More romantic or more realistic: trajectories and influencing factors of romantic love among Chinese college students from entering college to graduation

Love is a central human emotion discussed across disciplines. College years represent emerging adulthood, a period of intense psychosocial development when romantic relationships are salient and can reduce risks of depression, antisocial behavior, and other health issues. Understanding college students’ attitudes toward love can predict partner choice standards, romantic/marital development, and inform interventions to prevent mental health problems. Existing research largely reflects Western contexts; cultural differences suggest the need to examine Eastern (Chinese) students. Given shifting Chinese social values and lowered marriage rates amid demographic transition, examining trajectories of romantic and realistic attitudes and their determinants among Chinese college students can inform guidance to foster healthy attitudes toward love and marriage and potentially support marriage intentions and policies.

Theories of love distinguish styles and components: Lee’s six styles (eros, ludus, storge, pragma, mania, agape); attachment-based styles (secure, anxious-ambivalent, avoidant); passionate vs. companionate love (Hatfield & Walster); and Sternberg’s triangular theory (intimacy, passion, commitment). Based on Knox and Sporakowski, this study differentiates romantic attitudes (idealized love as central criterion) and realistic attitudes (love as calm, reliable, incorporating practical considerations). Chinese cultural context historically linked love to marriage, with strong pragmatic tendencies; reforms increased acceptance of free choice and campus romance, yet pragmatic views persist alongside more romantic orientations. Developmental trajectories research shows stage-like romantic development with cultural timing differences; Asian adolescents often start later. Few studies have tracked trajectories of romantic vs. realistic attitudes specifically among Chinese college students. Influencing factors previously implicated include gender (men often more romantic, women more realistic), age (developmental changes across components), hometown (rural/urban differences with mixed findings), personality (extroversion linked to relationship formation), academic major differences in psychosocial profiles, ethnicity/race, siblings, BMI, self-esteem, and family background (parents’ education, socioeconomic status), with theoretical frameworks such as postmaterialism and social closure offering competing predictions. Identified gaps include lack of longitudinal trajectory analyses of love attitudes in Chinese cultural context and unclear mechanisms of multiple sociodemographic and psychological predictors.

Design and participants: Longitudinal panel survey over five consecutive academic years using probability proportional to size sampling. Data from 2,473 students (1,307 males, 1,166 females) who enrolled in 2008 were analyzed across freshman to senior years. Measures: Attitudes toward love assessed using a 23-item, 5-point Romantic Love Scale. Principal component analysis (oblique rotation) extracted five factors (romantic, realistic, unique, mysterious, resistant). Two common factors were analyzed: - Romantic attitudes: 4 items (e.g., “Love is the most important, and everything else can be ignored”), total score 4–20; Cronbach’s alpha across years: 0.68, 0.68, 0.73, 0.76. - Realistic attitudes: 4 items (e.g., considering partner’s financial condition, career impact, genetic influence, future development), total score 4–20; Cronbach’s alpha across years: 0.67, 0.67, 0.71, 0.73. Additional variables: gender (men=1, women=0), ethnicity (Han=1, minority=0), hometown (urban=1, rural=0), siblings (yes=1, no=0), major (engineering/science/agriculture/medicine=1; humanities/social sciences=0), BMI (low=0, normal=1, high=2), extroversion (1–9), age (years), father’s and mother’s education (years), family economic status (1–5, high to low), family social status (1–5, high to low), and self-esteem (Rosenberg 10-item; 10–50 higher=greater self-esteem). Analysis: Stata 15.0 for descriptive and Pearson correlations. Mplus 8.3 for growth mixture modeling (GMM) to identify latent trajectory classes (model comparison via AIC, BIC, SABIC, LMR-LRT, B-LRT, entropy, and theoretical interpretability). Multinomial logistic regression (Stata 15.0) assessed predictors of class membership for romantic and realistic attitudes.

Sample: 52.85% male (n=1,307), 47.15% female (n=1,166); mean age 19.6. Descriptive trends: Romantic attitudes increased from 9.12 (freshman) to 9.695 (junior) then slightly decreased to 9.622 (senior). Realistic attitudes >12 throughout, increasing from 12.183 (freshman) to 12.794 (junior) then decreasing to 12.527 (senior). Overall, realistic scores exceeded romantic scores. Correlations: Romantic and realistic attitudes were negatively correlated (r = -0.1492, p<0.05). Romantic attitudes positively correlated with gender (men=1; r=0.1886, p<0.05), major (STEM=1; r=0.0791, p<0.05), age (r=0.0542, p<0.05), and BMI (r=0.0608, p<0.05); negatively with hometown (urban=1; r=-0.0622, p<0.05), father’s education (r=-0.0348, p<0.05), mother’s education (r=-0.0374, p<0.05), and self-esteem (r=-0.1240, p<0.05). Realistic attitudes showed significant associations with multiple variables (text reports positive associations with age, siblings, extroversion, urban hometown, parents’ education; and negative with gender, major, ethnicity, family economic and social status; BMI not significant). GMM trajectories: - Romantic attitudes (3-class model; entropy=0.68): • High-increasing (1.05%): from ~7.81 to ~18.46; slope=3.043, p<0.001. • Low-decreasing (13.79%): from ~14.27 to ~10.54; slope=-0.766, p<0.05. • Low-increasing (85.16%): from ~8.31 to ~9.36; slope=0.347, p<0.001. - Realistic attitudes (3-class model; entropy=0.86): • High-increasing (1.58%): from ~7.56 to ~16.97; slope=2.212, p<0.001. • Low-increasing (96.97%): from ~12.20 to ~12.53; slope=0.124, p<0.05. • High-decreasing (1.46%): from ~15.94 to ~7.34; slope=-2.213, p<0.001. Predictors of class membership (multinomial logistic regression): - Romantic attitudes: Compared to the low-increasing group, men were more likely to be in the low-decreasing group (RRR=1.92, p=0.000). No other predictors significantly distinguished high-increasing vs low-increasing or high-increasing vs low-decreasing. - Realistic attitudes: Using high-decreasing as reference, major significantly predicted membership: high-increasing vs high-decreasing (RRR=0.21, p=0.007) and low-increasing vs high-decreasing (RRR=0.42, p=0.036), indicating STEM students were more likely in the high-decreasing group than humanities/social sciences students. Using low-increasing as reference, extroversion (RRR=0.76, p=0.009) and urban hometown (RRR=2.83, p=0.050) predicted high-increasing membership (more introverted and urban students more likely high-increasing).

Both romantic and realistic attitudes tended to increase across college, with realistic attitudes consistently higher—suggesting these orientations are not mutually exclusive and can co-exist. Cultural context likely shapes these patterns: in China’s collectivist society, partner choice often integrates family expectations and practical considerations, elevating realistic attitudes. Near graduation, pragmatic concerns (e.g., careers) may further reinforce realism. The small high-increasing subgroups for both attitudes indicate that a minority undergo marked changes. Gender differences aligned with prior literature: men were more likely to show declining romantic attitudes across college relative to low-increasing peers. Academic major differentiated realistic-attitude trajectories: STEM students were more likely to be in the high-decreasing realistic group than humanities/social sciences students. Personality and context mattered: lower extroversion and urban background increased the likelihood of high-increasing realistic trajectories. Some associations diverged from prior Western-centric findings (e.g., extroversion’s positive relation with realistic attitudes), potentially reflecting contemporary Chinese social dynamics and warranting further investigation. Overall, findings underscore the need for universities and families to monitor shifts in love attitudes and tailor interventions to gender, major, personality, and hometown differences to support healthy relationship development and mental health.

The study delineated three developmental trajectories for both romantic and realistic love attitudes among Chinese college students and identified key predictors of class membership. Realistic attitudes were generally higher than romantic attitudes. Gender influenced romantic-attitude trajectories (men more likely low-decreasing), while major, extroversion, and hometown influenced realistic-attitude trajectories (STEM students more likely high-decreasing; lower extroversion and urban background associated with high-increasing). These insights can guide universities and families in providing targeted education, counseling, and empowerment strategies (e.g., mental health courses, case simulations, emotional counseling) to foster mature, healthy attitudes toward love and inform policies addressing broader social concerns about marriage intentions and rates. Future research should expand intervenable variables and validate mechanisms across diverse sociocultural contexts.

- Generalizability may be limited due to selective sampling and cultural specificity; applicability to other populations/countries is uncertain. - Self-reported Romantic Love Scale may introduce bias; internal consistency (Cronbach’s alpha) was modest. - Limited set of intervenable variables due to data constraints; future work should include additional modifiable predictors.