Evaluating Emotional Intelligence, Authentic Leadership, Self-Efficacy, and Perceived Stress Among Diagnostic Radiography Technology Students in Clinical Practice

Abstract

Background: Research on emotional intelligence and its associations with authentic leadership, general self-efficacy, and perceived stress among allied health students, particularly radiography students, remains scarce. This study described the levels of these constructs among radiography students during clinical practice, examined their interrelationships, and identified the emotional intelligence facets predicting perceived stress and authentic leadership. Methods: A cross-sectional study was conducted among 135 radiography students enrolled in clinical practice at Jazan University, Saudi Arabia, from March to May 2026 (response rate 61.3%). Data were collected using the Schutte Self-Report Emotional Intelligence Test, Authentic Leadership Questionnaire, General Self-Efficacy Scale, and Perceived Stress Scale. Spearman correlations and hierarchical multiple linear regression were performed. Results: Emotional intelligence was positively correlated with authentic leadership (p < 0.001) and general self-efficacy (p = 0.035) and negatively correlated with perceived stress (p = 0.012). Managing Own Emotions was the strongest negative predictor of perceived stress (β = −0.650, p < 0.001), whereas Managing Others’ Emotions was the strongest predictor of authentic leadership (β = 0.463, p < 0.001). General self-efficacy predicted neither outcome. Female students and those with higher GPAs reported higher emotional intelligence and authentic leadership scores. Conclusions: Specific emotional intelligence facets showed distinct predictive roles, with self-directed emotion regulation predicting lower perceived stress and interpersonal emotion management predicting authentic leadership. These findings support the integration of targeted, facet-specific emotional intelligence training within radiography clinical curricula as a potential approach to support stress management and leadership development.

1. Introduction

Clinical practice is a formative phase in healthcare education, marking the transition from classroom learning to direct patient care. During this period, students are exposed to multiple stressors, including high-stakes decision-making, complex interpersonal interactions, and continuous performance evaluation [1,2]. Radiography students face additional discipline-specific demands, such as mastering technical imaging skills, adhering to radiation safety protocols, and communicating effectively with patients in time-constrained environments [3]. Successfully managing these demands requires more than technical competence; it also depends on interpersonal and psychological resources that support adaptation, confidence, and stress regulation. Although the clinical environment is widely recognized as demanding across healthcare disciplines [4,5], the extent to which radiography students possess these resources and how these resources relate to one another warrants empirical investigation.

Emotional intelligence, authentic leadership, self-efficacy, and perceived stress are particularly relevant for understanding clinical readiness among healthcare students. Emotional intelligence (EI) refers to the ability to perceive, understand, manage, and use emotions in oneself and others [6,7], and it supports effective communication, emotional regulation, and teamwork in clinical settings [8,9]. Authentic leadership (AL), which is characterized by self-awareness, internalized moral perspective, balanced processing, and relational transparency [10,11], may help students develop ethical judgment and psychologically safe professional behavior [12,13]. General self-efficacy (GSE) reflects an individual’s belief in their ability to manage challenging situations [14,15], and it influences persistence, problem-solving, and confidence during clinical training [16]. Perceived stress (PSS) refers to the extent to which individuals appraise their circumstances as unpredictable, uncontrollable, and overwhelming [17], and in healthcare education it is shaped by workload, clinical expectations, and interpersonal tensions [18,19]. Together, these constructs represent complementary emotional, motivational, leadership, and stress-related dimensions of the student clinical experience.

Theoretical and empirical evidence suggests meaningful relationships among these constructs. Bandura’s social cognitive theory [20] explains how self-efficacy shapes appraisal and coping, while Lazarus and Folkman’s transactional model of stress [21] describes how emotional and cognitive resources influence responses to demanding situations. In healthcare and student populations, higher EI has been linked to lower PSS and stronger self-efficacy, and prior work in medical students also reported positive associations between EI, leadership capacity, self-efficacy, and lower stress [22,23,24,25]. In radiography, studies have shown that clinical stressors are common and that emotional competencies may be important during training [1,2]. However, existing studies have largely examined these constructs separately or in different populations, and evidence from radiography students remains limited. These two frameworks guided the present study in three ways. Social cognitive theory motivated the inclusion of general self-efficacy and the emotion-management facets as appraisal and coping resources, the transactional model framed perceived stress as the outcome of how those resources shape appraisal of clinical demands, and together they generated the expectation that self-directed and interpersonal emotional competencies would relate differentially to stress and leadership, which guided interpretation of the facet-specific regression findings.

Despite these theoretical and empirical foundations, few studies have examined EI, AL, self-efficacy, and PSS simultaneously. The only known investigation to measure these four constructs together found that EI was positively associated with leadership capacity and self-efficacy and negatively associated with PSS among medical students [26]. However, that study did not focus on allied health students or examine the unique contribution of EI facets. Within radiography research, previous studies have mainly described clinical stressors or measured EI alone, leaving an important gap in understanding how these variables interact in clinical practice [1,2]. Therefore, the levels of EI, AL, GSE, and PSS among radiography students in clinical practice were described, the bivariate relationships among these variables were examined, and the EI facets uniquely predicting PSS and AL were determined. It was hypothesized that EI would be significantly associated with AL, self-efficacy, and PSS in this population.

2. Materials and Methods

2.1. Study Design, Setting, and Population

A cross-sectional, descriptive-correlational study was conducted at the College of Nursing and Health Sciences, Jazan University, Saudi Arabia, from March to May 2026.

The target population comprised undergraduate radiography students who had completed at least two years of the Diagnostic Radiography Technology (DRT) program and had commenced clinical practice training. Eligibility criteria required enrollment in the third academic year or higher with active participation in clinical practice courses at the time of the survey. Students with no clinical practice experience were excluded.

2.2. Sample Size Justification

An a priori power analysis was conducted using Cohen’s [27] parameters for hierarchical multiple regression with eight predictors (two demographic controls and six psychological variables): significance level (α) of 0.05, statistical power (1 − β) of 0.80, and a medium effect size (f² = 0.15). The minimum required sample size was calculated as 133 participants. A total of 220 eligible students were invited to participate, of whom 135 completed the survey and met inclusion criteria, yielding an effective response rate of 61.3%.

2.3. Instruments

The survey comprised a demographic information section followed by four standardized instruments, namely the SSEIT, the ALQ, the GSE-10, and the PSS-4. All instruments were administered using previously validated Arabic translations [26]. All four instruments were selected as generic, validated measures of broad psychological constructs rather than domain-specific scales. The SSEIT captures overall emotional intelligence as a trait-based capacity applicable across settings, rather than clinical-specific emotional competencies. The ALQ assesses self-perceived authentic leadership qualities relevant to emerging professionals, consistent with its validated use among students without formal leadership roles. The GSE-10 measures generalized confidence in coping with novel and demanding situations rather than task-specific radiographic competencies, aligning with the study’s focus on broad self-efficacy appraisals. The PSS-4 assesses global perceived stress, consistent with the transactional stress model, rather than enumerating clinical-specific stressors, while its brevity minimized respondent burden within a multi-scale survey battery. These generic instruments facilitate comparability with the broader health professions education literature, where they are widely used.

Participants provided information on sex, age, academic year (3rd year, 4th year), academic GPA (above 4.5, 4.01–4.5, 3.51–4.0, 3.5 or below), cigarette/vape/shisha use (yes/no), and regular physical activity engagement (yes/no).

Schutte Self-Report Emotional Intelligence Test (SSEIT).

EI was assessed using the 33-item SSEIT [7]. Items are rated on a 5-point Likert scale ranging from 1 (Strongly Disagree) to 5 (Strongly Agree). Three items (5, 28, and 33) are reverse-coded. The instrument yields four subscale scores: Perception of Emotion (10 items, range 10–50), Managing Own Emotions (9 items, range 9–45), Managing Others’ Emotions (8 items, range 8–40), and Utilization of Emotion (6 items, range 6–30). The total score ranges from 33 to 165, with higher scores indicating greater EI. In the present study, Cronbach’s alpha for the total scale was 0.931, and subscale alphas ranged from 0.770 to 0.817.

Authentic Leadership Questionnaire (ALQ).

AL was measured using the 16-item ALQ [10]. Items are rated on a 5-point Likert scale from 1 (Strongly Disagree) to 5 (Strongly Agree). The ALQ comprises four subscales: Self-Awareness (items 1, 5, 9, 13), Internalized Moral Perspective (items 2, 6, 10, 14), Balanced Processing (items 3, 7, 11, 15), and Relational Transparency (items 4, 8, 12, 16). Each subscale score ranges from 4 to 20, and the total score ranges from 16 to 80, with higher scores reflecting stronger AL attributes. Cronbach’s alpha for the total scale was 0.845; subscale alphas ranged from 0.530 to 0.631, which is expected for four-item scales and consistent with published norms for the ALQ [11]. The comparatively low subscale coefficients are consistent with evidence that the four-item ALQ subscales show substantial content overlap and modest internal consistency, a psychometric characteristic documented in prior structural analyses of the instrument [28].

General Self-Efficacy Scale (GSE-10).

GSE was assessed using the 10-item GSE-10 [15]. Items are rated on a 4-point Likert scale from 1 (Not at All True) to 4 (Exactly True). The scale is unidimensional, producing a single total score ranging from 10 to 40, with higher scores indicating greater self-efficacy. Cronbach’s alpha in this study was 0.911.

Perceived Stress Scale—4-Item Version (PSS-4).

PSS was measured using the PSS-4 [17]. Items are rated on a 5-point scale from 0 (Never) to 4 (Very Often). Items 2 and 3 are reverse-scored. The total score ranges from 0 to 16, with higher scores indicating greater PSS.

2.4. Pilot Study

A pilot study was conducted with 30 DRT students prior to main data collection to evaluate instrument clarity and comprehension. No modifications were deemed necessary. Pilot participants were not included in the final analytic sample.

2.5. Procedure

Data were collected through an anonymous online survey administered via Google Forms. The survey link was distributed to eligible students through official academic communication channels. Completion time was approximately 15–20 min. All survey items were configured as required fields; consequently, the dataset contained no missing values, and all 135 responses were complete and included in the analysis.

2.6. Ethical Considerations

This study was approved by the Institutional Review Board (IRB) of Jazan University (Approval No. REC-47-10-1857). The study adhered to the ethical principles outlined in the Declaration of Helsinki. Participation was voluntary, and informed consent was obtained electronically from all participants prior to data collection. All responses were collected anonymously; no identifying information was recorded.

2.7. Data Analysis

Descriptive statistics including frequencies, percentages, means, standard deviations, medians, and interquartile ranges were calculated for all demographic variables and scale/subscale scores. Internal consistency reliability was evaluated using Cronbach’s alpha for each scale and subscale. Normality of score distributions was assessed using skewness and kurtosis values and the Kolmogorov–Smirnov test. As most score distributions departed from normality (K–S p < 0.05), non-parametric tests were employed for group comparisons and bivariate correlations. Group comparisons were performed using Mann–Whitney U tests for two-group variables (sex, smoking status, exercise status) and Kruskal–Wallis H tests for multi-group variables (academic year, GPA). Spearman’s rank-order correlation coefficients (ρ) were computed to examine bivariate associations among all scale and subscale scores.

Hierarchical multiple linear regression analysis (ordinary least squares) was conducted to identify predictors of PSS (PSS total) and AL (AL total). For each dependent variable, predictors were entered in two steps. Variables were entered hierarchically to distinguish the contribution of psychological factors from that of background characteristics. Step 1 included the demographic controls (sex and GPA), entered to account for their potential confounding influence, where GPA was coded on an ordinal scale (higher values indicating higher GPA). Step 2 added the psychological predictor variables (four EI subscales, GSE total, and the complementary outcome variable) to estimate their incremental contribution beyond demographics. Statistical significance was set at p < 0.05 (two-tailed) for all analyses. Although several score distributions departed from normality, ordinary least-squares regression was retained because it is robust to moderate departures from residual normality at this sample size [29], and the assumption relevant to the model concerns the residuals rather than the marginal distributions of individual variables. Regression assumptions were therefore examined directly. Multicollinearity was assessed using the variance inflation factor (VIF) and tolerance for every predictor. Normality of residuals was evaluated with the Shapiro–Wilk test, homoscedasticity with the Breusch–Pagan test, and influential observations with Cook’s distance. Variance inflation factors for each predictor, and the remaining diagnostic results are summarised in Section 3.4. All statistical procedures were performed using R statistical software (version 4.5.3). Variance inflation factors were obtained using the car package, and the Breusch–Pagan test was conducted using the lmtest package.

3. Results

3.1. Sample Characteristics

Demographic characteristics are presented in Table 1. The study included 135 DRT students (mean age 20.96 ± 1.84), with the majority being female (63.0%), in their 4th academic year (51.9%), holding a GPA above 4.5 (42.2%), non-smokers (94.8%), and not regularly exercising (67.4%).

Table 1. Sample demographics and group comparisons.

Variable	n (%)	EI Total Median (IQR)	AL Total Median (IQR)	GSE Total Median (IQR)	PSS Total Median (IQR)
Sex		p < 0.001 ***	p < 0.001 ***	p = 0.551	p = 0.268
Female	85 (63.0)	128 (17)	58 (12)	32 (8)	7 (4)
Male	50 (37.0)	120.5 (16.75)	52 (9)	30 (8.75)	7 (3)
Age (M ± SD)	20.96 ± 1.84
Academic Year		p = 0.422	p = 0.996	p = 0.276	p = 0.195
3rd Year	65 (48.1)	126 (12)	56 (11)	32 (7)	7 (3)
4th Year	70 (51.9)	126 (18)	54 (12)	30 (9.75)	6 (4)
Academic GPA		p < 0.001 ***	p = 0.010 *	p = 0.540	p = 0.910
Above 4.5	57 (42.2)	128 (19)	59 (12)	32 (7)	6 (4)
4.01–4.5	45 (33.3)	126 (13)	54 (10)	32 (8)	7 (4)
3.51–4.0	22 (16.3)	117.5 (24.25)	51 (5.75)	29.5 (7.5)	7.5 (3.75)
3.5 or below	11 (8.1)	126 (15.5)	51 (13)	30 (10.5)	7 (2.5)
Smoking		p = 0.177	p = 0.512	p = 0.173	p = 0.370
No	128 (94.8)	126 (16.25)	55.5 (11.25)	31 (8.25)	7 (4)
Yes	7 (5.2)	122 (13.5)	50 (9)	36 (8.5)	5 (3.5)
Exercise		p = 0.586	p = 0.148	p = 0.667	p = 0.346
Yes	44 (32.6)	126 (10.75)	52 (13)	31 (9.25)	6.5 (4)
No	91 (67.4)	126 (18)	56 (11)	32 (7)	7 (4)

Note. Mann–Whitney U for two-group; Kruskal–Wallis H for multi-group comparisons. * p < 0.05, *** p < 0.001.

Female students reported significantly higher SSEIT total scores, Median = 128 vs. 120.5, p < 0.001, and AL total scores, Median = 58 vs. 52, p < 0.001, than males. Academic GPA was significantly associated with EI (p < 0.001) and AL (p = 0.010), with higher-GPA students reporting higher scores.

3.2. Descriptive Statistics

Descriptive statistics for all scale and subscale scores are presented in Table 2. Students reported moderate-to-high EI (123.38 ± 16.53), moderate AL (55.39 ± 8.76), moderately high GSE (30.73 ± 6.79), and moderate PSS (6.51 ± 3.09).

Table 2. Descriptive statistics: scale and subscale scores.

Scale/Subscale	Items	Range	Mean	SD	Median	Min	Max
EI Total	33	33–165	123.38	16.53	126	59	156
Perception	10	10–50	35.97	5.34	36	22	50
Managing Own Emotions	9	9–45	34.70	5.00	35	16	45
Managing Others’ Emotions	8	8–40	29.81	4.54	30	12	40
Utilization	6	6–30	22.89	3.67	23	9	30
AL Total	16	16–80	55.39	8.76	55	25	80
Self-Awareness	4	4–20	13.61	2.75	13	6	20
Moral Perspective	4	4–20	14.33	2.47	14	7	20
Balanced Processing	4	4–20	13.89	2.68	14	4	20
Relational Transparency	4	4–20	13.55	2.77	13	4	20
GSE Total	10	10–40	30.73	6.79	31	10	40
PSS Total	4	0–16	6.51	3.09	7	0	15

3.3. Bivariate Correlations

Spearman correlations among all scale and subscale scores are presented in Table 3. At the total score level, EI showed a strong positive correlation with AL (ρ = 0.630, p < 0.001), a weak positive correlation with GSE (ρ = 0.181, p = 0.035), and a weak negative correlation with PSS (ρ = −0.217, p = 0.012). AL was weakly negatively correlated with PSS (ρ = −0.172, p = 0.046). Neither AL nor PSS was significantly associated with GSE. At the subscale level, all four EI subscales were significantly positively correlated with AL total (ρ = 0.473–0.633, all p < 0.001), with Managing Others’ Emotions showing the strongest association (ρ = 0.633). Managing Own Emotions was the only EI facet significantly negatively correlated with PSS (ρ = −0.353, p < 0.001). Relational Transparency was the only AL subscale significantly negatively correlated with PSS (ρ = −0.181, p < 0.05). GSE was weakly positively correlated with two EI facets: Perception of Emotion (ρ = 0.171, p < 0.05) and Managing Own Emotions (ρ = 0.205, p < 0.05).

Table 3. Spearman’s correlation matrix.

Variable	1	2	3	4	5	6	7	8	9	10	11	12
1. Perception of Emotion	—
2. Managing Own Emotions	0.632 ***	—
3. Managing Others’ Emotions	0.641 ***	0.685 ***	—
4. Utilization of Emotion	0.560 ***	0.711 ***	0.675 ***	—
5. EI total	0.827 ***	0.868 ***	0.872 ***	0.815 ***	—
6. Self-Awareness	0.375 ***	0.415 ***	0.528 ***	0.463 ***	0.506 ***	—
7. Internalized Moral Perspective	0.569 ***	0.501 ***	0.527 ***	0.389 ***	0.577 ***	0.607 ***	—
8. Balanced Processing	0.453 ***	0.342 ***	0.477 ***	0.308 ***	0.454 ***	0.466 ***	0.602 ***	—
9. Relational Transparency	0.424 ***	0.417 ***	0.516 ***	0.362 ***	0.502 ***	0.511 ***	0.561 ***	0.569 ***	—
10. AL total	0.556 ***	0.520 ***	0.633 ***	0.473 ***	0.630 ***	0.792 ***	0.829 ***	0.805 ***	0.803 ***	—
11. GSE Total	0.171 *	0.205 *	0.112	0.100	0.181 *	0.062	0.097	0.031	−0.003	0.031	—
12. PSS Total	−0.144	−0.353 ***	−0.149	−0.111	−0.217 *	−0.135	−0.106	−0.136	−0.181 *	−0.172 *	−0.036	—

Note. * p < 0.05, *** p < 0.001.

3.4. Hierarchical Multiple Regression

Table 4 presents the results of hierarchical multiple linear regression analyses examining predictors of PSS and AL. In the first model predicting PSS, demographics did not significantly predict PSS (R² = 0.008, p = 0.569). Adding psychological predictors significantly improved the model (ΔR² = 0.152, F-change(6, 126) = 3.813, p = 0.002; full model R² = 0.161, adjusted R² = 0.108, F(8, 126) = 3.020, p = 0.004), with Managing Own Emotions emerging as the strongest predictor (β = −0.650, p < 0.001). Managing Others’ Emotions was a borderline positive predictor (β = 0.303, p = 0.048), a pattern that may reflect a statistical suppressor effect, given that the bivariate correlation between managing others’ emotions and PSS was non-significant (ρ = −0.149, p > 0.05); however, additional analyses would be required to confirm this interpretation. No other predictors were significant.

Table 4. Hierarchical multiple linear regression.

Panel A: DV = Perceived Stress (PSS Total)
Predictor	B	SE	β	t	p	95% CI	VIF
Step 1: R2 = 0.008, adjusted R2 = −0.007, F(2, 132) = 0.566, p = 0.569
Sex (Female = 1)	−0.543	0.749	−0.085	−0.724	0.470	[−2.025, 0.939]	1.96
GPA (ordinal)	−0.032	0.381	−0.010	−0.085	0.932	[−0.786, 0.721]	1.90
Step 2: R2 = 0.161, adjusted R2 = 0.108, ΔR2 = 0.152, F-change(6, 126) = 3.813, p = 0.002
Sex (Female = 1)	−0.442	0.728	−0.069	−0.607	0.545	[−1.883, 0.999]	1.96
GPA (ordinal)	0.167	0.364	0.051	0.457	0.648	[−0.554, 0.887]	1.90
EI-Perception of Emotion	0.032	0.074	0.055	0.435	0.664	[−0.113, 0.178]	2.43
EI-Managing Own Emotions	−0.402	0.097	−0.650	−4.151	<0.001	[−0.593, −0.210]	3.68
EI-Managing Others’ Emotions	0.206	0.103	0.303	1.996	0.048	[0.002, 0.411]	3.46
EI-Utilization of Emotion	0.150	0.121	0.178	1.241	0.217	[−0.089, 0.389]	3.09
AL Total	−0.041	0.040	−0.118	−1.025	0.307	[−0.122, 0.039]	1.98
GSE Total	0.016	0.038	0.035	0.422	0.674	[−0.059, 0.091]	1.03
Panel B: DV = Authentic Leadership (AL Total)
Predictor	B	SE	β	t	p	95% CI	VIF
Step 1: R2 = 0.070, adjusted R2 = 0.055, F(2, 132) = 4.934, p = 0.009
Sex (Female = 1)	4.285	2.058	0.237	2.082	0.039	[0.214, 8.356]	1.96
GPA (ordinal)	0.343	1.046	0.037	0.328	0.743	[−1.727, 2.413]	1.90
Step 2: R2 = 0.499, adjusted R2 = 0.467, ΔR2 = 0.430, F-change(6, 126) = 18.021, p < 0.001
Sex (Female = 1)	0.484	1.597	0.027	0.303	0.762	[−2.677, 3.645]	1.96
GPA (ordinal)	0.347	0.798	0.038	0.435	0.664	[−1.232, 1.927]	1.90
EI-Perception of Emotion	0.284	0.159	0.173	1.782	0.077	[−0.031, 0.599]	2.37
EI-Managing Own Emotions	0.116	0.226	0.066	0.515	0.608	[−0.331, 0.563]	4.17
EI-Managing Others’ Emotions	0.892	0.216	0.463	4.135	<0.001	[0.465, 1.319]	3.14
EI-Utilization of Emotion	0.069	0.266	0.029	0.258	0.796	[−0.458, 0.596]	3.13
GSE Total	−0.104	0.082	−0.081	−1.264	0.208	[−0.267, 0.059]	1.02
PSS Total	−0.199	0.194	−0.070	−1.025	0.307	[−0.584, 0.185]	1.18

Note. B = unstandardized coefficient; β = standardized coefficient; Sex: Male = 0, Female = 1. GPA: ordinal (higher values indicate higher GPA). VIF = variance inflation factor; all values were below 5.0, indicating no problematic multicollinearity. Significant predictors (p < 0.05) are shown in bold.

In the second model predicting AL, demographics significantly predicted AL (R² = 0.070, p = 0.009), with sex as a significant predictor. Adding psychological predictors substantially improved the model (ΔR² = 0.430, F-change(6, 126) = 18.021, p < 0.001; full model R² = 0.499, adjusted R² = 0.467, F(8, 126) = 15.687, p < 0.001), and Managing Others’ Emotions emerged as the strongest predictor (β = 0.463, p < 0.001). Once EI facets were entered, the effect of sex became non-significant (β = 0.027, p = 0.762), indicating that the sex difference observed in Step 1 overlaps statistically with the EI facets. Because the design is cross-sectional and no formal mediation analysis was conducted, mediation cannot be inferred from this pattern. No other predictors were significant.

Regression assumptions were satisfactorily met for both models. All variance inflation factors were below 5.0 (range 1.02–4.17), indicating no problematic multicollinearity among predictors. Residuals were normally distributed for the stress model (Shapiro–Wilk p = 0.384) and showed only mild non-normality for the leadership model (Shapiro–Wilk p = 0.021, skewness = −0.60), which is within acceptable limits for ordinary least-squares regression at this sample size. Homoscedasticity was supported for both models (Breusch–Pagan p > 0.05), and no influential observations were identified (maximum Cook’s distance = 0.084 and 0.152 for the stress and leadership models, respectively). Ninety-five percent confidence intervals for all regression coefficients are reported in Table 4.

4. Discussion

This study evaluated the relationships among EI, AL, GSE, and PSS among radiography students during clinical practice. Students reported moderate-to-high EI (123.38 ± 16.53), which aligns closely with the global normative mean for the SSEIT, M ≈ 124 [30], is comparable to nursing students assessed with the same instrument, 120.59 ± 20.78 [31], and is similar to medical imaging students in Malaysia, 121.60 ± 11.23 [32]. Together, these comparisons suggest a consistent moderate-to-high EI profile across healthcare student populations. A recent scoping review of 163 studies on EI in allied health students found that most studies focused on nursing and medicine, with very limited evidence from radiography or medical imaging programs [33]. Among EI facets, managing own emotions and utilization of emotion received the highest per-item scores, indicating that students’ emotional strengths lie primarily in self-regulation rather than interpersonal emotion management, a distinction that carries implications for the regression findings discussed below.

AL scores were moderate (55.39 ± 8.76), with internalized moral perspective scoring highest among the four AL dimensions. This suggests that students’ ethical orientation is more developed than their relational transparency or balanced processing skills, a pattern also reported among healthcare professionals, where relational transparency tends to score lowest [34]. These patterns indicate that radiography students enter clinical practice with a reasonable foundation of emotional awareness and ethical judgment. The comparatively lower scores in interpersonal emotional competencies and relational leadership dimensions, however, suggest that skills involving others, such as reading emotions, sharing perspectives, and engaging transparently, require deliberate development during clinical training.

Students reported moderately high GSE, 30.73 ± 6.79 out of 40. This exceeded the cross-national normative mean of 29.6 reported across 25 countries [35] and was higher than that of Chinese nursing students, 24.3 ± 4.3 [35], and nursing students at a comparable institution, 28.56 ± 8.02 [31]. This suggests that radiography students who have progressed to clinical rotations possess reasonable confidence in their ability to manage challenging situations, likely shaped by accumulated mastery experiences during training [16]. PSS was moderate, 6.51 ± 3.09 out of 16, comparable to nursing students assessed with the same PSS instrument, 7.15 ± 2.11 [31]. While this level does not indicate acute distress, it suggests that clinical demands are perceived as meaningfully stressful, reinforcing the need to identify which psychological resources most effectively buffer stress during training.

At the bivariate level, EI was strongly correlated with AL (ρ = 0.630, p < 0.001), weakly correlated with self-efficacy (ρ = 0.181, p = 0.035), and negatively correlated with PSS (ρ = −0.217, p = 0.012). AL was also weakly negatively correlated with PSS (ρ = −0.172, p = 0.046), but neither AL nor PSS was significantly associated with GSE. This pattern positions EI as the central connecting variable in the network, significantly related to all three constructs, while self-efficacy appears to operate independently. The strong EI–AL association is consistent with meta-analytic evidence, ρ = 0.54 [24], and with longitudinal findings among nursing students showing that changes in EI track closely with changes in leadership [36]. The negative EI–PSS correlation aligns with findings across healthcare student populations [8,37]. The weak but significant total EI–GSE correlation was driven by two specific facets (perception of emotion and managing own emotions), suggesting that self-efficacy overlaps with self-oriented emotional competencies rather than with interpersonal facets such as managing others’ emotions. The absence of meaningful GSE associations with stress or leadership contrasts with nursing studies where self-efficacy played a mediating role [25], and may reflect the limited sensitivity of a generalized self-efficacy measure in a domain-specific clinical context. Whereas the nursing studies that reported a mediating role for self-efficacy typically used task-specific or clinical self-efficacy scales, the generalized measure used here captures broad confidence rather than radiography-specific competence, which plausibly explains why it did not relate to stress or leadership in this sample. This interpretation is explored further in the regression findings below. At the subscale level, relational transparency was the only AL subscale significantly negatively correlated with PSS. This suggests that students who communicate more openly and authentically in clinical settings may experience lower stress. A plausible reason is that transparency reduces the psychological burden of impression management and facilitates supportive interpersonal exchanges.

Hierarchical regression revealed that specific EI facets served distinct predictive roles. In the stress model, managing own emotions was the sole significant negative predictor of PSS (β = −0.650, p < 0.001), with the full model explaining 16.1% of the variance. This finding aligns with the theoretical proposition that individuals who effectively regulate their own emotional states appraise clinical stressors as manageable rather than overwhelming [21], and is consistent with evidence that self-management is the strongest EI facet predictor of academic stress among medical students [38]. Similar patterns have been reported among occupational therapy students [37] and healthcare professionals [39]. The borderline positive coefficient for managing others’ emotions (β = 0.303, p = 0.048) may reflect a statistical suppressor effect, as the bivariate correlation between managing others’ emotions and PSS was non-significant (ρ = −0.149, p > 0.05); however, additional analyses would be required to confirm this interpretation. This pattern may also suggest that high interpersonal emotional attunement carries costs through vicarious stress or compassion fatigue [40].

In the leadership model, managing others’ emotions was the dominant predictor of AL (β = 0.463, p < 0.001), with the full model explaining 49.9% of the variance, a larger effect than reported in prior studies examining EI and leadership at the total score level. This stronger association probably arises because the present model isolated a single interpersonal facet, managing others’ emotions, rather than relating a global EI score to leadership, and facet-level predictors typically track relational leadership more closely than composite scores [24]. This is theoretically coherent, because AL dimensions such as relational transparency and balanced processing require the ability to read and respond to others’ emotional states. GSE, PSS, and the remaining EI facets predicted neither outcome, indicating that the predictive power of EI resides in specific facets rather than in global scores. By contrast, the stress model accounted for only 16.1% of the variance, leaving most of the variation unexplained. This is consistent with stress being multiply determined. Unmeasured factors such as academic workload, financial strain, sleep quality, social support, and placement characteristics plausibly account for much of the remainder, while the portion unexplained in the leadership model points to dispositional, motivational, and prior leadership-experience factors. These findings also position AL within the wider literature on leadership in learning environments. Unlike ethical leadership, which centres on normative conduct, AL emphasises self-awareness and relational transparency. Unlike emotionally intelligent leadership, it treats emotional competence as one contributor to genuine self-expression rather than its defining feature. It also contrasts with destructive leadership climates, in which impression management and guarded communication predominate, a contrast consistent with relational transparency being the AL facet most closely linked to lower stress here. Because the design is cross-sectional, these distinctions are interpreted as conceptual positioning rather than evidence of causal precedence.

Female students reported significantly higher EI and AL scores than males, consistent with a broader pattern observed across healthcare education. Multiple studies using the SSEIT have reported higher EI among female nursing students [36,41], although some evidence suggests that male students who self-select into nursing may exhibit comparable or higher EI when scores are normalized [42]. In the Saudi educational context, where gender-segregated clinical environments may shape interpersonal skill development differently, these findings may reflect both socialization patterns and self-selection effects. The association between higher GPA and higher EI and AL scores suggests that academic success and psychosocial competencies are positively related, consistent with a recent study that found that EI predicted academic achievement through positive psychological characteristics [43].

4.1. Implications

These findings carry practical implications for radiography education, and each recommendation is aligned with the strength of the corresponding association. Because managing own emotions was the only facet associated with lower perceived stress, the most strongly supported recommendation is to embed structured self-regulation practice, through cognitive reappraisal exercises, brief mindfulness-based techniques, and reflective debriefing after clinical sessions, at the points in the clinical rotation where stress exposure is highest. Because managing others’ emotions was the dominant predictor of authentic leadership, simulation-based exercises targeting interpersonal emotion management and relational transparency, together with leadership development activities during clinical placements, are indicated for students moving toward supervisory roles. Faculty development that prepares clinical supervisors to model and coach these competencies would support both aims. Recommendations extending beyond the measured associations, such as claims about long-term career outcomes, are not warranted by the present cross-sectional data and are therefore not made.

4.2. Limitations

Several limitations should be considered. The cross-sectional design measures all variables at one time point and therefore cannot establish the temporal ordering or directionality of the observed associations. Recruitment by convenience sampling from a single institution limits generalizability to other radiography programmes and to allied health students more broadly, and reduces statistical power for detecting small effects. Because all four constructs were measured with self-report instruments administered in a single survey, common-method variance may have inflated some associations, and social desirability may have affected self-ratings of leadership and emotional intelligence, for which students may overrate themselves. The 61.3% response rate leaves open the possibility of response bias if participants differed systematically from non-respondents. The GSE scale was selected for its theoretical alignment with broad coping appraisals and cross-study comparability, but its generalized nature may have attenuated associations with clinical-specific outcomes. Future research should employ longitudinal, multi-site designs with larger samples, domain-specific self-efficacy measures tailored to radiography practice, and mediation analyses to clarify the pathways among these constructs.

5. Conclusions

This study demonstrated that EI is significantly associated with AL, self-efficacy, and PSS among radiography students in clinical practice. The key contribution lies in identifying the differential roles of specific EI facets. Managing own emotions was associated with lower PSS, whereas managing others’ emotions was associated with higher AL. These findings extend the evidence base from medical education to an under-represented allied health population and suggest that EI functions as a multidimensional competency with distinct facet-specific pathways rather than a single global trait. Radiography programs should consider embedding structured, facet-specific EI training within clinical curricula as a potential approach to support stress management and leadership development.