It is evident that higher education plays a pivotal role in the competitive and sustainable development of a nation. This assertion is corroborated by research findings that demonstrate a positive correlation between university education and higher economic income, in comparison to those who do not attain this level of education. ¹ The impact of this phenomenon is twofold: it is beneficial to the individual, and it is also conducive to the development of a skilled workforce that is able to drive innovation and productivity. Nevertheless, the efficacy of the system is undermined when educational institutions are unable to maintain a high retention rate of their students. This indicates a discrepancy between the system's design and the actual requirements of the population. ^{2– 4}

In order to optimise their potential, it is essential that systems guarantee not only access but also retention, by integrating mechanisms that identify and mitigate early barriers. The economic return on higher education has encouraged its promotion as an engine of competitiveness. However, this process faces a critical challenge in the limited functionality of education systems in retaining students. ^{5, 6}

This problem is indicative of a multifaceted nature, manifesting as a consequence of numerous underlying factors. These include, but are not limited to, inadequate curricular adaptation and an absence of sufficient consideration for psychosocial elements such as academic stress and identity disconnection. These elements are further compounded in contexts characterised by inequality, where institutional entities perpetuate exclusionary dynamics rather than addressing them. This phenomenon is indicative of the inefficiencies inherent in systems that prioritise access over retention, without addressing the underlying causes, such as a lack of prior academic preparation or the absence of emotional support networks. First-generation university scholarship recipients, in particular, encounter a “culture shock” when entering academic environments that fail to consider their sociocultural backgrounds. This can result in feelings of isolation and stress. Furthermore, the presence of institutional bureaucracy, manifesting in the form of protracted processes for the resolution of administrative issues, has been demonstrated to intensify feelings of frustration and result in elevated rates of attrition. This issue has been identified as a priority challenge for global agendas, as demonstrated by the case of students who, despite being recipients of state scholarships, discontinue their studies due to economic, psychological or institutional factors. ^{7, 8}

In Peru, the proportion of young people who have access to higher education is only 30%, which is lower than the proportion in countries such as Chile and Colombia. ⁹ This disparity is further accentuated when analysing expenditure per individual: while two out of ten young people in the lowest quintile manage to enrol, this figure rises to five out of ten in the highest quintile. The geographical concentration of educational resources in the capital, Lima, has been identified as a contributing factor to the exacerbation of the educational disparities between regions. A significant proportion of university enrolment is concentrated in Lima, with 41% of all enrolments being located within the city. This has the effect of limiting access to educational opportunities in regions that may not possess the same level of educational infrastructure as Lima. This phenomenon of territorial inequality is indicative of systems that demonstrate a bias towards urban and higher-income populations, thereby marginalising those inhabiting rural or peri-urban areas. Furthermore, the absence of public universities in the provinces compels a significant number of students to migrate, resulting in expenses that, when combined with job insecurity, heighten the likelihood of attrition. It is evident that, in the absence of effective decentralisation and regional inclusion policies, the system is likely to continue exhibiting exclusionary tendencies.

Conversely, 65% of the university-age population encounter economic impediments to accessing higher education, a phenomenon that has given rise to policies such as scholarships and educational loans. ^{10– 12} Nevertheless, these mechanisms frequently prove to be inadequate when not accompanied by academic and psychosocial support. Consequently, scholarship recipients from public educational institutions may encounter deficiencies in their prior learning, which can result in their falling behind and, ultimately, dropping out of their studies. Furthermore, the paucity of information regarding the requirements for maintaining a scholarship, such as minimum grade point averages, has been shown to engender feelings of anxiety and demotivation among scholarship recipients.

Consequently, the phenomenon of university dropout can be considered as multi-causal, as posited by, ¹³ thus necessitating interdisciplinary approaches to facilitate comprehension. From a psychological standpoint, factors such as academic self-concept, resilience to failure, and disconnection from the educational project have been demonstrated to influence the decision to discontinue one's education. ¹⁴ Conversely, the sociological approach to understanding the phenomenon of dropout, as articulated in the seminal work of, ¹⁵ undertakes a multifaceted examination of salient variables including family income, parental educational attainment, and gender.

A comprehensive understanding of the phenomenon of scholarship students dropping out of higher education necessitates an examination of both socioeconomic and psychological factors. As asserted by Refs. ( 16 and 17), economic precariousness imposes limitations on both access and retention. To illustrate this point, consider a scholarship student who is obliged to contribute financially to their family. In such a scenario, they may find themselves compelled to prioritise securing informal employment over their academic pursuits. From a psychological standpoint, the pressure to maintain high grades in order to retain scholarship funding has been shown to engender chronic stress, which in turn has a detrimental effect on performance. The interaction of these factors within education systems that lack safety nets, such as supplementary subsidies or academic flexibility, results in scholarships serving merely as temporary remedies rather than comprehensive solutions. In order to break this cycle, there is a necessity for policies that recognise the multiple dimensions of vulnerability. Such policies must integrate financial, academic and emotional support in a synergistic manner.

The Peruvian university sector is currently undergoing a series of reforms to its National Policy on Higher University Education. A key element of these reforms is the identification of the most pertinent factors to inform the development of strategies to address the issue of increasingly restricted access to university life. This assertion is supported by empirical evidence derived from rigorous research. The present study seeks to explore the psychosocial factors that explain academic dropout among scholarship students at a public university in 2024. The primary motivation underlying this research endeavour is to make a substantial scientific and practical contribution to policy makers in the domain of strategy design.

The results of the study are presented sequentially, commencing with a descriptive analysis of the characteristics of the respondents to identify the perceptions of those students and the SEM analysis to address the research objective.

The majority of the students surveyed are female (53%) and male (47%), and are enrolled in the following degree programmes: human medicine (18%), administration (16%), accounting (16%), and economics (15%).

In relation to social factors, the majority of students expressed disagreement with the notion that family members exert influence over their decision-making processes (42%). Conversely, a significant proportion of students concurred with the assertion that members of their immediate social circle tend to offer mutual support (43%) and acknowledged the presence of considerable challenges within their family environment (37%) ( Figure 2).

With regard to the psychological factors under consideration, it is evident that the majority of students (46%) adopt a proactive stance towards their university studies, while a significant proportion (44%) express a strong motivation to persevere in their academic pursuits. Furthermore, there is a notable consensus among the student body that their approach to academic endeavours is consistently positive (54%).

With regard to the matter of academic integration, it is evident that students have expressed their disagreement with inflexible schedules (47%), as well as the notion that university teachers are more demanding than school teachers (44%), and the perception that the university environment is more arduous than the school environment (43%).

With regard to social integration, students indicate that they strongly disagree with the assertion that they encounter any difficulty in living alone (53%), and also with the assertion that their family has experienced any financial difficulties (50%). Furthermore, they express disagreement with the statements that they are not entirely certain about their chosen career (35%) and that they have experienced any health problems (35%) in Table 2.

The present study aims to evaluate the psychosocial factors that explain academic dropout among scholarship students at a public university. To this end, a Structural Equation Model (SEM) analysis will be performed in order to pay close attention to construct validity, reliability, and the testing of hypotheses that contribute to this research.

With regard to the individual reliability of the items, all weights are above 0.707, with the exception of fifteen items belonging to the social, psychological, academic integration, and social integration constructs. Nevertheless, these items were retained on the basis that their content is also important in defining each of the constructs. In addition, the individual reliability of the items is adequate, as demonstrated in Table 3. ⁴⁸ Initially, an evaluation was conducted to ascertain the potential for multicollinearity between the constructs. ⁴⁹

In each of the partial multiple regressions, the VIF index for each of the exogenous constructs is less than 5. Therefore, we can affirm that there are no multicollinearity problems between the exogenous constructs of each endogenous variable, as shown in Table 3.

With regard to the internal consistency of the measurement scales, it is evident from Table 4 that the composite reliabilities of the constructs are greater than 0.7, with the exception of the psychological constructs. This observation serves to affirm the importance of the items or manifest variables for the model. In addition, the elevated Cronbach's alpha values (IA: α = 0.757; IS: α = 0.682; PS: α = 0.616; SO: α = 0.789) substantiate the inherent consistency and reliability of the instruments.

The AVE (Average Variance Extracted) coefficient represents the proportion of variance in the construct that can be explained by its indicators, and it has been verified that this coefficient is greater than 0.5 for all constructs, as shown in Table 5. We can therefore affirm that there is convergent validity for each of the latent variables. ⁵⁰

Discriminant validity analysis between constructs by calculating HTMT coefficients ⁵¹ and using the criteria of. ⁵⁰ Table 6 shows that all constructs achieve discriminant validity according to the Forner-Lacker criterion and the stricter HTMT criterion ⁵² therefore, all constructs measure different aspects.

In order to analyse the structural model and conduct hypothesis testing, it is hypothesised that a relationship exists between the constructs. The findings indicate that social factors related to integration with the environment and the family environment do not provide a satisfactory explanation for academic dropout as measured by academic integration (β = 0.228, t = 0.534, p < 0.593). Similarly, these factors do not provide a satisfactory explanation for academic dropout measured by social integration (β = 0.227, t = 0.579, p < 0.563). Contrary to the psychological factors that have been measured in previous studies on behavioural and subjective norms, it has been demonstrated that attitudes and subjective norms are more effective in explaining academic dropout when measured by academic integration (β = 0.836, t = 2.009, p < 0.045) and social integration (β = 0.747, t = 1.909, p < 0.056) ( Table 7).

Conversely, the results of the SEM analysis are presented for each of the 30 items of the four constructs considered: social integration, psychological integration, academic integration, and social integration. The results indicate that four items pertaining to the psychological constructs PS11, PS21, and PS32, as well as the social integration item IS12, are non-significant ( Table 8).

On the other hand, the structural equation model uses the results of confirmatory factor analysis, whose indices yield the following results: χ2 = CMIN/df = 1.64, GFI = 0.923, AGFI = 0.955, NFI = 0.906, CFI = 0.996, RMSEA = 0.081, and SRMR = 0.100, demonstrating that the indices of the constructs in the estimated model are adequate to explain the social and psychological factors that account for academic dropout among scholarship students at a Peruvian public university ( Table 9 and Figure 3).

The findings of the present study demonstrate that psychosocial factors are a contributing element in the decision of scholarship students to withdraw from their academic studies. These results are in alignment with those of previous research conducted by ^{27, 34} The present study identifies academic self-regulation, self-efficacy, and social support as key determinants of university persistence. These findings correspond with those of the aforementioned authors, who also identified the importance of social integration and the family environment in student retention. The present study's findings demonstrate a robust correlation between the capacity to establish amicable relationships and the capacity to coexist harmoniously with individuals who have chosen to discontinue their academic pursuits. This observation aligns with the identification of institutional commitment as a protective factor in the systematic review conducted by the aforementioned authors.

In addition, the results of the present study are consistent with those of the research conducted by Ramos ³¹ and Tlalajoe-Mokhatla. ²³ These researchers demonstrated that low academic performance, workload, and negative perceptions of the family environment have a detrimental effect on student motivation and increase the risk of dropout. The findings of the present study demonstrate that the family environment is a significant predictor of student attrition, thus indicating that a paucity of support from family members can directly influence students' decisions to abandon their studies.

In contrast, the findings of Vega et al. and Vilchez ^{30, 53} serve to reinforce the evidence that psychological factors, including depressive symptoms, suicidal ideation, insomnia, and insufficient parental support, are associated with an elevated risk of academic disengagement. The present study has identified a correlation between psychological well-being and emotional skills on the one hand, and university retention on the other. This finding highlights the importance of psychosocial interventions within higher education institutions in order to mitigate the adverse effects identified.

Kocsis & Molnár ⁵⁴ The psychological, realm our findings on mental health as a protective factor are in accordance with those of García et al. ¹³ who observe that “the pressure to maintain high grades in order to retain the scholarship generates chronic stress, affecting performance” (p. 12). Nonetheless, there is a divergence of opinion regarding the magnitude of the effect. Whilst the aforementioned study Garcia et al. ¹³ attributes greater weight to economic variables, the present study demonstrates that “knowledge of institutional norms” is the most influential predictor. This finding is in accordance with that of the study Kocsis and Molnár ⁵⁴ conducted by the second author, who identified that clarity in academic expectations reduces uncertainty among scholarship students.

In contrast to the findings presented above, the study by Nuñez ²⁷ emphasised the role of institutional and academic factors in dropout rates. This differs partially from the results obtained in this study, since the model developed herein highlighted the influence of subjective norms in the education system as the most influential factor. This finding suggests that enhancing comprehension and adaptability to institutional stipulations may substantially mitigate the inclination to withdraw among scholarship students.

In the Peruvian context, Salazar ³⁹ dentified that factors such as rural origin and mother tongue have an impact on scholarship loss. This finding is at odds with our results, which indicate that the family environment, rather than demographic variables, is the critical factor. The observed discrepancy may be attributed to methodological disparities, given that Salazar ³⁹ employed probit models, whereas our quantitative methodology placed greater emphasis on psychosocial constructs. Nevertheless, the consensus of both studies is that there is a requirement for comprehensive policies, as suggested by ⁵⁵ : “Scholarships must be complemented by mentoring and curricular flexibility to address multidimensional vulnerabilities” (p. 320).

Finally, the work of Franz & Paetsch ²⁵ on social integration in working students corroborates our observation that “those employees with strong university networks are less likely to drop out” (p. 5), although their study focuses on teaching careers, while the present study covers multiple disciplines. This convergence underscores the universality of social integration as a retention factor, irrespective of the academic context.

The present study identified the psychosocial factors that explain academic dropout among scholarship students at a Peruvian public university, thereby highlighting the relevance of psychological aspects over social ones. The findings indicated that prior behaviours, attitudes towards studies, and knowledge of institutional norms exert a substantial influence on academic and social integration. This suggests that intrinsic motivation and clarity in academic expectations function as protective factors against dropout. In contrast, factors such as difficulty in establishing social networks or family environment did not demonstrate a statistically significant relationship. This calls into question traditional models that prioritise social integration as a central axis.

Theoretically, the present findings serve to reinforce the necessity of integrating psychological perspectives into models of university dropout, a requirement that is in alignment with studies emphasising self-regulation and mental health as pivotal factors in the retention of students. Methodologically, the use of structural equations enabled the unravelling of complex interactions between latent variables, thus validating the usefulness of quantitative approaches to analyse multi-causal phenomena. Nevertheless, the restricted impact of social elements may be indicative of deficiencies in the operationalisation of constructs such as family assistance or institutional coexistence, domains that necessitate further qualitative investigation.

Among the study's limitations, the exclusive focus on a single university and a small sample of scholarship recipients with low academic performance is particularly salient, thereby restricting the generalisation of results. Moreover, the exclusion of contextual variables, such as geographical origin or workloads, may have resulted in the omission of determining factors in contexts of economic vulnerability. It is recommended that future research endeavours encompass a more diverse sample and adopt a mixed methods approach to encompass the subjective dimensions that are not amenable to quantification through Likert scales.

The practical implications of this study indicate a necessity to formulate educational policies that incorporate ongoing psychological support, academic mentoring, and emotional skills training for scholarship recipients. The implementation of institutional adaptation programmes, with a focus on clarifying normative expectations and reducing academic anxiety, has the potential to mitigate the risk of dropout. These efforts should be complemented by strategies that foster institutional belonging, even if their direct impact was not statistically significant in this study.

In future research, it is recommended that the role of socioeconomic variables not previously considered be explored. Such variables may include family financial stability and access to technological resources, which could modulate the relationship between psychosocial factors and dropout. Conversely, longitudinal research would facilitate analysis of how these factors evolve throughout the university trajectory, thereby identifying critical windows for preventive interventions.

The aggregate contribution of this study is that it highlights that academic drop-out among scholarship students is a multifaceted phenomenon, with psychological components emerging as primary determinants. Whilst the findings challenge established socially-centred paradigms, they nevertheless pave the way for innovation in the field of evidence-based retention strategies. The combination of methodological rigour and sensitivity to student realities remains pivotal to the transformation of educational systems into environments that are genuinely inclusive.

The study involved an anonymous online questionnaire administered via Google Forms (non-interventional; no clinical procedures or experimental manipulation; no directly identifiable data collected). This work originated as a formative assignment within a postgraduate research methods course at the Graduate School of the National University Pedro Ruiz Gallo (UNPRG), ⁵⁶ and it was not submitted as a thesis- or degree-related protocol. Therefore, no formal ethics committee/IRB approval or waiver reference number is available. In the manuscript we now cite the relevant UNPRG institutional documents (Code of Ethics and Scientific Integrity for Research; Regulation of the Research Ethics Committee) and we attach an updated sworn statement explaining this context. The study also adhered to the UNPRG Code of Ethics and Scientific Integrity for Research (Resolution No. 1134-2018-R).

Informed consent was obtained in written electronic form (e-consent). The consent statement was displayed on the first page of the Google Form, and respondents could proceed only after actively selecting the agreement option (“I agree”). This has been clarified in the manuscript, and an updated e-consent form is included as a supplementary file.