Learning engagement and its associated factors among nursing students in China: a cross-sectional study based on latent profile analysis Authors: Yu-Lu Xie, Xu-Hua Zhou, Ying-Jun Zhang, Yan-Jia Li, Lin Chen Published: 2026 ABSTRACT AbstractObjectivesTo examine the latent profiles of nursing students' learning engagement and clarify the factors related to these distinct profiles.DesignA cross-sectional study.SettingsFour nursing schools in Sichuan Province, China.ParticipantsA convenience sample of 1250 full-time nursing students.MethodsDemographic characteristics, future time perspective, academic delay of gratification and learning engagement of nursing students were assessed using self-reported questionnaires. A latent profile analysis was used to identify the latent profiles of learning engagement among nursing students and use multivariate logistic regression to examine the factors associated with these identified traits.ResultsThree profiles of learning engagement were identified based on the nursing students’ responses to the learning engagement assessment, designated as ‘low learning engagement group’ (8.4%), ‘moderate learning engagement group’ (59.8%) and ‘high learning engagement group’ (31.8%). Multinomial logistic regression showed that, relative to the high learning engagement group, junior college qualification (OR=0.505, 95% CI 0.268 to 0.951, p=0.034) and higher academic delay of gratification (OR=0.959, 95% CI 0.927 to 0.991, p=0.013) were associated with lower odds of low engagement, while student leadership (OR=0.801, 95% CI 0.675 to 0.954, p=0.011) and higher future time perspective (OR=0.987, 95% CI 0.978 to 0.997, p=0.008) were associated with lower odds of moderate engagement.ConclusionMost nursing students were categorised into the moderate learning engagement group. Compared with the high learning engagement group, junior college qualification and academic delay of gratification were associated with lower odds of being classified into the low engagement group. Likewise, student leadership and future time perspective showed lower odds of being classified into the moderate engagement group. Future time perspective and academic delay of gratification show positive correlations with learning engagement, suggesting that these factors may be relevant targets for further research. Strengths and limitations of this study This study used the latent profile analysis to identify distinct subgroups of learning engagement among nursing students. This study used the latent profile analysis to identify distinct subgroups of learning engagement among nursing students. The large sample size recruited from four nursing schools across Sichuan Province enhances the statistical power and robustness of the findings. The large sample size recruited from four nursing schools across Sichuan Province enhances the statistical power and robustness of the findings. By examining both future time perspective and academic delay of gratification as predictors, this study provides a comprehensive understanding of the motivational factors associated with different learning engagement profiles. By examining both future time perspective and academic delay of gratification as predictors, this study provides a comprehensive understanding of the motivational factors associated with different learning engagement profiles. The cross-sectional design limits the ability to infer causal relationships between the identified predictors and learning engagement profiles. The cross-sectional design limits the ability to infer causal relationships between the identified predictors and learning engagement profiles. The convenience sampling reduced the generalisability of our findings. The convenience sampling reduced the generalisability of our findings. Introduction With the ageing population and the growing prevalence of chronic diseases worldwide, the demand for high-quality care services in society is constantly expanding.1 However, the healthcare system worldwide is currently confronted with a severe nursing shortage, driven by factors such as excessive workloads, elevated job stress, burnout, and high turnover rates.2 3 Nursing students constitute the essential reserve force for the future of primary healthcare provision, and enhancing their competency development is anticipated to play a pivotal role in addressing the ongoing shortage of nursing human resources.4 5 Learning engagement is a sustained, fulfilling and focused behavioural state manifested through active participation in educational activities.4 Recently, the engagement of nursing students in their learning has garnered increased attention from scholars, being widely acknowledged as a critical factor in fostering both the professional competence and personal development essential to the nursing profession.6 7 It extends beyond the mere acquisition of professional nursing knowledge, fostering critical thinking and self-directed problem-solving abilities through experiential learning within clinical practice.7 This competence is essential in nursing because it has a direct link to the quality of patient care in clinical settings. However, there is still some controversy about the findings of different studies on nursing students’ learning engagement. A prior study conducted in Hubei Province, China, revealed that nursing students exhibited moderate levels of learning engagement.8 Another study conducted among undergraduate nursing students in Saudi Arabia showed that the learning engagement of nursing students was at a low level.9 In addition, earlier studies have determined that nursing students’ learning engagement is affected by a variety of demographic characteristics such as gender, age and educational background.10 11 However, current studies on nursing students’ learning engagement do not take into account population heterogeneity and fail to accurately identify categories of students with various levels of learning engagement. Therefore, it is essential to determine categories of nursing students’ learning engagement to prepare the ground for effective interventions to improve their learning engagement and facilitate good academic performance. Generally regarded as a key predictor of learning engagement, the future time perspective represents a cognitive and emotional orientation that significantly shapes how individuals foresee and strategise for future endeavours.12 A previous study concluded that students who possess a strong future time perspective tend to adopt a more positive outlook on their destiny, demonstrate enhanced goal-setting abilities, exhibit greater engagement in their academic pursuits and achieve higher levels of productivity in their studies.13 Multiple studies have pointed to a positive interaction between future time perspective and learning engagement.14 15 Furthermore, nursing students, faced with substantial academic pressure, are often required to proactively establish reasonable future learning objectives, adapt their study plans with flexibility and intensify their study efforts to attain favourable academic outcomes.4 However, existing research predominantly focuses on exploring the impact of future time perspective on the overall learning engagement of nursing students. At the same time, insufficient attention has been given to its effects on the distinct profiles of learning engagement among nursing students. Thus, future time perspective was selected as a potential determinant in the analysis of the varied learning engagement profiles among nursing students in this study. Academic delay of gratification is defined as a student’s tendency to forgo immediate pleasures to attain more meaningful long-term academic goals, highlighting their self-discipline and self-regulatory capabilities within the learning environment.16 17 Empirical evidence suggests that academic delay of gratification makes a crucial contribution to learning engagement and functions as an indispensable protective factor against academic burnout and dropout.18 Specifically, students who demonstrate a propensity for delaying gratification in their academic endeavours are apt to excel in learning engagement through the use of effective learning strategies, increased effort and reflective practices on their prior learning experiences and methods.19 20 Thus, it is essential to investigate the impact of academic delay of gratification on the diverse profiles of learning engagement among nursing students. To conclude, most current research has focused mainly on the overall level of nursing students’ learning engagement, ignoring group heterogeneity. Further research is warranted to delineate the various categories of academic engagement among nursing students and to explore their associated factors, including demographic characteristics, future time perspective and academic delay of gratification. Therefore, this study aimed to address the following research question: what are the distinct latent profiles of learning engagement among nursing students and how are demographic characteristics, future time perspective and academic delay of gratification associated with these profiles? Latent profile analysis (LPA) provides a statistical approach for classifying research subjects on an individual basis according to their reaction patterns to extraneous variables.21 Accordingly, the purpose of this study was to clarify the potential profiles and influencing factors of learning engagement among nursing students through LPA. Methods Study design and participants This study is a secondary analysis of a previously conducted cross-sectional survey.22 A convenience sampling method was used to recruit full-time nursing students between May and July 2024 from four nursing schools in Sichuan Province, China, namely Sichuan Nursing Vocational College, Zigong Vocational and Technical College, Southwest Medical University and North Sichuan Medical College. These four schools were selected based on accessibility and willingness to participate, comprising two public vocational colleges and two public university-affiliated nursing schools, representing both junior college and undergraduate nursing education programmes in the region. This selection was intended to capture a diverse range of nursing students across different educational levels and institutional settings. The inclusion criteria included: (1) full-time enrolment in nursing programmes; and (2) voluntary consent to participate in the survey. The exclusion criteria applied were: (1) suspension from studies for personal reasons; and (2) incomplete questionnaire responses. To ensure variability across different year groups, we intentionally recruited nursing students from all 4-year levels rather than restricting our sample to a single grade. In this study, we conducted a power analysis using G*Power software (Germany) for the planned multinomial logistic regression with 11 independent variables. With a medium effect size, alpha of 0.05 and power of 0.80, the minimum required sample was 154. However, LPA typically requires a larger sample size; previous methodological studies have recommended a minimum of 500 cases for stable aggregation and accurate profile identification, as smaller samples may lead to convergence issues and unstable solutions.23 Given these additional demands, we targeted a substantially larger sample. Ultimately, 1250 valid questionnaires were collected, exceeding both the GPower-derived requirement and the recommended threshold for LPA and proving sufficient for all planned analyses. Measures General information form The form designed by the researchers encompassed: age, gender, education level, whether they were an only child, residence, per capita monthly household income, whether they served as a student leader, academic achievement and attitudes towards nursing. Utrecht Work Engagement Scale-Student The Utrecht Work Engagement Scale-Student (UWES-S) was originally compiled by Schaufeli and subsequently adapted into Chinese by Li and Huang.24 25 This scale comprises 17 items distributed across three dimensions: vigour, dedication and absorption. Each item is scored from 1 to 7 (1=never; 7=always). Higher scores indicate a stronger level of learning engagement. Reliability indices for the UWES-S were excellent (α=0.964). General Future Time Perspective Scale The General Future Time Perspective Scale (GFTPS) designed by Song was used to assess nursing students’ future time perspective.26 This 20-item self-report instrument assesses five dimensions: future imagery, future efficacy, purpose consciousness, behavioural commitment and far-target orientation. Responses are measured on a 4-point Likert scale, ranging from 1 (strongly disagree) to 4 (strongly agree), with higher total scores indicating a more favourable FTP. The GFTPS demonstrated adequate reliability indices (α=0.886). Academic Delay Of Gratification Scale Nursing students’ academic delay of gratification was evaluated using the Academic Delay Of Gratification Scale (ADOGS) developed by Bembenutty and revised into a Chinese version by Du.16 27 The scale consists of two dimensions including 10 items: classroom academic delay of gratification and after-class academic delay of gratification. The ADOGS uses a 4-point Likert scale, where each item presents a dilemma scenario. All items are scored from 1 to 4, with higher total scores reflecting a greater capacity for academic delay of gratification among nursing students. The ADOGS demonstrated adequate internal consistency (α=0.896). Data collection All questionnaires were disseminated via the Questionnaire Star online platform. Although data collection was fully online, participants completed the questionnaire during a scheduled class session with a trained liaison officer present on-site to provide standardised instructions and supervise the process. The liaison officers did not influence students’ responses; their role was limited to ensuring participants understood the instructions and completed the survey independently. Questionnaires with a completion time exceeding 15 min were considered invalid. We set a minimum completion time threshold of 15 min to ensure data quality, as the questionnaire was estimated to require approximately 15 min to complete based on our pilot test. Questionnaires completed in less than 15 min were considered indicative of careless or rushed responding and were excluded from the final sample. To assess potential bias, we compared the demographic characteristics of excluded participants with those of the final sample and found no significant differences in age, gender or year of study (all p>0.05), suggesting that this exclusion did not introduce substantial selection bias. To avoid duplication of responses, IP addresses were restricted so that each IP could only answer once, and participants were only allowed to submit the questions after they had finished completing all of them. Finally, the investigator and research team members double-checked the returned questionnaires for errors or omissions and asked participants to make corrections or provide additional information if necessary, and no additional exclusions were made at this stage. Of the 1396 nursing students invited, 1382 accessed the survey, 1315 consented and 1250 valid questionnaires were ultimately included in the final analysis (65 excluded: 57 with completion time <15 min and 8 with regular responses), yielding a response rate of 89.5%. The participant flow diagram is presented in figure 1. Participant enrolment process. Data analysis Data were analysed using SPSS V.27.0 software and Mplus V.8.3 (USA). Descriptive statistics were applied for participants’ demographic characteristics and questionnaire scores consisted of means, SD, frequencies and proportions. Latent profiles of nursing students’ learning engagement were analysed using Mplus 8.3. The 17-item scores of the UWES-S were used as indicators in the LPA. Although the UWES-S comprises three established dimensions, item-level scores were chosen because LPA aims to detect nuanced response patterns across individual items, whereas dimensional scores would obscure individual-level heterogeneity. All 17 items showed acceptable skewness and kurtosis, supporting their use in LPA. Models with 1 to 5 profiles were sequentially estimated, and the optimal number was determined based on fit indices and theoretical interpretability. The following model fit indices were applied to determine the optimal number of profiles: Akaike information criterion (AIC), Bayesian information criterion (BIC), adjusted BIC (aBIC), Lo-Mendell-Rubin test (LMRT), bootstrap likelihood ratio test (BLRT) and entropy. Lower values of AIC, BIC and aBIC indicate better model fit.28 The LMRT and BLRT are significance tests used to compare models with k profiles against those with k-1 profiles, and a significant p value indicates that the model with k profiles offers a more favourable fit.28 The value of entropy ranges from 0 to 1, with the value approaching 1 indicating greater classification accuracy. Additionally, information entropy ≥0.80 signifies that at least 90% of participants have been accurately categorised, serving as a robust indicator of classification precision. The χ2 test was used to compare the potential profiles of nursing students’ learning engagement with different demographic characteristics. The differences in future time perspective and academic delay of gratification on different profiles were analysed by analysis of variance (ANOVA). Post hoc analysis was conducted to facilitate comparisons between profiles. Multiple logistic regression was adopted to explore the predictors of the profile membership of learning engagement among nursing students. A p value of <0.05 was considered statistically significant (two-sided). Results Demographic characteristics and scores of the main variables Detailed demographic characteristics and descriptive statistics of the main variables are presented in online supplemental appendix 1, as these descriptive findings derived from the same dataset have been previously reported elsewhere.22 Latent profile analysis of learning engagement among nursing students The 17 entries of the UWES-S were used as exogenous indicators, and 1–5 candidate patterns were consecutively identified for latent profile analyses of nursing students’ learning engagement (table 1). To determine the optimal number of latent profiles, we considered multiple criteria simultaneously, including AIC, BIC, aBIC, entropy, LMR, BLRT, class size, interpretability and parsimony. Although AIC, BIC and aBIC continued to decrease from the four-profile and five-profile solutions, the rate of decline decelerated substantially after three profiles, indicating diminishing returns in model fit. The three-profile model achieved the highest entropy, suggesting superior classification accuracy. Importantly, the four-profile model yielded one class accounting for only 6.1% of the sample, which falls below the recommended threshold for reliable and replicable class identification. The five-profile model further exacerbated this issue, producing two small classes. Moreover, the additional profiles in the four-profile and five-profile solutions did not reveal qualitatively distinct characteristics beyond those captured by the three-profile model, making their substantive interpretation difficult. Given the optimal balance of entropy, parsimony, class size adequacy and theoretical meaningfulness, the three-profile model was retained as the final solution. Table 1Latent profile analysis of learning engagement with fit indicesModelAICBICaBICEntropyLMRTBLRTCategorical probabilityAverage posterior probability172 800.51372 974.96372 866.964––– 264 521.08964 788.63564 623.4600.9310.00010.00010.422/0.5780.941/0.938 3 60 292.41260 651.57560 429.223 0.954 0.0040 0.0043 0.084/0.598/0.318 0.931/0.892/0.927 457 902.24758 353.76758 074.2390.9400.00540.00570.061/0.194/0.328/0.4170.915/0.887/0.903/0.921557 048.80657 592.68157 255.9770.9380.16600.11650.050/0.277/0.400/0.218/0.0540.902/0.876/0.914/0.889/0.923Bold values indicate the key indicators of Model 3 selected for their significance.aBIC, adjusted Bayesian information criterion; AIC, Akaike information criterion; BIC, Bayesian information criterion; BLRT, bootstrap likelihood ratio test ; LMRT, Lo-Mendell-Rubin test. Latent profile analysis of learning engagement with fit indices Bold values indicate the key indicators of Model 3 selected for their significance. aBIC, adjusted Bayesian information criterion; AIC, Akaike information criterion; BIC, Bayesian information criterion; BLRT, bootstrap likelihood ratio test ; LMRT, Lo-Mendell-Rubin test. As depicted in figure 2, nursing students’ learning engagement was categorised into three latent profiles. Profile 1, comprising 8.4% of the total sample, exhibited low learning engagement and was referred to as the ‘low learning engagement group’. Profile 2, representing 59.8% of the sample, demonstrated moderate learning engagement and was identified as the ‘moderate engagement group’. Profile 3, encompassing 31.8% of the sample, showed high learning engagement and was classified as the ‘high engagement group’. Latent profile analysis result of learning engagement. The three profiles differed statistically significantly in terms of the total learning engagement score and the vigour and absorption dimensions (table 2). Post-hoc analyses demonstrated that profiles 1–3 had progressive levels of vigour and absorption and overall learning engagement. Whereas, the three profiles were similar in terms of dedication with no statistical difference. Table 2Profiles differences in learning engagement and the results of post hoc analysisVariablesProfile 1Mean±SDProfile 2Mean±SDProfile 3Mean±SDFP valuePost hoc analysisVigour25.60±6.2627.14±6.9427.67±6.765.4200.0041<2<3Dedication12.46±4.1711.90±4.0011.72±3.872.0200.1331, 2, 3Absorption27.12±7.1128.54±6.9029.27±6.277.478<0.0011<2<3Overall learning engagement75.85±17.9480.08±18.5181.81±17.426.983<0.0011<2<3 Profiles differences in learning engagement and the results of post hoc analysis Predictors of latent profiles of learning engagement There were significant variations in education level, future time perspective, and academic delay of gratification among nursing students with different learning engagement profiles (table 3). A multiple logistic regression analysis was conducted with learning engagement patterns being the outcome variable, demographic characteristics and future time perspective and academic delay of gratification being the independent variables, and the ‘high learning engagement group’ serving as the reference. The results of multiple logistic regression analysis revealed that education level, serving as a student leader, the level of future time perspective, and academic delay of gratification were the predictive variables of the learning engagement patterns. Specifically, compared with the high learning engagement group, those with junior college qualifications (OR=0.505, 95% CI 0.268 to 0.951) and higher academic delay of gratification scores (OR=0.959, 95% CI 0.927 to 0.991) were associated with lower odds of being classified into the low learning engagement group. In addition, nursing students who served as student leaders (OR=0.801, 95% CI 0.675 to 0.954) and those with higher future time perspective scores (OR=0.987, 95% CI 0.978 to 0.997) also showed lower odds of being in the low learning engagement group compared with the high engagement group. Table 3Logistic regression of different learning engagement profilesVariablesLow learning engagement group (ref. high learning engagement group)Moderate learning engagement group (ref. high learning engagement group)BSEOR95% CIP valueBSEOR95% CIP valueAge (years) <18−0.7410.6510.477(0.433 to 1.708)0.255−0.0690.2500.933(0.572 to 1.522)0.781 ≥18Ref.Ref.Gender Male0.0390.2941.039(0.584 to 1.851)0.8950.0800.1151.084(0.865 to 1.358)0.486 FemaleRef.Ref.Education level Technical secondary school0.2180.6451.244(0.351 to 4.403)0.735−0.2240.2550.799(0.485 to 1.318)0.380 Junior college−0.6830.3230.505(0.268 to 0.951)0.034*−0.1820.1080.833(0.674 to 1.030)0.091 UndergraduateRef.Ref.Only child Yes−0.9360.2910.585(0.330 to 1.035)0.0660.0100.0991.010(0.832 to 1.226)0.918 NoRef.Ref.Residence Urban−0.0400.2450.961(0.595 to 1.553)0.871−0.0690.0930.933(0.778 to 1.119)0.456 RuralRef.Ref.Per capita monthly household income (¥) ≤30000.0680.3501.070(0.539 to 2.124)0.846−0.0870.1340.917(0.705 to 1.192)0.516 3001–5000−0.0760.3430.926(0.473 to 1.816)0.824−0.1090.1300.897(0.696 to 1.156)0.401 >5000Ref.Ref.Serve as a student leader Yes0.0100.2251.010(0.650 to 1.571)0.964−0.2220.0880.801(0.675 to 0.954)0.011* NoRef.Ref.Academic achievement Excellent0.0030.3321.003(0.523 to 1.923)0.9940.1620.1361.175(0.901 to 1.533)0.233 Favourable−0.2830.2380.753(0.473 to 1.200)0.2330.0060.0941.006(0.837 to 1.210)0.947 OrdinaryRef.Ref.Attitudes towards nursing Approved−0.0450.6380.956(0.273 to 3.340)0.9430.3950.2801.485(0.858 to 2.569)0.158 Disapprove0.4490.6601.567(0.430 to 5.719)0.4960.4280.2961.534(0.859 to 2.738)0.148 Not to matterRef.Ref.Future time perspective0.0070.0131.007(0.981 to 1.023)0.614−0.0130.0050.987(0.978 to 0.997)0.008*Academic delay of gratification−0.0420.0170.959(0.927 to 0.991)0.013*0.0000.0071.000(0.986 to 1.014)0.975*p<0.05ref., reference. Logistic regression of different learning engagement profiles p<0.05 ref., reference. Discussion As far as we are aware, this study constitutes the first attempt to identify the potential profiles of learning engagement and their associated factors among nursing students using LPA. Unlike our earlier publication,22 which used variable-centred regression methods to examine linear associations, the present study adopts a person-centred approach to uncover heterogeneous subgroups. Our findings indicated that nursing students’ learning engagement can be categorised into three latent profiles based on their total learning engagement scores: low learning engagement group (8.4%), moderate learning engagement group (59.8%) and high learning engagement group (31.8%). Education level, serving as a student leader, future time perspective and academic delay of gratification were predictors of the profiles of learning engagement among nursing students. These findings extend our prior variable-centred research by delineating distinct engagement profiles and their predictors, enabling targeted strategies that address nursing students’ heterogeneous needs. Three distinct potential profiles of learning engagement were discerned, illustrating the individual heterogeneity among nursing students. The majority of these students were classified into the moderate learning engagement group. This moderate level of engagement may be attributed to the expanding scope of nursing knowledge and clinical skills that students are required to master, which significantly burdens their academic workload. The observed level of learning engagement among nursing students aligns with findings from a prior study, which reported that nursing students typically exhibit a moderate level of engagement in learning activities.8 The rapidly varying domain of knowledge in nursing and the gradual growth of social demand for high-quality nursing services have contributed to the fact that nursing students are required to master more and more content and skills, which significantly burdened their academic workload, thus inducing academic burnout and ultimately leading to a reduction in motivation to learn.9 This situation may be further exacerbated by the increasing proportion of younger nursing faculty in colleges, who may lack sufficient expertise in effectively applying advanced pedagogical methods to sustain student interest and engagement.29 30 Finally, it is worth noting that the low engagement group exhibited a low level of commitment to learning, which was mainly reflected in the low scores of vigour and absorption. The low scores on vigour and absorption among this group suggest that these students lack the psychological energy and cognitive focus necessary for deep learning. This finding is particularly concerning, as these students may be at risk of academic underperformance and professional unpreparedness, potentially affecting their future clinical competence. From a practical standpoint, the identification of this at-risk group enables nursing educators to implement early targeted interventions, such as individualised academic counselling, peer mentoring programmes and motivational interviewing, to prevent these students from falling further behind. Early detection and tailored support for this group could be a cost-effective strategy to improve overall student outcomes and ultimately enhance the quality of the future nursing workforce. Therefore, nursing educators should precisely identify the heterogeneity in nursing students’ learning engagement, enabling them to implement tailored interventions that foster the development of effective study habits, encourage the setting of ambitious learning objectives, and ultimately enhance students’ overall engagement in their educational pursuits. In this study, education level was associated with learning engagement among nursing students. Specifically, junior college students had lower odds of being in the low learning engagement group than undergraduate students, compared with the high engagement group. This finding aligns with previous studies, which have revealed a negative correlation between educational attainment and learning engagement.31 32 A plausible explanation is that junior college students, facing greater employment pressure due to lower academic qualifications in a competitive healthcare job market, are compelled to establish clearer learning goals and invest more effort into their studies compared with their undergraduate counterparts.19 In contrast, undergraduate nursing students may experience less immediate employment pressure, which could lead to a reduced sense of urgency in their academic pursuits. These findings carry important implications for nursing educators and curriculum designers: rather than assuming that students in higher education programmes are inherently more motivated, institutions should proactively monitor engagement levels across all academic programmes and allocate additional support resources to undergraduate nursing students who may be experiencing reduced academic urgency. Moreover, career development interventions that highlight the long-term benefits of academic excellence, even for those with higher qualifications, may help sustain engagement among this group. Our study demonstrated that student leadership experience was associated with lower odds of low learning engagement, suggesting a greater likelihood of being in the high engagement group. This indicated that nursing students who hold leadership roles are more inclined to devote greater energy and attention to actively participating in learning activities. This phenomenon may be explained by the fact that student leaders have more opportunities for interaction and dialogue with their teachers and supervisors in academic and professional contexts.33 These interactions enable them to develop more effective study methods, better plan and allocate their study time, and ultimately improve their overall academic efficiency. Furthermore, in areas requiring a range of competencies, such as writing, time management and presentation skills, student leaders typically gain more practical experience and more keenly recognise the benefits of learning, which subsequently reinforces their commitment to academic pursuits.34 Finally, in their efforts to set a positive example, student leaders tend to be more diligent in self-monitoring and display a greater motivation for learning.35 It is noteworthy that nursing students with better future time perspective were unlikely to have lower levels of learning engagement. This is in line with a previous study, which pointed out that future time perspective was positively related to learning engagement.36 This situation may be attributed to the fact that a future time perspective cultivates a positive attitude toward career development among nursing students, enabling them to establish clear goals and objectively analyse their future prospects to plan effectively. As a result, they are more inclined to dedicate greater effort to their current studies, enhancing their skills and increasing their level of engagement in learning.37 These findings suggest that leadership development may serve as a protective factor against low academic engagement. Practically, nursing schools should consider integrating leadership training into the formal curriculum through peer tutoring, student representative roles and research assistant opportunities, while ensuring equitable access for all students, particularly those with early signs of disengagement. Creating structured leadership pathways may simultaneously enhance students’ academic engagement, professional identity and career readiness. Our finding revealed that academic delay of gratification was positively associated with learning engagement, with higher scores linked to greater odds of being in the high engagement group. In this study, academic delay of gratification was conceptually defined as students’ tendency to forgo immediate pleasurable alternatives in pursuit of more valuable long-term academic objectives, and operationally measured using the ADOGS, with higher scores indicating a greater capacity for self-regulation in academic contexts. Previous research has demonstrated that nursing students with a high level of academic delay of gratification possess a strong perseverance that enables them to achieve their goals.38 39 They are skilled at devising suitable study plans and objectives and consistently putting them into practice, which fosters a high degree of enthusiasm for learning.18 In contrast, students with low levels of delayed gratification are more prone to external temptations and are likely to sacrifice long-term goals for immediate pleasure, often leading to academic procrastination and burnout.19 The significant association between academic delay of gratification and learning engagement highlights the importance of self-regulation. Nursing educators should integrate self-regulation training into the curriculum for low-engagement students through goal-setting, time management and self-monitoring. A supportive environment with regular feedback may further strengthen these skills, ultimately enhancing both academic engagement and professional competence. Drawing from these insights, fostering students’ academic delay of gratification is recommended as an effective approach to enhancing learning engagement among nursing students. Limitations Some limitations should be considered. First, convenience sampling was used to recruit participants, which may introduce non-response bias and limit the generalisability of our findings beyond the specific sample. Second, reliance on self-report questionnaires rather than objective records may introduce recall and social desirability biases, and the use of self-reported academic achievement may further affect the accuracy of related estimates. Third, the cross-sectional design precludes any causal or directional inferences. Therefore, all findings should be interpreted as associational rather than causal. Fourth, although participants were recruited from multiple nursing schools, we did not adjust for potential clustering effects of school-level characteristics in our analysis, which may have underestimated SEs and limited the generalisability of our results beyond the participating institutions. Fifth, the LPA used to identify engagement groups involves inherent classification uncertainty, as individuals are assigned to profiles based on estimated probabilities rather than observed memberships, which may affect the stability and replicability of the profiles. Finally, the main variables included in this study are predominantly dynamic indicators, and their trajectories and correlations over time remain to be examined using longitudinal designs in future research. In addition, given that all participants were recruited from Sichuan Province, the findings may not be generalisable to nursing students in other regions or educational systems. These limitations should be considered when interpreting the results of this study. Implications It is crucial to consider the level of learning engagement and associated factors to facilitate the efficiency of nursing students’ academic endeavours and achievement. Nursing educators should proactively assess potential patterns of learning engagement among nursing students and develop targeted strategies to enhance learning outcomes. These strategies should be based on a comprehensive analysis of key factors, such as the student’s education level, leadership experience, future time perspective and capacity for academic delay of gratification. Additionally, owing to the dynamic and diversified nature of learning engagement, students ought to be evaluated appropriately on a regular basis so that nursing educators can provide them with prompt assistance. Conclusion This study used LPA to identify three distinct profiles of learning engagement among nursing students: ‘low learning engagement group’ (8.4%), ‘moderate learning engagement group’ (59.8%) and ‘high learning engagement group’ (31.8%). Most nursing students were categorised in the moderate learning engagement group. Junior college qualification and academic delay of gratification were associated with lower odds of low learning engagement, while student leadership and future time perspective were associated with lower odds of moderate learning engagement, compared with the high engagement group. Given these observed associations, interventions targeting academic delay of gratification may be a promising avenue for future research, although causal conclusions await longitudinal or experimental designs. Supplementary material