A Study on the Latent Classes of Adolescent Risk Behaviors

A Study on the Latent Classes of Adolescent Risk Behaviors | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article A Study on the Latent Classes of Adolescent Risk Behaviors Jiawen Pu, Zhaoyan Luo, Hailin Xu, Aiqing Han, Yan Tang, Li Wang This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6616552/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Purpose: Adolescent health is gaining increased attention, necessitating systematic analysis of the diversity and complexity of adolescent risk behaviors among middle school students. This study aims to examine the data characteristics and latent categorical distributions of these behaviors, identify high-risk populations and associated factors, and provide evidence to inform targeted interventions and prevention strategies. Methods: The data used in our study are derived from China’s first public dataset on adolescent health. It is produced by a project that has carried out a multiwave survey of 99,327 middle and high school students from 2015–2021[ 1 ][ 2 ]. In our study, taxometrics and latent class analysis (LCA) were used to study the latent classes of risk behaviors of middle school students (n = 14,662) in a cross-sectional study conducted in 2020. Results: The results of taxometrics indicate that adolescent risk behaviors of middle school students present a continuous dimension. Additionally, we further explored the latent classes of the risk–behavior groups. There are four latent classes: (1) high-risk behavior group (13.62%); (2) moderate-risk behavior group (athletics sport) (17.17%); (3) moderate-risk behavior group (electronic products) (14.16%); and (4) low-risk behavior group (55.05%). The high-risk group is driven by multiple cooccurring factors, whereas the moderate-risk subgroup is correlated primarily with physical inactivity or electronic overuse, which are linked to environmental, familial, and cognitive influences. Conclusions: Adolescent risk behaviors in middle school students form four latent classes: the predominant low-risk group, a high-risk group requiring prioritized intervention for multiple factors, and a moderate-risk group where physical inactivity outweighs electronic product use as a risk. Additionally, smoking and drinking were significantly related. adolescence risk behavior taxometrics latent class analysis Chinese adolescents Figures Figure 1 Figure 2 Figure 3 Introduction Mental health problems directly or indirectly affect adolescents’ physical and mental state and quality of life. With the development of science and technology and the complexity of the social environment, the factors affecting the mental health of adolescents are increasing. Adolescents are in a critical period of development. Adolescent risk behaviors are closely related to family, school, social and personal factors [ 3 ]. These factors include both risks and protective elements. Mental illness and adverse consequences such as violence and drug abuse may arise when these elements are imbalanced [ 4 ][ 5 ][ 6 ][ 7 ]. Adolescents with different cognitive levels take different self-protection measures. However, external support is generally needed to make their development safe and sound [ 8 ][ 9 ]. In the field of adolescent risk behavior research, the Youth Risk Behavior Surveillance System (YRBSS) conducts a survey every two years, including six kinds of priority health risk behaviors: intentional injury, tobacco use, alcohol and other drug use, sexual behaviors, unhealthy diet and insufficient physical exercise [ 10 ][ 11 ][ 12 ][ 13 ]. This provides powerful data support for subsequent research. Previous studies have demonstrated that psychological symptoms such as anxiety and depression are associated with mobile phone use, unhealthy weight-control behavior, drinking and smoking [ 13 ][ 14 ][ 15 ][ 16 ]. There are also studies that have used LCA to explore the characteristics of specific risk factors [ 17 ]. However, most studies concentrate on single or similar behaviors [ 18 ]. Although an increasing number of studies have used LCA to identify latent classes of adolescent risk behavior, studies focused on middle school students, especially those using LCA after group type identification in combination with taxometrics, are still rare. For example, a study has shown that adolescent risk behaviors can be classified into several major classes. These have been classified into four latent types: low-risk class, moderate-risk class 1 (smoking/alcohol use (AU)/screen time (ST)), moderate-risk class 2 (unhealthy weight loss (ULW)/problematic mobile phone use (PMPU)), and high-risk class [ 18 ]. This comprehensive method can describe the behavior characteristics of each group more accurately. Our study focuses on the adolescent risk behaviors of middle school students. On the basis of the cross-sectional data collected from 14,662 middle school students, we constructed risk subgroups by data type identification and LCA. Our study aims to (1) analyze the dimension structure characteristics of risk behaviors and (2) reveal the correlation mode of risk factors by latent class probability distribution and the risk characteristics of risk behaviors in middle school students by determining their heterogeneity to provide an evidence-based basis for targeted interventions. Methods Study sample and measures Study data and samples are collected from the Database of Youth Health (DYH), and a total of 99,327 students from 186 secondary schools in 17 cities of Shandong Province participated in the survey [ 2 ]. Our study focuses on the analysis of the risk behavior data of 14,662 middle school students in the 2020 risk behavior questionnaire. The questionnaire is adapted from the questionnaires of state, national and local schools covered by the YRBSS [ 10 ]. It has undergone strict reliability and validity tests and accords with the relevant research standards [ 19 ]. The questionnaire uses the method of state and local adolescent risk behavior surveys to measure risk behaviors. The method has been revised, adjusted and translated into Chinese, which is suitable for Chinese adolescents. In our study, the number of outliers and missing values was very small. The total number of samples is less than 10. Owing to the large sample size, our study deleted the outlier and missing value samples. Ultimately, we preserved 14,662 valid samples for later analysis. This process does not significantly affect the accuracy or reliability of the research results. Statistical analysis Our study adopts descriptive statistical methods to preliminarily organize, summarize and present the data. We understand the basic characteristics and distribution rules of the data comprehensively and systematically. We check the data quality, identify outliers and extreme situations and complete data cleaning. These findings provide the basis and reference for subsequent analysis. Taxometric analysis: In the process of taxonomic analysis, R version 4.4.2 was selected as the analysis tool and was carried out via the RTaxometrics package [ 25 ]. Additionally, we follow the default settings of Ruscio to a certain extent [ 25 ][ 22 ]. Specifically, we set the random number of seeds to 123. Limited by equipment conditions, we adjust the parameters of the default settings slightly during the process of MAXEIG analysis. Nevertheless, this adjustment does not affect the overall analysis structure or conclusions. Owing to the simplicity of the measured data, our study uses these variables as indicators for taxometric analysis. Taxometric analysis aims to clarify whether the studied structure is essentially a dimensional structure or a categorical structure [ 20 ]. To ensure the accuracy and reliability of the results, it is usually necessary to comprehensively use 2–3 methods for in-depth exploration. We adopt three methods, MAMBAC, MAXEIG and L-MODE, to conduct a detailed analysis of the internal structure of risk behaviors in middle school students. Specifically, if the output graphs of MAMBAC and MAXEIG are peak curves and L-MODE generates a double- or multipeak curve, the analysis results tend to be categorical; in contrast, they are more inclined to be dimensional [ 21 ]. In addition, when evaluating the structure type, the comparison curve fit index (CCFI) and base rate estimation indicators generated by the analysis results need to be considered comprehensively [ 22 ][ 23 ][ 24 ]. Latent class analysis In the second stage, we use SPSS 22.0 to convert the data format and carry out classification and assignment according to the degree of influence of risk behavior factors: 0 for options that do not cause risk behaviors or reduce them, and 1 for options that may cause risk behaviors or increase them [ 24 ][ 26 ]. Then, we carry out LCA via MPLUS 8.3. An LCA is a method used to describe the latent classes of variables. It is widely used in the analysis of latent characteristics and groups. Additionally, it can analyze the group characteristics of dimensional data according to the required research direction. In particular, in cross-sectional studies, it is commonly used in group classification under observation indices. In this study, the sample size was large enough for the LCA to yield stable results. For this exploratory method, it is necessary to increase the number of fitting classes so that more models will be generated. Then, we select the best model for discussion on the basis of the corresponding evaluation indices of the model fitting effect [ 27 ]. In the evaluation of the model fitting effect, the smaller the values of the AIC (Akaike information criterion), BIC (Bayesian information criterion) and aBIC (corrected BIC) are, the better the model fitting effect [ 18 ][ 28 ][ 29 ]. Among them, the BIC is considered to be the most suitable evaluation index in most cross-sectional studies. The bootstrapped likelihood ratio test (BLRT) and Vuong-Lo‒Mendell‒Rubin (VLMR) are used to evaluate the accuracy of classification. Specifically, an entropy < 0.60 means that more than 20% of individuals have classification errors; an entropy ≥ 0.80 means that the classification accuracy is greater than 90%, and the individual classification in the sample is good [ 28 ]. P values < 0.05 for BLRT and VLMR suggest that the K classification may perform better than the K-1 classification does, but it is not necessarily significant and should be interpreted in conjunction with practical considerations [ 16 ]. In our study, each class sample size was ≥ 50; therefore, the aBIC is the information index with the highest classification accuracy [ 28 ]. Results Indicator analyses The results of the descriptive statistics in Table 1 show that there are significant differences in the mean value, discrete degree and distribution pattern of each variable. Variables with high means and low dispersion (such as Q3–Q6) reflect the universality and stability of the behaviors. Variables with low means and high dispersion (such as Q2 and Q7–Q12) suggest that these behaviors are not common in the sample, but there are extreme cases. The analysis of skewness and kurtosis further reveals the morphological characteristics of the data distribution, which provides an important reference for subsequent statistical modeling and analysis. Because the variables may belong to the possible influencing factors of risk behavior, all the indicators are included in the following study. Table 1 Descriptive Statistics of the Variables (n = 14662) Variable Mean Standard Deviation Variance Skewness Kurtosis Statistics Standard Error Statistics Statistics Statistics Statistics Q1 3.75 0.012 1.433 2.054 -0.732 -0.909 Q2 1.34 0.007 0.891 0.794 2.773 6.850 Q3 1.85 0.003 0.357 0.127 -1.961 1.845 Q4 1.87 0.003 0.341 0.116 -2.151 2.625 Q5 1.85 0.003 0.361 0.130 -1.922 1.693 Q6 1.90 0.002 0.301 0.091 -2.648 5.015 Q7 1.48 0.011 1.275 1.625 2.844 7.312 Q8 1.36 0.009 1.096 1.202 3.462 11.830 Q9 1.35 0.009 1.082 1.170 3.462 11.800 Q10 1.48 0.011 1.313 1.725 3.007 8.295 Q11 1.71 0.013 1.585 2.514 2.192 3.555 Q12 1.47 0.011 1.305 1.702 2.908 7.595 Q13 3.30 0.018 2.162 4.676 0.718 -0.547 Q14 1.93 0.013 1.529 2.339 1.748 2.124 Q15 2.11 0.013 1.526 2.327 1.614 1.932 Q16 3.16 0.011 1.328 1.764 0.276 -0.275 Q17 1.65 0.007 0.868 0.753 1.218 0.632 Q18 1.77 0.011 1.278 1.633 1.599 1.243 Q19 3.23 0.012 1.489 2.217 -0.214 -1.331 Q20 3.19 0.012 1.455 2.117 -0.179 -1.294 Q21 2.64 0.011 1.385 1.918 0.339 -1.097 Q22 1.73 0.010 1.178 1.388 1.565 1.349 Q23 2.54 0.012 1.508 2.275 0.388 -1.382 Note: Q1: frequency of seat belt use in vehicles; Q2: number of trips in a vehicle with a drunk driver; Q3: whether the driver was bullied on campus; Q4: whether the driver was bullied on the Internet; Q5: whether the driver was sad and helpless; Q6−Q9: smoking; Q10−Q12: drinking; Q13: number of exercise days in the past 7 days; Q14: time spent watching TV each week; Q15: time spent playing games each week; Q16−Q18: time spent on PE lessons within or outside the curriculum; Q19: frequency of fathers exercising each week; Q20: frequency of exercise with parents; Q22: frequency of dizziness in physical exercise; and Q23: degree of tooth cleaning. Taxometrics Standard classification quantitative analysis was the primary procedure of our study. Owing to the default setting of some parameters, each variable is used as the input variable in turn, and the remaining variables are used as the output variables to ensure that the results are convincing during the MAMBAC and MIXEIG analyses. If there are k variables, k×(k − 1) input‒output variable pairs are generated. In the case of a large sample size of 14662, the empirical basis rate is set to 50% to obtain more stable results and explanations. The results are shown in Fig. 1. Both the average MAMBAC curve and MAXEIG curve do not present obvious peaks; meanwhile, the L-MODE curve presents a single peak. In addition, the shape of each curve is similar to the dimension curve. The objective CCFI indices of MAMBAC, MAXEIG and L-MODE are shown in the Supplementary Materials. The CCFI values of the three programs fall into the fuzzy [0.45, 0.55] interval. The CCFI indices of MAMBAC and L-MODE are less than 0.50. The classification ratios generated by base rate estimates are relatively low, which provides additional evidence for the dimensional structure. Latent class analysis Through the evaluation of the relevant indicators of the fitting effect, four types of latent structure classification are found to be the optimal scheme. As shown in Table 2 , as the number of classes increases, the LL, AIC, BIC, aBIC and entropy indicators gradually decrease. The entropy values are greater than 0.8, and the P values of BLRT and VLMR are less than 0.05, indicating an ideal model fitting effect. However, in terms of the downward trend of the aBIC index and the steep slope figure of the BIC, after the four types of models, the change in the BIC tends to be gradual. In addition, in the classification process, the class probability should not be too small (greater than 10%); otherwise, the class will be of no practical significance. Considering the above factors, we select the four-class model whose class probability has no subclass, and all the indicators reach the standard. Table 2 Identification of the latent class model Class LL AIC BIC aBIC Entropy BLRT VLMR Class Probability (%) P P 2 -1513 80.846 30285 5.693 30321 2.565 30306 3.203 0.963 < 0.001 < 0.001 17.27/82.73 3 -1443 75.984 28889 3.967 28943 3.071 28920 7.439 0.929 < 0.001 < 0.001 65.65/16.17 /18.18 4* -1423 13.315 28481 6.63 28553 7.966 28523 6.064 0.863 < 0.001 < 0.001 13.62/17.17/ 14.16/55.05 5 -1408 66.916 28197 1.832 28287 5.401 28249 7.229 0.864 < 0.001 < 0.001 14.41/14.29/ 12.50/54.32/4.48 Note: LL: log likelihood; AIC: Akaike information criterion; BIC: Bayesian information criterion; aBIC: corrected BIC; BLRT: bootstrapped likelihood ratio test (K−1 vs. K classes); VLMR: Vuong−Lo−Mendell−Rubin (K−1 vs. K classes); *: optimal fitting model . Figure 2 depicts the conditional probability distribution of four latent classes of adolescent risk behaviors in middle school students. Moreover, it can intuitively show the distribution of risk factors for risk behaviors. Among the four classes, the major class can be divided into a high-risk group, a moderate-risk group and a low-risk group. The radar map clearly shows that the risk factors for the two moderate-risk groups are different. The characteristic variable of one of the moderate-risk groups is closely related to physical exercise and the number of parental exercises. The conditional probabilities are Q17 (0.801), Q18 (0.837), Q19 (0.839), Q20 (0.851) and Q21 (0.928). However, the characteristic variable of another moderate-risk group is related to the use of electronic products, and its conditional probability is Q15 (0.726). In addition, the probabilistic estimation range of the high-risk group is mostly above 0.69, and each variable tends to have a certain risk impact. The probabilistic estimation range of the low-risk group is mostly less than 0.3, and the risk factors are relatively small. Discussion Our study is the first to explore the latent structure of adolescent risk behaviors among middle school students in China through the combination of taxometrics and LCA. First, the analysis of the specific variables in Table 1 reveals that the safety-related behaviors are as follows: the mean value of Q1 (frequency of seat belt use in vehicles) is relatively high (3.75), the distribution is left skewed (skewness: -0.732), and the kurtosis is relatively low (-0.909), indicating that most respondents wear seat belts in vehicles, but some of them still do not obey this rule. The mean value of Q2 (the number of riders riding in a vehicle with a drunk driver) is low (1,34), the distribution is right skewed (skewness: 2.773), and the kurtosis is high (6.850), indicating that most respondents rarely ride in a vehicle with a drunk driver, but there are extreme values. Mental health and bullying: The mean values of Q3–Q5 are low (1.85–1.87), and the skewness and kurtosis are close to 0, indicating that these phenomena are not common in the sample and that the distribution is relatively symmetrical. Smoking and drinking: The mean values of Q6–Q12 are low (1.35–1.71), but the skewness and kurtosis are high, indicating that these behaviors are not common in the sample, but there are a few extreme cases. Physical exercise and recreational activities: The mean values of Q13-Q15 are high (1.93–3.30), and the skewness and kurtosis are close to 0, indicating that the distribution of these variables is close to a normal distribution. Family exercise participation: The mean value of Q19‒Q21 is high (2.64‒3.23), and the skewness and kurtosis are low, indicating that parents’ exercise is relatively common and symmetrically distributed in the sample. Exercise and health: The mean value of Q22 is low (1.73), and the skewness and kurtosis are low, indicating that the respondents seldom feel dizzy during exercise. The mean value of Q23 is high (2.54), and the skewness and kurtosis are low, indicating that the respondents have good tooth cleaning habits. Second, the existing state of risk behaviors in Chinese middle school students is evaluated via taxometric analysis. The results demonstrate that the average CCFI value is less than 0.5 and that the base rate estimate value is relatively low, indicating that the risk behaviors of middle school students present continuous dimensional distribution characteristics. This result is consistent with a previous study on the risk behaviors of Chinese high school students [ 30 ]. Therefore, it may support latent classification studies. On the basis of the above results, LCA divides the research samples into four latent classes: approximately 13.62% of the research samples are assigned to class 1, 17.17% are assigned to class 2, 14.16% are assigned to class 3, and 55.05% are assigned to class 4. As shown in Fig. 2, class 1 has a high conditional probability for most variables, and only variables related to physical exercise have a low risk level, so it is defined as a high-risk group. Class 4 has a low conditional probability for all the variables, so it is classified into a low-risk group. Although both class 2 and class 3 belong to the moderate-risk group, there are distinct differences in their risk behavior tendencies. Among them, class 2 is related mainly to physical exercise-related risk behaviors, such as the frequency of participating in extracurricular PE lessons and exercising with parents. This result reflects the influence of the living environment and the school and social environments on the psychological and cognitive development of middle school students. Moreover, this study highlights the importance of family companionship, which is consistent with existing research conclusions [ 31 ][ 32 ][ 33 ]. Class 3 is related mainly to the risk associated with electronic product use, in which playing video games is the major risk factor. While the duration of watching TV can also be a risk factor [ 34 ] (57.2%), it is significantly lower than similar indicators in the high-risk group (69.2%). Therefore, the variable of watching TV is included in the high-risk group for analysis. In the high-risk group, risk behavior variables with a conditional probability greater than 60% include the following: Q1 (frequency of seat belt use in vehicles), Q2 (number of vehicles riding with a drunk driver), Q7–Q9 (smoking), Q10–Q12 (drinking), Q14 (time spent watching TV each week) and Q22 (frequency of dizziness in physical exercise). Combined with the distribution characteristics of risk factors in the two moderate-risk groups, our study revealed that the risk behaviors of middle school students are related mainly to accidents, smoking and drinking, the use of electronic products, physical exercise and the family environment. This finding is highly consistent with previous research results [ 35 ][ 18 ][ 19 ]. Compared with high school students [ 30 ][ 36 ][ 37 ][ 3 ], middle school students are more vulnerable to environmental factors and material temptations (such as electronic products, tobacco and alcohol) because of their younger age, immature psychological cognition level and incompletely established mental health standards. In addition, our study revealed that there is a certain correlation between the risk behavior variables. As shown in Fig. 3, the correlation coefficient between Q7‒Q9 (smoking-related variables) and Q10‒Q12 (drinking-related variables) is 0.4‒0.7, indicating a moderate degree of correlation. These findings indicate that smoking and drinking may be related to material temptations in early adolescence and that these temptations may increase the probability of risk behaviors [ 18 ]. Moreover, the correlation coefficient between Q19 (frequency of fathers’ exercise each week) and Q20 (frequency of mothers’ exercise each week) was 0.7, indicating a relatively high degree of correlation. This finding indicates that there is a significant correlation between fathers’ exercise and mothers’ exercise. Additionally, these findings suggest that a positive family environment plays an important role in promoting the sound growth of adolescents. Conclusions Notably, our study uses a sample of middle school students in a single province in China to carry out dimensional analysis and latent class identification of risk behaviors. Therefore, the universality of these conclusions may be limited because they fail to fully reflect the characteristics of adolescent risk behaviors in other cultural backgrounds. Although our study does not explore the influence mechanism and causality between the variables further, it still has the following value: (1) Through multidimensional analysis, the heterogeneity of risk behaviors in middle school students is revealed, which provides a methodological reference for the classification and identification of risk behaviors. (2) The distribution rules of risk factors such as smoking, drinking and electronic product use are clarified, which provides an empirical basis for the adjustment of risk behavior prevention strategies for middle school students. Declarations Ethics approval and consent to participate Ethical approval was not needed, as this study used publicly available data. Consent for publication This manuscript has not been published or presented elsewhere in part or in entirety, and is not under consideration by another journal. All the authors have approved the manuscript and agree with submission to your esteemed journal. Availability of data and materials Database: Shandong University. Database of Youth Health. Population Health Data Archive PHDA,2021. https://doi.org/10.12213/11.A0031.202107.209.V1.0. Competing Interests I declare that the authors have no competing interests as defined by BMC, or other interests that might be perceived to influence the results and/or discussion reported in this paper. Funding This work was supported by the funds of two horizontal projects from Beijing University of Chinese Medicine(BUCM-2021-JS-FW-031)and(BUCM-2024-JS-FW-137) Authors' contributions Jiawen Pu: Writing-review & editing, Writing-original draft, Visualization, Software, Methodology, Formal analysis, Data curation, Conceptualization. Zhaoyan Luo: Writing-review & translating. Hailin Xu: Methodology, Conceptualization. Aiqing Han: Writing-review & editing, Supervision, Methodology. Yan Tang: Writing-review & editing, Supervision, Methodology, Funding acquisition, Conceptualization. Li wang: Writing-review & editing, Supervision, Methodology, Funding acquisition, Conceptualization. Acknowledgments This work was supported by the funds of two horizontal projects from Beijing University of Chinese Medicine(BUCM-2021-JS-FW-031)and(BUCM-2024-JS-FW-137). We express our sincere gratitude to the National Population Health Data Center (NPHDC) for the data. The data is collected from Population Health Data Archive (PHDA). The research is instructed by Li Wang, Yan Tang, and Aiqing Han. And we thank Zhaoyan Luo for the translation. References Database. Shandong University. Database of Youth Health. Population Health Data Archive PHDA,2021. https://doi.org/10.12213/11.A0031.202107.209.V1.0 Zhang S, Luo W, Dong X, Chen W, Yi X, Zhou W, Zhang Y, Zhao Y. A Dataset on the Status Quo of Health and Health-Related Behaviors of Chinese Youth: A Longitudinal Large-Scale Survey in the Secondary School Students of Shandong Province, Chinese Medical Sciences Journal, Volume 37, Issue 1,2022, Pages 60–66, ISSN 1001–9294. https://doi.org/10.24920/004051 Resnick MD, Bearman PS, Blum RW, Bauman KE, Harris KM, Jones J, Tabor J, Beuhring T, Sieving RE, Shew M, Ireland M, Bearinger LH, Udry JR. Protecting adolescents from harm. Findings from the National Longitudinal Study on Adolescent Health. JAMA. 1997;278(10):823–32. 10.1001/jama.278.10.823 . Vidal C, Meshi D. Behavioral Addictive Disorders in Children and Adolescents. J Am Acad Child Adolesc Psychiatry. 2023;62(5):512–4. 10.1016/j.jaac.2022.07.819 . Epub 2022 Aug 22. Mellins CA, Elkington KS, Bauermeister JA, Brackis-Cott E, Dolezal C, McKay M, Wiznia A, Bamji M, Abrams EJ. Sexual and drug use behavior in perinatally HIV-infected youth: mental health and family influences. J Am Acad Child Adolesc Psychiatry. 2009;48(8):810–9. 10.1097/CHI.0b013e3181a81346 . Prinstein MJ, Boergers J, Spirito A. Adolescents' and their friends' health-risk behavior: factors that alter or add to peer influence. J Pediatr Psychol. 2001 Jul-Aug;26(5):287–98. 10.1093/jpepsy/26.5.287 . Hojat M. Protecting adolescents from harm. JAMA. 1998;279(5):353–4. 10.1001/jama.279.5.353 . Fergus S, Zimmerman MA. Adolescent resilience: a framework for understanding healthy development in the face of risk. Annu Rev Public Health. 2005;26:399–419. 10.1146/annurev.publhealth.26.021304.144357 . Rew L, Horner SD. Youth Resilience Framework for reducing health-risk behaviors in adolescents. J Pediatr Nurs. 2003;18(6):379–88. 10.1016/s0882-5963(03)00162-3 . Youth Risk Behavior Surveillance System. (YRBSS) https://www.cdc.gov/yrbs/data/index.html Brener ND, Kann L, Kinchen SA, Grunbaum JA, Whalen L, Eaton D, Hawkins J, Ross JG. Methodology of the youth risk behavior surveillance system. MMWR Recomm Rep. 2004;53(RR–12):1–13. https://pubmed.ncbi.nlm.nih.gov/15385915/ . Youth Risk Behavior Surveillance: National College Health Risk Behavior Survey–United States. 1995. MMWR CDC Surveill Summ. 1997;46(6):1–56. https://pubmed.ncbi.nlm.nih.gov/9393659/ Tingting Gao J, Li H, Zhang J, Gao Y, Kong Y, Mei HS. The influence of alexithymia on mobile phone addiction: The role of depression, anxiety and stress. J Affect Disord. 2018;225:761–6. https://doi.org/10.1016/j.jad.2017.08.020 . Li Y, Li G, Liu L, Wu H. Correlations between mobile phone addiction and anxiety, depression, impulsivity, and poor sleep quality among college students: A systematic review and meta-analysis. J Behav Addict. 2020;9(3):551–71. 10.1556/2006.2020.00057 . Wall DN-SM, Larson NI, Eisenberg ME. Katie Loth, Dieting and Disordered Eating Behaviors from Adolescence to Young Adulthood: Findings from a 10-Year Longitudinal Study. J Am Diet Assoc, 111, Issue 7, 2011, Pages 1004–1011, ISSN 0002-8223, https://doi.org/10.1016/j.jada.2011.04.012 Tomczyk S, Isensee B. Reiner Hanewinkel,Latent classes of polysubstance use among adolescents—a systematic review,Drug and Alcohol Dependence,160, 2016, Pages 12–29, ISSN 0376–8716, https://doi.org/10.1016/j.drugalcdep.2015.11.035 Choi AY, Israel T, Nylund-Gibson K. Syndemic behavioral risk and suicidality among bisexual adolescents: A latent class analysis. J Consult Clin Psychol. 2020;88(7):597–612. 10.1037/ccp0000500 . Jie H, Yang R, Li D, Zhao S, Wan Y, Tao F, Fang J, Zhang S. A latent class analysis of psychological symptoms and health risk behaviors among Chinese adolescents. Asian J Psychiatry 55. 2021;102518:1876–2018. https://doi.org/10.1016/j.ajp.2020.102518 . Yi X, Liu Z, Qiao W, et al. Clustering effects of health risk behavior on mental health and physical activity in Chinese adolescents. Health Qual Life Outcomes. 2020;18(1):211. 10.1186/s12955-020-01468-z . Ruscio J. Shirley B. Wang,3.11 - Taxometric Analysis,Editor(s): Gordon J.G. Asmundson,Comprehensive Clinical Psychology (Second Edition),Elsevier,2022,Pages 148–175,ISBN 9780128222324. https://doi.org/10.1016/B978-0-12-818697-8.00217-X Fen Ren R, Zhou X, Zhou SC, Schneider EA. Storch,The latent structure of olfactory reference disorder symptoms: A taxometric analysis,Journal of Obsessive-Compulsive and Related Disorders,27,2020,100583,ISSN 2211–3649, https://doi.org/10.1016/j.jocrd.2020.100583 Ruscio J, Ruscio AM, Meron M. Applying the Bootstrap to Taxometric Analysis: Generating Empirical Sampling Distributions to Help Interpret Results. Multivar Behav Res. 2007;42(2):349–86. https://doi.org/10.1080/00273170701360795 . Walters GD, Ruscio J. Trajectories of Youthful Antisocial Behavior: Classes or Continua? J Abnorm Child Psychol. 2013;41:653–66. https://doi.org/10.1007/s10802-012-9700-1 . Sinha P, Calfee CS, Delucchi KL. Practitioner's Guide to Latent Class Analysis: Methodological Considerations and Common Pitfalls. Crit Care Med. 2021;49(1):e63–79. 10.1097/CCM.0000000000004710 . https:// cran.r-project.org/web/packages/RTaxometrics/index.html Weller BE, Bowen NK, Faubert SJ. Latent Class Analysis: A Guide to Best Practice. J Black Psychol. 2020;009579842093093. 10.1177/0095798420930932 . McCutcheon AL. Latent class analysis. SAGE Publications, Inc.; 1987. https://doi.org/10.4135/9781412984713 . Ruscio J, Carney LM, Dever L, Pliskin M, Wang SB. Using the comparison curve fix index (CCFI) in taxometric analyses: Averaging curves, standard errors, and CCFI profiles. Psychol Assess. 2018;30(6):744–54. https://doi.org/10.1037/pas0000522 . Natacha Carragher G, Adamson B, Bunting. Siobhan McCann,Subtypes of depression in a nationally representative sample, Journal of Affective Disorders, 113, Issues 1–2, 2009, Pages 88–99, ISSN 0165–0327, https://doi.org/10.1016/j.jad.2008.05.015 Mengcheng Wang. Measurement and Latent Structure of Adolescent Health Risk Behaviors [D]. Cent South Univ. 2013. 10.16128/j.cnki.1005-3611.2013.05.031 . Becker SP, Luebbe AM, Langberg JM. Co-occurring mental health problems and peer functioning among youth with attention-deficit/hyperactivity disorder: a review and recommendations for future research. Clin Child Fam Psychol Rev. 2012;15(4):279–302. 10.1007/s10567-012-0122-y . Braun D, Lascelles K. Involving and supporting families, friends, and carers during a mental health crisis. Lancet Psychiatry. 2024;11(8):586–7. 10.1016/S2215-0366(24)00165-2 . Epub 2024 Jun 11. Mollica RF, Donelan K, Tor S, Lavelle J, Elias C, Frankel M, Blendon RJ. The effect of trauma and confinement on functional health and mental health status of Cambodians living in Thailand-Cambodia border camps. JAMA. 1993;270(5):581–6. https://pubmed.ncbi.nlm.nih.gov/8331755/ . Lemola S, Perkinson-Gloor N, Brand S, Dewald-Kaufmann JF, Grob A. Adolescents' electronic media use at night, sleep disturbance, and depressive symptoms in the smartphone age. J Youth Adolesc. 2015;44(2):405–18. 10.1007/s10964-014-0176-x . Epub 2014 Sep 10. Riesch SK, Kedrowski K, Brown RL, Temkin BM, Wang K, Henriques J, Jacobson G, Giustino-Kluba N. Health-risk behaviors among a sample of US preadolescents: types, frequency, and predictive factors. Int J Nurs Stud. 2013;50(8):1067-79. 10.1016/j.ijnurstu.2012.10.012 . Epub 2012 Nov 21. Erratum in: Int J Nurs Stud. 2020;109:103739. doi: 10.1016/j.ijnurstu.2020.103739. Schulenberg J, Bachman JG, O'Malley PM, Johnston LD. High school educational success and subsequent substance use: a panel analysis following adolescents into young adulthood. J Health Soc Behav. 1994;35(1):45–62. https://pubmed.ncbi.nlm.nih.gov/8014429/ . O'Malley PM, Johnston LD. Driving after drug or alcohol use by US high school seniors, 2001–2011. Am J Public Health. 2013;103(11):2027–34. 10.2105/AJPH.2013.301246 . Epub 2013 Sep 12. Additional Declarations No competing interests reported. Supplementary Files BICSteepSlopeFigure.xlsx SummarizeCCFIandBaseRateEstimates.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6616552","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":475130984,"identity":"fdefb97b-faed-46e8-89b5-c0c2be60b2c3","order_by":0,"name":"Jiawen Pu","email":"","orcid":"","institution":"Beijing University of Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Jiawen","middleName":"","lastName":"Pu","suffix":""},{"id":475130985,"identity":"db97de1d-ecee-4265-89cf-16a8b5bf3759","order_by":1,"name":"Zhaoyan Luo","email":"","orcid":"","institution":"Beijing University of Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Zhaoyan","middleName":"","lastName":"Luo","suffix":""},{"id":475130986,"identity":"ab240ffd-13c2-4071-a053-706ae0ce906c","order_by":2,"name":"Hailin Xu","email":"","orcid":"","institution":"Beijing University of Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Hailin","middleName":"","lastName":"Xu","suffix":""},{"id":475130987,"identity":"38974a3f-0730-4d6f-8119-1bebd57d0893","order_by":3,"name":"Aiqing Han","email":"","orcid":"","institution":"Beijing University of Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Aiqing","middleName":"","lastName":"Han","suffix":""},{"id":475130988,"identity":"06c0aec2-e677-4e15-8e70-d9cf1f8c4a91","order_by":4,"name":"Yan Tang","email":"","orcid":"","institution":"Beijing University of Chinese Medicine","correspondingAuthor":false,"prefix":"","firstName":"Yan","middleName":"","lastName":"Tang","suffix":""},{"id":475130989,"identity":"80ab301c-fa69-4926-aab5-32b29bbeb5c1","order_by":5,"name":"Li Wang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAx0lEQVRIiWNgGAWjYLCCBwY2CRAWG7FaEgzSSNbCcJgELQbHzx5+kVBwPs/g+OkEhg9lhxn4ZzcQ0HImL80iweB2scGZ3A2MM84dZpC4cwC/FrMDOWYGQC2JG27wbmDmbTvMYCCRQEDL+TcgLecgWv4SpeVGjvGDBIMDEC2MxGixv/HGDBjIycWSQL8c7DmXziNxg4AWyf4c4w8f/tjl8R0/u/HBjzJrOf4ZBLQAAZsEjHUAiHkIqgcC5g/EqBoFo2AUjIIRDADoYEkw/PFyhwAAAABJRU5ErkJggg==","orcid":"","institution":"Beijing University of Chinese Medicine","correspondingAuthor":true,"prefix":"","firstName":"Li","middleName":"","lastName":"Wang","suffix":""}],"badges":[],"createdAt":"2025-05-08 04:23:21","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6616552/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6616552/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":85394468,"identity":"a290c4b7-241d-4c6c-acf1-6532714fedfa","added_by":"auto","created_at":"2025-06-25 10:57:16","extension":"jpeg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":488297,"visible":true,"origin":"","legend":"\u003cp\u003eTaxometric analysis results\u003c/p\u003e\n\u003cp\u003eNote: The average MAMBAC curve, MAXEIG curve and L-MODE curve (dark curve) of the risk behavior data, simulated classification data (left) and simulated dimension data (right). In each graph, the gray curve represents the middle 50% of the simulated value, and the two light color curves represent the maximum and minimum simulated classification and dimension values.\u003c/p\u003e","description":"","filename":"floatimage1.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6616552/v1/0aee97a399fc13bf15c0d352.jpeg"},{"id":85396310,"identity":"5cd3c241-6ffa-4d6b-9c5c-bfb9ad66f38e","added_by":"auto","created_at":"2025-06-25 11:13:16","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":63175,"visible":true,"origin":"","legend":"\u003cp\u003eProbability distributions of the four classes\u003c/p\u003e\n\u003cp\u003eNote: Q1: frequency of seat belt use in vehicles; Q2: number of trips in a vehicle with a drunk driver; Q3: whether the driver was bullied on campus; Q4: whether the driver was bullied on the Internet; Q5: whether the driver was sad and helpless; Q6-Q9: smoking; Q10-Q12: drinking; Q13: number of exercise days in the past 7 days; Q14: time spent watching TV each week; Q15: time spent playing games each week; Q16-Q18: time spent on PE lessons within or outside the curriculum; Q19: frequency of fathers exercising each week; Q20: frequency of exercise with parents; Q22: frequency of dizziness in physical exercise; and Q23: degree of tooth cleaning.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-6616552/v1/0e593a35fc9ee5f21272b2b1.png"},{"id":85394474,"identity":"1e63dc53-122c-487f-98c4-03a76a791113","added_by":"auto","created_at":"2025-06-25 10:57:16","extension":"jpeg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":310498,"visible":true,"origin":"","legend":"\u003cp\u003eHeatmap of the Variables\u003c/p\u003e\n\u003cp\u003eNote: Q1: frequency of seat belt use in vehicles, Q2: number of riding in a vehicle with a drunk driver, Q3: whether being bullied on campus, Q4: whether being bullied on the Internet, Q5: whether feeling sad and helpless, Q6-Q9: smoking, Q10-Q12: drinking, Q13: number of exercise days in the past 7 days, Q14: time spent on watching TV each week, Q15: time spent on playing games each week, Q16-Q18: time spent on PE lessons within or out of the curriculum, Q19: frequency of father’s exercise each week, Q20: frequency of mother’s exercise each week, Q21: frequency of exercise with parents, Q22: frequency of dizziness in physical exercise, Q23 : tooth cleaning\u003c/p\u003e","description":"","filename":"floatimage3.jpeg","url":"https://assets-eu.researchsquare.com/files/rs-6616552/v1/9412d670b408084b273a0abf.jpeg"},{"id":105565452,"identity":"ca746026-2e6b-4c96-81e6-4cf2ede357cf","added_by":"auto","created_at":"2026-03-27 12:53:17","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1583014,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6616552/v1/2f23c3d9-fa6d-4397-931c-ea9080501e9e.pdf"},{"id":85395015,"identity":"cafd321b-6ec8-4bfb-9850-84553dd31458","added_by":"auto","created_at":"2025-06-25 11:05:16","extension":"xlsx","order_by":0,"title":"","display":"","copyAsset":false,"role":"supplement","size":22752,"visible":true,"origin":"","legend":"","description":"","filename":"BICSteepSlopeFigure.xlsx","url":"https://assets-eu.researchsquare.com/files/rs-6616552/v1/5ffa7adfbafc4460ec0b1c62.xlsx"},{"id":85395013,"identity":"c34eb795-fda4-4f88-8582-ba98be2879e7","added_by":"auto","created_at":"2025-06-25 11:05:16","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":16572,"visible":true,"origin":"","legend":"","description":"","filename":"SummarizeCCFIandBaseRateEstimates.docx","url":"https://assets-eu.researchsquare.com/files/rs-6616552/v1/abf8629140171c7ddc2fdc68.docx"}],"financialInterests":"No competing interests reported.","formattedTitle":"A Study on the Latent Classes of Adolescent Risk Behaviors","fulltext":[{"header":"Introduction","content":"\u003cp\u003eMental health problems directly or indirectly affect adolescents\u0026rsquo; physical and mental state and quality of life. With the development of science and technology and the complexity of the social environment, the factors affecting the mental health of adolescents are increasing. Adolescents are in a critical period of development. Adolescent risk behaviors are closely related to family, school, social and personal factors [\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e]. These factors include both risks and protective elements. Mental illness and adverse consequences such as violence and drug abuse may arise when these elements are imbalanced [\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e][\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e][\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e][\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e]. Adolescents with different cognitive levels take different self-protection measures. However, external support is generally needed to make their development safe and sound [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e][\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn the field of adolescent risk behavior research, the Youth Risk Behavior Surveillance System (YRBSS) conducts a survey every two years, including six kinds of priority health risk behaviors: intentional injury, tobacco use, alcohol and other drug use, sexual behaviors, unhealthy diet and insufficient physical exercise [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e][\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e][\u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e][\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e]. This provides powerful data support for subsequent research. Previous studies have demonstrated that psychological symptoms such as anxiety and depression are associated with mobile phone use, unhealthy weight-control behavior, drinking and smoking [\u003cspan citationid=\"CR13\" class=\"CitationRef\"\u003e13\u003c/span\u003e][\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e][\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e][\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. There are also studies that have used LCA to explore the characteristics of specific risk factors [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. However, most studies concentrate on single or similar behaviors [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eAlthough an increasing number of studies have used LCA to identify latent classes of adolescent risk behavior, studies focused on middle school students, especially those using LCA after group type identification in combination with taxometrics, are still rare. For example, a study has shown that adolescent risk behaviors can be classified into several major classes. These have been classified into four latent types: low-risk class, moderate-risk class 1 (smoking/alcohol use (AU)/screen time (ST)), moderate-risk class 2 (unhealthy weight loss (ULW)/problematic mobile phone use (PMPU)), and high-risk class [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. This comprehensive method can describe the behavior characteristics of each group more accurately. Our study focuses on the adolescent risk behaviors of middle school students. On the basis of the cross-sectional data collected from 14,662 middle school students, we constructed risk subgroups by data type identification and LCA. Our study aims to (1) analyze the dimension structure characteristics of risk behaviors and (2) reveal the correlation mode of risk factors by latent class probability distribution and the risk characteristics of risk behaviors in middle school students by determining their heterogeneity to provide an evidence-based basis for targeted interventions.\u003c/p\u003e "},{"header":"Methods","content":" \u003cp\u003eStudy sample and measures\u003c/p\u003e \u003cp\u003eStudy data and samples are collected from the Database of Youth Health (DYH), and a total of 99,327 students from 186 secondary schools in 17 cities of Shandong Province participated in the survey [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Our study focuses on the analysis of the risk behavior data of 14,662 middle school students in the 2020 risk behavior questionnaire. The questionnaire is adapted from the questionnaires of state, national and local schools covered by the YRBSS [\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e]. It has undergone strict reliability and validity tests and accords with the relevant research standards [\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e]. The questionnaire uses the method of state and local adolescent risk behavior surveys to measure risk behaviors. The method has been revised, adjusted and translated into Chinese, which is suitable for Chinese adolescents.\u003c/p\u003e \u003cp\u003eIn our study, the number of outliers and missing values was very small. The total number of samples is less than 10. Owing to the large sample size, our study deleted the outlier and missing value samples. Ultimately, we preserved 14,662 valid samples for later analysis. This process does not significantly affect the accuracy or reliability of the research results.\u003c/p\u003e \u003cdiv id=\"Sec2\" class=\"Section2\"\u003e \u003ch2\u003eStatistical analysis\u003c/h2\u003e \u003cp\u003eOur study adopts descriptive statistical methods to preliminarily organize, summarize and present the data. We understand the basic characteristics and distribution rules of the data comprehensively and systematically. We check the data quality, identify outliers and extreme situations and complete data cleaning. These findings provide the basis and reference for subsequent analysis.\u003c/p\u003e \u003cp\u003eTaxometric analysis:\u003c/p\u003e \u003cp\u003eIn the process of taxonomic analysis, R version 4.4.2 was selected as the analysis tool and was carried out via the RTaxometrics package [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]. Additionally, we follow the default settings of Ruscio to a certain extent [\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e][\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]. Specifically, we set the random number of seeds to 123. Limited by equipment conditions, we adjust the parameters of the default settings slightly during the process of MAXEIG analysis. Nevertheless, this adjustment does not affect the overall analysis structure or conclusions.\u003c/p\u003e \u003cp\u003eOwing to the simplicity of the measured data, our study uses these variables as indicators for taxometric analysis. Taxometric analysis aims to clarify whether the studied structure is essentially a dimensional structure or a categorical structure [\u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]. To ensure the accuracy and reliability of the results, it is usually necessary to comprehensively use 2\u0026ndash;3 methods for in-depth exploration. We adopt three methods, MAMBAC, MAXEIG and L-MODE, to conduct a detailed analysis of the internal structure of risk behaviors in middle school students. Specifically, if the output graphs of MAMBAC and MAXEIG are peak curves and L-MODE generates a double- or multipeak curve, the analysis results tend to be categorical; in contrast, they are more inclined to be dimensional [\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]. In addition, when evaluating the structure type, the comparison curve fit index (CCFI) and base rate estimation indicators generated by the analysis results need to be considered comprehensively [\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e][\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e][\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eLatent class analysis\u003c/p\u003e \u003cp\u003eIn the second stage, we use SPSS 22.0 to convert the data format and carry out classification and assignment according to the degree of influence of risk behavior factors: 0 for options that do not cause risk behaviors or reduce them, and 1 for options that may cause risk behaviors or increase them [\u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e24\u003c/span\u003e][\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e]. Then, we carry out LCA via MPLUS 8.3.\u003c/p\u003e \u003cp\u003eAn LCA is a method used to describe the latent classes of variables. It is widely used in the analysis of latent characteristics and groups. Additionally, it can analyze the group characteristics of dimensional data according to the required research direction. In particular, in cross-sectional studies, it is commonly used in group classification under observation indices. In this study, the sample size was large enough for the LCA to yield stable results. For this exploratory method, it is necessary to increase the number of fitting classes so that more models will be generated. Then, we select the best model for discussion on the basis of the corresponding evaluation indices of the model fitting effect [\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eIn the evaluation of the model fitting effect, the smaller the values of the AIC (Akaike information criterion), BIC (Bayesian information criterion) and aBIC (corrected BIC) are, the better the model fitting effect [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e][\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e][\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e]. Among them, the BIC is considered to be the most suitable evaluation index in most cross-sectional studies. The bootstrapped likelihood ratio test (BLRT) and Vuong-Lo‒Mendell‒Rubin (VLMR) are used to evaluate the accuracy of classification. Specifically, an entropy\u0026thinsp;\u0026lt;\u0026thinsp;0.60 means that more than 20% of individuals have classification errors; an entropy\u0026thinsp;\u0026ge;\u0026thinsp;0.80 means that the classification accuracy is greater than 90%, and the individual classification in the sample is good [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]. P values\u0026thinsp;\u0026lt;\u0026thinsp;0.05 for BLRT and VLMR suggest that the K classification may perform better than the K-1 classification does, but it is not necessarily significant and should be interpreted in conjunction with practical considerations [\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e]. In our study, each class sample size was \u0026ge;\u0026thinsp;50; therefore, the aBIC is the information index with the highest classification accuracy [\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e].\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eIndicator analyses\u003c/p\u003e\n\u003cp\u003eThe results of the descriptive statistics in Table \u003cspan class=\"InternalRef\"\u003e1\u003c/span\u003e show that there are significant differences in the mean value, discrete degree and distribution pattern of each variable. Variables with high means and low dispersion (such as Q3\u0026ndash;Q6) reflect the universality and stability of the behaviors. Variables with low means and high dispersion (such as Q2 and Q7\u0026ndash;Q12) suggest that these behaviors are not common in the sample, but there are extreme cases. The analysis of skewness and kurtosis further reveals the morphological characteristics of the data distribution, which provides an important reference for subsequent statistical modeling and analysis. Because the variables may belong to the possible influencing factors of risk behavior, all the indicators are included in the following study.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab1\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eDescriptive Statistics of the Variables (n\u0026thinsp;=\u0026thinsp;14662)\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"7\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eVariable\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" colspan=\"2\"\u003e\n \u003cp\u003eMean\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStandard Deviation\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eVariance\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eSkewness\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eKurtosis\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStatistics\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStandard Error\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStatistics\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStatistics\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStatistics\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eStatistics\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ1\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.75\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.433\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.054\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.732\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.909\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.34\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.891\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.794\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.773\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e6.850\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.357\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.127\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-1.961\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.845\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ4\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.87\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.341\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.116\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-2.151\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.625\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.85\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.003\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.361\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.130\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-1.922\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.693\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ6\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.90\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.002\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.301\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.091\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-2.648\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e5.015\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ7\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.275\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.625\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.844\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e7.312\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ8\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.36\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.009\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.096\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.202\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.462\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e11.830\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ9\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.35\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.009\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.082\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.170\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.462\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e11.800\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ10\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.48\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.313\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.725\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.007\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e8.295\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.71\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.585\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.514\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.192\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.555\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ12\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.47\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.305\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.702\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.908\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e7.595\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ13\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.30\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.018\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.162\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e4.676\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.718\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.547\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ14\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.93\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.529\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.339\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.748\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.124\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ15\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.11\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.013\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.526\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.327\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.614\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.932\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.16\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.328\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.764\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.276\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.275\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ17\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.65\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.007\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.868\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.753\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.218\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.632\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ18\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.77\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.278\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.633\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.599\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.243\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.489\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.217\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.214\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-1.331\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ20\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e3.19\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.455\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.117\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-0.179\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-1.294\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ21\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.64\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.011\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.385\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.918\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.339\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-1.097\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ22\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.73\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.010\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.178\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.388\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.565\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.349\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003eQ23\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.54\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.012\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e1.508\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e2.275\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.388\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e-1.382\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"7\"\u003eNote: Q1: frequency of seat belt use in vehicles; Q2: number of trips in a vehicle with a drunk driver; Q3: whether the driver was bullied on campus; Q4: whether the driver was bullied on the Internet; Q5: whether the driver was sad and helpless; Q6\u0026minus;Q9: smoking; Q10\u0026minus;Q12: drinking; Q13: number of exercise days in the past 7 days; Q14: time spent watching TV each week; Q15: time spent playing games each week; Q16\u0026minus;Q18: time spent on PE lessons within or outside the curriculum; Q19: frequency of fathers exercising each week; Q20: frequency of exercise with parents; Q22: frequency of dizziness in physical exercise; and Q23: degree of tooth cleaning.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eTaxometrics\u003c/p\u003e\n\u003cp\u003eStandard classification quantitative analysis was the primary procedure of our study. Owing to the default setting of some parameters, each variable is used as the input variable in turn, and the remaining variables are used as the output variables to ensure that the results are convincing during the MAMBAC and MIXEIG analyses. If there are k variables, k\u0026times;(k\u0026thinsp;\u0026minus;\u0026thinsp;1) input‒output variable pairs are generated. In the case of a large sample size of 14662, the empirical basis rate is set to 50% to obtain more stable results and explanations. The results are shown in Fig. 1. Both the average MAMBAC curve and MAXEIG curve do not present obvious peaks; meanwhile, the L-MODE curve presents a single peak. In addition, the shape of each curve is similar to the dimension curve.\u003c/p\u003e\n\u003cp\u003eThe objective CCFI indices of MAMBAC, MAXEIG and L-MODE are shown in the Supplementary Materials. The CCFI values of the three programs fall into the fuzzy [0.45, 0.55] interval. The CCFI indices of MAMBAC and L-MODE are less than 0.50. The classification ratios generated by base rate estimates are relatively low, which provides additional evidence for the dimensional structure.\u003c/p\u003e\n\u003cp\u003eLatent class analysis\u003c/p\u003e\n\u003cp\u003eThrough the evaluation of the relevant indicators of the fitting effect, four types of latent structure classification are found to be the optimal scheme.\u003c/p\u003e\n\u003cp\u003eAs shown in Table\u0026nbsp;\u003cspan class=\"InternalRef\"\u003e2\u003c/span\u003e, as the number of classes increases, the LL, AIC, BIC, aBIC and entropy indicators gradually decrease. The entropy values are greater than 0.8, and the P values of BLRT and VLMR are less than 0.05, indicating an ideal model fitting effect. However, in terms of the downward trend of the aBIC index and the steep slope figure of the BIC, after the four types of models, the change in the BIC tends to be gradual. In addition, in the classification process, the class probability should not be too small (greater than 10%); otherwise, the class will be of no practical significance. Considering the above factors, we select the four-class model whose class probability has no subclass, and all the indicators reach the standard.\u003c/p\u003e\n\u003cdiv class=\"gridtable\"\u003e\u0026nbsp;\u003ctable id=\"Tab2\" border=\"1\"\u003e\n \u003ccaption language=\"En\"\u003e\n \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e\n \u003cdiv class=\"CaptionContent\"\u003e\n \u003cp\u003eIdentification of the latent class model\u003c/p\u003e\n \u003c/div\u003e\n \u003c/caption\u003e\n \u003ccolgroup cols=\"9\"\u003e\u003c/colgroup\u003e\n \u003cthead\u003e\n \u003ctr\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eClass\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eLL\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eAIC\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eBIC\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eaBIC\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eEntropy\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eBLRT\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eVLMR\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\" rowspan=\"2\"\u003e\n \u003cp\u003eClass Probability (%)\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eP\u003c/p\u003e\n \u003c/th\u003e\n \u003cth align=\"left\"\u003e\n \u003cp\u003eP\u003c/p\u003e\n \u003c/th\u003e\n \u003c/tr\u003e\n \u003c/thead\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e2\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-1513\u003c/p\u003e\n \u003cp\u003e80.846\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e30285\u003c/p\u003e\n \u003cp\u003e5.693\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e30321\u003c/p\u003e\n \u003cp\u003e2.565\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e30306\u003c/p\u003e\n \u003cp\u003e3.203\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.963\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e17.27/82.73\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e3\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-1443\u003c/p\u003e\n \u003cp\u003e75.984\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e28889\u003c/p\u003e\n \u003cp\u003e3.967\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e28943\u003c/p\u003e\n \u003cp\u003e3.071\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e28920\u003c/p\u003e\n \u003cp\u003e7.439\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.929\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e65.65/16.17\u003c/p\u003e\n \u003cp\u003e/18.18\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e4*\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-1423\u003c/p\u003e\n \u003cp\u003e13.315\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e28481\u003c/p\u003e\n \u003cp\u003e6.63\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e28553\u003c/p\u003e\n \u003cp\u003e7.966\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e28523\u003c/p\u003e\n \u003cp\u003e6.064\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.863\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e13.62/17.17/\u003c/p\u003e\n \u003cp\u003e14.16/55.05\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e5\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"left\"\u003e\n \u003cp\u003e-1408\u003c/p\u003e\n \u003cp\u003e66.916\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e28197\u003c/p\u003e\n \u003cp\u003e1.832\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e28287\u003c/p\u003e\n \u003cp\u003e5.401\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e28249\u003c/p\u003e\n \u003cp\u003e7.229\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e0.864\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e\u0026lt;\u0026thinsp;0.001\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd align=\"char\"\u003e\n \u003cp\u003e14.41/14.29/\u003c/p\u003e\n \u003cp\u003e12.50/54.32/4.48\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003ctfoot\u003e\n \u003ctr\u003e\n \u003ctd colspan=\"9\"\u003e\u003csup\u003eNote: LL: log likelihood; AIC: Akaike information criterion; BIC: Bayesian information criterion; aBIC: corrected BIC; BLRT: bootstrapped likelihood ratio test (K\u0026minus;1 vs. K classes); VLMR: Vuong\u0026minus;Lo\u0026minus;Mendell\u0026minus;Rubin (K\u0026minus;1 vs. K classes); *: optimal fitting model\u003c/sup\u003e.\u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tfoot\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003eFigure 2 depicts the conditional probability distribution of four latent classes of adolescent risk behaviors in middle school students. Moreover, it can intuitively show the distribution of risk factors for risk behaviors.\u003c/p\u003e\n\u003cp\u003eAmong the four classes, the major class can be divided into a high-risk group, a moderate-risk group and a low-risk group. The radar map clearly shows that the risk factors for the two moderate-risk groups are different. The characteristic variable of one of the moderate-risk groups is closely related to physical exercise and the number of parental exercises. The conditional probabilities are Q17 (0.801), Q18 (0.837), Q19 (0.839), Q20 (0.851) and Q21 (0.928). However, the characteristic variable of another moderate-risk group is related to the use of electronic products, and its conditional probability is Q15 (0.726). In addition, the probabilistic estimation range of the high-risk group is mostly above 0.69, and each variable tends to have a certain risk impact. The probabilistic estimation range of the low-risk group is mostly less than 0.3, and the risk factors are relatively small.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eOur study is the first to explore the latent structure of adolescent risk behaviors among middle school students in China through the combination of taxometrics and LCA.\u003c/p\u003e \u003cp\u003eFirst, the analysis of the specific variables in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e reveals that the safety-related behaviors are as follows: the mean value of Q1 (frequency of seat belt use in vehicles) is relatively high (3.75), the distribution is left skewed (skewness: -0.732), and the kurtosis is relatively low (-0.909), indicating that most respondents wear seat belts in vehicles, but some of them still do not obey this rule. The mean value of Q2 (the number of riders riding in a vehicle with a drunk driver) is low (1,34), the distribution is right skewed (skewness: 2.773), and the kurtosis is high (6.850), indicating that most respondents rarely ride in a vehicle with a drunk driver, but there are extreme values. Mental health and bullying: The mean values of Q3\u0026ndash;Q5 are low (1.85\u0026ndash;1.87), and the skewness and kurtosis are close to 0, indicating that these phenomena are not common in the sample and that the distribution is relatively symmetrical. Smoking and drinking: The mean values of Q6\u0026ndash;Q12 are low (1.35\u0026ndash;1.71), but the skewness and kurtosis are high, indicating that these behaviors are not common in the sample, but there are a few extreme cases. Physical exercise and recreational activities: The mean values of Q13-Q15 are high (1.93\u0026ndash;3.30), and the skewness and kurtosis are close to 0, indicating that the distribution of these variables is close to a normal distribution. Family exercise participation: The mean value of Q19‒Q21 is high (2.64‒3.23), and the skewness and kurtosis are low, indicating that parents\u0026rsquo; exercise is relatively common and symmetrically distributed in the sample. Exercise and health: The mean value of Q22 is low (1.73), and the skewness and kurtosis are low, indicating that the respondents seldom feel dizzy during exercise. The mean value of Q23 is high (2.54), and the skewness and kurtosis are low, indicating that the respondents have good tooth cleaning habits.\u003c/p\u003e \u003cp\u003eSecond, the existing state of risk behaviors in Chinese middle school students is evaluated via taxometric analysis. The results demonstrate that the average CCFI value is less than 0.5 and that the base rate estimate value is relatively low, indicating that the risk behaviors of middle school students present continuous dimensional distribution characteristics. This result is consistent with a previous study on the risk behaviors of Chinese high school students [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e]. Therefore, it may support latent classification studies.\u003c/p\u003e \u003cp\u003eOn the basis of the above results, LCA divides the research samples into four latent classes: approximately 13.62% of the research samples are assigned to class 1, 17.17% are assigned to class 2, 14.16% are assigned to class 3, and 55.05% are assigned to class 4. As shown in Fig.\u0026nbsp;2, class 1 has a high conditional probability for most variables, and only variables related to physical exercise have a low risk level, so it is defined as a high-risk group. Class 4 has a low conditional probability for all the variables, so it is classified into a low-risk group. Although both class 2 and class 3 belong to the moderate-risk group, there are distinct differences in their risk behavior tendencies. Among them, class 2 is related mainly to physical exercise-related risk behaviors, such as the frequency of participating in extracurricular PE lessons and exercising with parents. This result reflects the influence of the living environment and the school and social environments on the psychological and cognitive development of middle school students. Moreover, this study highlights the importance of family companionship, which is consistent with existing research conclusions [\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e][\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e][\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]. Class 3 is related mainly to the risk associated with electronic product use, in which playing video games is the major risk factor. While the duration of watching TV can also be a risk factor [\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e] (57.2%), it is significantly lower than similar indicators in the high-risk group (69.2%). Therefore, the variable of watching TV is included in the high-risk group for analysis.\u003c/p\u003e \u003cp\u003eIn the high-risk group, risk behavior variables with a conditional probability greater than 60% include the following: Q1 (frequency of seat belt use in vehicles), Q2 (number of vehicles riding with a drunk driver), Q7\u0026ndash;Q9 (smoking), Q10\u0026ndash;Q12 (drinking), Q14 (time spent watching TV each week) and Q22 (frequency of dizziness in physical exercise). Combined with the distribution characteristics of risk factors in the two moderate-risk groups, our study revealed that the risk behaviors of middle school students are related mainly to accidents, smoking and drinking, the use of electronic products, physical exercise and the family environment. This finding is highly consistent with previous research results [\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e][\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e][\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e].\u003c/p\u003e \u003cp\u003eCompared with high school students [\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e][\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e][\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e][\u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e], middle school students are more vulnerable to environmental factors and material temptations (such as electronic products, tobacco and alcohol) because of their younger age, immature psychological cognition level and incompletely established mental health standards. In addition, our study revealed that there is a certain correlation between the risk behavior variables. As shown in Fig.\u0026nbsp;3, the correlation coefficient between Q7‒Q9 (smoking-related variables) and Q10‒Q12 (drinking-related variables) is 0.4‒0.7, indicating a moderate degree of correlation. These findings indicate that smoking and drinking may be related to material temptations in early adolescence and that these temptations may increase the probability of risk behaviors [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. Moreover, the correlation coefficient between Q19 (frequency of fathers\u0026rsquo; exercise each week) and Q20 (frequency of mothers\u0026rsquo; exercise each week) was 0.7, indicating a relatively high degree of correlation. This finding indicates that there is a significant correlation between fathers\u0026rsquo; exercise and mothers\u0026rsquo; exercise. Additionally, these findings suggest that a positive family environment plays an important role in promoting the sound growth of adolescents.\u003c/p\u003e"},{"header":"Conclusions","content":"\u003cp\u003eNotably, our study uses a sample of middle school students in a single province in China to carry out dimensional analysis and latent class identification of risk behaviors. Therefore, the universality of these conclusions may be limited because they fail to fully reflect the characteristics of adolescent risk behaviors in other cultural backgrounds. Although our study does not explore the influence mechanism and causality between the variables further, it still has the following value: (1) Through multidimensional analysis, the heterogeneity of risk behaviors in middle school students is revealed, which provides a methodological reference for the classification and identification of risk behaviors. (2) The distribution rules of risk factors such as smoking, drinking and electronic product use are clarified, which provides an empirical basis for the adjustment of risk behavior prevention strategies for middle school students.\u003c/p\u003e "},{"header":"Declarations","content":"\u003ch2\u003eEthics approval and consent to participate\u003c/h2\u003e\n\u003cp\u003eEthical approval was not needed, as this study used publicly available data.\u003c/p\u003e\n\u003ch2\u003eConsent for publication\u003c/h2\u003e\n\u003cp\u003eThis manuscript has not been published or presented elsewhere in part or in entirety, and is not under consideration by another journal. All the authors have approved the manuscript and agree with submission to your esteemed journal.\u003c/p\u003e\n\u003ch2\u003eAvailability of data and materials\u003c/h2\u003e\n\u003cp\u003eDatabase: Shandong University. Database of Youth Health. Population Health Data Archive PHDA,2021. https://doi.org/10.12213/11.A0031.202107.209.V1.0.\u003c/p\u003e\n\u003ch2\u003eCompeting Interests\u003c/h2\u003e\n\u003cp\u003eI declare that the authors have no competing interests as defined by BMC, or other interests that might be perceived to influence the results and/or discussion reported in this paper.\u003c/p\u003e\n\u003ch2\u003e\u0026nbsp;Funding\u003c/h2\u003e\n\u003cp\u003eThis work was supported by the funds of two horizontal projects from Beijing University of Chinese Medicine(BUCM-2021-JS-FW-031)and(BUCM-2024-JS-FW-137)\u003c/p\u003e\n\u003ch2\u003eAuthors\u0026apos; contributions\u003c/h2\u003e\n\u003cp\u003eJiawen Pu: Writing-review \u0026amp; editing, Writing-original draft, Visualization, Software, Methodology, Formal analysis, Data curation, Conceptualization. Zhaoyan Luo: Writing-review \u0026amp; translating. Hailin Xu: Methodology, Conceptualization. Aiqing Han: Writing-review \u0026amp; editing, Supervision, Methodology. Yan Tang: Writing-review \u0026amp; editing, Supervision, Methodology, Funding acquisition, Conceptualization. Li wang: Writing-review \u0026amp; editing, Supervision, Methodology, Funding acquisition, Conceptualization.\u003c/p\u003e\n\u003ch2\u003eAcknowledgments\u003c/h2\u003e\n\u003cp\u003eThis work was supported by the funds of two horizontal projects from Beijing University of Chinese Medicine(BUCM-2021-JS-FW-031)and(BUCM-2024-JS-FW-137). We express our sincere gratitude to the National Population Health Data Center (NPHDC) for the data. The data is collected from Population Health Data Archive (PHDA). The research is instructed by Li Wang, Yan Tang, and Aiqing Han. And we thank Zhaoyan Luo for the translation.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eDatabase. Shandong University. Database of Youth Health. Population Health Data Archive PHDA,2021. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.12213/11.A0031.202107.209.V1.0\u003c/span\u003e\u003cspan address=\"10.12213/11.A0031.202107.209.V1.0\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang S, Luo W, Dong X, Chen W, Yi X, Zhou W, Zhang Y, Zhao Y. A Dataset on the Status Quo of Health and Health-Related Behaviors of Chinese Youth: A Longitudinal Large-Scale Survey in the Secondary School Students of Shandong Province, Chinese Medical Sciences Journal, Volume 37, Issue 1,2022, Pages 60\u0026ndash;66, ISSN 1001\u0026ndash;9294. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.24920/004051\u003c/span\u003e\u003cspan address=\"10.24920/004051\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eResnick MD, Bearman PS, Blum RW, Bauman KE, Harris KM, Jones J, Tabor J, Beuhring T, Sieving RE, Shew M, Ireland M, Bearinger LH, Udry JR. Protecting adolescents from harm. Findings from the National Longitudinal Study on Adolescent Health. JAMA. 1997;278(10):823\u0026ndash;32. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jama.278.10.823\u003c/span\u003e\u003cspan address=\"10.1001/jama.278.10.823\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eVidal C, Meshi D. Behavioral Addictive Disorders in Children and Adolescents. J Am Acad Child Adolesc Psychiatry. 2023;62(5):512\u0026ndash;4. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.jaac.2022.07.819\u003c/span\u003e\u003cspan address=\"10.1016/j.jaac.2022.07.819\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2022 Aug 22.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMellins CA, Elkington KS, Bauermeister JA, Brackis-Cott E, Dolezal C, McKay M, Wiznia A, Bamji M, Abrams EJ. Sexual and drug use behavior in perinatally HIV-infected youth: mental health and family influences. J Am Acad Child Adolesc Psychiatry. 2009;48(8):810\u0026ndash;9. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/CHI.0b013e3181a81346\u003c/span\u003e\u003cspan address=\"10.1097/CHI.0b013e3181a81346\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePrinstein MJ, Boergers J, Spirito A. Adolescents' and their friends' health-risk behavior: factors that alter or add to peer influence. J Pediatr Psychol. 2001 Jul-Aug;26(5):287\u0026ndash;98. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1093/jpepsy/26.5.287\u003c/span\u003e\u003cspan address=\"10.1093/jpepsy/26.5.287\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHojat M. Protecting adolescents from harm. JAMA. 1998;279(5):353\u0026ndash;4. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1001/jama.279.5.353\u003c/span\u003e\u003cspan address=\"10.1001/jama.279.5.353\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFergus S, Zimmerman MA. Adolescent resilience: a framework for understanding healthy development in the face of risk. Annu Rev Public Health. 2005;26:399\u0026ndash;419. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1146/annurev.publhealth.26.021304.144357\u003c/span\u003e\u003cspan address=\"10.1146/annurev.publhealth.26.021304.144357\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRew L, Horner SD. Youth Resilience Framework for reducing health-risk behaviors in adolescents. J Pediatr Nurs. 2003;18(6):379\u0026ndash;88. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/s0882-5963(03)00162-3\u003c/span\u003e\u003cspan address=\"10.1016/s0882-5963(03)00162-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYouth Risk Behavior Surveillance System. (YRBSS)\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.cdc.gov/yrbs/data/index.html\u003c/span\u003e\u003cspan address=\"https://www.cdc.gov/yrbs/data/index.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBrener ND, Kann L, Kinchen SA, Grunbaum JA, Whalen L, Eaton D, Hawkins J, Ross JG. Methodology of the youth risk behavior surveillance system. MMWR Recomm Rep. 2004;53(RR\u0026ndash;12):1\u0026ndash;13. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pubmed.ncbi.nlm.nih.gov/15385915/\u003c/span\u003e\u003cspan address=\"https://pubmed.ncbi.nlm.nih.gov/15385915/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYouth Risk Behavior Surveillance: National College Health Risk Behavior Survey\u0026ndash;United States. 1995. MMWR CDC Surveill Summ. 1997;46(6):1\u0026ndash;56. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pubmed.ncbi.nlm.nih.gov/9393659/\u003c/span\u003e\u003cspan address=\"https://pubmed.ncbi.nlm.nih.gov/9393659/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTingting Gao J, Li H, Zhang J, Gao Y, Kong Y, Mei HS. The influence of alexithymia on mobile phone addiction: The role of depression, anxiety and stress. J Affect Disord. 2018;225:761\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jad.2017.08.020\u003c/span\u003e\u003cspan address=\"10.1016/j.jad.2017.08.020\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi Y, Li G, Liu L, Wu H. Correlations between mobile phone addiction and anxiety, depression, impulsivity, and poor sleep quality among college students: A systematic review and meta-analysis. J Behav Addict. 2020;9(3):551\u0026ndash;71. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1556/2006.2020.00057\u003c/span\u003e\u003cspan address=\"10.1556/2006.2020.00057\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWall DN-SM, Larson NI, Eisenberg ME. Katie Loth, Dieting and Disordered Eating Behaviors from Adolescence to Young Adulthood: Findings from a 10-Year Longitudinal Study. J Am Diet Assoc, 111, Issue 7, 2011, Pages 1004\u0026ndash;1011, ISSN 0002-8223, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jada.2011.04.012\u003c/span\u003e\u003cspan address=\"10.1016/j.jada.2011.04.012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTomczyk S, Isensee B. Reiner Hanewinkel,Latent classes of polysubstance use among adolescents\u0026mdash;a systematic review,Drug and Alcohol Dependence,160, 2016, Pages 12\u0026ndash;29, ISSN 0376\u0026ndash;8716, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.drugalcdep.2015.11.035\u003c/span\u003e\u003cspan address=\"10.1016/j.drugalcdep.2015.11.035\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChoi AY, Israel T, Nylund-Gibson K. Syndemic behavioral risk and suicidality among bisexual adolescents: A latent class analysis. J Consult Clin Psychol. 2020;88(7):597\u0026ndash;612. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1037/ccp0000500\u003c/span\u003e\u003cspan address=\"10.1037/ccp0000500\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJie H, Yang R, Li D, Zhao S, Wan Y, Tao F, Fang J, Zhang S. A latent class analysis of psychological symptoms and health risk behaviors among Chinese adolescents. Asian J Psychiatry 55. 2021;102518:1876\u0026ndash;2018. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ajp.2020.102518\u003c/span\u003e\u003cspan address=\"10.1016/j.ajp.2020.102518\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYi X, Liu Z, Qiao W, et al. Clustering effects of health risk behavior on mental health and physical activity in Chinese adolescents. Health Qual Life Outcomes. 2020;18(1):211. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1186/s12955-020-01468-z\u003c/span\u003e\u003cspan address=\"10.1186/s12955-020-01468-z\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRuscio J. Shirley B. Wang,3.11 - Taxometric Analysis,Editor(s): Gordon J.G. Asmundson,Comprehensive Clinical Psychology (Second Edition),Elsevier,2022,Pages 148\u0026ndash;175,ISBN 9780128222324. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/B978-0-12-818697-8.00217-X\u003c/span\u003e\u003cspan address=\"10.1016/B978-0-12-818697-8.00217-X\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eFen Ren R, Zhou X, Zhou SC, Schneider EA. Storch,The latent structure of olfactory reference disorder symptoms: A taxometric analysis,Journal of Obsessive-Compulsive and Related Disorders,27,2020,100583,ISSN 2211\u0026ndash;3649, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jocrd.2020.100583\u003c/span\u003e\u003cspan address=\"10.1016/j.jocrd.2020.100583\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRuscio J, Ruscio AM, Meron M. Applying the Bootstrap to Taxometric Analysis: Generating Empirical Sampling Distributions to Help Interpret Results. Multivar Behav Res. 2007;42(2):349\u0026ndash;86. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/00273170701360795\u003c/span\u003e\u003cspan address=\"10.1080/00273170701360795\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWalters GD, Ruscio J. Trajectories of Youthful Antisocial Behavior: Classes or Continua? J Abnorm Child Psychol. 2013;41:653\u0026ndash;66. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s10802-012-9700-1\u003c/span\u003e\u003cspan address=\"10.1007/s10802-012-9700-1\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSinha P, Calfee CS, Delucchi KL. Practitioner's Guide to Latent Class Analysis: Methodological Considerations and Common Pitfalls. Crit Care Med. 2021;49(1):e63\u0026ndash;79. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1097/CCM.0000000000004710\u003c/span\u003e\u003cspan address=\"10.1097/CCM.0000000000004710\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ehttps://\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ecran.r-project.org/web/packages/RTaxometrics/index.html\u003c/span\u003e\u003cspan address=\"http://cran.r-project.org/web/packages/RTaxometrics/index.html\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWeller BE, Bowen NK, Faubert SJ. Latent Class Analysis: A Guide to Best Practice. J Black Psychol. 2020;009579842093093. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1177/0095798420930932\u003c/span\u003e\u003cspan address=\"10.1177/0095798420930932\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMcCutcheon AL. Latent class analysis. SAGE Publications, Inc.; 1987. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.4135/9781412984713\u003c/span\u003e\u003cspan address=\"10.4135/9781412984713\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRuscio J, Carney LM, Dever L, Pliskin M, Wang SB. Using the comparison curve fix index (CCFI) in taxometric analyses: Averaging curves, standard errors, and CCFI profiles. Psychol Assess. 2018;30(6):744\u0026ndash;54. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1037/pas0000522\u003c/span\u003e\u003cspan address=\"10.1037/pas0000522\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eNatacha Carragher G, Adamson B, Bunting. Siobhan McCann,Subtypes of depression in a nationally representative sample, Journal of Affective Disorders, 113, Issues 1\u0026ndash;2, 2009, Pages 88\u0026ndash;99, ISSN 0165\u0026ndash;0327, \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jad.2008.05.015\u003c/span\u003e\u003cspan address=\"10.1016/j.jad.2008.05.015\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMengcheng Wang. Measurement and Latent Structure of Adolescent Health Risk Behaviors [D]. Cent South Univ. 2013. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.16128/j.cnki.1005-3611.2013.05.031\u003c/span\u003e\u003cspan address=\"10.16128/j.cnki.1005-3611.2013.05.031\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBecker SP, Luebbe AM, Langberg JM. Co-occurring mental health problems and peer functioning among youth with attention-deficit/hyperactivity disorder: a review and recommendations for future research. Clin Child Fam Psychol Rev. 2012;15(4):279\u0026ndash;302. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10567-012-0122-y\u003c/span\u003e\u003cspan address=\"10.1007/s10567-012-0122-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBraun D, Lascelles K. Involving and supporting families, friends, and carers during a mental health crisis. Lancet Psychiatry. 2024;11(8):586\u0026ndash;7. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/S2215-0366(24)00165-2\u003c/span\u003e\u003cspan address=\"10.1016/S2215-0366(24)00165-2\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2024 Jun 11.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMollica RF, Donelan K, Tor S, Lavelle J, Elias C, Frankel M, Blendon RJ. The effect of trauma and confinement on functional health and mental health status of Cambodians living in Thailand-Cambodia border camps. JAMA. 1993;270(5):581\u0026ndash;6. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pubmed.ncbi.nlm.nih.gov/8331755/\u003c/span\u003e\u003cspan address=\"https://pubmed.ncbi.nlm.nih.gov/8331755/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLemola S, Perkinson-Gloor N, Brand S, Dewald-Kaufmann JF, Grob A. Adolescents' electronic media use at night, sleep disturbance, and depressive symptoms in the smartphone age. J Youth Adolesc. 2015;44(2):405\u0026ndash;18. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1007/s10964-014-0176-x\u003c/span\u003e\u003cspan address=\"10.1007/s10964-014-0176-x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2014 Sep 10.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eRiesch SK, Kedrowski K, Brown RL, Temkin BM, Wang K, Henriques J, Jacobson G, Giustino-Kluba N. Health-risk behaviors among a sample of US preadolescents: types, frequency, and predictive factors. Int J Nurs Stud. 2013;50(8):1067-79. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.ijnurstu.2012.10.012\u003c/span\u003e\u003cspan address=\"10.1016/j.ijnurstu.2012.10.012\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2012 Nov 21. Erratum in: Int J Nurs Stud. 2020;109:103739. doi: 10.1016/j.ijnurstu.2020.103739.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSchulenberg J, Bachman JG, O'Malley PM, Johnston LD. High school educational success and subsequent substance use: a panel analysis following adolescents into young adulthood. J Health Soc Behav. 1994;35(1):45\u0026ndash;62. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://pubmed.ncbi.nlm.nih.gov/8014429/\u003c/span\u003e\u003cspan address=\"https://pubmed.ncbi.nlm.nih.gov/8014429/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e.\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eO'Malley PM, Johnston LD. Driving after drug or alcohol use by US high school seniors, 2001\u0026ndash;2011. Am J Public Health. 2013;103(11):2027\u0026ndash;34. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.2105/AJPH.2013.301246\u003c/span\u003e\u003cspan address=\"10.2105/AJPH.2013.301246\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e. Epub 2013 Sep 12.\u003c/span\u003e\u003c/li\u003e \u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"adolescence, risk behavior, taxometrics, latent class analysis, Chinese adolescents","lastPublishedDoi":"10.21203/rs.3.rs-6616552/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6616552/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003ePurpose:\u003c/h2\u003e \u003cp\u003eAdolescent health is gaining increased attention, necessitating systematic analysis of the diversity and complexity of adolescent risk behaviors among middle school students. This study aims to examine the data characteristics and latent categorical distributions of these behaviors, identify high-risk populations and associated factors, and provide evidence to inform targeted interventions and prevention strategies.\u003c/p\u003e\u003ch2\u003eMethods:\u003c/h2\u003e \u003cp\u003eThe data used in our study are derived from China\u0026rsquo;s first public dataset on adolescent health. It is produced by a project that has carried out a multiwave survey of 99,327 middle and high school students from 2015\u0026ndash;2021[\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e][\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. In our study, taxometrics and latent class analysis (LCA) were used to study the latent classes of risk behaviors of middle school students (n\u0026thinsp;=\u0026thinsp;14,662) in a cross-sectional study conducted in 2020.\u003c/p\u003e\u003ch2\u003eResults:\u003c/h2\u003e \u003cp\u003eThe results of taxometrics indicate that adolescent risk behaviors of middle school students present a continuous dimension. Additionally, we further explored the latent classes of the risk\u0026ndash;behavior groups. There are four latent classes: (1) high-risk behavior group (13.62%); (2) moderate-risk behavior group (athletics sport) (17.17%); (3) moderate-risk behavior group (electronic products) (14.16%); and (4) low-risk behavior group (55.05%). The high-risk group is driven by multiple cooccurring factors, whereas the moderate-risk subgroup is correlated primarily with physical inactivity or electronic overuse, which are linked to environmental, familial, and cognitive influences.\u003c/p\u003e\u003ch2\u003eConclusions:\u003c/h2\u003e \u003cp\u003eAdolescent risk behaviors in middle school students form four latent classes: the predominant low-risk group, a high-risk group requiring prioritized intervention for multiple factors, and a moderate-risk group where physical inactivity outweighs electronic product use as a risk. Additionally, smoking and drinking were significantly related.\u003c/p\u003e","manuscriptTitle":"A Study on the Latent Classes of Adolescent Risk Behaviors","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-06-25 10:57:11","doi":"10.21203/rs.3.rs-6616552/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"edf9cf5b-cc1d-4985-b579-d94d2c65e996","owner":[],"postedDate":"June 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2026-03-24T14:42:11+00:00","versionOfRecord":[],"versionCreatedAt":"2025-06-25 10:57:11","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6616552","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6616552","identity":"rs-6616552","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}