Personality traits in female entrepreneurship: AI-based assessment vs. self-perception biases and tailored educational intervention

Abstract

Introduction:

Understanding the entrepreneurial mindset and its psychological foundations has become a priority in education aimed at sustainability and innovation. Personality traits and socioemotional competencies play a key role in shaping entrepreneurial dispositions; however, self-report measures may be affected by subjectivity and self-perception biases.

Methods:

This study examines the relationship between self-confidence, entrepreneurial self-perception, and OCEAN/Big Five personality traits in a multinational sample of university students. A total of 409 participants from six countries participated in the study. Self-confidence was measured using the Self-Confidence subscale of the Entrepreneurial Mindset Profile (EMP), while entrepreneurial self-perception was assessed using a single self-report item. In parallel, the 35 Big Five facets were inferred using a natural language processing (NLP) and artificial intelligence tool (Human AI) applied to free-text narratives (approximately 800–1,000 words). Pearson correlations, linear regressions, and Random Forest models with 10-fold cross-validation were conducted.

Results:

Self-confidence showed a significant negative correlation with Neuroticism (r = −0.33, p < .01), suggesting that emotional stability is associated with higher entrepreneurial self-confidence. Random Forest analyses identified Modesty, Artistic Interests, Neuroticism, Outgoing, and Adventurousness as the most relevant predictors of entrepreneurial self-perception. No overall gender differences were found in entrepreneurial self-perception levels, although different dispositional patterns emerged in the predictors.

Discussion:

The findings highlight the role of emotional stability and socioemotional competencies in shaping entrepreneurial profiles and suggest that AI-based personality inference can complement traditional self-report measures by mitigating self-perception biases. These results provide relevant implications for educational leadership and support the design of tailored educational interventions to foster entrepreneurial competencies in higher education contexts.

1 Introduction

In the face of growing social, economic, and environmental challenges, the 2030 Agenda and Education for Sustainable Development (ESD) position education as a lever for transformation (Leicht et al., 2018; UNESCO, 2020). This is because it functions as an engine of social development by driving transformations in the economy, equality, and cohesion (OECD, 2021, 2023, 2025; UNESCO, 2015). From this perspective, educational leadership is key to articulating a shared vision of sustainability (Al-khamaiseh et al., 2024) and translating it into coherent institutional practices at different levels and settings (Elkaleh et al., 2025; Ghamrawi et al., 2025). However, for this sustainability vision to materialize in educational institutions, leadership must translate into concrete student capabilities (Novy and Wetmore, 2021).

This has prompted education systems to face accelerated transformations enabling them to address new labor and social demands (Bartlett, 2022; Wang et al., 2024), which require preparing students with tools for action and resilience (Jennings, 2018; OECD, 2023; World Economic Forum, 2023). Responding to this demand involves fostering a broad entrepreneurial mindset, that is, preparing students to transform their environment responsibly and sustainably beyond business creation, promoting their participation in social, economic, and environmental challenges (Rosário and Raimundo, 2024; Sottini et al., 2024). This approach aligns with trends toward inclusion, lifelong learning, and the integration of digital technologies to strengthen students' holistic development (Navas-Bonilla et al., 2025; OECD, 2023; Samaniego López et al., 2025; UNESCO, 2021). Within this comprehensive focus, socioemotional competencies are essential because they predict educational achievement, employability, and well-being beyond cognitive ability. Moreover, they are malleable and therefore subject to educational intervention (Chernyshenko et al., 2018; Kankaraš and Suárez-Álvarez, 2019; OECD, 2015, 2024). Their contribution to emotional regulation, collaboration, and perseverance makes them direct levers for entrepreneurial action in uncertain contexts (De Cock et al., 2020; Mueller et al., 2017; Sarasvathy, 2001).

2 Theoretical framework and literature review

2.1 Socio-emotional competencies, personality, and the psychological meaning of self-confidence

Socioemotional competencies can be coherently described and assessed through Big Five personality traits (Costa and McCrae, 1992; McCrae and Costa, 1999; Walton et al., 2023). Evidence shows that these competencies correspond to relatively stable personality dimensions and can be reliably measured as emotional, social, and cognitive facets (OECD, 2015, 2024; Soto and John, 2017). This conceptual alignment has been consolidated by the OECD, which organizes competencies around the Big Five and documents their structure, development, malleability, and relevance to educational, labor, and life outcomes across the lifespan (Chernyshenko et al., 2018; Kankaraš and Suárez-Álvarez, 2019; OECD, 2024). Specifically, the Assessment Framework of the OECD Study on Social and Emotional Skills (SSES) operationalizes socioemotional competencies explicitly aligned with the Big Five model, unifying the language of assessment and intervention and improving comparability across ages and contexts (Kankaraš and Suárez-Álvarez, 2019; OECD, 2024).

Multimethod psychometric developments have shown validity and reliability (including convergent and criterion validity) by combining self-report, situational judgment, and forced-choice formats to capture competencies such as emotional regulation, collaboration, and perseverance within the Big Five framework (Murano et al., 2021; Walton et al., 2022). Thus, adopting a single measure based on 35 Big Five facets enables the inference of competency profiles with conceptual clarity and assessment efficiency, reinforcing their usefulness for policy and program design (OECD, 2024).

From a broader psychological perspective, certain socio-emotional competencies, particularly self-confidence, have been conceptualized not only as relatively stable dispositions, but also as psychologically meaningful indicators that are sensitive to processes of adaptation and vulnerability in mental health. Longitudinal studies show that loss of self-confidence may emerge as a consequence of adverse mental health trajectories and, in turn, act as a predictor of subsequent dysfunction, suggesting a bidirectional relationship between self-confidence, psychological well-being, and emotional regulation (Kang and Malvaso, 2024). This perspective allows self-confidence to be situated beyond domain-specific contexts, integrating it into explanatory models of general psychological functioning.

2.2 Personalidad, mentalidad emprendedora, autopercepción y sostenibilidad

Within this competency framework, entrepreneurial personality can be understood as a set of dispositions and skills linked to the Big Five that foster a positive and resilient action pattern.

In this study, the entrepreneurial mindset is conceptualized as a cross-cutting competence for responsible and sustainable action, extending beyond business creation and aligned with approaches to social and impact-oriented entrepreneurship (Rosário and Raimundo, 2024; Sottini et al., 2024). It is clearly distinguished both from self-confidence, understood as the perception of personal capability or self-efficacy (Bandura, 1997), and from entrepreneurial self-perception, which refers to the subjective evaluation of one's own entrepreneurial identity (Postigo et al., 2021). This conceptual distinction is essential to avoid construct overlap and to ensure a precise interpretation of the empirical results (Soto and John, 2017).

Scientific evidence shows systematic relationships between personality traits and entrepreneurial intentions and performance: neuroticism and agreeableness tend to be negatively associated, while openness to experience, conscientiousness, and extraversion are positively associated (Freiberg and Matz, 2023; Hagenauer and Zipko, 2024; Howard and Boudreaux, 2024; Li et al., 2022; Zhao et al., 2010). At the facet level, these dispositions account for differences in creativity and innovation (Li et al., 2022; Peljko and Auer, 2022), as well as perseverance and the ability to mobilize resources (Baron and Tang, 2009; Becker et al., 2023; OECD, 2024) key aspects of both economic and social/sustainable entrepreneurship (Postigo et al., 2021; Sottini et al., 2024). Similarly, lower levels of neuroticism are associated with greater resilience and emotional stability in the face of uncertainty, facilitating persistence in entrepreneurial activities (Soto and John, 2017; Walton et al., 2023) and acting as a protective factor (Freiberg and Matz, 2023; Hagenauer and Zipko, 2024; Li et al., 2022; Soto and John, 2017; Zhao et al., 2010). In this regard, the Evidence-Based Entrepreneurship approach argues that personality traits constitute consistent predictors of entrepreneurial success, reinforcing the need to design educational interventions grounded in scientific evidence (Rauch and Gielnik, 2021).

The relevance of self-perception and personality traits is not confined to educational or entrepreneurial contexts. Across a range of applied fields, the combination of dispositional traits, self-perception, and the analysis of complex patterns has been shown to provide additional explanatory value for understanding individual trajectories and supporting decision-making processes. In this regard, recent studies illustrate how personality trait–based approaches and artificial intelligence techniques have been used to identify differentiated profiles and meaningful patterns in diverse applied contexts, reinforcing the translational potential of these methods beyond traditional self-report measures, including settings in which social desirability bias and distorted self-perception may compromise assessment validity (Angelini et al., 2024; Cao et al., 2024; Kang and Malvaso, 2023a, b, 2024; Vecchiarini and Somià, 2023).

In the field of entrepreneurial personality, recent evidence shows that both general Big Five traits and specific facets provide differential explanatory value in the prediction of entrepreneurship (intentions and behavior), supporting the use of dispositional frameworks in educational contexts (Postigo et al., 2021). In parallel, the field of social entrepreneurship, closely linked to social innovation, has proposed a research agenda oriented toward creating social value and sustainability (Sottini et al., 2024). The convergence of entrepreneurial dispositional profiles, entrepreneurial passion, and responsible innovation capabilities supports the design of educational interventions aligned with leadership oriented toward Education for Sustainable Development (ESD), in which each project and decision contributes to enabling students to understand complex problems, collaborate, innovate, and act responsibly (Newman et al., 2021). At a macro level, entrepreneurship, and sustainable entrepreneurship in particular, acts as a lever for economic dynamism, job creation, and the transition toward solutions with positive social and environmental impact, in line with the goals of the 2030 Agenda (Decker et al., 2014; Rosário et al., 2022; Rosário and Raimundo, 2024; Shui et al., 2025).

2.3 Contribución del estudio y objetivos

Despite the above, substantial gaps persist in the literature. On one hand, motivations and personality traits influencing entrepreneurship have been shown to differ between men and women, requiring differentiated analyses to avoid neutral solutions that reproduce inequalities (Hagenauer and Zipko, 2024; Laouiti et al., 2022). On the other hand, the study of entrepreneurial self-perception, understood as individuals' evaluation of their entrepreneurial characteristics and attitudes, has received limited explicit attention despite its potential to modulate predisposition to act. In this regard, recent work integrates mindset and self-efficacy as mediators of entrepreneurial intention (Jiatong et al., 2021). Studies integrating trait-based personality measures (Big Five) with self-reported self-confidence and entrepreneurial self-perception in multi-country university samples remain limited. Likewise, few studies have compared traditional self-report instruments with third-generation approaches based on artificial intelligence and natural language processing, which are capable of mitigating social desirability bias through the analysis of extended texts (Hussain et al., 2025; Somià and Vecchiarini, 2024; Uriarte et al., 2025).

Beyond applying AI-based personality inference to a new educational context, the present study advances prior research by integrating a fine-grained, facet-level mapping of personality with measures of entrepreneurial self-perception in a multinational sample. In contrast to previous AI-based approaches that primarily focus on trait prediction accuracy, this study emphasizes the comparison between self-reported perceptions and AI-inferred dispositional profiles in order to address self-report bias and enhance interpretability in applied educational research.

In this context, text analysis using language models has proven particularly effective for accurately inferring personality traits, as demonstrated in recent studies based on extended personal narratives (Hussain et al., 2025). The present study aims to contribute to this field by comparing entrepreneurial self-perception with AI-inferred personality trait profiles, incorporating sex-based analyses in a university sample drawn from six countries. To this end, three specific objectives are proposed: (1) to assess self-confidence using self-report measures and examine its relationship with OCEAN/Big Five personality traits, taking into account the role of subjectivity in self-evaluation; (2) to analyze entrepreneurial self-perception and identify associated personality predictors; and (3) to compare predictors of entrepreneurial self-perception as a function of sex in a multi-country university sample. In doing so, the study provides relevant evidence for educational leadership by informing the prioritization of competencies and the design of personalized training interventions.

3 Method

3.1 Participants

The sample used was part of several university ecosystems across six countries within the WINN Network (Argentina, Bolivia, Colombia, Ecuador, Spain, and Peru). In the project cohort, 727 participants began the assessment battery between 2021 and 2023; of these, 601 completed all the measures required in the general study (≈ 58 percent women, 42 percent men; M_age ≈ 23.8 years). Recruitment was carried out in academic environments at the universities participating in the WINN Network, and participation was voluntary.

The inclusion criteria were: (a) belonging to the university community of one of the six countries in the WINN Network; (b) agreeing to participate and completing the assessment tests; and (c) providing the minimum demographic data (sex and year of birth). Exclusion criteria included incomplete questionnaires or texts preventing calculation of key variables (EMP self-confidence, self-perception, and Big Five facets) and duplicate or inconsistent cases after verification and prior anonymization using an alphanumeric code.

After applying these criteria, an analytical sample of 409 cases was obtained, with complete and usable information for all variables of interest. In this subsample, the mean age was 23.8 years (SD = 4.8), with an approximate distribution of 58 percent women and 42 percent men. Case selection was aimed at ensuring data integrity for correlational and predictive analyses (bivariate and multivariate regressions and Random Forest with cross-validation).

The study was conducted in accordance with the ethical principles applicable to research involving human participants, ensuring informed participation, anonymity, and the use of data exclusively for purposes related to the research objective. Institutional ethical approval was managed by participating institutions.

3.2 Instruments

Self-confidence (self-report: Entrepreneurial Mindset Profile, EMP): Self-confidence was assessed using the Self-Confidence scale of the Entrepreneurial Mindset Profile (EMP), a self-report instrument developed to measure the entrepreneurial mindset, whose development, reliability, and validity have been documented by Davis et al. (2016). The EMP organizes its results into 14 scales grouped into two domains: (1) personality characteristics (Independence, Preference for Limited Structure, Nonconformity, Risk Acceptance, Action Orientation, Passion, Need to Achieve) and (2) skills (Future Focus, Idea Generation, Execution, Self-Confidence, Optimism, Persistence, Interpersonal Sensitivity) (EMP, 2019). Score interpretation is performed with comparative norms (for example, entrepreneurs and corporate managers), which makes it possible to contextualize the individual profile relative to reference groups (Davis et al., 2024). Psychometrically, the EMP has evidence of reliability and convergent and discriminant validity, as well as expected differences between entrepreneurs and managers (Davis et al., 2016). For the purposes of this study, the EMP Self-Confidence scale is conceptualized as a general belief in one's ability to leverage skills and talents to achieve goals and overcome challenges, in accordance with the developer's technical documentation (EMP, 2019).

3.2.1 Entrepreneurial mindset self-perception (single item)

Entrepreneurial mindset self-perception was obtained using the question: “Would you describe yourself as an entrepreneurial person?” with three response options: “yes”, “no”, “maybe, I’m not sure.” This item was administered together with the EMP and is part of the self-report block in the data collection protocol (see Procedure).

3.2.2 Socioemotional competencies/personality traits (objective AI-based assessment)

A third-generation artificial intelligence tool (Human AI) was used to infer the 35 OCEAN/Big Five facets from a Spanish text of approximately 800–1,000 words. The approach is based on texts written in natural language and analyzes large-scale linguistic patterns (vocabulary, syntax, semantics, and paralinguistic cues) to infer traits and socioemotional competencies. From a technical point of view, the system combines Transformer-type representations with sequential LSTM modeling and a supervised regressor for each facet within a multioutput regression framework. During training, 1,778 texts automatically labeled with IBM Watson Personality Insights (PI) were used as an instrumental proxy signal for the 35 facets, enabling the construction of a high-dimensional supervised dataset. This decision was driven by engineering considerations aimed at obtaining a structured linguistic signal and does not imply the use of Watson PI as a psychological gold standard. The corpus was divided into training, validation, and test sets (70/15/15), incorporating hyperparameter tuning and cross-validation. On the test set, a high convergence was observed between the model outputs and the algorithmic labels used during training (r = .93; p < .02). This result should be interpreted as an indicator of algorithmic consistency and learning stability, rather than as direct evidence of psychometric or construct validity (IBM, 2018; Park et al., 2015). In a complementary manner, the model was subjected to additional validation procedures using an independent sample (n = 451), aimed at examining its structural coherence with the Big Five framework and its conceptual convergence with standardized psychometric self-reports (NEO PI-R, IPIP-NEO-120), as well as its operational stability across different types of writing tasks. The detailed results of these validation analyses are not reported in the present manuscript, whose focus is the application of the model in an educational and entrepreneurial context, but they inform its use as a complementary tool for text-based dispositional assessment. Overall, the system constitutes a hybrid model (Transformer, LSTM, and supervised regression) that learns associations between natural language and the 35 Big Five facets from labeled examples and has been evaluated using multiple technical and psychometric criteria. The model was developed by the research team responsible for this study and is currently undergoing scientific validation through peer review. It is available as a web-based application and API for academic and professional use.

3.3 Procedure

A cross-sectional, correlational, and ex post facto study was conducted, using non probabilistic convenience sampling in university ecosystems across six countries of the WINN Network (Argentina, Bolivia, Colombia, Ecuador, Spain, and Peru) between 2021 and 2023. Participation was voluntary and no incentives were offered. National nodes disseminated the invitation within their university communities and coordinated the administration of the common protocol. Before beginning, participants were informed about the objectives of the study and data confidentiality; each record was assigned an alphanumeric code to ensure anonymization during processing and analysis.

The assessment battery was administered in three blocks and completed in a single session. In Block 1, demographic data were collected: sex, year of birth, and country of academic affiliation. The data were linked exclusively to the participant's alphanumeric code. In Block 2, the self-report Entrepreneurial Mindset Profile (EMP) was administered, with a focus on the Self-Confidence scale, following the specifications of the instrument (Davis et al., 2016). In the same block, the entrepreneurial mindset self-perception item was administered: “Would you describe yourself as an entrepreneurial person?”, with the response options “yes”, “no”, or “maybe, I’m not sure.” In Block 3, text production for linguistic assessment with Human AI was carried out. Each participant wrote a text in Spanish on a free topic, with a target length of 800–1,000 words. The texts were processed with an artificial intelligence assistant (described in the Instruments section) to infer the 35 OCEAN/Big Five facets. The blocks were presented in the indicated order: demographics → EMP + self-perception → text. Text writing was individual and without a strict time limit within the session, with the instruction to reach the target length. No linguistic assistance or feedback was provided during text production.

3.4 Data analysis

To clean and shape the analytical sample, prespecified quality criteria were applied. The following were excluded: (a) incomplete records (missing minimum demographic data, incomplete EMP, and/or missing or manifestly insufficient text), and (b) duplicates or inconsistencies detected during the preprocessing verification prior to anonymization. Of the 727 individuals who began the assessment battery, 601 completed all required measures; after further analytical screening according to the exclusion steps described, a final sample of 409 cases with complete information on all variables of interest was used.

Regarding variable definition and coding, self-confidence (self-report) was assessed using the Self-Confidence scale of the EMP and coded as a continuous variable. Entrepreneurial mindset self-perception was assessed using a single item with the options “yes/no/maybe.” For the main analyses, it was coded as an ordinal variable (no = 0; maybe = 1; yes = 2). In sensitivity analyses, a binary version was used (yes = 1; no/maybe = 0). Lastly, the 35 OCEAN/Big Five facets were inferred through natural language processing and coded as continuous variables.

For descriptive statistics, means, medians, standard deviations, and ranges were calculated for continuous variables, and frequencies/percentages for categorical ones. Distributions and outliers were inspected (criterion ±3 SD and graphical checks). Linearity and homoscedasticity assumptions were tested for linear models. For the analysis of bivariate relationships, Pearson correlations were estimated between EMP Self-Confidence and each of the 35 facets. Correlation coefficients (r), 95 percent confidence intervals, p values, and descriptive statistics were reported.

For explanatory and predictive models, bivariate linear regression was applied, modeling entrepreneurial self-perception (ordinal) from each facet separately, reporting standardized β coefficients, 95 percent confidence intervals, p values, and R². Multivariate linear regression was also applied, including all facets simultaneously to estimate the combined effect and compare the multivariate R² with the bivariate models. Finally, a Random Forest (RF) regression model was used to evaluate the nonlinear, multivariate predictive capacity of the facets, as well as the relative importance of each predictor. Ten-fold cross-validation was used to estimate R²_CV and RMSE_CV and to mitigate overfitting (Breiman, 2001).

To explore sex differences, Random Forest with 10-fold CV was applied to male and female subsamples. Patterns of variable importance were compared (correlation between importance vectors and descriptive contrasts of rankings). This analysis was complemented with stratified models by sex.

To assess the consistency and integration of findings, a convergence-of-methods strategy was applied: facets found in common across (a) correlations, (b) regressions (β), and (c) RF (importance scores) were identified. Convergence across techniques was interpreted as strengthened evidence of robustness; discrepancies were discussed in terms of nonlinearity, interactions, and multicollinearity.

Although additional sensitivity analyses were considered (robust models and alternative codings of the dependent variable), they were not included in this manuscript because they did not provide substantial additional information beyond the reported models.

Analyses were performed in Python. Pandas and NumPy were used for data management, statsmodels for linear models, and scikit-learn for Random Forest and cross-validation; visualization was carried out with matplotlib. A random seed was set to ensure reproducibility, and versions and parameters are documented in the supplementary material (Pedregosa et al., 2011; Seabold and Perktold, 2010).

4 Results

The first objective examined the relationship between self-confidence, assessed using the Self-Confidence scale of the Entrepreneurial Mindset Profile (EMP), and the personality traits of the OCEAN/Big Five model, inferred through automated linguistic analysis with Human AI. Pearson correlations were calculated to estimate the linear association between self-reported self-confidence and the 35 personality facets derived from natural language processing. Descriptive analyses showed that scores on the Self-Confidence scale fell within a medium-high range, with slight right skewness and moderate interindividual variability, as shown in Figure 1.

Figure 1

Table 1 presents the correlation coefficients corresponding to a subset of Big Five dimensions and facets, selected due to their theoretical relevance and/or greater magnitude of association with self-confidence.

The correlational analysis revealed both positive and negative associations between self-confidence and personality traits (see Figure 2). Positive correlations were observed with Extraversion (r = 0.16, p < .05), Conscientiousness (r = 0.15, p = .078), and Hostility (r = 0.17, p = .066). Conversely, negative correlations were found with Neuroticism (r = –0.33, p < .01), Emotionality (r = –0.18, p = .092), and Stress Susceptibility (r = –0.13, p = .104). Among all associations, only the inverse relationship with Neuroticism reached statistical significance (p = .004), with a moderate effect size (r = –0.33), suggesting that higher emotional stability may be associated with higher levels of self-confidence within the entrepreneurial profile.

Figure 2

Taken together, the results suggest that higher levels of self-confidence tend to be associated with higher scores in Extraversion and Conscientiousness, and with lower levels of Neuroticism and emotional reactivity. However, only the inverse relationship between Neuroticism and self-confidence reached statistical significance (p = 0.004), indicating that a greater presence of neurotic traits (such as anxiety, insecurity, or emotional vulnerability) corresponds to lower perceived self-confidence. This finding reinforces the role of emotional stability as a potential central factor in the self-perception of personal competence within the entrepreneurial profile.

These results provide an empirical basis for subsequent analyses aimed at exploring how entrepreneurial self-perception and personality profiles interact in predicting perceived entrepreneurial mindset across different contexts and genders.

Regarding the second objective of the study, entrepreneurial self-perception was analyzed as the dependent variable. Unlike self-confidence, which was assessed as a continuous variable using a standardized scale (EMP), entrepreneurial self-perception was operationalized as a nominal categorical variable based on a single dichotomous question. This approach required the use of models appropriate to the non-continuous nature of the variable, beginning with bivariate analyses and extending to multivariate and machine learning techniques. As an initial step, bivariate linear regressions were conducted between each of the 35 personality facets and entrepreneurial self-perception in order to explore potential individual associations. Given the multifaceted nature of the traits involved, a stepped analytical strategy was implemented, including bivariate regressions, multiple regression, and a machine learning model (Random Forest) to capture both linear and nonlinear relationships among variables.

Descriptive analyses of entrepreneurial self-perception show that 63.3 percent of participants described themselves as entrepreneurial, 24.2 percent indicated “maybe, I’m not sure,” and 12.5 percent responded “no.” Moderate differences were observed between sexes: 66.8 percent of women and 58.1 percent of men identified themselves as entrepreneurial, although this difference did not reach statistical significance [χ²(2) = 4.12, p = 0.127]. The distribution of responses therefore tends toward the affirmative category, reflecting an overall positive self-perception of entrepreneurial mindset in the sample (see Table 2).

The results of the bivariate linear regressions indicated that three personality facets initially showed a statistically significant relationship (p < .05) with entrepreneurial self-perception in the preliminary analysis (without correction for multiple comparisons): Emotionality (R² = 0.10), Neuroticism (R² = 0.07), and Adventurousness (R² = 0.06). Although individual coefficients of determination were low (R² < .10), the pattern of results suggested that lower levels of Neuroticism and higher levels of Emotionality and Adventurousness were associated with greater entrepreneurial self-perception (see Table 3 and Figure 3).

Figure 3

Of the 35 statistical contrasts conducted simultaneously, none of the associations between the facets and entrepreneurial self-perception reached statistical significance (q ≥ 0.05), indicating that no individual predictor substantially explains the dependent variable when analyzed in isolation. This methodological limitation is consistent with the multifactorial nature of the construct under evaluation and underscores the need to employ more integrative analytical models capable of capturing joint effects, interactions, and nonlinear latent structures.

4.1 Exploratory analysis: multiple linear regression

To preliminarily explore the multivariate structure of the associations between personality facets and entrepreneurial self-perception, a multiple linear regression model was estimated including all 35 OCEAN/Big Five facets simultaneously as predictors. Given the high number of predictors relative to the sample size and the exploratory nature of this analysis, the model was included exclusively for descriptive and methodological purposes, and not for confirmatory inference. This approach allows for the description of combined patterns among facets and the examination of the relative distribution of coefficients under mutual control, without attributing independent effects or causal conclusions to the model. Table 4 presents the unstandardized coefficients, standard errors, p values, and 95 percent confidence intervals.

Although the model explained a high proportion of variance (R² = 0.92), the low adjusted R² (0.22) and the non-significant F-test (F = 1.31) indicate substantial overfitting and limited statistical stability. This pattern is likely related to the high number of predictors relative to the sample size and the presence of multicollinearity among personality facets. Consistently, information criteria (AIC = 17.64; BIC = 78.44) penalized model complexity without indicating a clear improvement in model fit. Consequently, this model does not provide reliable inferential evidence regarding the independent contribution of individual predictors and is therefore reported for exploratory and methodological purposes only. These limitations highlight the need for analytical approaches better suited to high-dimensional and interrelated predictor spaces, which motivated the use of Random Forest models as the primary integrative method in this study.

Given the limitations observed in multivariate linear models, a non-linear and regularized machine learning approach was adopted as a complementary analytical strategy. Specifically, a Random Forest algorithm with cross-validation (10-fold cross-validation) was applied, using entrepreneurial self-perception as the dependent variable and the 35 personality facets inferred by Human AI as predictors.

The model enabled the estimation of permutation-based variable importance (mean decrease in accuracy) for each facet within a multivariate framework. Figure 4 presents the variables ranked from highest to lowest relative importance. The five facets with the highest relative weights were Modesty, Artistic Interests, Neuroticism, Extraversion, and Adventurousness, whereas traits such as Sensation Seeking, Intellect, Cheerfulness, and Self-efficacy showed values close to zero, indicating a marginal contribution within the model.

Figure 4

In terms of predictive performance, the Random Forest model demonstrated limited performance under cross-validation (10-fold cross-validation), with an approximate cross-validated coefficient of determination of R² ≈ 0.11, a mean absolute error (MAE) of 0.13, and a root mean squared error (RMSE) of 0.25. These results indicate that the model explains a small but non-trivial proportion of variance in entrepreneurial self-perception, which is consistent with the subjective, multifactorial, and context-dependent nature of the construct under study.

It is important to note that variable importance measures derived from Random Forest reflect the relative contribution of each facet within the multivariate model, and should not be interpreted as independent effects, causal relationships, or standalone predictors. Moreover, given the high degree of conceptual and empirical correlation among Big Five facets, importance rankings should be interpreted with caution, as shared variance among related traits may redistribute importance across predictors. Consequently, the relative ordering of facets should be understood as indicative of broader dispositional patterns, rather than precise estimates of the individual weight of each trait.

Overall, the analyses suggest that entrepreneurial self-perception is associated with a complex dispositional configuration integrating facets related to emotional stability, cognitive or intellectual openness, and social modulation. In this regard, models that integrate multiple predictors and accommodate high-dimensional spaces, non-linear relationships, and complex interactions are particularly useful as exploratory tools, providing a more appropriate basis for the integrative interpretation of perceived entrepreneurial profiles than traditional linear approaches.

Complementarily, the relative importance of personality facets was examined by sex, based on an exploratory analysis of the importance coefficients generated by the Random Forest model applied to each subgroup. Among men, the facets Modesty, Impulsiveness, and Self-discipline showed the highest relative importance values (> 0.06), whereas among women, Artistic interests, Dutifulness, and Trust stood out. Both Modesty and Emotionality remained among the relevant predictors in both groups, although with different magnitudes. Although these results point to potential differences in the personality profiles associated with entrepreneurial self-perception, they should be interpreted with caution, since independent predictive models for each sex were not validated.

5 Discussion

The purpose of this study was to examine the relationship between the OCEAN/Big Five personality facets and entrepreneurial self-confidence, as well as to evaluate the predictive capacity of these facets for entrepreneurial mindset self-perception, using both traditional self-report measures and linguistic analysis through artificial intelligence. This approach responds to the need to understand the psychological profile underpinning educational leadership oriented toward sustainability (Liao et al., 2022), at a time when educational institutions face the urgent task of preparing agents capable of driving social, economic, and environmental innovation (Arbués et al., 2025; Leal Filho et al., 2020; Trevisan et al., 2024; UNESCO, 2020).

The discussion is organized around two research objectives: first, the dispositional correlates of entrepreneurial self-confidence are examined; second, the predictive and exploratory patterns associated with entrepreneurial self-perception are analyzed. Importantly, these findings should be interpreted in light of the conceptual distinction between self-confidence, understood as a general perception of personal competence, entrepreneurial self-perception as an identity-based self-classification, and entrepreneurial mindset as a broader dispositional orientation. Although related, these constructs operate at different psychological levels and should not be treated as interchangeable.

The results are consistent with the presence of a shared dispositional basis between entrepreneurial mindset and sustainable leadership, linked to self-regulation, cognitive openness, and emotional stability, and suggest a potentially relevant educational leadership profile for advancing the objectives of the 2030 Agenda (Freiberg and Matz, 2023; George et al., 2021; OECD, 2024).

The findings of this study confirm that neuroticism, evaluated inversely, is the principal inhibitor of entrepreneurial self-confidence, reinforcing the idea that emotional stability operates as a protective factor in uncertainty-driven contexts inherent to entrepreneurship (Li et al., 2022; López-Núñez et al., 2022; Şahin et al., 2019; Zhao et al., 2010). This evidence aligns with results from studies such as Freiberg and Matz (2023), who found that individuals with higher levels of neuroticism tend to obtain poorer outcomes in early stages of business development, suggesting a weaker ability to sustain entrepreneurial action under pressure. In line with this, recent studies show that neuroticism not only reduces self-confidence but also increases risk avoidance, both in the entrepreneurial intention phase, where a reduction in willingness to assume financial risk is observed (Salameh et al., 2022), and in execution phases, where lower emotional stability is associated with reduced propensity for innovation and risk-taking among active entrepreneurs (Hagenauer and Zipko, 2024). Taken together, these data reinforce the interpretation that emotional stability acts as a psychological prerequisite for sustaining viable entrepreneurial trajectories in educational contexts and is therefore an essential component of sustainability-oriented leadership.

In contrast, facets linked to openness to experience, such as Artistic interests and Adventurousness, showed positive associations with entrepreneurial self-perception, consistent with evidence highlighting the role of cognitive openness in innovative thinking and transformation-oriented leadership for sustainability (Xu, 2023; Zhang et al., 2022). These results suggest that perceived entrepreneurship in education is not grounded merely in economic drive, but in a dispositional pattern that facilitates creativity, the pursuit of novel solutions, and socio-environmental sensitivity, characteristics identified as central to current models of sustainable educational leadership (Barnett-Itzhaki et al., 2025; Nguyen et al., 2025; Peng et al., 2024; Trevisan et al., 2024).

Likewise, the Random Forest model suggested a relatively greater contribution of facets such as Modesty and Trust within the multivariate set of predictors, pointing to a possible relational and ethical dimension in entrepreneurial self-perception. Since both facets fall under Agreeableness, this pattern suggests a prosocial relational profile that facilitates cooperation and trust-building, pillars of responsible leadership aligned with sustainability (Maak and Pless, 2006). This prosocial foundation is consistent with evidence linking Agreeableness with prosocial and cooperative behavior (Thielmann et al., 2020). In this regard, various proposals conceptualize sustainable leadership as leadership based on integrity, relationship-building, and collaboration (Al-Khamaiseh et al., 2024; Peng et al., 2024). Therefore, the results help transcend a traditional view of entrepreneurship centered on competitiveness and individual self-efficacy, proposing instead a paradigm of sustainable entrepreneurship that integrates community values and social responsibility (Rosário and Raimundo, 2024). This conceptual transition is especially relevant in education, where entrepreneurial leadership does not primarily manifest in business creation, but in the generation of innovative initiatives aimed at equity, inclusion, and environmental sustainability.

One of the main methodological contributions of this study lies in the incorporation of artificial intelligence based on linguistic analysis for personality assessment. The use of natural language processing models makes it possible to capture dispositional manifestations through free-text production, overcoming certain limitations of self-report measures such as social desirability bias and automatic response tendencies (Hickman et al., 2019; Kjell et al., 2022). Recent studies have highlighted the potential of AI as a valid tool for psychological assessment, provided that it relies on transparent models subjected to external validation (Eichstaedt et al., 2021; Kjell et al., 2024). Moreover, international organizations such as UNESCO (2023), the European Commission (2022), and the European Union (2024) have emphasized the importance of using AI as a strategic support for educational leadership, always under principles of ethics, equity, and human oversight (European Commission, 2022; European Union, 2024; UNESCO, 2023). In this study, AI facilitated the identification of nonlinear interaction patterns among personality facets, suggesting predictive structures that did not emerge in traditional linear models. This evidence positions AI not as a substitute but as a methodological complement for comprehensive assessment of entrepreneurial leadership.

From a methodological standpoint, it is important to distinguish between the algorithmic convergence observed during model training and psychometric validity in the strict sense. The high agreement between the system outputs and the algorithmic labels used as training signals reflects technical consistency and supervised learning stability but does not in itself constitute direct evidence of psychological construct validity. Validation against psychometric instruments administered to human participants represents a distinct and conceptually independent level of evidence, which should be evaluated using its own criteria and lies beyond the scope of the present applied study. Nevertheless, the model used in this study is grounded in prior validation processes conducted with independent samples, aimed at examining its structural coherence with the Big Five framework and its conceptual convergence with standardized psychometric measures, as well as identifying systematic discrepancies with traditional self-report instruments. These analyses, conducted at both trait and facet levels, allowed for the assessment of patterns of theoretical alignment and effects associated with method variance, thereby supporting a complementary, rather than substitutive, interpretation of language-based personality inference. The detailed results of these validation analyses are not reported in the present manuscript, whose focus is the application of the model in an educational and entrepreneurial context.

Regarding the gender dimension, although the analysis did not reveal significant differences in entrepreneurial mindset self-perception between men and women, the predictive models suggest distinct dispositional configurations: among women, entrepreneurial self-perception is relatively more associated with aesthetic openness and prosocial orientation, while among men self-discipline carries greater weight. This pattern aligns with evidence describing gender-related differences in the configuration of entrepreneurship, not in terms of capability, but in motivational and relational mechanisms shaped by context and norms (Cardella et al., 2020; Global Entrepreneurship Monitor [GEM], 2023; Hagenauer and Zipko, 2024; Laouiti et al., 2022). From the perspective of sustainable educational leadership, this finding reinforces the need to avoid neutral or homogeneous training approaches and to design differentiated strategies that foster diverse leadership styles, respect student heterogeneity, and promote equitable access to institutional leadership roles (Rosário and Raimundo, 2024; UNESCO, 2020).

The results of this study have strategic implications for strengthening sustainability-oriented educational leadership. The identification of an entrepreneurial profile characterized by emotional stability, cognitive openness, and prosocial orientation provides an empirical basis for designing training programs that integrate socioemotional competencies, innovative thinking, and ethical responsibility, consistent with the principles of Education for Sustainable Development (Arbués et al., 2025; Leal Filho et al., 2020, 2023; UNESCO, 2020) and with dispositional evidence associated with emotional stability and openness in university populations (Liao et al., 2022; Salameh et al., 2022; Xu 2023; Zhang et al., 2022; Zhao et al., 2010) and with the socioenvironmental orientation inherent to sustainable entrepreneurship in educational contexts (Rosário and Raimundo, 2024). Likewise, the validation of AI-based linguistic analysis as an assessment tool suggests new avenues for data-driven educational governance, supporting the personalization of learning pathways and the early detection of leadership potential (European Commission, 2022; European Union, 2024; Hickman et al., 2019; Kjell et al., 2022). Overall, these findings position entrepreneurial mindset as a key competency for transformative leadership in higher education, acting as a bridge between technological innovation, human development, and global sustainability goals (Nguyen et al., 2025; Rosário and Raimundo, 2024).

Finally, although some authors are affiliated with the company that developed the AI ​​system evaluated in this study, the algorithm operates autonomously, and its results were not affected by the authors' involvement. Data collection was performed by authors not affiliated with the company, while the affiliated authors participated exclusively in the study design, data analysis, and manuscript preparation. This separation of roles was intended to minimize potential bias and ensure the independence of the empirical results.

6 Limitations and future directions

Although this study provides novel evidence on the dispositional profile associated with entrepreneurial self-perception in higher education and introduces methodological innovation through AI/NLP, it presents several limitations that must be considered when interpreting its findings.

First, in relation to the conceptual and measurement framework, we acknowledge ongoing debate regarding the explanatory power of the Big Five model in entrepreneurship (Laouiti et al., 2022), largely attributable to measurement strategies rather than the theoretical framework itself. Self-report instruments may introduce method variance (social desirability, response styles) and offer limited granularity for capturing interdependencies among facets, contributing to inconclusive results in some studies. In this work, we employed the Big Five as a common language and conceptual organizer (Chernyshenko et al., 2018; Kankaraš and Suárez-Álvarez, 2019; OECD, 2024) and triangulated measures combining self-report (EMP) with AI/NLP-based inference from extensive text (35 facets), aiming to (i) attenuate biases inherent to self-reporting, (ii) model multivariate patterns, and (iii) better approximate the relationship between traits and entrepreneurial self-perception, also considering potential gender differences reported in the literature (Laouiti et al., 2022). However, this approach does not fully eliminate measurement limitations and should be interpreted with caution.

Second, regarding the artificial intelligence component, an important limitation concerns the use of IBM Watson Personality Insights as the labeling source during model training. This system was not employed as a psychological gold standard, but rather as an instrumental, algorithmic proxy signal selected based on engineering criteria. Given its discontinuation and the ongoing debate surrounding its construct validity, the observed convergence with Watson PI should be interpreted exclusively as evidence of the model's technical consistency and learning stability, rather than as a demonstration of psychometric validity. Furthermore, although the model builds on prior validation processes conducted with independent samples, aimed at examining its structural coherence with the Big Five model and its conceptual convergence with psychometric self-report measures, the detailed results of these validations are not reported in the present manuscript. The focus of this article is instead on the application of the model within an educational and entrepreneurial context.

Third, the study design presents inherent limitations. The cross-sectional design, convenience sampling, and the use of a single-item measure of entrepreneurial self-perception constrain causal inference, generalizability, and the sensitivity of the dependent variable. Nevertheless, single-item measures have been considered appropriate in exploratory contexts when the construct of interest is global, highly salient, and conceptually unidimensional, and when the goal is to reduce respondent burden or avoid overlap with closely related variables (Bergkvist and Rossiter, 2007; Wanous et al., 1997). Prior methodological studies have shown that single-item measures can provide valid estimates for self-perception constructs, provided that their results are interpreted with caution and complemented by other dispositional indicators (Fisher et al., 2016). Even so, the use of an ordinal single-item measure may increase measurement error and attenuate observed effect sizes, leading to more conservative estimates of the relationships examined. Accordingly, the results should be interpreted as referring specifically to perceived entrepreneurial identity, rather than to a multidimensional entrepreneurial mindset.

Finally, from a statistical and technical perspective, linear models with multiple predictors showed indications of multicollinearity or overfitting, and the importance measures derived from Random Forest models should be interpreted as associative patterns rather than as causal relationships. Within the AI/NLP component, the use of external labels for model training may introduce model dependency and inherited biases; moreover, the possibility of external assistance in the production of the analyzed texts cannot be entirely ruled out, which constitutes an additional source of variability that should be considered when interpreting the results. In addition, language-based trait inference may be influenced by formal characteristics of the texts, such as length, degree of verbosity, narrative style, or specific discourse conventions, as well as by potential differences between groups. These factors introduce further sources of variance that need to be taken into account when drawing conclusions from the findings.

Future research should: (a) incorporate longitudinal or quasi-experimental designs, probabilistic sampling, and multilevel and multi-group models for the rigorous analysis of group differences; (b) replace the single-item self-perception measure with validated multi-item scales and analyze the dependent variable using appropriate models (ordinal/multinomial logistic), applying dimensionality reduction and regularization to improve stability; (c) broaden gender measurement, integrate intersectional variables, and include behavioral criteria (e.g., participation in real innovation/SDG projects) to anchor inferences in observable outcomes.

Finally, although the present study is not designed for clinical purposes and does not allow diagnostic inferences, its findings may inform future research in applied and clinically oriented contexts. In particular, the comparison between self-perceived traits and AI/NLP-inferred personality profiles may help identify systematic discrepancies in self-evaluation, which could be relevant for the study of self-perception biases in populations where subjective assessment of functioning is critical. Longitudinal and clinically focused studies could explore this approach as a complementary assessment tool, while carefully considering ethical, validity, and contextual constraints.

7 Conclusions

The study examined the relationship between OCEAN/Big Five traits and facets and two indicators of entrepreneurial mindset—self-confidence (Self-Confidence, EMP) and entrepreneurial self-perception—in a multinational university sample. The results consistently indicate that emotional stability (inverse neuroticism) is negatively associated with entrepreneurial self-confidence, while facets related to openness (e.g., artistic interests, adventurousness) and relational/ethical components (e.g., modesty, trust) tend to appear as relatively more informative within exploratory non-linear models.

Methodologically, the integration of AI/NLP to infer 35 Big Five facets from extended texts provided information complementary to self-reports, allowing for the capture of nonlinear predictive structures that did not emerge in traditional linear models. The triangulation strategy offers a promising avenue for reducing method bias and enriching the dispositional characterization of perceived entrepreneurship in higher education.

In essence, the findings support the existence of a dispositional entrepreneurial profile characterized by emotional stability, cognitive openness, and prosocial orientation, consistent with a form of leadership aligned with Education for Sustainable Development that guides decision-making, institutional culture, and practices toward sustainability-oriented thinking and action. Moreover, although no overall sex differences were observed in self-perception, the predictive analyses suggest differentiated dispositional configurations between women and men, encouraging the avoidance of neutral training approaches and the design of differentiated leadership development strategies that respect the diversity of trajectories and styles.

From an applied perspective, the results guide the design of curricular programs and institutional actions integrating (i) socioemotional competencies, (ii) innovative thinking associated with openness, and (iii) ethical and social responsibility as core components of entrepreneurial leadership in universities. In parallel, they support the use of data-driven decision-support systems (including third-generation AI tools) for personalized learning pathways and early detection of entrepreneurial potential, always under principles of transparency, validation, and human oversight.

Finally, in light of the identified limitations, it is recommended that future research adopt longitudinal or quasi-experimental designs, ensure psychometric robustness through multi-item scales and external validations of the linguistic component, and deepen multilevel analyses. Additionally, it is essential to mitigate biases derived from self-perception by incorporating multiple sources of information and social desirability controls. This agenda will refine the understanding of how traits and their facets relate to entrepreneurial mindset in higher education, strengthening the relevance of evidence-based interventions for developing the competencies necessary to achieve the goals of sustainable development.

References

• 1

  Al-khamaisehR. A. M.BaileyR. P.JarvisA. (2024). Definitions of sustainable leadership in education: a systematic review and analysis. J. Res. Leadersh. Educ.19 (3), 280–297. 10.1177/19427751241283029. Original work published 2025.

  • CrossRef
  • Google Scholar
  • View reference in article

• 2

  AngeliniG.MalvasoA.SchirripaA.CampioneF.D’AddarioS. L.ToschiN.et al (2024). Unraveling sex differences in Parkinson’s disease through explainable machine learning. J. Neurol. Sci.462, 123091. 10.1016/j.jns.2024.123091

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 3

  ArbuésE.Abad-VillaverdeB.Costa-ParísA.BalaguerÁConesa-LareoM.-D.BeltramoC. (2025). Socio-emotional competencies for sustainable development: an exploratory review. Educ. Sci.15 (7), 831. 10.3390/educsci15070831

  • CrossRef
  • Google Scholar
  • View reference in article

• 4

  BanduraA. (1997). Self-efficacy: The Exercise of Control. New York: W. H. Freeman and Company.

  • Google Scholar
  • View reference in article

• 5

  Barnett-ItzhakiZ.TifferetS.EtsteinY.GefenI.RavidO.BarokasG.et al (2025). A holistic approach to sustainability in higher education institutes: social, economic, educational, and mobility perspectives. Front. Educ.10, 1588223. 10.3389/feduc.2025.1588223

  • CrossRef
  • Google Scholar
  • View reference in article

• 6

  BaronR. A.TangJ. (2009). Entrepreneurs’ social skills and new venture performance: mediating mechanisms and cultural generality. J. Manag.35, 282–306. 10.1177/0149206307312513

  • CrossRef
  • Google Scholar
  • View reference in article

• 7

  BartlettL. (2022). Specifying hybrid models of teachers’ work during COVID-19. Educ. Res.51 (2), 152–5. 10.3102/0013189X211069399

  • CrossRef
  • Google Scholar
  • View reference in article

• 8

  BeckerK.EbbersJ. J.EngelY. (2023). Network to passion or passion to network?J. Business Ventur.38 (4), 106299. 10.1016/j.jbusvent.2023.106299

  • CrossRef
  • Google Scholar
  • View reference in article

• 9

  BergkvistL.RossiterJ. R. (2007). The predictive validity of multiple-item versus single-item measures of the same constructs. J. Mark. Res.44 (2), 175–184. 10.1509/jmkr.44.2.175

  • CrossRef
  • Google Scholar
  • View reference in article

• 10

  BreimanL. (2001). Random forests. Mach .Learn.45 (1), 5–32. 10.1023/A:1010933404324

  • CrossRef
  • Google Scholar
  • View reference in article

• 11

  CaoX.KosinskiM.StillwellD. (2024). Large language models know how the personality of public figures is perceived by the general public. Sci. Rep.14, 57271. 10.1038/s41598-024-57271-z

  • CrossRef
  • Google Scholar
  • View reference in article

• 12

  CardellaG. M.Hernández-SánchezB. R.Sánchez-GarcíaJ. C. (2020). Women entrepreneurship: a systematic review to outline the boundaries of scientific literature. Front. Psychol.11, 1557. 10.3389/fpsyg.2020.01557

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 13

  ChernyshenkoO. S.KankarašM.DrasgowF. (2018). “Social and emotional skills for student success and well-being: conceptual framework for the OECD study on social and emotional skills,” OECD Education Working Papers, No. 173 (Paris: OECD Publishing). 10.1787/db1d8e59-en

  • CrossRef
  • Google Scholar
  • View reference in article

• 14

  CostaP. T.McCraeR. R. (1992). Four ways five factors are basic. Pers. Individ. Differ.13 (6), 653–665. 10.1016/0191-8869(92)90236-I

  • CrossRef
  • Google Scholar
  • View reference in article

• 15

  DavisM. H.HallJ. A.MayerP. S. (2016). Developing a new measure of entrepreneurial mindset: reliability, validity, and implications for practitioners. Consult. Psychol. J. Pract. Res.68 (1), 21–48. 10.1037/cpb0000045

  • CrossRef
  • Google Scholar
  • View reference in article

• 16

  DavisM.HallJ.MayerP. (2024). Individual feedback report sample: Individual entrepreneur [Report, PDF]. Entrepreneurial Mindset Profile (Leadership Development Institute, Eckerd College). Available online at:https://www.emindsetprofile.com/about/wp-content/uploads/2025/01/Blurred-Susan-Sample-EMP-2.0-Feedback-Report.pdf (Accessed October 25, 2025).

  • Google Scholar
  • View reference in article

• 17

  De CockR.DenooL.ClarysseB. (2020). Surviving the emotional rollercoaster called entrepreneurship: the role of emotion regulation. J. Bus. Vent.35 (2), 105936. 10.1016/j.jbusvent.2019.04.004

  • CrossRef
  • Google Scholar
  • View reference in article

• 18

  DeckerR. A.HaltiwangerJ.JarminR. S.MirandaJ. (2014). The role of entrepreneurship in US job creation and economic dynamism. J. Econ. Perspect.28 (3), 3–24. 10.1257/jep.28.3.3

  • CrossRef
  • Google Scholar
  • View reference in article

• 19

  EichstaedtJ. C.KernM. L.YadenD. B.SchwartzH. A.GiorgiS.ParkG.et al (2021). Closed- and open-vocabulary approaches to text analysis: a review, quantitative comparison, and recommendations. Psychol. Methods26 (4), 398–427. 10.1037/met0000349

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 20

  ElkalehE.AliN.Abu KhurmaO.El SherifH. M. (2025). Towards a model for sustainable leadership in educational contexts: a moderated mediated analysis of UAE and Hong Kong. Soc. Sci. Humanit. Open11, 101478. 10.1016/j.ssaho.2025.101478

  • CrossRef
  • Google Scholar
  • View reference in article

• 21

  Entrepreneurial Mindset Profile (EMP). (2019). EMP: Scale definitions and tips [PDF]. Available online at:https://www.emindsetprofile.com/wp-content/uploads/2019/10/EMP-Scale-Definitions-and-Tips.pdf (Accessed October 25, 2025).

  • Google Scholar
  • View reference in article

• 22

  European Commission. (2022). Ethical guidelines on the use of AI and data in teaching and learning for educators. Publications Office of the EU. Available online at:https://op.europa.eu/en/publication-detail/-/publication/d81a0d54-5348-11ed-92ed-01aa75ed71a1 (Accessed October 25, 2025).

  • Google Scholar
  • View reference in article

• 23

  European Union. (2024). Regulation (EU) 2024/1689… (Artificial Intelligence Act). Official Journal of the European Union, L 2024/1689 (12 July 2024). Available online at:https://eur-lex.europa.eu/eli/reg/2024/1689/oj (Accessed October 25, 2025).

  • Google Scholar
  • View reference in article

• 24

  FisherG. G.MatthewsR. A.GibbonsA. M. (2016). Developing and investigating the use of single-item measures in organizational research. J. Occup. Health Psychol.21 (1), 3–23. 10.1037/a0039139

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 25

  FreibergB.MatzS. C. (2023). Founder personality and entrepreneurial outcomes: a large-scale field study of technology startups. Proc. Natl. Acad. Sci.120 (19), e2215829120. 10.1073/pnas.2215829120

  • CrossRef
  • Google Scholar
  • View reference in article

• 26

  GeorgeG.MerrillR. K.SchillebeeckxS. J. D. (2021). Digital sustainability and entrepreneurship: how digital innovations are helping tackle climate change and sustainable development. Entrep. Theory Pract.45 (5), 999–1027. 10.1177/1042258719899425

  • CrossRef
  • Google Scholar
  • View reference in article

• 27

  GhamrawiN.ShalT.GhamrawiN. A. R.Abu-TinehA.AlshaboulY. (2025). Unleashing the potential of teacher leadership for ESD. Front. Educ.10, 1614623. 10.3389/feduc.2025.1614623

  • CrossRef
  • Google Scholar
  • View reference in article

• 28

  Global Entrepreneurship Monitor (GEM). (2023). Women’s entrepreneurship: 2022/2023 report-Challenging bias and stereotypes. GEM Consortium. Available online at:https://www.gemconsortium.org/report/gem-20222023-womens-entrepreneurship-challenging-bias-and-stereotypes-2 (Accessed October 25, 2025).

  • Google Scholar
  • View reference in article

• 29

  HagenauerW.ZipkoH. T. (2024). The relationship between entrepreneurial personality patterns linked to risk, innovation and gender across industrial sectors. Sci. Rep.14 (1), 20864. 10.1038/s41598-024-71794-5

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 30

  HickmanL.TayL.WooS. E. (2019). Validity evidence for off-the-shelf language-based personality assessment using video interviews: convergent and discriminant relationships with self and observer ratings. Pers. Assess. Decis.5 (3), 3. 10.25035/pad.2019.03.003

  • CrossRef
  • Google Scholar
  • View reference in article

• 31

  HowardM.BoudreauxM. (2024). A systematic literature review and meta-analysis of entrepreneurial personality. Entrep. Res. J.14 (1), 283–312. 10.1515/erj-2021-0322

  • CrossRef
  • Google Scholar
  • View reference in article

• 32

  HussainR.MaJ.KhandelwalR.OltmannsJ., and GuptaM. (2025). Personality prediction from life stories using language models. arXiv [Preprint]. arXiv:2506.19258. 10.48550/arXiv.2506.19258

  • CrossRef
  • Google Scholar
  • View reference in article

• 33

  IBM. (2018). Watson PI documentation. Available online at:https://cloud.ibm.com/docs/services/personality-insights/science.html (Accessed October 25, 2025).

  • Google Scholar
  • View reference in article

• 34

  JenningsP.A. (2018). The Trauma-sensitive classroom: Building Resilience with Compassionate Teaching. New York, NY: W. W. Norton & Company. Available online at:https://wwnorton.com/books/9780393711868 (Accessed October 25, 2025).

  • Google Scholar
  • View reference in article

• 35

  JiatongW.MuradM.BajunF.TufailM. S.MirzaF.RafiqM. (2021). Impact of entrepreneurial education, mindset, and creativity on entrepreneurial intention: mediating role of entrepreneurial self-efficacy. Front. Psychol.12, 724440. 10.3389/fpsyg.2021.724440

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 36

  KangW.MalvasoA. (2023a). Personality traits predict self-rated health (SRH) in coronary heart disease (CHD) patients and healthy controls. Healthcare11 (11), 1645. 10.3390/healthcare11111645

  • CrossRef
  • Google Scholar
  • View reference in article

• 37

  KangW.MalvasoA. (2023b). Mental health in coronary heart disease (CHD) patients: findings from the UK household longitudinal study (UKHLS). Healthcare11 (10), 1364. 10.3390/healthcare11101364

  • CrossRef
  • Google Scholar
  • View reference in article

• 38

  KangW.MalvasoA. (2024). Understanding the longitudinal associations between e-cigarette use and general mental health, social dysfunction and anhedonia, depression and anxiety, and loss of confidence in a sample from the UK: a linear mixed-effect examination. J. Affect. Disord.346, 200–205. 10.1016/j.jad.2023.11.013

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 39

  KankarašM.Suárez-ÁlvarezJ. (2019). The OECD Study on Social and Emotional Skills: Assessment framework. OECD Education Working Paper No. 207. 10.1787/5007adef-en

  • CrossRef
  • Google Scholar
  • View reference in article

• 40

  KjellO. N. E.KjellK.SchwartzH. A. (2024). Beyond rating scales: with targeted evaluation, large language models are poised for psychological assessment. Psychiatry Res.333, 115667. 10.1016/j.psychres.2023.115667

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 41

  KjellO. N. E.SikströmS.KjellK.SchwartzH. A. (2022). Natural language analyzed with AI-based transformers predict traditional subjective well-being measures approaching the theoretical upper limits in accuracy. Sci. Rep.12, 3918. 10.1038/s41598-022-07520-w

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 42

  LaouitiR.VitelJ.CastañoM. S. (2022). A gender-based approach to the influence of personality traits on entrepreneurial intention. J. Bus. Res.142, 819–829. 10.1016/j.jbusres.2022.01.018

  • CrossRef
  • Google Scholar
  • View reference in article

• 43

  Leal FilhoW.DibbernT.Viera TrevisanL.Coggo CristofolettiE.DinisM. A. P.MatandirotyaN.et al (2023). Mapping universities–communities partnerships in the delivery of the sustainable development goals. Front. Environ. Sci.11, 1246875. 10.3389/fenvs.2023.1246875

  • CrossRef
  • Google Scholar
  • View reference in article

• 44

  Leal FilhoW.EustachioJ. H. P. P.CaldanaA. C. F.WillM.Lange SalviaA.RampassoI. S.et al (2020). Sustainability leadership in higher education institutions: an overview of challenges. Sustainability12 (9), 3761. 10.3390/su12093761

  • CrossRef
  • Google Scholar
  • View reference in article

• 45

  LeichtA.HeissJ.ByunW. J. (2018). Issues and Trends in Education for Sustainable Development.Paris: UNESCO Publishing. Available online at:https://unesdoc.unesco.org/ark:/48223/pf0000261445 (Accessed October 25, 2025).

  • Google Scholar
  • View reference in article

• 46

  LiL.-N.YuJ.WangH. (2022). The relationship between personality traits and entrepreneurial intention: the mediating role of creativity. Front. Psychol.13, 822206. 10.3389/fpsyg.2022.822206

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 47

  LiaoY.ZhangX.LuoY. (2022). Sustainable leadership: a literature review and prospects. Front. Psychol.13, 1045570. 10.3389/fpsyg.2022.1045570

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 48

  López-NúñezM. I.Rubio-ValdehitaS.Díaz-RamiroE. M.Aparicio GarcíaM. E. (2022). The role of individual variables as antecedents of entrepreneurship processes: emotional intelligence and self-efficacy. Front. Psychol.13, 978313. 10.3389/fpsyg.2022.978313

  • CrossRef
  • Google Scholar
  • View reference in article

• 49

  MaakT.PlessN. M. (2006). Responsible leadership in a stakeholder society: a relational perspective. J. Bus. Ethics66 (1), 99–115. 10.1007/s10551-006-9047-z

  • CrossRef
  • Google Scholar
  • View reference in article

• 50

  McCraeR. R.CostaP. T.Jr. (1999). "A five-factor theory of personality," in Handbook of Personality: Theory and Research. 2nd ed. eds PervinL. A.JohnO. P. (New York, NY: Guilford Press), 139–153.

  • Google Scholar
  • View reference in article

• 51

  MuellerB. A.WolfeM. T.SyedI. (2017). Passion and grit: an exploration of the pathways leading to venture success. J. Bus. Vent.32 (3), 260–279. 10.1016/j.jbusvent.2017.02.001

  • CrossRef
  • Google Scholar
  • View reference in article

• 52

  MuranoD.LipnevichA. A.WaltonK. E.BurrusJ.WayJ. D.Anguiano-CarrascoC. (2021). Measuring social and emotional skills in elementary students: development of self-report Likert, situational judgment test, and forced-choice items. Pers. Individ. Differ.169, 110012. 10.1016/j.paid.2020.110012

  • CrossRef
  • Google Scholar
  • View reference in article

• 53

  Navas-BonillaC.Guerra-ArangoJ. A.Oviedo-GuadoD. A.Murillo-NoriegaD. E. (2025). Inclusive education through technology: a systematic review of types, tools and characteristics. Front. Educ.10, 1527851. 10.3389/feduc.2025.1527851

  • CrossRef
  • Google Scholar
  • View reference in article

• 54

  NewmanA.ObschonkaM.MoellerJ.ChandanG. G. (2021). Entrepreneurial passion: a review, synthesis, and agenda for future research. Appl. Psychol.70 (2), 816–60. 10.1111/apps.12236

  • CrossRef
  • Google Scholar
  • View reference in article

• 55

  NguyenP.TimilsinaB.ShamsuzzohaA. (2025). Higher education as a driver of green innovation and entrepreneurship: a systematic literature review and future research agenda. J. Clean. Prod.516, 145820. 10.1016/j.jclepro.2025.145820

  • CrossRef
  • Google Scholar
  • View reference in article

• 56

  NovyJ. W.WetmoreJ. M. (2021). A core curriculum for sustainability leadership. Sustainability13 (19), 10557. 10.3390/su131910557

  • CrossRef
  • Google Scholar
  • View reference in article

• 57

  OECD. (2015). Skills for Social Progress: The Power of Social and Emotional Skills. Paris: OECD Publishing. 10.1787/9789264226159-en

  • CrossRef
  • Google Scholar
  • View reference in article

• 58

  OECD. (2021). Education at a Glance 2021: OECD Indicators. Paris: OECD Publishing. 10.1787/b35a14e5-en

  • CrossRef
  • Google Scholar
  • View reference in article

• 59

  OECD. (2023). Digital Equity and Inclusion in Education (EDU/WKP(2023)14).Paris: OECD Publishing. Available online at:https://one.oecd.org/document/EDU/WKP(2023)14/en/pdf (Accessed October 25, 2025).

  • Google Scholar
  • View reference in article

• 60

  OECD. (2024) Social and Emotional Skills for Better Lives: Findings from the OECD Survey on Social and Emotional Skills 2023.Paris: OECD Publishing. 10.1787/35ca7b7c-en

  • CrossRef
  • Google Scholar
  • View reference in article

• 61

  OECD. (2025). “How are social outcomes related to education?” (Chap. 6), in Education at a Glance 2025: OECD Indicators, ed. OECD (Paris: OECD Publishing), 143–158. Available online at:https://www.oecd.org/en/publications/education-at-a-glance-2025_1c0d9c79-en/full-report/how-are-social-outcomes-related-to-education_1f5e60da.html (Accessed October 25, 2025).

  • Google Scholar
  • View reference in article

• 62

  ParkG.SchwartzH. A.EichstaedtJ. C.KernM. L.KosinskiM.StillwellD. J.et al (2015). Automatic personality assessment through social media language. J. Pers. Soc. Psychol.108 (6), 934–952. 10.1037/pspp0000020

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 63

  PedregosaF.VaroquauxG.GramfortA.MichelV.ThirionB.GriselO.et al (2011). Scikit-learn: machine learning in python. J. Mach. Learn. Res.12, 2825–2830.

  • Google Scholar
  • View reference in article

• 64

  PeljkoŽAuer AntončičJ. (2022). Impacts of entrepreneurial openness and creativity on company growth. Front. Psychol.13, 860382. 10.3389/fpsyg.2022.860382

  • Pubmed Abstract
  • CrossRef
  • Google Scholar
  • View reference in article

• 65

  PengY.AliasB. S.WanX.MansorA. N. (2024). The impact of sustainable leadership on education quality: a systematic review. J. Law Sustain. Dev.12 (6), e3760. 10.55908/sdgs.v12i6.3760

  • CrossRef
  • Google Scholar
  • View reference in article

• 66

  PostigoÁCuestaM.García-CuetoE. (2021). Personalidad emprendedora, escrupulosidad, autocontrol y determinación: el lado psicológico del trabajo por cuenta propia. Ann. Psychol.37 (2), 361–370. 10.6018/analesps.453711

  • CrossRef
  • Google Scholar
  • View reference in article

• 67

  RauchA.GielnikM. M. (2021). "Evidence-based entrepreneurship: an extended approach," in The Psychology of Entrepreneurship: New Perspectives, eds GielnikM. M.CardonM. S.FreseM. (Abingdon: Routledge), 323–339. 10.4324/9781003137573-18

  • CrossRef
  • Google Scholar
  • View reference in article

• 68

  RosárioA. T.RaimundoR. (2024). Sustainable entrepreneurship education: a systematic bibliometric literature review. Sustainability16 (2), 784. 10.3390/su16020784

  • CrossRef
  • Google Scholar
  • View reference in article

• 69

  RosárioA. T.RaimundoR. J.CruzS. P. (2022). Sustainable entrepreneurship: a literature review. Sustainability14 (9), 5556. 10.3390/su14095556

  • CrossRef
  • Google Scholar
  • View reference in article

• 70

  ŞahinF.KaradağH.TuncerB. (2019). Big five personality traits, entrepreneurial self-efficacy and entrepreneurial intention: a configurational approach. Int. J. Entrep. Behav. Res.25 (6), 1188–1211. 10.1108/IJEBR-07-2018-0466

  • CrossRef
  • Google Scholar
  • View reference in article

• 71

  SalamehA. A.AkhtarH.GulR.OmarA. B.HanifS. (2022). Personality traits and entrepreneurial intentions: financial risk-taking as mediator. Front. Psychol.13, 927718. 10.3389/fpsyg.2022.927718

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 72

  Samaniego LópezM. V.Orrego RiofríoM. C.Barriga-FrayS. F.ViteriP.SB. (2025). Technologies in inclusive education: solution or challenge? A systematic review. Educ. Sci.15 (6), 715. 10.3390/educsci15060715

  • CrossRef
  • Google Scholar

• 73

  SarasvathyS. D. (2001). Causation and effectuation: toward a theoretical shift from economic inevitability to entrepreneurial contingency. Acad. Manag. Rev.26 (2), 243–263. 10.2307/259121

  • CrossRef
  • Google Scholar

• 74

  SeaboldS.PerktoldJ. (2010). “Statsmodels: econometric and statistical modeling with python.” in Proceedings of the 9th Python in Science Conference (SciPy 2010), 57–61. 10.25080/Majora-92bf1922-011

  • CrossRef
  • Google Scholar

• 75

  ShuiX.KorostelevaJ.NguyenB. (2025). From green to growth: the effect of go green on entrepreneurial growth aspirations. J. Bus. Ventur.40 (4), 106494. 10.1016/j.jbusvent.2025.106494

  • CrossRef
  • Google Scholar

• 76

  SomiàT.VecchiariniM. (2024). Navigating the new frontier: the impact of artificial intelligence on students’ entrepreneurial competencies. Int. J. Entrep. Behav. Res.30 (11), 236–260. 10.1108/IJEBR-08-2023-0788

  • CrossRef
  • Google Scholar

• 77

  SotoC. J.JohnO. P. (2017). The next big five inventory (BFI-2): developing and assessing a hierarchical model with 15 facets to enhance bandwidth, fidelity, and predictive power. J. Pers. Soc. Psychol.113 (1), 117–143. 10.1037/pspp0000096

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 78

  SottiniA.ZupicI.GiudiciA. (2024). Social entrepreneurship and social innovation: a bibliometric review and research agenda. Eur. Manag. J.43 (2), 182–193. 10.1016/j.emj.2024.03.006

  • CrossRef
  • Google Scholar

• 79

  ThielmannI.SpadaroG.BallietD. (2020). Personality and prosocial behavior: a theoretical framework and meta-analysis. Psychol. Bull.146 (1), 30–90. 10.1037/bul0000217

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 80

  TrevisanL. V.Leal FilhoW.Ávila PedrozoE. (2024). Transformative organisational learning for sustainability in higher education: a literature review and an international multi-case study. J. Clean. Prod.447, 141634. 10.1016/j.jclepro.2024.141634

  • CrossRef
  • Google Scholar

• 81

  UNESCO. (2015). Rethinking Education: Towards a Global Common Good?Paris: UNESCO Publishing. Available online at:https://unevoc.unesco.org/e-forum/RethinkingEducation.pdf (Accessed October 25, 2025).

  • Google Scholar

• 82

  UNESCO. (2020). Education for Sustainable Development: A Roadmap (ESD for 2030).Paris: UNESCO Publishing. Available online at:https://unesdoc.unesco.org/ark:/48223/pf0000374802 (Accessed October 25, 2025).

  • Google Scholar

• 83

  UNESCO. (2021). Reimagining our Futures Together: A new Social Contract for Education.Paris: UNESCO Publishing. Available online at:https://unesdoc.unesco.org/ark:/48223/pf0000379707 (Accessed October 25, 2025).

  • Google Scholar

• 84

  UNESCO. (2023). Guidance for Generative AI in Education and Research.Paris: UNESCO Publishing. Available online at:https://unesdoc.unesco.org/ark:/48223/pf0000386693 (Accessed October 25, 2025).

  • Google Scholar

• 85

  UriarteS.Baier-FuentesH.Espinoza-BenavidesJ.Inzunza-MendozaW. (2025). Artificial intelligence technologies and entrepreneurship: a hybrid literature review. Rev. Manag. Sci.20, 251–299. 10.1007/s11846-025-00839-4

  • CrossRef
  • Google Scholar

• 86

  VecchiariniM.SomiàT. (2023). Redefining entrepreneurship education in the age of artificial intelligence: an explorative analysis. Int. J. Manag. Educ.21 (3), 100879. 10.1016/j.ijme.2023.100879

  • CrossRef
  • Google Scholar

• 87

  WaltonK. E.AllenJ.BoxM. J.MuranoD.BurrusJ. (2023). Social and emotional skills predict postsecondary enrollment and retention. J. Intell.11 (10), 186. 10.3390/jintelligence11100186

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 88

  WaltonK. E.BurrusJ.MuranoD.Anguiano-CarrascoC.WayJ.RobertsR. D. (2022). A big five-based multimethod social and emotional skills assessment: the mosaic™ by ACT® social emotional learning assessment. J. Intell.10 (4), 72. 10.3390/jintelligence10040072

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 89

  WangX.LiuJ.JiaS.ZhangY.LiY.ChenH.et al (2024). Hybrid teaching after COVID-19: advantages, challenges and future directions. BMC Med. Educ.24, 763. 10.1186/s12909-024-05745-z

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 90

  WanousJ. P.ReichersA. E.HudyM. J. (1997). Overall job satisfaction: how good are single-item measures?J. Appl. Psychol.82 (2), 247–252. 10.1037/0021-9010.82.2.247

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 91

  World Economic Forum. (2023). The Future of Jobs Report 2023. Geneva: World Economic Forum. Available online at:https://www.weforum.org/reports/the-future-of-jobs-report-2023 (Accessed October 25, 2025).

  • Google Scholar

• 92

  XuM. (2023). The effects of personality traits on entrepreneurial intention and creativity among Chinese and African college students in era of industry 4.0/5.0. Front. Psychiatry14, 1098148. 10.3389/fpsyt.2023.1098148

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 93

  ZhangY.WangP.ZhaoY. (2022). Big five personality, academic entrepreneurial motivation, and academic entrepreneurial intention: a research method based on fuzzy set qualitative comparative analysis. Front. Psychol.12, 799770. 10.3389/fpsyg.2021.799770

  • Pubmed Abstract
  • CrossRef
  • Google Scholar

• 94

  ZhaoH.SeibertS. E.LumpkinG. T. (2010). The relationship of personality to entrepreneurial intentions and performance: a meta-analytic review. J. Manag.36 (2), 381–404. 10.1177/0149206309335187

  • CrossRef
  • Google Scholar