their education, 

reflected in time spent, participation in activities, and engagement with fac-

ulty and peers (Astin, 1984). Studies have found that greater student in-

volvement positively affects academic performance (Bao & Zhang, 2012). 

Although the I-E-O model was developed for higher education, its frame-

work applies broadly, as both higher and basic education involve similar de-

velopmental processes shaped by Input, Environment, and Outcome. There-

fore, drawing on the I-E-O model and using national survey data, this study 

examines ninth-grade junior high school students to better understand the 

factors influencing academic performance and to expand research on this 

population. 

All along, the “scores-only theory” has led schools and parents to 

overemphasize written examination results, neglecting process evaluation 

and overall quality assessment. This manner could cause student disengage-

ment and psychological stress, both closely linked to academic performance. 

Effectively evaluating academic performance and identifying its key influ-

encing factors have thus become critical issues. Using data from the China 

Education Panel Survey (CEPS), this study examines junior high school stu-

dents’ academic performance from input and output perspectives and devel-

ops a predictive model to provide empirical insights for improvement. 

Methods 

Data Source 

The data utilized in this study are derived from the China Education Panel 

Survey (CEPS). The CEPS is a large-scale, nationally representative longi-

tudinal survey designed and implemented by a top-tier university in China 

with an outstanding academic reputation. This survey aims to analyze how 

factors such as family, school, community, and broader social structures col-

lectively influence individual educational outcomes. It also seeks to deeply 

uncover the role and mechanisms of these educational outcomes throughout 

an individual’s life course, thereby providing a reliable empirical basis for 

the formulation of educational public policy and for academic research in 

related fields. 

To ensure the national representativeness of the sample, CEPS em-

ployed a multi-stage probability proportional to size (PPS) sampling design. 

Using the population’s average education level and the proportion of the mi-

grant population as stratification variables, this method first randomly select-

ed 28 county-level units nationwide as survey sites. Subsequently, 112 

schools were randomly drawn from these selected county-level units. Finally, 

438 classes were sampled from these schools, and all students in the chosen 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2014 

classes were included in the sample. The baseline survey, conducted in the 

2013-2014 academic year, comprised approximately 20,000 students. 

CEPS employs questionnaire surveys as its primary data collection 

method, targeting a wide range of respondents including students, parents, 

homeroom teachers, main subject teachers, and school administrators. The 

content of the questionnaires is tailored to each respondent group: the stu-

dent questionnaire focuses on basic information, academic growth, physical 

and mental health, and social-behavioral development; the parent question-

naire centers on the home educational environment, perspectives on school 

education, and educational expectations for their children; the teacher ques-

tionnaire concentrates on educational philosophies, daily teaching practices, 

work-related stress, and job satisfaction; and the school administrator ques-

tionnaire is primarily concerned with the school’s basic information, educa-

tional mission, teaching facilities, and daily instructional management. 

CEPS is a thirty-year longitudinal survey, with its baseline survey 

launched in the 2013-2014 academic year. It began tracking two cohorts of 

students, who were then in the seventh and ninth grades. These students were 

surveyed annually during their middle school years, and follow-ups are 

planned at multiple subsequent time points after their graduation, with the 

final survey scheduled for completion in 2043. This study uses data from the 

first follow-up wave conducted in the 2014-2015 academic year. This wave 

targeted the baseline cohort of 10,279 seventh-grade students, of whom 

9,449 were successfully re-interviewed, resulting in a high follow-up rate of 

91.9%. Because CEPS has not released new data since 2015, this wave con-

stitutes the most recent dataset available. For this research, variables were 

selected from the student, parent, and school administrator questionnaires. 

After data cleaning, a total of 5,998 valid cases were obtained. Concurrently, 

in strict compliance with ethical and legal standards, CEPS ensures the con-

fidentiality of all identifiable information, such as county (district), school, 

and administrative codes, to effectively protect respondent privacy. 

Description of the Variables 

Dependent Variable 

The dependent variable in this study is student academic performance. The 

data for this variable consist of student midterm examination scores from the 

fall 2014 semester, which were provided directly by the surveyed schools.  



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2015 

 
Figure 1. Indicator System Based on the IEO Model. 

 

 

 

These are administrative data, not self-reported variables. Specifically, the 

data include the raw scores for three subjects—Chinese, Mathematics, and 

English—and the corresponding maximum possible score for each subject. 

(The CEPS survey exclusively covers academic performance in these three 

core subjects). 

As Chinese, Mathematics, and English are core subjects in middle 

school (and notably, these three subjects have the same maximum score in 

the high school entrance examination in most provinces), they were assigned 

equal weight to comprehensively reflect student academic performance. The 

raw scores from these three subjects were directly summed to create a raw 

total score. To eliminate inter-school differences, this study then conducted a 

within-school standardization of the students’ total scores. On a school-by-

school basis, the arithmetic mean of all students’ total scores was calculated, 

with this school-level mean serving as a benchmark for each school’s aca-

demic level. Finally, based on the individual student’s score and their 

school’s mean score, a core binary variable was generated to represent the 

student’s performance level relative to their in-school peer group. If a stu-

dent’s total score was higher than the school mean, the variable was assigned 

a value of 1; conversely, if the score was equal to or lower than the mean, it 

was assigned a value of 0. 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2016 

Independent Variables 

The independent variables are the influences related to students’ academic 

performance, with a total of 14 variables divided into two dimensions, Input 

and Environment, and four sub-dimensions. See Figure 1 for details. 

Input Dimension 

Input variables refer to students’ pre-institutional characteristics and are di-

vided into two sub-dimensions: personal background and psychological 

characteristics. Personal background includes Gender, Ethnicity, Hukou type, 

and Students’ self-rated health. Psychological characteristics include stu-

dents’ self-educational expectations, parental educational expectations, 

learning motivation, and learning persistence. 

Gender is a binary nominal variable. The data were sourced from the 

‘Personal Background’ module of the student questionnaire and represent a 

single-item measure self-reported by students during the 2013-2014 baseline 

survey. As gender is a stable attribute, it did not require repeated measure-

ment; therefore, this information was not collected again in the 2014-2015 

follow-up survey. The complete gender information was obtained by merg-

ing the baseline data with the follow-up data using the unique student identi-

fication code (variable name: ids). The original coding for this variable, 

“1=male, 2=female,” was used without modification. 

Data on ethnicity were also sourced from the ‘Personal Background’ 

module of the student questionnaire, representing a single-item measure self-

reported by students in the 2013-2014 baseline survey. Due to its stability, 

this question was not repeated in the follow-up survey; the ethnicity infor-

mation for students was obtained by merging the datasets. The original vari-

able was a multi-category nominal variable with eight categories (1=Han, 

2=Mongol, 3=Manchu, 4=Hui, 5=Tibetan, 6=Zhuang, 7=Uygur, and 

8=Other). To avoid the potential interference of small-sample categories in 

the statistical analysis, this study transformed it. The original eight categories 

were consolidated into a binary nominal variable, with the coding scheme 

0=Ethnic minority (including Mongol, Manchu, and other groups) and 

1=Han. 

The student’s hukou (household registration) type is a single-item 

measure self-reported by students in the 2013-2014 baseline survey. This 

information was obtained by merging the baseline and follow-up data using 

the student ID. The original variable was a four-category nominal variable 

with the following codes and meanings: 1=Agricultural hukou, 2=Non-

agricultural hukou, 3=Resident hukou (a type issued in some regions to all 

residents, without distinguishing between agricultural and non-agricultural 

status), and 4=No hukou. Based on this study’s focus on the traditional ur-



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2017 

ban-rural hukou disparity, the original variable was recoded. Considering 

that the “Resident hukou” is a product of hukou system reforms aimed at re-

ducing this binary distinction and is more closely aligned with the non-

agricultural hukou in terms of public services and social rights (urban attrib-

utes), “2=Non-agricultural hukou” and “3=Resident hukou” were merged to 

form the urban hukou category, which was assigned a value of 2. The 

“1=Agricultural hukou” category was retained as is. Furthermore, the “4=No 

hukou” category was excluded from the analysis due to its ambiguity and 

extremely small proportion. The resulting transformed variable is a binary 

nominal variable. 

Data on self-rated health were derived from the ‘Physical and Mental 

Health’ module of the student questionnaire. The corresponding original 

question was: “Which one of the following best describes your general 

health condition AT PRESENT?” This is a single-item, self-reported meas-

ure. The variable was originally a five-category ordinal variable (1=Very 

poor, 2=Not very good, 3=Moderate, 4=Good, 5=Very good). For this study, 

it was recoded into a three-category ordinal variable with the following 

structure: the ‘Very poor’ (1) and ‘Not very good’ (2) categories were 

merged into ‘poor’ and assigned a value of 1; the ‘Moderate’ (3) category 

retained its original meaning but was reassigned a value of 2; and the ‘Good’ 

(4) and ‘Very good’ (5) categories were combined into ‘good’ and assigned a 

value of 3. The transformed variable is thus a three-category ordinal variable. 

Data on students’ self-educational expectations were sourced from 

the ‘Academic Development’ module of the student questionnaire. The vari-

able is a single-item, self-reported measure based on the question: “What is 

the highest level of education you expect yourself to receive?” The original 

variable was a 10-category ordinal scale: (1) Drop out now, (2) Graduate 

from junior high school, (3) Go to technical secondary school or technical 

school, (4) Go to vocational high school, (5) Go to senior high school, (6) 

Graduate from junior college, (7) Get a bachelor degree, (8) Get a Master 

degree, (9) Get a Doctor degree, and (10) I don’t care. Based on a hierar-

chical logic of ascending educational expectations, this study simplified the 

variable into a three-category ordinal measure. The first category, “1 = Sec-

ondary Education and Below,” includes the original codes 2 through 5. The 

“Drop out now” (1) and “I don’t care” (10) options were also assigned to this 

category because they largely reflect low educational expectations. The se-

cond category, “2 = University Degree,” represents the foundational level of 

higher education and combines the original codes for “Graduate from junior 

college” (6) and “Get a bachelor degree” (7). The third category, “3 = Post-

graduate Degree,” corresponds to the advanced level of higher education and 

includes the original codes for “Get a Master degree” (8) and “Get a Doctor 

degree” (9). 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2018 

The raw data for parental educational expectations were derived from 

the ‘Family Education’ module of the parent questionnaire. This variable is a 

single-item, self-reported measure based on the question: “What is the high-

est level of education do you expect this child to receive?” The original cod-

ing and the meanings of the categories for this variable were identical to 

those for student self-educational expectations. For the analysis, this study 

applied the same consolidation and simplification logic to parental educa-

tional expectations as was used for the student self-educational expectations 

variable. 

Data on learning motivation were derived from a sub-item within a 

scale in the student questionnaire’s ‘Physical and Mental Health’ module, 

which begins with the prompt: “How much do you agree with each of the 

following statements about your experiences in GRADE 7?” It is a single-

item, self-reported measure corresponding to the specific statement: “I would 

try my best to finish even the homework I dislike.” The variable was origi-

nally a four-point Likert-type ordinal variable, with the following codes and 

meanings: 1=Strongly disagree, 2=Somewhat disagree, 3=Somewhat agree, 

and 4=Strongly agree. To simplify the analysis, this study merged and trans-

formed it. The categories ‘1=Strongly disagree’ and ‘2=Somewhat disagree,’ 

which had smaller sample proportions, were combined into ‘Not really agree’ 

and assigned a value of 1. The ‘3=Somewhat agree’ category retained its 

original meaning, with its code adjusted to 2, and the ‘4=Strongly agree’ cat-

egory was recoded to 3. The transformed variable is a three-category ordinal 

variable. 

Data on learning persistence were also sourced from the student 

questionnaire’s ‘Physical and Mental Health’ module. It is a single-item, 

self-reported measure from the same scale as the learning motivation varia-

ble, corresponding to the specific statement: “I would try my best to finish 

my homework, even if it would take me quite a long time.” As a four-point 

Likert-type ordinal variable, its original coding and meanings were identical 

to those of the ‘learning motivation’ variable. The logic used for its consoli-

dation and transformation was also the same as that applied to ‘learning mo-

tivation’, resulting in a three-category ordinal variable. 

Environment Dimension 

Environmental variables reflect learning resources and influence inside and 

outside school, central to the educational process. It’s divided into two sub-

dimensions: student involvement and school characteristics. Student in-

volvement includes weekly hours spent on school homework, cram school 

attendance, extra homework (assigned by parents or cram school), and Exer-

cise. School characteristics include School type, Availability of library, and 

Per-student Expenditure. 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2019 

Regarding school characteristics, three variables were included: 

school type, whether the school has a library, and per-student expenditure. 

All three items were sourced from the ‘Basic Information of the School’ 

module of the school administrator questionnaire. 

School type is a single-item measure self-reported by the principal. 

The variable was originally a five-category nominal variable with the follow-

ing codes and meanings: (1) Public school, (2) Private school subsidized by 

the government, (3) Ordinary private school, (4) Private school for children 

of migrant workers, and (5) Other. Because this study’s primary aim is to 

investigate the difference in impact between public and private schools on 

student achievement, this variable was recoded. Specifically, categories 2 

through 4 were merged to form the ‘private school’ category, which was as-

signed a value of 0. The ‘public school’ category (1) was kept unchanged. 

The transformed variable is a binary nominal variable. Additionally, among 

the 112 schools in the sample, none were classified as ‘Other’ (5). 

Whether the school has a library is a single-item measure self-

reported by the principal, derived from a sub-item under the scale “Does 

your school have the following facilities?”. The original variable was a 

three-category ordinal variable: (1) No, (2) Yes, but need to be improved, 

and (3) Yes, and well equipped. For this study, categories 2 and 3 were 

merged to form the “has a library” category, which was assigned a code of 2. 

The “No” category remained unchanged. The transformed variable is a bina-

ry nominal variable. 

Per-student expenditure was a self-reported item completed by the 

principal, representing objective administrative statistical data. The corre-

sponding original question was: “How much fiscal appropriation per student 

has your school received THIS YEAR?”. The variable was originally a con-

tinuous variable, taking non-negative values (unit: Yuan/student) with a 

range of 0 and above. For this study, the data were recoded into a three-

category ordinal variable with the following codes and meanings: 1=Low 

funding (<800 Yuan), 2=Medium funding (800-1,800 Yuan), which repre-

sents the primary distribution range of the data, and 3=High funding (≥1800 

Yuan). 

Data on the weekly time students spent on school homework, cram 

school, and extra homework were all derived from the ‘Academic Develop-

ment’ module of the student questionnaire. They were sub-items under the 

scale asking, “How much time ON AVERAGE EVERYDAY did you spend 

on the following extra-curricular activities?”. The corresponding original 

items were: ‘Doing homework assigned by teacher,’ ‘Taking cram school 

courses (related to schoolwork),’ and ‘Doing homework assigned by parents 

or cram school.’ The time spent was measured separately for weekdays and 

weekends. The response options for the weekday scale were: (1) 0 hours, (2) 

Less than 1 hour, (3) About 1-2 hours, (4) About 2-3 hours, (5) About 3-4 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2020 

hours, and (6) More than 4 hours. The response options for the weekend 

scale were: (1) None, (2) Less than 2 hours, (3) About 2-4 hours, (4) About 

4-6 hours, (5) About 6-8 hours, and (6) More than 8 hours. Because the orig-

inal options represented interval data, the midpoint of each interval was tak-

en and multiplied by the corresponding number of days. The resulting values 

for weekdays and weekends were then summed to calculate the total weekly 

time spent (unit: hours). 

Data on students’ physical exercise were obtained from the ‘Physical 

and Mental Health’ module of the student questionnaire. This was a self-

reported, two-part measure corresponding to the question: “How often do 

you do physical exercise? Usually [ ] days a week, [ ] minutes a day.” Exist-

ing research indicates that “ensuring two hours of daily physical activity for 

middle school students” is a national-level reform goal that is being vigor-

ously promoted but has not yet been fully realized, and that insufficient ex-

ercise among students remains a common phenomenon (Liu & Shan, 2025). 

Based on this, we concluded that some students’ self-reported daily exercise 

durations (e.g., over 150 minutes) likely deviated significantly from the 

prevalent reality. Therefore, to ensure data validity and the robustness of the 

analysis, this study treated cases reporting more than 150 minutes of daily 

exercise as outliers and excluded them. Subsequently, for the remaining val-

id sample, the ‘days of exercise per week’ was multiplied by the ‘minutes of 

exercise per day,’ and the result was converted into hours to generate the to-

tal weekly physical exercise duration. The transformed variable is a continu-

ous variable with non-negative values. 

Methods of Analysis 

Data cleaning was carried out via Stata 17 to retain relevant variables, recode 

categorical variables, and remove outliers. Statistical analysis was performed 

using SPSS 26.0; chi-square tests were applied to categorical variables relat-

ed to student background, psychological and school characteristics, while 

one-way ANOVA was used for continuous variables on student involvement, 

with significance set at p < 0.05. Variables with p < 0.05 were entered into 

binary logistic regression to assess their association with academic perfor-

mance. Multicollinearity was assessed using the variance inflation factor 

(VIF), with VIF < 5 indicating acceptable independence among predictors. 

Based on logistic regression results, a nomogram was constructed using R 

4.4.2 to visualize the prediction model and improve interpretability. Model 

performance was evaluated using ROC curves, area under the curve (AUC), 

and calibration curves. 

 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2021 

Table 1. Effects of Input Characteristics on Academic Performance. 

Categorization N High score achiever Percentage  ² p value 

Gender    273.078 ＜0.001 

Male  3,024 1,408 46.56   

Female  2,974 2,013 67.69   

Ethnicity     0.954 0.329 

Han 5,430 3,108 57.24   

Minority 568 313 55.10   

Hukou type    12.958 <0.001 

Rural  3,182 1,746 54.90   

Urban  2,816 1,675 59.48   

Self-rated health    4.603 0.1 

Poor  387 215 55.56   

Fair 1,749 964 55.12   

Good  3,862 2,242 58.05   

Self-educational expectation    785.385 <0.001 

Secondary Education & Below 1,160 289 24.91   

University Degree 3,058 1,757 57.46   

Postgraduate Degree 1,780 1,375 77.25   

Parental education expectation    702.429 <0.001 

Secondary Education & Below 723 129 17.84   

University Degree 3,222 1,771 54.97   

Postgraduate Degree 2,053 1,521 74.09   

Learning motivation    183.38 <0.001 

Not really agree 1,197 491 41.02   

Fair 2,536 1,459 57.53   

Strongly agree 2,265 1,471 64.94   

Learning persistence    192.135 <0.001 

Not really agree 1,251 518 41.41   

Fair 2,476 1,414 57.11   

Strongly agree 2,271 1,489 65.57   

Total 5,998 3,421 57.04   

 

 

Results 

Differences in Junior High School Students’ Academic Performance on 

Inputs 

The result showed that of the 5,998 students, 3,421(57.04% of the total) had 

academic performance above the average of their schools. 

In terms of demographic characteristics, girls’ academic performance 

(67.7%) is better than boys’ (46.6%), and the difference is statistically signif-

icant (² ＝ 273.078, p < 0.001). This result indicates that gender is one of 

the factors that significantly affects students’ performance. The percentage 

of Han Ethnicity and ethnic minority students who performed better than the 

school average was 57.24% and 55.10%, respectively. From the results, eth-

nicity is not a significant influence on academic performance. In terms of 

Hukou type, urban household register students outperformed rural household 

regis ter  students  with a significant  difference (² ＝  12.958,  



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2022 

Table 2. The Impact of School Characteristics on Academic Performance. 

Categorization N Numbers of above-average Percentage  ² p value 

School type    9.517 0.002 

Private 410 204 49.76   

Public 5,588 3,217 57.57   

Library    0.011 0.915 

No 902 513 56.87   

Yes 5,096 2,908 57.06   

Per-student expenditure    16.609 < 0.001 

Low 2,209 1,187 53.73   

Medium 2,751 1,608 58.45   

High 1,038 626 60.31   

 

 

 

p < 0.001). The highest percentage of students who perceived themselves as 

physically good achieved good grades (58.10%), followed by those who per-

ceived themselves as physically poor and fair (55.60% and 55.12%, respec-

tively), but there was no significant difference in grades between students 

with different health conditions (p ＝ 0.1). In terms of psychological charac-

teristics, students’ and parents’ educational expectations have a significant 

effect on students’ academic performance, with higher educational expecta-

tions being associated with better academic performance. Students’ learning 

motivation and persistence are also significant influences on academic per-

formance; the stronger the learning motivation and persistence, the better the 

students’ academic performance. The details are shown in Table 1. 

The Impact of School Characteristics on Students’ Aca-

demic Performance 

Regarding school characteristics, the proportion of students with above-

average academic performance was significantly higher in public schools 

(57.57%) than in private schools (49.76%) (² = 9.517, p = 0.002). In terms 

of school material resources, whether a school had a library was not signifi-

cantly associated with student academic performance (² = 0.011, p = 0.915). 

In fact, the proportion of students who achieved above-average scores was 

nearly identical in schools with libraries (57.06%) compared to schools 

without them (56.87%). Per-student expenditure had a significant effect on 

academic performance(²＝16.609，P<0.001), with students’ academic per-

formance gradually improving as per-student expenditure increased, as 

shown in Table 2. 

Regarding student involvement, this study found that the amount of 

time students spent per week writing school and extra homework, attending 

cram school, and exercising all had a significant effect on students’ academic  



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2023 

Table 3. The Impact of Student Involvement on Academic Performance. 

Categorization Square sum Degrees of freedom Mean square F p value 

Sch_ homework 48.122 32 1.504 6.310 ＜0.001 

Extra_ homework  19.316 32 0.604 2.482 ＜0.001 

Cram_ school  23.076 30 0.769 3.173 ＜0.001 

Exercise 50.513 109 0.463 1.923 ＜0.001 

 

 

 

Table 4. Descriptive Analysis of Student Engagement (N = 5,998). 

Categorization Min Max Mean SD 

Sch_ homework 0 40.5 16.07 8.390 

Extra_ homework  0 40.5 4.14 6.144 

Cram_ school  0 40.5 4.30 7.535 

Exercise 0 17.5 2.30 2.148 

 

 

 

performance (p < 0.001). The effect of hours of writing school homework on 

academic performance was relatively high (F ＝ 6.310), followed by hours 

of cram school (F = 3.173) and extra work (F = 2.482), and exercise had the 

smallest effect (F ＝ 1.923), but still had a significant effect on performance 

(see Table 3 for details). In addition, we used descriptive statistics to analyze 

student involvement, and the results are presented in Table 4. The results 

showed that students spent the longest amount of time per week on school 

homework, with an average of 16.07 hours per week, followed by attending 

cram school and writing extra homework. In contrast, students exercise an 

average of only 2.30 hours per week, dramatically less than the time spent on 

academics. 

 

Logistic Regression Analysis of Factors Affecting Aca-

demic Performance 

Binary Logistic regression analyses were conducted using the statistically 

significant factors from the univariate analyses as independent variables and 

students’ academic performance as the dependent variable. In order to miti-

gate potential bias in the results due to high correlation between variables, 

VIF was used to check the data for multicollinearity. The results showed that 

the VIF values of all the variables in this study were less than 5 (range 1.019 

- 2.147), proving that there is no problem of multicollinearity among the var-

iables.  

 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2024 

Table 5. Logistic Regression Results. 

Independent Variable B SE Wald ² p Value OR 

95%CI 

Lower 
Limit 

Upper 
Limit 

Intercept -2.876 0.173 277.671 <0.001 0.056 0.040 0.079 

Exercise -0.015 0.014 1.172 0.279 0.985 0.958 1.012 

Extra_ homework -0.007 0.006 1.483 0.223 0.993 0.982 1.004 

Cram_ school -0.016 0.005 12.286 <0.001 0.984 0.976 0.993 

Sch_ homework 0.010 0.004 7.807 0.005 1.011 1.003 1.018 

Gender        

Female 0.756 0.060 160.300 <0.001 2.129 1.894 2.393 

Male        

Hukou type        

Urban  -0.097 0.063 2.386 0.122 0.908 0.803 1.026 

Rural        

Self-educational expectation         

Secondary Education & Below        

University Degree 0.809 0.090 81.000 <0.001 2.245 1.882 2.677 

Postgraduate Degree 1.450 0.107 184.118 <0.001 4.262 3.457 5.255 

Parental educational expectation        

Secondary Education & Below        

University Degree 1.114 0.118 89.633 <0.001 3.046 2.419 3.836 

Postgraduate Degree 1.692 0.128 174.621 <0.001 5.428 4.224 6.976 

Motivation        

Not really agree        

Fair 0.260 0.095 7.482 0.006 1.297 1.076 1.563 

Strongly agree 0.273 0.114 5.776 0.016 1.314 1.052 1.642 

Persistence        

Not really agree        

Fair 0.200 0.095 4.444 0.035 1.221 1.014 1.470 

Strongly agree 0.304 0.114 7.149 0.007 1.355 1.085 1.694 

Per-student expenditure        

Low        

Medium 0.339 0.066 26.531 <0.001 1.404 1.234 1.598 

High 0.215 0.088 5.991 0.014 1.240 1.044 1.473 

School type        

Public 0.209 0.118 3.149 0.076 1.232 0.978 1.552 

Private        

 

 

The logistic regression analysis revealed that in terms of student in-

volvement, the weekly time spent on exercise and extra homework had no 

significant effect on academic performance. Attending cram school had a 

significant negative impact. School-assigned homework was the only varia-

ble in the student involvement dimension that positively influenced academ-

ic performance; for each additional hour spent on school homework, the 

odds of performing above the school average increased by 1.1%. Regarding 

personal background, gender differences were significant, with female stu-

dents significantly outperforming male students. The effect of hukou type on 

academic performance was not significant. With respect to psychological 

characteristics, the student’s own educational expectations, parental educa-

tional expectations, learning motivation, and learning persistence all had a 

significant impact on academic performance. Among these, the student’s 

o w n  a n d  p a r e n t a l  e d u c a t i o n a l  e x p e c t a t i o n s  w e r e  t h e  



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2025 

 
Figure 2. Nomogram. 

 

 

 

strongest predictors of academic performance, with their odds ratios (ORs) 

being substantially greater than 1 and significantly higher than those of other 

independent variables. The regression results are detailed in Table 5. 

Predictive Modelling of Academic Performance of Jun-

ior High School Students 

A Nomogram prediction model was developed based on the factors affecting 

students’ academic performance, the details of which are shown in Figure 2. 

The names of the variables are displayed on the left side; each variable cor-

responds to a line segment, and the scale of the line segment indicates the 

range of scores of the variable. The larger the range of scores, the greater the 

impact of the variable on the predicted results of the model, and the more 

important the variable is. Individual scores on each variable are summed to 

obtain the TOTAL POINT, which represents the total score of an individual 

in the model. The probability that a student’s academic performance is high-

er than the school average is obtained by plotting the Total Point against the 

Prob of outcome axis. According to the nomogram, the significance of the 

predictor variables is: parental education expectation > self-education  



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2026 

 
Figure 3. ROC Curve. 

 

 

 

Figure 4. Calibration Curve. 
 

 

expectation > cram school > gender > school homework > per-student ex-

penditure >learning persistence > learning motivation. 

The results of the ROC curve test showed that the average AUC val-

ue was 0.756, and the sensitivity and specificity were 0.709 and 0.667, re-

spectively. These results indicate that the model is effective in predicting 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2027 

students’ academic performance, and the predictions are within acceptable 

limits (Mandrekar, 2010). The specific details of the ROC curve are shown 

in Figure 3. The calibration curve fluctuates around the ideal curve and al-

most overlaps with the ideal curve, which fully indicates that the model pre-

dicts well and is repeatable. The mean absolute error (MAE) is 0.003, and 

the lower error rate once again indicates the prediction accuracy of the model, 

details of which are shown in Figure 4. 

Discussion 

Based on the Input-Environment-Outcome (I-E-O) theoretical framework, 

we utilized nationally representative large-scale survey data to investigate 

the influence of multidimensional factors on the academic performance of 

ninth-grade students. The findings indicate that the formation of academic 

performance is a complex process involving the synergistic effect of multiple 

factors. A student’s individual and psychological characteristics serve as the 

internal drivers influencing academic performance. At the same time, the 

educational environment in which students are situated, and their academic 

involvement within it, also have a significant impact on their academic per-

formance. 

Slightly more than half of the total number of students in this study, 

57.04%, performed above the school average. Overall, students’ academic 

performance tended to be better than average. In terms of the distribution of 

the number of students with high and low academic performance, this result 

is also in line with the reality that 50% of students enter regular senior sec-

ondary schools under the mechanism of ‘50-50 streaming in the high school 

entrance examination’. 

The Dominant Role of Educational Expectations 

This study found that both parental and student self-educational expectations 

significantly and positively predict academic achievement, with the role of 

parental expectations being more critical. Higher levels of expectation are 

associated with a greater probability of achieving high academic perfor-

mance, a finding that is consistent with existing research (Fishamn, 2022). 

Furthermore, according to the results from the nomogram model and the re-

gression analysis, the importance of parental educational expectations for 

student achievement is slightly greater than that of student self-expectations, 

highlighting the crucial role played by external family expectations. 

This finding aligns with the classic Wisconsin Model of Status At-

tainment, which posits that educational expectations, as a key 

sociopsychological variable, serve as the core mediator linking family socio-

economic background to an individual’s ultimate educational achievements 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2028 

(Sewell et al., 1969). Our study demonstrates that this model remains appli-

cable within the Chinese educational context. However, this study also finds 

that when we incorporate the practical application of these expectations, spe-

cifically the weekly time spent in cram school, into the model, this factor 

exhibits a significant negative effect on academic performance. This indi-

cates that merely holding high expectations is insufficient. How these expec-

tations are translated into concrete educational behaviors is key to moderat-

ing their ultimate effect. This finding thereby adds a new interpretive dimen-

sion to the Wisconsin Model’s application in contemporary China. 

Specifically, the positive influence of high parental educational ex-

pectations is effective because, through high-quality interaction and support, 

these expectations are effectively internalized by students and transformed 

into autonomous self-expectations (Pinquart & Ebeling, 2020a). According 

to Self-Determination Theory, the positive influence of high parental expec-

tations is realized when these expectations are conveyed in a supportive, un-

derstanding, and encouraging manner. In such a context, students feel their 

autonomy is respected and are thus more likely to integrate external goals as 

intrinsic motivation, fostering greater learning persistence and self-efficacy. 

This process of internalization aligns parental and student expectations, syn-

ergistically driving academic success. However, this study’s finding that 

weekly time spent in cram school negatively impacts academic performance 

illustrates an alienated pathway for the transmission of expectations. When 

high parental expectations fail to be internalized through effective family 

communication and emotional support, and are instead transformed into a 

form of external, coercive behavioral control, the positive effect can be di-

minished or even reversed. In such cases, an ‘expectation gap’ emerges be-

tween the parents’ high expectations and the students’ actual feelings, abili-

ties, or willingness (Cheng et al., 2022). An expectation gap transforms edu-

cational expectations from a motivational goal into an oppressive source of 

external control, negatively affecting academic performance through several 

mechanisms. First, the expectation gap directly undermines a student’s learn-

ing autonomy, shifting learning from intrinsic exploration to passive coping, 

which in turn weakens learning motivation (Li & Hu, 2021). Second, persis-

tent pressure and the negative feedback that may accompany failure to meet 

high expectations can erode a student’s academic self-efficacy, making them 

feel that they can never satisfy their parents, regardless of their efforts. Final-

ly, long-term psychological stress, academic burden, and negative emotions 

significantly increase the risk of academic burnout, leading students to cog-

nitively and emotionally disengage from learning, which results in a substan-

tial negative impact on their academic achievement (Cheng et al., 2022). 

Therefore, the influence of educational expectations on the academic 

achievement of middle school students is not a simple linear positive rela-

tionship but rather a complex process characterized by dynamic interplay. 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2029 

While high expectations from parents and students are undoubtedly valuable 

psychological capital for academic success, the realization of their positive 

effects hinges on the manner of their transmission and application. Only 

when the communication of these expectations maintains and promotes a 

student’s autonomy and intrinsic motivation, thereby achieving high-quality 

internalization, can they serve as a catalyst for academic development. Con-

versely, they will erode a student’s learning motivation, be detrimental to 

their physical and mental well-being, and ultimately impede academic pro-

gress. 

The Negative Impact of Cram Schools on Academic 

Performance 

The influence of student involvement on academic performance presents a 

contradiction: although the time spent completing school homework is posi-

tively associated with high academic performance, the duration of participa-

tion in cram school shows a significant negative association. This seemingly 

conflicting result indicates that the sheer ‘quantity of time invested’ is not 

the key to academic success; rather, the nature of that engagement may play 

a more crucial role. We argue that school homework and cram school cours-

es are fundamentally different in how they affect student learning autonomy. 

From the perspective of Self-Determination Theory (SDT), school home-

work, as an extension of the school curriculum, is typically designed by 

teachers in accordance with the syllabus. Its content is synchronized with the 

pace of instruction, and its difficulty is often within the student’s zone of 

proximal development, which helps students consolidate knowledge and ef-

fectively satisfies their need for competence. In contrast, cram school cours-

es are often arranged by parents and are more representative of external con-

trol rather than an autonomous choice by the student. These courses fre-

quently aim to “teach ahead” or “raise the bar,” and their learning content 

may exceed the student’s current cognitive abilities, leading to persistent 

frustration and undermining their sense of competence. This passive partici-

pation can thwart the student’s need for autonomy. Even if a student is in-

trinsically motivated, such a controlling learning context may lead to ineffi-

cient, surface-level learning, thereby negatively impacting academic 

achievement. As existing research has indicated, excessive extracurricular 

tutoring intensity is closely associated with students’ psychological stress 

and academic burnout (Fu et al., 2023). 

Additionally, although descriptive statistics indicate that students’ 

physical exercise time (mean = 2.30 hours/week) is severely constrained by 

numerous academic activities, the multiple regression model reveals that, 

within the analytical framework of this study, exercise duration was not a 

direct predictor of academic achievement (p = 0.279). Therefore, the nega-



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2030 

tive effect of cram school is not primarily realized through the indirect path-

way of crowding out time for physical exercise, but is more likely to stem 

from its direct impact on learning psychology and the learning process. Spe-

cifically, although existing research has indicated an association between 

attending cram school and negative emotions such as an aversion to learning 

(Li & Liu, 2022), this study finds that learning motivation and learning per-

sistence remain among the most critical psychological characteristics for 

achieving high performance. In other words, the negative impact of attending 

cram school may not be that it simply and completely destroys students’ mo-

tivation to learn; rather, it likely operates through other mechanisms while 

their motivation is still present. For example, the intensive, test-oriented 

training in cram school may hinder the development of students’ deep learn-

ing and critical thinking skills (Guill et al., 2022). Even if a student is moti-

vated to complete assignments, their learning efficiency and ultimate out-

comes could be substantially compromised by a shift toward surface-level 

learning. 

Indirect Effects of Hukou and School Type 

In the univariate analysis, hukou type and school type were initially found to 

be significantly associated with academic achievement. However, when mi-

cro-level variables such as individual student characteristics, the psychologi-

cal characteristics, and specific school resources were incorporated into the 

binary logistic regression model, the predictive effects of these two variables 

were no longer significant. This does not suggest that hukou type and school 

type have no impact on academic performance. Instead, it reveals that their 

influence is indirect: after the inclusion of more substantive micro-level vari-

ables, their effects are mediated by these variables.  

In the analysis, the effect of hukou type on academic performance 

exhibited a significant transformation. In the univariate analysis, students 

with a non-agricultural hukou demonstrated significantly better academic 

performance than students with an agricultural hukou. However, in the bina-

ry logistic regression model that incorporated more control variables, the di-

rect predictive effect of hukou type became non-significant. This shift indi-

cates that hukou type does not directly determine academic performance but 

more likely operates through more specific mediating pathways. Specifically, 

from the perspective of family cultural capital, educational expectations 

demonstrated the strongest predictive power among all independent variables. 

This points to a pathway for the influence of hukou type. A reasonable ex-

planation is that families with a non-agricultural hukou are typically situated 

in urban environments with more advantaged socioeconomic status and cul-

tural resources, and this environment may naturally translate into higher ed-

ucational expectations among family members (Sun et al., 2025). These high 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2031 

expectations, acting as a potent form of “cultural capital,” not only directly 

motivate students but also positively influence their “learning motivation and 

learning persistence,” and these two psychological qualities are also signifi-

cant predictors in the model. Consequently, when the model directly 

measures this core driver of educational expectations, a portion of the ex-

planatory power held by hukou, a relatively macro-level background label, is 

absorbed by these more targeted psychological variables. While educational 

expectations are the most central of these variables, the model’s other signif-

icant predictors, learning motivation and learning persistence, also likely 

stem from the family environment. Together, they constitute a psychological 

mediation chain through which hukou type influences academic performance. 

On the other hand, from the perspective of educational resource allocation, 

the regression analysis shows that “per-student expenditure” is another sig-

nificant environmental factor influencing student achievement. China’s long-

standing urban-rural dual structure has created vast disparities in educational 

resources, resulting in significant inequality in accessing financial support 

between rural schools associated with agricultural hukou and urban schools 

associated with non-agricultural hukou (Zhao, 2023). Urban schools typical-

ly receive higher per-student funding, which translates to higher-quality 

teachers, better facilities, and richer educational resources. Because this 

study’s model directly controls for “per-student expenditure” which is the  

core variable reflecting the school environment, it also accounts for the re-

source advantages attached to hukou status, rendering the independent effect 

of hukou type no longer statistically prominent. Therefore, the influence of 

hukou type on the academic performance of middle school students is indi-

rect; its effect is realized through two pathways: shaping the “family soft en-

vironment” and the “school hard environment.” When the model simultane-

ously controls for both the family soft environment and the school hard envi-

ronment, the influence of hukou type as a macro-level social identity is no 

longer significant. 

Similarly, the school type variable, which was significant in the 

univariate analysis, was no longer significant in the binary logistic analysis. 

This disappearance of significance may be due to the presence of selection 

bias. The “public school advantage” observed in the univariate analysis may 

not stem from the “public” status itself, but is rather a combined result of the 

school’s resource endowments and the student population it serves. Existing 

research has pointed out that schools of different types have significant dif-

ferences in their student populations, and a school’s educational effective-

ness is largely influenced by the selection mechanisms related to its student 

quality (Yao, 2023). First, from the perspective of resource endowments, the 

binary logistic regression model included “per-student expenditure” as a key 

variable in the analysis. The results show that the level of per-student ex-

penditure has a significant positive impact on students’ academic perfor-



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2032 

mance, which indicates that the financial resources a school receives are an 

important factor in student academic development. When the model controls 

for this variable, which more directly reflects a school’s resource level, the 

general labels of “public” or “private” lose their independent predictive 

power. In other words, for both public and private schools, being able to ob-

tain adequate resource investment is the key to promoting student academic 

development, not the nature of their governance. Second, from the perspec-

tive of student body composition, this study, proceeding from the ‘input’ di-

mension, selected a large number of variables related to students’ individual 

backgrounds, including educational expectations, learning motivation, and 

persistence. All of these variables were found to be powerful predictors of 

academic performance. In Chinese educational practice, schools of different 

types or reputations often attract student groups from diverse backgrounds, a 

phenomenon known as ‘student sorting’ (Tan & Zhang, 2025; Yao, 2023). 

High-quality public schools with strong reputations may naturally attract 

students who come from families with higher educational expectations and 

more ample support. Therefore, the advantage of public schools observed in 

the univariate analysis may simply be a concentrated reflection of the pre-

existing advantages of its student population. Consequently, when the multi-

variate regression model takes these advantageous factors into account, the 

independent effect of the school type subsequently disappears, its explanato-

ry power having been absorbed by the more substantive variables. 

Model Evaluation and Reflection 

Building upon the identification of key factors influencing the academic per-

formance of middle school students, this study further developed a 

nomogram model with strong predictive performance. In terms of model per-

formance, its AUC was 0.756, with a sensitivity and specificity of 0.709 and 

0.667, respectively. This indicates that the model can effectively discrimi-

nate between student groups performing above or below the in-school aver-

age, demonstrating predictive power within an acceptable range. Further-

more, the model’s calibration curve nearly coincided with the ideal curve, 

and its MAE was only 0.003 This clearly demonstrates a high degree of con-

cordance between the predicted probabilities and the observed frequencies of 

high achievement, showcasing good calibration performance and thereby 

enhancing the model’s reliability for practical application. The strong per-

formance of the model in this study can be attributed to its precise capture of 

key influencing factors, particularly through the quantification and success-

ful integration of variables such as student and parental educational expecta-

tions, gender, and academic involvement into a unified predictive framework. 

Compared to traditional regression equations, the greatest advantage 

of the nomogram model constructed in this study lies in its intuitive nature 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2033 

and ease of use. The model transforms complex statistical results into a sim-

ple, graphical scoring system, enabling educators without a statistical back-

ground to use it with ease. On a practical level, schools and teachers can uti-

lize this nomogram to conduct prospective assessments of student academic 

performance. By inputting a student’s specific information on the various 

indicators, one can quickly obtain a quantified probability of that student’s 

achievement being above the in-school average. This helps educators identi-

fy students who may face academic challenges, thereby allowing for early 

intervention and the provision of personalized academic support or psycho-

logical counseling to achieve targeted instruction. 

The model in this study also has certain limitations. On the one hand, 

there are limitations in the scope of variable selection. Although this study, 

guided by the I-E-O model, included 14 variables from the ‘Input’ and ‘En-

vironment’ dimensions, it may still have overlooked other potentially im-

portant factors. These include teacher-level variables such as teaching style, 

peer effects, and other deeper psychological variables like student self-

efficacy and resilience. These unmeasured variables may have constrained 

the upper limit of the model’s predictive accuracy. On the other hand, there 

are boundaries at the methodological level. First, the model constructed in 

this study is a predictive model; it reveals correlations between variables and 

the outcome, rather than strict causal relationships. Second, the model’s ex-

ternal validity requires further examination. Although the model demonstrat-

ed good internal validation, its generalizability to different regions or types 

of schools still needs to be verified through further empirical testing. 

Conclusion and Limitation 

This study systematically investigated the multifaceted factors influencing 

the academic performance of middle school students and constructed a 

nomogram model with good predictive performance (AUC = 0.756). The 

conclusions indicate that psychological capital that originates from the fami-

ly and is internalized by the student, especially educational expectations, 

serves as the core driving force that surpasses other variables, highlighting 

the decisive role of intrinsic motivation in academic achievement. The study 

also revealed the differential returns of various academic involvement be-

haviors and elucidated that the influence of macro-structural factors, such as 

hukou type and school type, is largely realized indirectly through more direct 

micro-level pathways like family cultural background and actual school re-

sources. Overall, this study not only identified the key predictors of academ-

ic performance but also provided deeper insights into the complex interac-

tion among these factors, offering a valuable empirical reference for educa-

tional practice and policy interventions. 



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2034 

This study has several limitations. First, the data used were collected 

from surveys conducted between 2014 and 2015, the most recent available 

from the CEPS dataset. However, to more accurately and comprehensively 

investigate factors influencing junior high school students’ academic per-

formance, future studies should employ more representative and up-to-date 

data that reflect changes in the educational context. Second, the study fo-

cused exclusively on Grade 9 students. As final-year students, they face 

unique curricular demands and psychological pressures. Therefore, further 

research is needed to examine the factors influencing academic performance 

among students in earlier grades of junior high school. 

 

  



Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2035 

References 

Astin, A. W. (1977). Four critical years: 

Effects of college on beliefs, attitudes, 

and knowledge. Jossey-Bass. 

Astin, A. W. (1984). Student involvement: 

A developmental theory for higher 

education. Journal of College Student 

Personnel, 25(4), 297-308. 

Bao, W., & Zhang, X. Y. (2012). The mul-

ti-dimensional structure and influence 

mechanism of student academic in-

volvement in China. Fudan Education 

Forum, 10(6), 20-28. DOI: 

https://doi.org/10.13397/j.cnki.fef.20

12.06.012  

Cheng, G., Du, S. H., & Yu, Q. (2022). Are 

your educational expectations ration-

al? Research on effects of parent-

child discrepancies in educational ex-

pectations on academic performance. 

Journal of East China Normal Uni-

versity (Educational Sciences), 40(1), 

74-87. DOI: 

https://doi.org/10.16382/j.cnki.1000-

5560.2022.01.006  

Entwistle, N.J., Thompson, J., & Wilson, 

J.D. (1974). Motivation and study 

habits. Higher Education, 3, 379-396. 

DOI: 

https://doi.org/10.1007/BF00153949  

Eysenck, H. (2018). The dynamics of anx-

iety and hysteria: An experimental 

application of modern learning theory 

to psychiatry. Routledge. 

Fishman, S. H. (2022). College-for-some or 

college-for-all? Inequality in the rela-

tionship between educational expec-

tations and educational attainment 

across academic achievement. Social 

Science Research, 107, 102747. DOI: 

https://doi.org/10.1016/j.ssresearch.2

022.102747  

Fu, C., Ou, H., Mo, T., & Liao, L. (2023). 

Effect mechanism of extracurricular 

tuition and implications on “double 

reduction” policy: Extracurricular tui-

tion intensity, psychological resili-

ence, and academic performance. Be-

havioral Sciences, 13(3), 217. DOI: 

https://doi.org/10.3390/bs13030217  

Guan, Z. (2022). The influence of private 

tutoring on academic performance of 

junior high school students: Based on 

PSM-DID method estimation. China 

Economics of Education Review, 

7(1), 57-74. DOI: 

https://doi.org/10.19512/j.cnki.issn20

96-2088.2022.01.004.  

Guill, K., Ömeroğulları, M., & Köller, O. 

(2022). Intensity and content of pri-

vate tutoring lessons during German 

secondary schooling: Effects on stu-

dents’ grades and test achievement. 

European Journal of Psychology of 

Education, 37(4), 1093-1114. DOI: 

https://doi.org/10.1007/s10212-021-

00581-x  

Huang, X. T., Guan, K. X., Xiong, G. H., 

Chen, H., & Lu, X. D. (2016). The 

impact of “super-high school” on ed-

ucational equity and efficiency: An 

empirical study based on undergradu-

ates academic performance data of 

University K. Education Research 

Monthly, 5, 32-37. DOI: 

https://doi.org/10.16477/j.cnki.issn16

74-2311.2016.05.005  

Iwuagwu, B. (2016). Environmental influ-

ence on students’ academic perfor-

mance in secondary school. Interna-

tional Journal of Fundamental Psy-

chology and Social Sciences, 6(1), 

10-14. DOI: 

https://doi.org/10.14331/ijfpss.2016.3

30058  

Kweon, B.-S., Ellis, C. D., Lee, J., & Ja-

cobs, K. (2017). The link between 

school environments and student aca-

demic performance. Urban Forestry 

& Urban Greening, 23, 35-43. DOI: 

https://doi.org/10.1016/j.ufug.2017.0

2.002  

Li, J. L., & Hu, Y. M. (2021). How does 

“shooting for the stars” come true? —

From the perspective of consistency 

https://doi.org/10.13397/j.cnki.fef.2012.06.012
https://doi.org/10.13397/j.cnki.fef.2012.06.012
https://doi.org/10.16382/j.cnki.1000-5560.2022.01.006
https://doi.org/10.16382/j.cnki.1000-5560.2022.01.006
https://doi.org/10.1007/BF00153949
https://doi.org/10.1016/j.ssresearch.2022.102747
https://doi.org/10.1016/j.ssresearch.2022.102747
https://doi.org/10.3390/bs13030217
https://doi.org/10.19512/j.cnki.issn2096-2088.2022.01.004
https://doi.org/10.19512/j.cnki.issn2096-2088.2022.01.004
https://doi.org/10.1007/s10212-021-00581-x
https://doi.org/10.1007/s10212-021-00581-x
https://doi.org/10.16477/j.cnki.issn1674-2311.2016.05.005
https://doi.org/10.16477/j.cnki.issn1674-2311.2016.05.005
https://doi.org/10.14331/ijfpss.2016.330058
https://doi.org/10.14331/ijfpss.2016.330058
https://doi.org/10.1016/j.ufug.2017.02.002
https://doi.org/10.1016/j.ufug.2017.02.002


Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2036 

and difference of parent-child educa-

tional expectation. Sociological Stud-

ies, 36(3), 204-224.  

Li, S. Y., & Liu, A. Y. (2022). “Melanchol-

ic Children”: Does cram school par-

ticipation lead to negative emotions? 

Causal inference based on Chinese 

Education Panel Survey (CEPS) data. 

Chinese Journal of Sociology, 42(2), 

60-93. DOI: 

https://doi.org/10.15992/j.cnki.31-

1123/c.2022.02.002  

Liu, H. Y., & Shan, S. (2025). Reflections 

on implementing the policy of “No 

less than 2 hours of comprehensive 

physical activity per day for primary 

and secondary school students”. 

Journal of Physical Education, 32(2), 

1-11. DOI: 

https://doi.org/10.16237/j.cnki.cn44-

1404/g8.20250221.001  

Mandrekar, J. N. (2010). Receiver operat-

ing characteristic curve in diagnostic 

test assessment. Journal of Thoracic 

Oncology, 5(9), 1315-1316. DOI: 

https://doi.org/10.1097/JTO.0b013e3

181ec173d  

Moreira-Morales, D. B., & García-Loor, M. 

I. (2024). Motivation in academic 

performance. International Research 

Journal of Management, IT & Social 

Sciences, 11(1), 30-38. DOI: 

https://doi.org/10.21744/irjmis.v11n1

.2403  

Nauzeer, S. & Jaunky, V.C. (2019). Moti-

vation and Academic Performance: A 

SEM Approach. International Journal 

of Environmental and Science Educa-

tion, 14(1), 41-60. Available at: 

http://www.ijese.net/makale/2102.ht

ml  

Pinquart, M., & Ebeling, M. (2020a). Pa-

rental educational expectations and 

academic achievement in children 

and adolescents—A meta-analysis. 

Educational Psychology Review, 

32(2), 463-480. DOI: 

https://doi.org/10.1007/s10648-019-

09506-z  

Pinquart, M., & Ebeling, M. (2020b). Stu-

dents’ expected and actual academic 

achievement - A meta-analysis. In-

ternational Journal of Educational 

Research, 100, 101524. DOI: 

https://doi.org/10.1016/j.ijer.2019.10

1524  

Podiya, J. K., Navaneetham, J., & Bhola, P. 

(2025). Influences of school climate 

on emotional health and academic 

achievement of school-going adoles-

cents in India: A systematic review. 

BMC Public Health, 25, 54. DOI: 

https://doi.org/10.1186/s12889-024-

21268-0  

Sanders, C. E., Field, T. M., & Diego, M. A. 

(2001). Adolescents’ academic ex-

pectations and achievement. Adoles-

cence, 36(144), 795-802. DOI: 

https://pubmed.ncbi.nlm.nih.gov/119

28883/  

Sewell, W. H., Haller, A. O., & Portes, A. 

(1969). The educational and early oc-

cupational attainment process. Amer-

ican Sociological Review, 34(1), 82-

92. DOI: 

https://doi.org/10.2307/2092789  

Sun, Y., Zhao, L., & Zhao, Z. (2025). 

Hukou status and children’s educa-

tion in China. Economic Develop-

ment and Cultural Change, 73(2), 

979-1021. DOI: 

https://doi.org/10.1086/729926  

Tan, L. F., & Zhang, Z. Y. (2025). The cur-

rent situation, problems, and optimi-

zation strategies for establishing new 

high-quality private schools: A case 

study of School X in Huanggang City. 

Asia-Pacific Education Research, 

1(2). DOI: 

https://doi.org/10.62177/aper.v1i2.33

3  

Yamamoto, Y., & Holloway, S. D. (2010). 

Parental expectations and children's 

academic performance in sociocul-

tural context. Educational Psychology 

https://doi.org/10.15992/j.cnki.31-1123/c.2022.02.002
https://doi.org/10.15992/j.cnki.31-1123/c.2022.02.002
https://doi.org/10.16237/j.cnki.cn44-1404/g8.20250221.001
https://doi.org/10.16237/j.cnki.cn44-1404/g8.20250221.001
https://doi.org/10.1097/JTO.0b013e3181ec173d
https://doi.org/10.1097/JTO.0b013e3181ec173d
https://doi.org/10.21744/irjmis.v11n1.2403
https://doi.org/10.21744/irjmis.v11n1.2403
http://www.ijese.net/makale/2102.html
http://www.ijese.net/makale/2102.html
https://doi.org/10.1007/s10648-019-09506-z
https://doi.org/10.1007/s10648-019-09506-z
https://doi.org/10.1016/j.ijer.2019.101524
https://doi.org/10.1016/j.ijer.2019.101524
https://doi.org/10.1186/s12889-024-21268-0
https://doi.org/10.1186/s12889-024-21268-0
https://pubmed.ncbi.nlm.nih.gov/11928883/
https://pubmed.ncbi.nlm.nih.gov/11928883/
https://doi.org/10.2307/2092789
https://doi.org/10.1086/729926
https://doi.org/10.62177/aper.v1i2.333
https://doi.org/10.62177/aper.v1i2.333


Yan & Li. (China). Factors Influencing Junior High School Students’ Academics. 

BECE, Vol.21, No.2, 2025 2037 

Review, 22(3), 189-214. DOI: 

https://doi.org/10.1007/s10648-010-

9121-z  

Yao, H., & Zhang, Y. (2022). “Screening” 

or “meritocracy”: Does the choice of 

a private school help students achieve 

higher academic performance?-Based 

on the PISA 2018 analysis of four 

provinces and municipalities in China. 

Frontiers of Education in China, 

17(2), 181-206. DOI: 

https://doi.org/10.3868/s110-007-

022-0009-6  

Zang, Y., Wu, H. B., & Cheng, H. Q. 

(2020). On the training path of pro-

fessional talents from the perspective 

of Astin’s I-E-O theory—In the case 

of the medical school of the Universi-

ty of Hong Kong and that of the Na-

tional University of Singapore. Edu-

cation Research Monthly, 2, 56-63. 

DOI: 

https://doi.org/10.16477/j.cnki.issn16

74-2311.2020.02.009  

Zhang, H. P., Liu, Y. C., Wu, X. H., & 

Zhang, Q. G. (2020). The influence 

of family-school partnerships on aca-

demic performance—The mediating 

role of student engagement. Educa-

tion Research Monthly, 1, 3-11. DOI: 

https://doi.org/10.16477/j.cnki.issn16

74-2311.2020.01.001  

Zhu, Y. (2023). Study of the household 

registration system and academic per-

formance of migrant workers’ chil-

dren. Academic Journal of Manage-

ment and Social Sciences, 3(3), 170-

173. DOI: 

https://doi.org/10.54097/ajmss.v3i3.1

1127  

Zysberg, L., & Schwabsky, N. (2020). 

School climate, academic self-

efficacy and student achievement. 

Educational Psychology, 41(4), 467-

482. DOI: 

https://doi.org/10.1080/01443410.202

0.1813690  

 

 

Received: September 16, 2025 

Revised: October 30, 2025 

Accepted: November 08, 2025

 

https://doi.org/10.1007/s10648-010-9121-z
https://doi.org/10.1007/s10648-010-9121-z
https://doi.org/10.3868/s110-007-022-0009-6
https://doi.org/10.3868/s110-007-022-0009-6
https://doi.org/10.16477/j.cnki.issn1674-2311.2020.02.009
https://doi.org/10.16477/j.cnki.issn1674-2311.2020.02.009
https://doi.org/10.16477/j.cnki.issn1674-2311.2020.01.001
https://doi.org/10.16477/j.cnki.issn1674-2311.2020.01.001
https://doi.org/10.54097/ajmss.v3i3.11127
https://doi.org/10.54097/ajmss.v3i3.11127
https://doi.org/10.1080/01443410.2020.1813690
https://doi.org/10.1080/01443410.2020.1813690

