Prevalence and correlates of unprotected casual sex among sexually experienced college students in Guangxi: a cross-sectional study Authors: Jiao Qin, Zhifeng Lin, Jinfeng He, Weiao Qin, Junhui Liu, Xinju Huang, Qiuqian Su, Li Ye, Hao Liang, Yuanjia Zheng, Bingyu Liang, Ping Cen Published: 2026 ABSTRACT AbstractBackgroundRisky sexual behavior significantly contributes to sexually transmitted infections (STIs) among Chinese college students, yet data on unprotected casual sex remain limited. This study investigated the prevalence and associated factors of unprotected casual sex among sexually experienced students in Guangxi.MethodsA cross‑sectional online survey was conducted from October 2020 to March 2021among students from 16 colleges in Guangxi, using multistage stratified cluster sampling. Firth’s penalized likelihood multivariable logistic regression was used to identify factors associated with unprotected casual sex. Stratified analyses were then conducted according to sexual orientation, and interaction terms were added to test for cross-group differences.ResultsAmong 12,847 sexually experienced students, 5.9% reported engaging in unprotected casual sex in the past year. In multivariate analysis, factors negatively associated with unprotected casual sex included higher level of HIV-related knowledge (OR = 0.55, 95% CI 0.38–0.80), neutral (OR = 0.57, 95% CI 0.37–0.88) or accepting attitudes (OR = 0.38, 95% CI 0.26–0.56) toward premarital cohabitation, and neutral attitudes on engaging in one-night stands (OR = 0.65, 95% CI 0.44–0.95). In contrast, having more than five sexual partners (OR = 1.60, 95% CI 1.10–2.31), seeking sexual partners online (OR = 5.04, 95% CI 4.00–6.38), and receiving HIV testing in the past year (OR = 1.70, 95% CI 1.30–2.22) were positively associated with unprotected casual sex.ConclusionThe prevalence of unprotected casual sex was low among sexually experienced college students in Guangxi. Unprotected casual sex was associated with HIV-related knowledge, HIV testing history, acceptance of premarital cohabitation, multiple sexual partners, and the behavior of seeking sexual partners online. These findings suggest that targeted interventions addressing the modifiable factors are of great significance for reducing unprotected casual sex and STIs risk in this population.Keywords: College students, Unprotected casual sex, HIV/AIDS, Factors, China Introduction Youth are sexually active and vulnerable to Human immunodeficiency virus (HIV). Global estimates indicate that approximately 3,600 new HIV infections occur daily, of which nearly 1,000 are among individuals aged 15–24 years [1]. The issue of HIV among college students has garnered widespread global attention [2]. In China, this epidemic is particularly concerning, as the annual number of new infections among college students has risen by 30%–50% in recent years [3]. Risky sexual behavior (RSB) is defined as actions that significantly increase the risk of HIV infection, specifically including unprotected sexual intercourse with high-risk populations (e.g., commercial sex partners, casual sex partners, or multiple sexual partners) [4]. Sexual transmission remains the primary route of HIV infection among college students [5]. Without adequate sexual health guidance, students are prone to engaging in RSB, thereby increasing their risk of HIV acquisition. Consequently, reducing RSB among college students has been recognized as a key strategy for controlling HIV spread in this population. Among various forms of RSB, unprotected casual sex is particularly sensitive to environmental changes, such as the widespread use of the internet. With the growth of online platforms, internet dating has become a primary mode of social interaction for college students, and an increasing number are using these platforms to seek sexual partners [6, 7]. Compared to offline environments, online social platforms facilitate unprotected casual sex due to their anonymity and convenience [8]. Data from China indicate that 57.98% of college students who engaged in casual sex found partners online [9], and over half (53.9%) of male university students using the internet to seek sexual partners reported unprotcted casual sex [10]. Engaging in unprotected casual sex not only elevates the risk of sexually transmitted infections (STIs) but also carries a range of adverse consequences, including disruptions in academic performance [11, 12], mental health issues [13], unintended pregnancy and abortions [14]. Therefore, assessing the current prevalence of unprotected casual sex among college students is urgently needed to inform interventions aimed at mitigating these outcomes. Although previous studies have extensively examined RSB and their determinants among college students, these established behavioral patterns and accosiated factors may have shifted under new social paradigms [15–17]. Furthermore, little data exist on the correlates of unprotected casual sex among sexually experienced college students. Meanwhile, most prior studies defined casual sex without considering consistent condom use during the encounter, potentially overestimating the associated risk level [18–20]. To address these limitations, the present study employed a more rigorous definition of casual sex—specifically, unprotected penetrative intercourse with non‑regular, non‑commercial partners—and focused exclusively on sexually experienced college students. This approach enables a more accurate assessment of the true risk level and its associated factors. Guangxi has the third highest HIV prevalence among all provinces in China. Between 1996 and 2019, Guangxi reported a total of 13,839 HIV infections among individuals aged 16–25 years [21]. Moreover, HIV infections among students in Guangxi have increased annually from 2007 to 2019 [22]. Despite this heavy burden, no study has investigated the prevalence and correlates of unprotected casual sex among sexually experienced college students in Guangxi. Understanding the current status and correlates of unprotected casual sex in this region is critical for designing evidence-based interventions. Thus, this study aims to: (1) assess the prevalence of unprotected casual sex among sexually experienced college students in Guangxi; (2) identify factors associated with unprotected casual sex. The findings will provide empirical evidence for identifying high‑risk populations and key modifiable factors, and will inform the development of targeted interventions to reduce unprotected casual sex and HIV/ STIs transmission among college students in Guangxi. Materials and methods Study design and participants A cross-sectional survey conducted by the Nanning Centre for Disease Control and Prevention utilized stratified and cluster sampling methods among college students from 16 colleges in Nanning City, spanning from October 2020 to March 2021. Based on the China National AIDS Sentinel Surveillance Questionnaire [23], and subsequent review of relevant literature and expert guidance, we developed a self-designed anonymous online questionnaire through the professional Chinese survey website Wenjuanxing [24]. A unique link or QR code for the questionnaire was distributed to the trained head teachers, who were then responsible for administering the survey in their respective classes. All potential respondents were informed that they had the right to refuse to participate and to withdraw from the survey at any time without any reason. Neither refusal to participate nor withdrawal from the survey affected their performance or course evaluations. Before the formal survey, we determined the sample size for a pre-survey based on the total sample size, using the principle that the number of pre-survey respondents would be 10% to 20% of the total sample size. After the pre-survey, the Cronbach’s alpha and the Kaiser-Meyer-Olkin (KMO) measure of sampling adequacy for the overall questionnaire were 0.751 and 0.820, respectively. For the HIV-related knowledge scale, the values were 0.770 and 0.867, respectively [25]. The inclusion criteria for college students were (1) studying in the colleges located in Nanning, (2) aged 18–25 years old, and (3) consent to participate in this survey. The exclusion criteria were as follows: (1) foreign students, and (2) college students with postgraduate degrees. Sample size and technique A cross-sectional survey of 49,728 college students in China indicated that 19.5% had engaged in casual sex [19]. We assumed that the prevalence of unprotected casual sexual behavior among college students in Guangxi was 20%. A sample size of 6,245 was calculated using PASS software (version 21.0.3). Taking into account a 20% drop-out rate, a minimum of 7,807 participants was required. This study employed a multistage stratified cluster sampling method. In the first stage, 16 universities were randomly selected from 35 universities in Nanning, Guangxi, using a random number table method (i.e., simple random sampling). In the second stage, four faculties/departments were randomly selected from each sampled university. In the third stage, students in the selected faculties/departments were stratified into six levels according to their academic year: first year through fifth year, and postgraduate year. Subsequently, within each department, five classes were sampled from each of the first, second, and third years, two classes were sampled from each of the fourth and fifth years, and one class was sampled from the postgraduate year. The proportion of each college student ranged from 5 to 10%. A self-administered, anonymous online questionnaire was used to conduct the survey. Sexually experienced students were defined as those who answered “Yes” to the question “Have you ever had sexual intercourse (vaginal or anal)?”. A total of 108,987 students completed the survey, of whom 15,098 answered “Yes”. The following ineligible questionnaires were eliminated: (1) Self-reported age at first sexual intercourse is older than the current reported age; (2) The current reported age was not consistent with the current stage of study (For example, the stage of study was a freshman, but the reported age was 40 years old); (3) Students who reported they were single but the marital status was cohabiting or married. Finally, 2,251 unqualified questionnaires were eliminated, leaving 12,847 valid questionnaires. The final valid response rate was 85.1% (12,847/15,098). Measure The questionnaire was composed of the following six sections: sociodemographic characteristics, HIV/sexual health-related knowledge, sexual attitudes, behavioral characteristics, and HIV/AIDS prevention service use. Sociodemographic characteristics Sociodemographic variables included sex (male, female), age (18–21 years, 22–25 years), marital status (unmarried, married/cohabiting), stage of study (professional college, bachelor), social relationship (single or not), and sexual orientation (heterosexual, non‑heterosexual). HIV-related knowledge HIV-related knowledge was assessed with 8 items. Each item was assessed using a yes/no/unsure format, with only the correct response scoring 1 point. Consistent with the National Center for HIV Control and Prevention [25], participants with a score of 6 or greater were considered aware of HIV-related knowledge. Sexual attitudes Attitude toward sex was measured through two questions by asking participants about their attitudes toward one-night stands and premarital cohabitation. Each response option was responded by “acceptable”, “neutral”, or “unacceptable”. HIV-related risk behavior In this study, unprotected casual sex was defined as unprotected penetrative sexual intercourse (including vaginal and anal intercourse) with non-regular, non-commercial opposite-sex partners. This behavior was assessed using the following two questions: (1) “In the past year, have you had sexual intercourse with a casual sexual partner?” (yes/no); (2) “In the past year, how often did you use condoms during sexual intercourse with casual sexual partners?” (never/sometimes/every time). Participants were classified as having engaged in unprotected casual sex if they met both of the following criteria: (1) a history of sexual intercourse with a casual sexual partner in the past year; and (2) reported “never” or “sometimes” using condoms in response to the condom use frequency question. Otherwise, they were classified as not having engaged in unprotected casual sex. Behavioral characteristics The number of lifetime sexual partners was classified according to one partner, two to five partners, and more than five partners. Seeking sexual partners online was defined as having sought sexual partners through the internet and engaged in penetrative sexual intercourse (vaginal or anal intercourse). Having ever sought sexual partners online was measured by the following two options: “yes”, “no”. The stages of student participation in the school-based sexual courses were categorized into: “college”, “high school”, and “junior high school or earlier”. Additionally, we collected data on the types of casual sex partners among college students who engaged in unprotected casual sex. HIV/AIDS prevention service use HIV/AIDS prevention service refers to HIV/AIDS prevention and control education, information, and services provided by universities (including university hospitals, faculties/departments, and student organizations) and local Centers for Disease Control and Prevention in Guangxi. The content includes information on routes of HIV/AIDS transmission, preventive measures (e.g., condom use, pre‑ and post‑exposure prophylaxis), HIV testing and treatment services, and stigma reduction. Educational formats include special lectures, class meetings, distribution of educational materials, peer education, and online courses. HIV/AIDS prevention service use was measured by three questions with yes/no options: (1) Have you received any HIV/AIDS education in the last year? (2) Have you participated in HIV/AIDS prevention volunteer activities in the last year? (3) Have you ever been tested for HIV in the last year? Statistical analysis Data were organized in Microsoft Excel 2019 and analyzed using R software (version 4.2.1). The chi-square test was used to compare the distribution of demographic characteristics between college students who engaged in unprotected casual sex and those who did not. Firth’s penalized likelihood logistic regression model was used to address a complete separation issue. Specifically, for the variable “Had ever sought sexual partners online,” no unprotected casual sex events occurred in the missing category. Under complete separation, standard maximum likelihood estimation fails to converge because parameter estimates approach infinity. Firth’s method introduces a penalty term that resolves this issue and produces stable estimates. Given the substantial differences in sexual networking patterns and risk exposure between heterosexual and sexual minority populations, we conducted a stratified analysis by sexual orientation to explore differences in factors associated with unprotected casual sex. To test for differences between the two subgroups, we also performed interaction analyses by adding product terms (sexual orientation × each covariate) to the logistic regression model. The results of the multivariable analyzes are reported as adjusted odds ratios (ORs) with 95% confidence intervals (CIs), and the significance level was set at p ≤ 0.05. Results Sociodemographic characteristics Table 1 presents the demographic characteristics of the participants. A total of 12,847 individuals completed the survey, with a median age (interquartile range) of 21 (20–22) years. Of these, 755 (5.9%) reported engaging in unprotected casual sexual behavior in the past year. Most participants were male (61.5%) and were not single (55.7%). The majority of participants identified as heterosexual (89.0%) and unmarried (97.9%). In addition, more than half (56.0%) were enrolled in a professional (non‑degree) college. Table 1Demographic characteristics of sexually experienced college students in Guangxi (N = 12,847)VariablesTotalHaving ever engaged in unprotected casual sex P n(%)Yes n(%)No n(%)N12,847 (100.0)755(5.9)12,092 (94.1)Sex< 0.001 Male7902 (61.5)649 (8.2)7253 (91.8) Female4945 (38.5)106 (2.1)4839 (97.9)Age< 0.001 18–218886 (69.2)601 (6.8)8285 (93.2) 22–253961 (30.8)154 (3.9)3807 (96.1)Marital status< 0.001 Unmarried12,580 (97.9)719 (5.7)11,861 (94.3) Married/Cohabiting267 (2.1)36 (13.5)231 (86.5)Stage of study< 0.001 Professional college7197 (56.0)516 (7.2)6681 (92.8) Bachelor5650 (44.0)239 (4.2)5411 (95.8)Sex orientation5693 (44.3)< 0.001 Heterosexual11,432 (89.0)598 (5.2)10,834 (94.8) Non-Heterosexual1415 (11.0)157 (11.1)1258 (88.9)Single< 0.001 No7154 (55.7)194 (2.7)6960 (97.3) Yes5693 (44.3)561 (9.9)5132 (90.1) Demographic characteristics of sexually experienced college students in Guangxi (N = 12,847) Types of casual sexual partners The types of casual sexual partners among college students who engaged in unprotected casual sex are shown in Table 2. Among participants who reported ever having engaged in unprotected casual sex (n = 755), the most common partner was a former boyfriend or girlfriend, identified by 648 participants (85.8%). The second most prevalent category was classmates or friends (excluding former partners), reported by 222 participants (29.4%). Online acquaintances were cited by 90 participants (11.9%), while school teachers were reported by 71 participants (9.4%). Relatives (including both older and same generation) were selected by 43 participants (5.7%). Sex workers and other strangers were each reported by 35 participants (4.6%). Additionally, 78 participants (10.3%) selected the “other” category. Table 2Types of casual sexual partners among participants who reported unprotected casual sex (N = 755)Partner type n %Former boyfriend/girlfriend64885.8Classmate/friend (excluding former partner)22229.4Online acquaintance9011.9School teacher719.4Relative435.7Sex worker354.6Other stranger354.6Other7810.3As multiple responses were permitted, percentages sum to more than 100% Types of casual sexual partners among participants who reported unprotected casual sex (N = 755) As multiple responses were permitted, percentages sum to more than 100% Factors associated with unprotected casual sex Table 3 shows the adjusted OR with 95%CI of associated factors on unprotected casual sexual behavior. Participants with higher levels of HIV‑related knowledge were significantly less likely to engage in unprotected casual sex than those with lower levels (OR = 0.55, 95% CI 0.38–0.80). Compared with participants who considered premarital cohabitation unacceptable, those with a neutral attitude (OR = 0.57, 95% CI 0.37–0.88) and those who accepted it (OR = 0.38, 95% CI 0.26–0.56) had significantly lower odds of engaging in unprotected casual sex. Compared to participants with one lifetime sexual partner, those with more than five partners had significantly higher odds of engaging in unprotected casual sex (OR = 1.60, 95% CI 1.10–2.31). Participants who had ever sought sexual partners online had substantially higher odds of engaging in unprotected casual sex compared to those who had never done so (OR = 5.04, 95% CI 4.00–6.38). Additionally, participants who had undergone HIV testing in the past year were more likely to have engaged in unprotected casual sex (OR = 1.70, 95% CI 1.30–2.22). Table 3Multivariable logistic regression analysis of unprotected casual sex among college students (N = 12,847)VariablesBSEOR95% CI P Sex Female0–1.00–– Male0.020.161.020.75–1.390.91Age 18–210–1.00–– 22–25-0.120.140.890.67–1.180.422Marital status Unmarried0–1.00–– Married/cohabiting-0.160.270.850.50–1.450.554Stage of study Professional college0–1.00–– Bachelor-0.20.130.820.64–1.050.117Sex orientation Heterosexual0–1.00–– Non-heterosexual-0.190.150.830.62–1.100.197Single No0–1.00–– Yes-0.020.130.980.76–1.260.855Have HIV-related knowledge No0–1.00–– Yes-0.60.190.550.38–0.800.002Attitude toward cohabiting before getting married Unacceptable0–1.00–– Neutral-0.560.220.570.37–0.880.011 Acceptable-0.960.20.380.26–0.56< 0.001Attitude on engaging in one night stand Unacceptable0–1.00–– Neutral-0.430.190.650.44–0.950.027 Acceptable-0.210.180.810.57–1.170.256Stage of school-based sex education College0–1.00–– High school-0.20.240.820.51–1.300.389 Other0.270.181.30.92–1.870.141Number of lifetime sexual partners 10–1.00–– 2–5-0.060.140.940.71–1.250.691 > 50.470.191.61.10–2.310.013 Missing0.490.161.631.18–2.260.003Had ever sought sexual partners online Never0–1.00–– Yes1.620.125.044.00-6.38< 0.001Had participated in HIV/AIDS prevention volunteer activities in the past year No0–1.00–– Yes-0.370.20.690.47–1.020.064Had received HIV/AIDS education in the past year No0–1.00–– Yes0.230.21.260.85–1.860.247Underwent HIV testing in the past year No0–1.00–– Yes0.530.141.71.30–2.22< 0.001OR: Odds ratio; CI: Confidence interval Multivariable logistic regression analysis of unprotected casual sex among college students (N = 12,847) OR: Odds ratio; CI: Confidence interval Subgroup analysis by sexual orientation The results of the stratified analysis are shown in Table 4. Among heterosexual participants, several factors were significantly associated with unprotected casual sex. Compared to those who found premarital cohabitation unacceptable, heterosexual participants with a neutral attitude (OR = 0.42, 95% CI 0.25–0.69) and those who found it acceptable (OR = 0.34, 95% CI 0.22–0.53) had significantly lower odds of engaging in unprotected casual sex. Relative to those with one lifetime sexual partner, heterosexual participants with more than five partners showed increased odds (OR = 1.64, 95% CI 1.09–2.48). Having ever sought sexual partners online was strongly associated with higher odds of unprotected casual sex (OR = 5.10, 95% CI 3.95–6.59). Additionally, heterosexual participants who had undergone HIV testing in the past year were more likely to have engaged in unprotected casual sex (OR = 1.85, 95% CI 1.36–2.51). Table 4Multivariable logistic regression analysis of unprotected casual sex stratified by sexual orientationCharacteristicsHeterosexualNon-HeterosexualP for interactionOR95% CI P OR95% CI P Sex Female1.00--1.00-- Male1.050.74–1.510.7720.950.47–1.910.8880.569Age <=211.00--1.00-- 22–250.860.62–1.180.3510.870.45–1.670.6840.913Marital status Unmarried1.00--1.00-- Married/cohabiting0.610.29–1.230.171.480.60–3.720.3970.071Stage of study Professional college1.00--1.00-- Bachelor0.910.68–1.210.5060.640.36–1.120.1170.222Single No1.00--1.00-- Yes1.090.82–1.440.5620.750.40–1.400.3610.05Have HIV-related knowledge No1.00--1.00-- Yes0.760.48–1.180.2210.280.13–0.59< 0.0010.006Attitude toward cohabiting before getting married Unacceptable1.00--1.00-- Neutral0.420.25–0.69< 0.0011.790.70–4.750.2270.095 Acceptable0.340.22–0.53< 0.0010.630.25–1.580.3240.976Attitude on engaging in one night stand Unacceptable1.00--1.00-- Neutral0.850.55–1.300.4520.230.09–0.560.0010.143 Acceptable0.890.60–1.340.580.480.19–1.210.1170.334Stage of school-based sex education College1.00--1.00-- High school0.840.50–1.440.5330.780.28–2.140.6270.804 Other1.320.88–2.010.1831.080.50–2.350.8440.866Number of lifetime sexual partners 11.00--1.00-- 2–51.060.78–1.450.7110.630.31–1.250.1860.079 > 51.641.09–2.480.0181.460.60–3.570.3990.529 Missing1.741.21–2.530.0031.120.56–2.250.7410.259Had ever sought sexual partners online No1.00--1.00-- Yes5.103.95–6.59< 0.0015.192.86–9.80< 0.0010.385Had participated in HIV/AIDS prevention volunteer activities in the past year No1.00--1.00-- Yes0.720.46–1.120.140.540.23–1.240.1440.176Had received HIV/AIDS education in the past year No1.00--1.00-- Yes1.280.82–1.990.2731.580.65–3.910.310.211Underwent HIV testing in the past year No1.00--1.00-- Yes1.851.36–2.51< 0.0011.230.69–2.220.4830.086OR: Odds ratio; CI: Confidence interval Multivariable logistic regression analysis of unprotected casual sex stratified by sexual orientation OR: Odds ratio; CI: Confidence interval Among non‑heterosexual participants, HIV-related knowledge was associated with significantly lower odds of unprotected casual sex (OR = 0.28, 95% CI 0.13–0.59). Regarding attitudes toward one‑night stands, non‑heterosexual participants with a neutral attitude had lower odds of unprotected casual sex compared to those who found it unacceptable (OR = 0.23, 95% CI 0.09–0.56). Similar to the heterosexual subgroup, having ever sought sexual partners online was strongly associated with increased odds of unprotected casual sex (OR = 5.19, 95% CI 2.86–9.80). A significant interaction was observed for HIV-related knowledge (p for interaction = 0.006), indicating that its protective effect on unprotected casual sex was significantly stronger among non‑heterosexual participants than among heterosexual participants. No other significant interactions were found (all p > 0.05), with the exception of a borderline significant interaction for single status (p = 0.050). Discussion This study used the largest sample size to investigate the prevalence and associated factors of unprotected casual sex among college students in Guangxi. The findings indicated that the prevalence of unprotected casual sex among college students in Guangxi was low. Notably, we demonstrated that unprotected casual sex was associated not only with cognitive factors such as HIV-related knowledge and sexual attitudes, but also with behavioral characteristics including multiple sexual partnerships, online seeking of sexual partners, and HIV testing. These findings provide crucial evidence to identify high-risk populations and reduce the risk of unprotected casual sex, allowing colleges to inform sexuality education and HIV prevention efforts better. Overall, the prevalence of unprotected casual sex in the present study (5.9%) is substantially lower than that reported by Liu et al. in six Chinese cities (29.7%) [26], and also markedly lower than findings from comparable populations in Zhejiang Province (17.9%) [20] and Chongqing City (7.13%) [27]. This discrepancy may be related to differences in sociocultural context and the definition of unprotected casual sex across studies. Despite the lower prevalence observed here, multiple studies consistently indicate that unprotected casual sex among college students is becoming increasingly common [26, 28], warranting continued attention and enhanced preventive interventions. The primary partners for unprotected casual sex were former boyfriends/girlfriends (85.8%) and classmates/friends (29.4%), suggesting that most casual sexual encounters occurred within acquaintance networks rather than as one‑night stands with strangers. This finding implies that college students may underestimate risk due to irrational trust in acquaintances. Therefore, interventions targeting college populations should emphasize that HIV/STI risk awareness should not be compromised by partner familiarity, aiming to reduce unprotected sexual activities. An unexpected but notable finding is that 15.1% of participants identified teachers or relatives as their casual sexual partners. Despite the self-reported nature of the data, this proportion remains concerning, as such encounters may involve power imbalances and potential sexual violence or abuse. A previous study has reported that among Chinese female college students who experienced physical contact abuse, 71.3% were from acquaintances, including 12.5% who were teachers and 21.3% who were relatives [29]. Another study found that among Chinese college students who experienced penetrative childhood sexual abuse, 13.6% of males and 17.4% of females reported a family member as the perpetrator [30]. This finding highlights the potential social issue of college students suffering sexual exploitation by authority figures in schools and families. Accordingly, universities should establish accessible reporting mechanisms and support services for victims of campus sexual violence or assault, while the government should increase investment in prevention policies and public awareness campaigns targeting sexual violence involving disparities in social or economic status. Consistent with previous research [16, 31], this study found that college students with higher levels of HIV-related knowledge were less likely to engage in unprotected casual sex, as greater awareness of HIV/STI risks discourages RSBs. While prior studies have shown that open sexual attitudes are associated with higher rates of RSBs [27, 32], our study found that students holding neutral or acceptable attitudes toward premarital cohabitation had a lower risk of unprotected casual sex. A possible explanation is that these students view cohabitation as a stable phase in romantic relationships or a natural step toward marriage, making them more inclined toward committed relationships rather than casual sex without commitment [32]. This finding highlights the importance of integrating content on building stable partner relationships into sex education and HIV prevention services for college students. Similarly, those who had received HIV testing in the past year were more likely to engage in unprotected casual sex. Interestingly, these findings deviated from previous studies [33, 34]. A plausible explanation is that individuals who engage in unprotected casual sex are aware of their higher risk and proactively seek HIV testing services. This interpretation is supported by studies showing that individuals with higher levels of RSBs were more likely to get tested [35, 36]. This finding suggests that college students with a history of unprotected casual sex should be considered a priority population for HIV testing services. Moreover, these services should integrate HIV/STI risk reduction counseling and interventions to promote safer sexual practices. Similar to previous studies [37, 38], this study observed that a higher number of sexual partners was associated with a greater likelihood of unprotected casual sex. Individuals with multiple sexual partners generally exhibit higher levels of sexual activity, which significantly increases the likelihood of engaging in unprotected casual sex. A study reported that among male college students who engaged in casual sex, the proportion of those with multiple sexual partners was as high as 61.33% [39]. This study also validated the role of online partner‑seeking in unprotected casual sex. Students who sought sexual partners online were more inclined to engage in unprotected casual sex, which is consistent with previous studies [40, 41]. This phenomenon may be attributed to the anonymity of the internet and the widespread use of social networking applications, which provide sexually active students with a low-threshold and efficient online platform for partner seeking [10]. However, while offering convenience, online platforms also carry public health risks. The unknown health status of online partners may contribute to the continued spread of HIV and other STIs among young student populations. Existing studies have identified online partner seeking as an independent associated factor for HIV infection [41, 42]. These findings suggest that college sexual education should focus on high-risk groups with a history of multiple sexual partners and online partner-seeking behaviors, delivering targeted interventions that address their sensation-seeking traits and the context of online social networking. Furthermore, intervention efforts should be extended into online spaces such as social networking apps by embedding risk alerts, testing services, and protective skills training to mitigate the undermining effect of anonymous social environments on safe sexual practices. The stratified analysis revealed distinct patterns of association between heterosexual and non-heterosexual participants. For heterosexual students, the findings were generally consistent with the general population. Notably, HIV-related knowledge was not independently associated with unprotected casual sex among heterosexual students but emerged as a robust protective factor among non‑heterosexual participants. As a key population in HIV prevention and control, non-heterosexual individuals have long received targeted interventions, resulting in higher perceived susceptibility than their heterosexual counterparts [43]. No other significant interactions were detected, indicating that the associations of other factors with unprotected casual sex did not differ significantly by sexual orientation. Therefore, caution is needed when interpreting such differences, as they may reflect sampling variability rather than true effect modification. A consistent and robust finding across both strata was the strong association between seeking partners online and unprotected casual sex engagement. This underscores that online platforms serve as a primary facilitator of sexual networking for all college students [41]. These findings underscore the necessity of implementing differentiated sexual health interventions on college campuses. For non‑heterosexual students, targeted HIV education should be maintained and strengthened. For heterosexual students, knowledge‑based approaches alone may be insufficient, and behavioral interventions should be prioritized. Across both groups, online platforms offer a strategic opportunity for sexual health promotion, including digital education and rapid HIV self‑testing. Some limitations of this study should be acknowledged. First, this was a cross-sectional study. The reported associations should be interpreted as correlations rather than causal relationships. Second, self-reported data based on online surveys suffer from reporting bias. For instance, the 11.0% proportion of non‑heterosexual participants is likely an underestimate due to social desirability bias in the Chinese cultural context. In addition, missing values in some sexual behavior variables warrant cautious interpretation. Furthermore, the online survey did not allow for fully random sampling, as we could not randomly select classes at each grade level. Finally, the sample was drawn from 16 colleges predominantly located in the capital of Guangxi Province, limiting generalizability to students from other provinces or regions with different economic development levels. Nevertheless, the sample size of this study represents 9.2% (108,987/1,184,167) of the university student population in Guangxi, which to some extent indicates that the findings are reliable. Future research should employ computer‑assisted self‑interviewing to minimize social desirability bias. In addition, the role of dating applications in facilitating unprotected casual sex warrants further investigation. Finally, intervention studies targeting modifiable factors identified in the present study are warranted (e.g., online partner-seeking, HIV‑related knowledge). Conclusions This study reveals a relatively low prevalence of unprotected casual sex among sexually experienced college students in Guangxi. Unprotected casual sex is significantly associated with multiple cognitive and behavioral factors. These findings suggest the need to strengthen HIV education with a focus on condom use, to prioritize online platforms for risk reduction messaging and outreach, and to integrate HIV testing with personalized risk counseling. Acknowledgements We thank all researchers involved in this study and the staff of the Nanning Center for Disease Control and Prevention. Abbreviations Confidence intervals Human immunodeficiency virus Odds ratios Risky sexual behavior Sexually transmitted infections Author contributions PC, BL, HL and LY conceived and designed the study. JQ, ZL and JH analyzed the data and wrote the manuscript. WQ, JL, XH and QS were responsible for data collection. PC, YZ and BL further edited the manuscript and all gave final approval. All authors have critically reviewed the paper. Funding This work was supported by the Guangxi Key Research and Development Program (AB25069008), the Guangxi Natural Science Foundation (2023GXNSFAA026287), the Thousands of Young and Middle age Key Teachers Training Program in Guangxi Colleges and Universities (To Bingyu Liang), Nanning Scientific Research and Technological Development Project (20223051), and the National Natural Science Foundation of China under Grant (No. 82060610 and 82103899). Data availability The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Declarations Ethics approval and consent to participate This study has been reviewed and approved by the medical ethics committee of Guangxi Medical University (no. 2019-SB-088). Electronic consent was obtained from all participants before the survey and participants were allowed to withdraw from the survey at any process of the questionnaire. No personally identifiable information was involved in this survey and the results were used for data analysis only and not for commercial purposes. Participation in the survey was voluntary and participants did not receive any incentives for participating. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Jiao Qin, Zhifeng Lin and Jinfeng He have contributed equally to this work and share first authorship. Contributor Information Yuanjia Zheng, Email: 303771755@qq.com. Bingyu Liang, Email: liangbingyu@gxmu.edu.cn. Ping Cen, Email: cenping877@163.com. References Associated Data Data Availability Statement The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.