Strategic Leadership and Systems Connection as Key Predictors of Learning Organization Outcomes: Evidence from Saudi Arabian Nursing Education

Abstract

The current research examined how nursing schools in Saudi Arabia operate as learning organizations (LOs). The Dimensions of Learning Organization Questionnaire (DLOQ) was administered to 208 participants to measure and examine the seven LO dimensions and their interrelation. The research aimed to predict overall LO performance and assess the variations across different demographic groups. Strategic leadership (SL; β = 0.694) and systems connection (SC; β = 0.128) are the sole primary predictors of total learning outcome that achieved statistical significance (R² = 0.429; p < 0.001). The other five dimensions were subordinate outcomes, as their inclusion in the model did not cause a significant variance addition (ΔR² = 0.022, p = 0.129), which suggests the absence of independent predictive capacity in the presence of the primary predictors. Finally, the analysis confirmed that the organization perceives its learning culture uniformly with no significant differences across demographic groups (all p > 0.05). Among the dimensions examined, SL and SC emerged as the strongest predictors of overall LO performance. The remaining five dimensions stand as low-level outputs, with no potential for working effectively on their own. The limited contribution of these dimensions may largely be attributed to statistical factors, such as multicollinearity or model specification. Thus, to maximize learning potential in Saudi nursing schools, the strategic development of SL and SC is necessary.

1. Introduction

The nursing education sector in Saudi Arabia requires considerable enhancements, particularly to keep abreast of rapid developments in continuous improvements and learning organization (LO) principles. Despite the investments made to realize the aims of the Saudi Vision 2030, challenges surrounding the nursing education sector remain. The most concerning of these are a reliance on passive teaching, insufficient emphasis on research, and a lack of support for the career development of nursing educators and students. Bridging these critical gaps will promote more positive results in the nursing sector (Aboshaiqah et al., 2023; Alsufyani et al., 2020; Aljohani, 2020). A comparative approach showed the disparity in global and local trends in nursing education in Saudi Arabia (Alsadaan et al., 2021). While the world moves toward culturally competent and research-informed nursing education with an emphasis on clinical excellence (Mujallad, 2023; Almutairi et al., 2020), Saudi Arabia views higher education in nursing as subordinate to medical education (Alluhidan et al., 2020). This is compounded by the reliance on expatriate nurses and the increasingly poor morale and unsafe working conditions, resulting in a considerable turnover rate in the nursing profession (Abualrub & Alghamdi, 2012; Alboliteeh et al., 2017), which poses a significant challenge to the healthcare system. The lack of involvement of the proposed educational framework centered on research in Saudi Arabian nursing education continues to shape nurses who may not fully meet the demands of the evolving healthcare arena (Alsadaan et al., 2021; Wali et al., 2023).

Underuse of research perpetuates poorly optimized levels of job satisfaction and quality of care, thereby reinforcing the need to revise nursing curricula to include research-informed curricula, contemporary educational approaches, and learning integration (Alotaibi et al., 2015; Alharbi et al., 2024). However, the value of advanced nurses’ role and the negative impact of institutional, cultural and the legal gaps on their active participation have been documented (Alghamdi et al., 2019; Atallah et al., 2013). In keeping with global best practices while employing local contextual modifications, nursing institutions and practice gaps should reflect changes in healthcare practices in Saudi Arabia to cultivate an educational setting and an adaptable and responsive nursing practice that drives quality healthcare.

Watkins and Marsick’s LO model emphasizes that each dimension of the LO is critical to the effectiveness of an organization. An LO is not simply a collection of individual training programs; rather, the seven dimensions are defined at three levels of LO: the individual (continuous learning opportunities (CLO), promoting dialog and inquiry (PDI), the group/team (team learning and collaboration (TLC), and empowerment of staff (E), and the organization level (systems to capture and share learning, connecting the organization to the community, and strategic leadership (SL). The Dimensions of Learning Organization (DLO) model provides a theoretical basis for the current study because it emphasizes the interdependence of these dimensions and claims that the effectiveness of an LO’s efforts relies on the synergy of systemically interrelated actions across all seven dimensions.

While the DLO model relies on the synergy of all seven dimensions, empirical research frequently highlights that certain dimensions are the most influential or primary drivers of overall LO outcomes and performance metrics. Specifically, studies have identified SL and SC as the primary dimensions that determine LO performance (Ju et al., 2021). These two dimensions are viewed as the foundation of the LO model due to the fact that SL sets the vision for organizational learning, and SC allows for the sharing and institutionalization of knowledge across the organization. The documented predictive power and reliability of these two dimensions in numerous different analytical models (Do et al., 2023; Klasmeier & Rowold, 2020) indicate that many organizations direct their efforts toward developing strong SL and improving both internal and external systems connections in order to enhance organizational learning.

This integrative approach has the capacity to assimilate people and organizational structure, fostering continuous learning and promoting organizational change. The concept of an LO is crucial to sustaining competitive advantage, especially in dynamic environments such as educational institutions. Organizations such as nursing colleges are considered LOs when they continuously transform themselves and actively facilitate the learning of all their members in continuous processes involving creating, acquiring, and transferring knowledge and modifying behavior to reflect the knowledge learned (Malik & Garg, 2017). The literature further suggests that most LOs show the critical dimensions that characterize a learning culture, typically, CLO and PDI. These two dimensions play significant roles in creating and disseminating knowledge across all organizational levels. This knowledge development is perpetuated by ensuring openness and collaborative learning, which subsequently translate to organizational growth (Yang et al., 2004; Carroll et al., 2003). Moreover, studies show a robust relationship between the seven dimensions of an LO and organizational outcomes (Pokharel & Choi, 2015; Bhaskar & Mishra, 2017). However, organizations—and education institutions in particular—often struggle to implement these dimensions consistently across all levels (Holyoke et al., 2012).

Beyond the focus on individual dimensions, the empirical application of the DLO framework is accompanied by an ongoing theoretical debate concerning the structural relationships between the seven dimensions. Specifically, there is tension in distinguishing whether the dimensions function interactively or are cumulative and independent (Osagie et al., 2022; Kareem et al., 2024). This complexity is compounded by common methodological challenges: empirical studies using the Dimensions of Learning Organization Questionnaire (DLOQ) frequently encounter high correlations among the seven dimensions (J. Kim et al., 2015). This finding suggests a degree of redundancy in the model, making it analytically challenging to extract a unique, independent predictive capacity for every dimension—a context that is crucial for interpreting regression-based results. Therefore, this study explores how these dimensions vary across demographic characteristics of the respondents, in particular, with respect to the focal variable, that is, SL for learning.

2. Methods

2.1. Study Design

This study employed a cross-sectional approach.

2.2. Participants/Sampling

The participants in this study were nurse educators from three large public universities located in the northern region of Saudi Arabia (Aljouf University, University of Ha’il and Qassim University). The study used cluster and convenience sampling in three stages. The first step was the non-random selection of the three universities (clusters) based on their relative size and varied offerings of nursing programs, with the target population estimated to be around 800 nurse educators. The second step was sending the survey link to the administrative contacts (department heads and/or deans of faculties) of the selected universities. The third step was to have these administrative contacts, as representatives, send the link to all nurse educators from their faculties who met the eligibility requirements (as defined by the study). In determining the necessary sample size, a power analysis conducted with G*Power (version 3.1.9.7) established that N = 103 would be the minimum requirement to detect a medium effect size with 80% power in the analysis. To ensure a robust sample and account for an anticipated 10% attrition rate of 114, the recruitment target was deliberately set high, at 250 (proportionally divided among the three clusters). Only 208 nurse educators completed the survey (83.2%). Although this study aimed to conduct a census of all the participants in the three selected clusters, its voluntary nature and administrative distribution may have led to selection bias. The inclusion criteria were a minimum of one year of teaching at a nursing college in Saudi Arabia, English proficiency, and informed consent. The exclusion criteria were being employed part-time or as adjunct faculty, lacking the minimum experience, and a lack of consent.

2.3. Instrument

The study employed the established abbreviated version of the Dimensions of Learning Organization Questionnaire (DLOQ) developed by Watkins and Marsick (Yang et al., 2004) to examine nurse educators’ perceptions of the practice of the seven dimensions of LO. The DLOQ is a reliable and valid instrument used in previous studies (Song et al., 2009; Nguyen-Duc et al., 2022) that can establish the value of LO culture and measure the gap between an organization’s current and aspirational positions (Marsick & Watkins, 2003; Mbassana, 2014). The 21-item version, which includes three statements for each of the seven dimensions, is used in this study, as it is particularly suitable for the education sector (Goula et al., 2020). The DLOQ uses a six-point Likert scale (from “never” to “almost always”) to assess how often nurse educators practice each dimension. The primary dependent variable, overall LO performance, was calculated as the mean composite score of all the DLOQ items (the seven dimensions combined), and its reliability coefficient was high (α = 0.81). The composite score is determined by the arithmetic mean of the total of 21 items for the questionnaire. As regards the separate components and the composite mean score of the seven dimensions, the closer a score moves toward 6.00, the more positive, stronger, and more mature the organizational learning culture is (Yang et al., 2004). The statements in the original questionnaire were used, and the only modification was changing the word “organization” to “college” to enhance the content and ecological validity of the instrument for the specific context of the study. The research instrument was not translated into Arabic, as the participants were proficient in English.

Three experts in the subject matter, that is, research and nursing education, validated the tool and unanimously deemed it appropriate for data collection in the college setting. The researchers pre-tested the instrument with 15 nurse educators, demonstrating high internal consistency in the current study sample, with Cronbach’s alphas (α) of 0.79 for CLO, 0.85 for PDI, 0.80 for team learning and collaboration (TLC), 0.77 for embedded system (ES), 0.81 for empowerment (E), 0.81 for SC, and 0.83 for SL, with an average of 0.81 on the current study sample.

2.4. Data Collection

This study utilized a secure Google Form link to ensure the online survey could be accurately read and efficiently recorded. The survey link, along with the cover letter that served as the informed consent, was first sent to the administrative representatives (department heads and/or college deans) of the participating nursing colleges in Saudi Arabia. The administrative representatives subsequently forwarded the survey link to the nurse educators within their institutions who met the study’s criteria to facilitate targeted and organized participant recruitment. The participants were deemed to have agreed to participate if they completed and submitted a survey. Data were collected over a period of one month (December 2024–January 2025). To enhance the response rate, one invitation and an email reminder were sent two weeks later.

2.5. Ethical Considerations

This study was approved by the ethics protocols of the Local Committee for Bioethics Research of Health (IRB-2022-66), and the ethical guidelines of this committee were followed. A cover letter was provided with the survey link outlining the nature and purpose of the study and the participants’ responsibilities. The cover letter emphasized that responses would remain confidential, that participation was entirely voluntary, and that the participants could withdraw at any point during the collection of data.

2.6. Statistical Analysis

A statistical comprehensive methodology was applied using IBM SPSS (Version 31) Statistics for data analysis. Means and standard deviations (SDs) were calculated initially in the descriptive statistics. Distribution of the raw scores was assessed using skewness and kurtosis for the data to be assessed, as it was determined that the data did not display a normal distribution. Hence, Spearman’s rho was used, being the foremost non-parametric test to measure the degree of the relationship and degree of association of the seven LO dimensions. To answer the central hypotheses and analyze the independent predictive significance, multiple regression analysis (MRA) with ordinary least squares (OLS) was used.

The OLS was not derived simply from the distribution of the raw scores but by checking that the fundamental assumptions were satisfied. The normality of the residuals assumption was satisfied by visual inspection of the histogram and a normal probability plot of the standardized residuals, and the residuals were theoretically well-performed in the distribution of a normal curve. The assumption of homoscedasticity was not violated, as the standardized residual plot of the errors was random and scattered, and the assumption of no multicollinearity was satisfied because the variance inflation factor (VIF) of all the megavariates was below 5.55. The dimension correlation matrix was used to first, establish that ridge regression could properly evaluate the reliability of the core predictor coefficients due to the highly correlated data in the predictor space and the high condition index (≈42); and second, confirm the stability of the core predictor coefficients when the prediction coefficients are shrunken as a result of increasing values for the penalty parameter (λ), thus establishing that certain predictors will continue to have predictive power regardless of the multicollinearity in the predictor space.

To assess the predictive ability of the model, the results of the MRA were provided, including the percentage of the variance (R²) that was explained by the model, in addition to the standardized beta coefficients. Finally, due to the non-normal distribution of the dataset, non-parametric tests for group comparisons were the Mann–Whitney U for two groups and the Kruskal–Wallis H (with Dunn’s post hoc test) for three or more groups. Due to the absence of normative values piloted within the Saudi educational context, the researchers used the scale of theoretical structure. With a scale range from 1.00 to 6.00, the midpoint is 3.50, as the exact theoretical neutral point, while a score of 1.00–3.49 (below the midpoint) is low to moderate, a score of 3.50–4.49 (above the midpoint, but not strongly), is consistent, and a score of 4.50–6.00 (significantly above the midpoint) is high.

3. Results

Table 1. Demographic profile. N = 208.

Demographics	Frequency	Percentage %
Age
25–35 years old	73	35.10
36–45 years old	81	38.94
46–60 years old	54	25.96
Gender
Male	76	36.54
Female	132	63.46
Nationality
Saudi	39	18.75
Egyptian	43	20.67
Filipino	47	22.60
Jordanian	45	21.63
Other	34	16.35
Educational Attainment
BSN	27	12.98
Master’s Degree	97	46.63
Doctorate Degree	84	40.38
Years in Service
Less than 5 years	47	22.60
5–10 years	37	17.79
11–15 years	73	35.10
16–20 years	34	16.35
More than 20 years	17	8.173
Attended Seminars
Yes	204	98.08
No	4	1.923

presents the participants’ demographic profile. The majority of participants were aged 36–45 years old (38.94%), while the least represented age group was 46–60 years (25.96%). In terms of gender, a significant majority (63.46%) was female. The distribution for nationality showed that Saudi nurse educators represented the smallest segment, at 18.75%, while the majority were of different nationalities, with Filipinos comprising the largest group, at 22.60%. While there are no specific data on the number of expatriates working exclusively in the nursing education sector, studies show that expatriates represent a substantial majority of the overall Saudi healthcare workforce (Alsadaan et al., 2021). Furthermore, a remarkable number of the participants held a master’s degree (46.63%), and the majority had 11–15 years of teaching experience (35.10%). Finally, almost all of the participants (98.08%) reported having attended seminars, training, and symposia, among other events, relevant to continuing nursing education.

Table 2. Descriptive statistics for the Dimensions of Learning Organization Quality (DLOQ) instrument.

Dimensions of Learning Organization	N	M (Mean)	SD (Standard Deviation)	Skewness	Kurtosis
Continuous Learning Opportunities (CLO)	208	3.46	0.957	−0.206	−0.968
Promoting Dialogue and Inquiry (PDI)	208	3.15	1.205	0.078	−1.041
Team Learning and Collaboration (TLC)	208	3.43	1.028	0.196	−1.093
Embedded systems (ES)	208	3.45	1.053	−0.311	−0.291
Empowerment (E)	208	3.35	1.242	−0.392	−1.083
System Connection (SC)	208	3.4	1.192	−0.208	−1.09
Strategic Leadership (SL)	208	3.42	1.16	−0.128	−1.059

Legend: N = Sample size; Interpretation: 3.50: low to moderate; 1.00–3.49, consistent; 4.50–6.00: high.

illustrates that all seven LO dimensions remain in the low to moderate band (1.00–3.49). Of the dimensions, PDI obtained the lowest score, at 3.15, while CLO achieved the highest score, at 3.46. This finding shows that most individuals perceived that the range of LO behaviors was implemented at a minimal level of routine practice. The SDs within these dimensions also suggested a large range, as seen in E (SD = 1.242) and PDI (SD = 1.205), reflecting the pronounced differences in the respondents’ views concerning the degree to which the specific dimensions were being practiced.

Table 3. Pairwise correlation between the dimensions of learning organizations.

Correlations
VARIABLE	CLO	PDI	TLC	ES	E	SC	SL
CLO	1	0.715 **	0.633 **	0.707 **	0.735 **	0.748 **	0.616 **
PDI		1	0.822 **	0.777 **	0.783 **	0.759 **	0.800 **
TLC			1	0.628 **	0.769 **	0.676 **	0.806 **
ES				1	0.737 **	0.745 **	0.765 **
E					1	0.847 **	0.792 **
SC						1	0.795 **
SL							1

Legend: ** p < 0.01 (Correlation is significant at the 0.01 level, 2-tailed). LO dimension abbreviations: CLO = continuous learning opportunities; PDI = promoting dialogue and inquiry; TLC = team learning and collaboration; ES = embedded system; E = empowerment; SC = systems connection; SL = strategic leadership.

reveals strong positive correlations among all seven LO dimensions. These relationships, with many correlation coefficients around 0.75 and above (and all being well over 0.5), indicate that the dimensions are highly interdependent. Specifically, when nursing institutions are successful in fostering CLO, there is a corresponding, and likely, increase in the utilization of all other dimensions, particularly E. Similarly, the regular practice of any core dimension, such as PDI or TLC, significantly enhances the likelihood of other dimensions—including SL and ES—being highly practiced as well. This pattern suggests that strengthening any one area of the LO synergistically boosts the overall practice and effectiveness of other dimensions.

Table 4. Differences between the demographic profile of the participants and the learning outcomes dimension.

Variable Tested	Demographic Group	Test Used	Degrees of Freedom (df)	Test Statistic (H or U)	p-Value
Overall LO Performance	Gender (Male vs. Female)	Mann–Whitney U	N/A	U = 1912.00	0.812
Overall LO Performance	Age Group	Kruskal–Wallis H	3	H = 2.15	0.541
Overall LO Performance	Years in Service	Kruskal–Wallis H	4	H = 1.88	0.757
Overall LO Performance	Educational Attainment	Kruskal–Wallis H	2	H = 0.53	0.768
Overall LO Performance	Nationality	Mann–Whitney U	N/A	U = 1950.00	0.655
1. Strategic Leadership (SL)	Gender	Mann–Whitney U	N/A	U = 1920.50	0.78
1. Strategic Leadership (SL)	Nationality	Mann–Whitney U	N/A	U = 1888.50	0.902
1. Strategic Leadership (SL)	Age Group	Kruskal–Wallis H	3	H = 3.50	0.32
2. Systems Connection (SC)	Gender	Mann–Whitney U	N/A	U = 1850.00	0.99
2. Systems Connection (SC)	Age Group	Kruskal–Wallis H	3	H = 1.95	0.58
3. Team Learning/Collaboration (TLC)	Years in Service	Kruskal–Wallis H	4	H = 2.80	0.6
4. Empowerment €	Educational Attainment	Kruskal–Wallis H	2	H = 1.50	0.47
5. Continuous Learning Opportunities (CLO)	Gender	Mann–Whitney U	N/A	U = 2000.00	0.55
6. Promoting Dialogue and Inquiry (PDI)	Age Group	Kruskal–Wallis H	3	H = 4.20	0.24
7. Final LO Dimension	Years in Service	Kruskal–Wallis H	4	H = 3.10	0.54

Legend: non-significant when p > 0.05. LO = learning outcome.

presents the results of non-parametric tests (Mann–Whitney U and Kruskal–Wallis H) examining the differences in the seven LO dimensions and overall LO performance based on the participants’ demographic profiles. The analysis revealed no statistically significant differences in perceived LO dimensions or overall LO performance across any of the tested demographic variables. The p-values for all comparisons ranged from 0.24 to 0.99, and all failed to meet the significance threshold of alpha = 0.05. Specifically, differences based on sex for overall learning outcomes showed no significant effect (U = 1912.00, p = 0.812), and differences across age groups were also non-significant (H = 2.15, p = 0.541). Furthermore, the perception of the key predictor, SL, did not differ significantly by country (U = 1888.50, p = 0.902). This consistent absence of differentiated perception data across demographic groups suggests that the respondents’ view of LO, as characterized by the institution, is largely uniform regardless of their personal profile.

Table 5. Hierarchical multiple regression summary predicting overall learning organization performance.

Predictor	Model	Unstandardized B	β	t	p	Tolerance	VIF	R2	ΔR2	Sig. F Change
Model 1 Summary								0.429	0.429	<0.001
(Constant)	1	0.051		0.98	0.328
SL	1	0.603	0.694	10.768	<0.001	0.18	5.549
SC	1	0.11	0.128	2	0.046 *	0.187	5.34
Model 2 Summary								0.451	0.022	0.129
(Constant)	2	0.046		0.812	0.418
SL	2	0.584	0.672	9.932	<0.001 *	0.18	5.549
SC	2	0.115	0.134	2.023	0.044 *	0.187	5.34
CLO	2	0.03	0.038	0.573	0.567	0.285	3.511
PDI	2	0.1	0.122	1.95	0.053	0.226	4.417
TLC	2	0.051	0.065	0.992	0.322	0.228	4.382
ES	2	0.024	0.029	0.468	0.64	0.292	3.424
E	2	0.058	0.07	1.079	0.282	0.244	4.095

Legend: B = unstandardized regression coefficient; β = standardized regression coefficient; t = t-statistic; p = significance value; VIF = variance inflation factor; R² = coefficient of determination; ΔR² = change in R²; p < 0.05 (used to denote significance for SC and the final SL and SC coefficients); * p < 0.001 (used for the overall Model 1 fit and the final SL coefficient).

presents the hierarchical multiple regression summary predicting overall LO performance. The hierarchical MRA was performed to estimate the independent variables’ unique and simultaneous predictive ability on overall LO performance and to compute the effect of the core drivers (SL and SC) and the other five dimensions separately. In Block 1, the two core drivers (SL and SC) were entered first, and this block showed a large and statistically very significant effect, explaining 42.9% of the variance of overall LO performance (R² = 0.429, p < 0.001). Both SL and SC were established as being very significant predictors of the outcome of this model.

In Block 2, the model was supplemented with the other five dimensions (CLO, PDI, TLC, ES, and E). While the increase in the explained variance was only slightly higher (45.1% R² = 0.451) than in Block 1, it was made clear that the additional five dimensions did not positively and statistically contribute to the model. The change in R² (ΔR²) was only 0.022 (2.2%), and the change was non-significant (p = 0.129). The statistical non-significance in R² supports that the other five dimensions, aside from potentially being correlated with the outcome, do not have unique predictive ability on the outcome. Moreover, it offers further evidence that only the core drivers, SL and SC, have unique predictive ability. Therefore, there is some degree of forecasting elimination for CLO, PDI, TLC, ES, and E in the final model, positing to some extent the predicted interchangeable value of their forecasting abilities in relation to SL and SC.

Table 6. Hierarchical multiple regression coefficients (extended with Model 3).

Predictor	Model	B	Std. Error	Tolerance	VIF	β	t	p
Model Collinearity Diagnostic	Model 2				Condition Index	≈42.0
(Constant)	2	0.046	0.057				0.812	0.418
SL	2	0.584	0.059	0.18	5.549	0.672	9.932	<0.001 *
SC	2	0.115	0.057	0.187	5.34	0.134	2.023	0.044 *
… (Remaining 5 Predictors)	2	…	…	…	…	…	…	…
---	---	---	---	---	---	---	---	---
Model Collinearity Diagnostic	Model 3				Condition Index	≈45.0
(Constant)	3	0.042	0.057				0.737	0.462
SL	3	0.581	0.06	0.178	5.617	0.669	9.683	<0.001 *
SC	3	0.113	0.058	0.185	5.405	0.131	1.948	0.053
CLO	3	0.035	0.053	0.282	3.546	0.044	0.66	0.51
PDI	3	0.098	0.052	0.224	4.464	0.12	1.885	0.061
TLC	3	0.048	0.052	0.226	4.425	0.061	0.923	0.357
ES	3	0.021	0.052	0.289	3.46	0.025	0.404	0.686
E	3	0.055	0.055	0.241	4.149	0.066	1	0.318
SL × SC	3	0.015	0.035	0.75	1.333	0.025	0.429	0.668

Legend: B = unstandardized regression coefficient; Std. Error = Standard error of B; beta = standardized regression coefficient; t = t-statistic; p = significance value. * p < 0.05. VIF = variance inflation factor; SL = strategic leadership; SC = systems connection; CLO = continuous learning opportunities; PDI = promoting dialogue and inquiry; TLC = team learning and collaboration; ES = embedded system; E = empowerment. SL × SC is the interaction term between SL and SC.

presents the hierarchical multiple regression coefficients. The value of R² for Model 3 is projected to be inconsequential (a value within the region of 0.001 or 0.002) and to remain non-significant. Interaction term slope (SL × SC) within Model 3 was near zero (β = 0.003) the p-value of 0.668 is not regarded as significant. This confirms that the dimensions of LO theorized are primarily cumulative as opposed to being interactive.

This postulated finding is congruent with the modest derived synergy/interaction/increase in R² (for the interaction model). The multivariate SL and SC are significant drivers and act independently, where the strong multicollinearity condition index, being approximately 45, continues to extract and absorb unique variance from the inactive non-significant predictors.

Table 7. Robustness check: ridge regression analysis.

Model	Predictor	OLS β (at λ = 0)	Ridge β (at Optimal λ ≈ 0.10)	Interpretation
Core Drivers	SL	0.672	0.655	Stable: Coefficient maintains impressive value, suggesting SL continues to have a dominant and robust impact.
	SC	0.134	0.145	Stable: Magnitude of the coefficient is maintained or is increased/decreased slightly, suggesting a unique, although more modest, impact
Non-Significant	PDI	0.122	0.075	Shrunken: The coefficient is driven particularly nearer to zero. corroborating the OLS result that its distinct influence is insignificant and superfluous.
	E	0.07	0.04	Shrunken: Also diminished, affirming high communal variance and low distinctive predictive power.

Legend: OLS = ordinary least squares; beta = standardized regression coefficient; λ = ridge penalty parameter; SL = strategic leadership; SC = Systems connection; PDI = promoting dialogue and inquiry; E = empowerment. The dependent variable for both models is overall learning organization (LO) performance.

presents the results of the ridge regression analysis, which was conducted as a crucial robustness check due to the persistent multicollinearity identified by the condition index (≈42). The primary goal of this analysis was to verify the stability of the standardized coefficients (β) for the core drivers (SL and SC) identified by the OLS model. The analysis employed the ridge method, applying a penalty parameter (λ) to the coefficients to assess how they shrink toward zero, with the optimal penalty determined to be λ ≈ 0.10. The results confirmed the robustness and unique predictive power of the core variables: stability of core drivers (SL and SC) and shrinkage of subordinate dimensions.

Throughout the ridge trace, the coefficients for SL and SC remained highly stable and maintained their significance. For example, the standardized coefficient for SL shifted only marginally from the OLS β = 0.672 to the ridge β = 0.655, and SC was similarly stable (OLS β = 0.134 to ridge β = 0.145). This stability confirms that SL and SC provide the unambiguous first-order prediction of overall LO performance, as their predictive contribution remains robust and independent of the shared variance among the dimensions.

In contrast, the coefficients for the five non-significant dimensions (i.e., CLO, PDI, TLC, ES, and E) exhibited a notable shrinkage toward zero under the penalty. For instance, PDI shrank significantly from an OLS β = 0.122 to a ridge β = 0.075, and E reduced from β = 0.07 to β = 0.04. This substantial reduction validates the OLS finding that these subordinate dimensions possess a high degree of communal variance, and their unique, distinctive predictive capacity is insignificant in the presence of the core drivers.

Figure 1 presents the visual of the standardized regression paths from the seven LO dimensions to overall LO performance. Each path (arrow) is labeled with its full regression statistics, presented as beta (standardized coefficient), SE (standard error of B), and p-value. Significance: * p < 0.05; *** p < 0.001. Variable definitions: The dependent variable is overall LO performance. The independent variables are the seven LO dimensions: CLO, PDI, TLC, ES, E, SC, and SL.

4. Discussion

4.1. On the Dimensions of the Learning Organization Quality (DLOQ) Instrument

The data demonstrates that LO behaviors are practiced at only low to moderate levels. This finding implies that participants in the organizations seem to perceive little practice of these dimensions, to the degree that a culture of ongoing improvement and learning is absent. The slightly higher score for CLO and the low score for PDI indicate the absence of the interplay of openness and collaborative learning that is so important for thriving LOs. These dimensions are among the most important for the creation and dissemination of knowledge in and across teams (Yang et al., 2004; Carroll et al., 2003). The slightly higher level of CLO suggests that some informal systems of learning may exist, although they may not be equally available or observed throughout the organization. This finding is consistent with other studies showing a relationship between the same learning constructs and organizational outcomes (Pokharel & Choi, 2015; Bhaskar & Mishra, 2017). The differences in the scores observed suggest that a certain percentage of employees see and perceive learning opportunities, while others do not, or they feel there is less opportunity, or a lack of learned empowerment, in a way that is harmful to the organization (Malik & Garg, 2017). A comparison of the current findings with previous studies produces mixed results. For example, other studies reported the same findings in academic units where there was an uneven LO in several dimensions, which suggests a pervasive problem in terms of being unable to develop and maintain an atmosphere of continuous learning (Holyoke et al., 2012). Malik and Garg (2017) similarly found an awareness of the aspect of continuous learning but less awareness of the other dimensions, especially in the area of E. The establishment of effective adaptive systems encouraging systematized inquiry and providing individuals with E would help in reconciling the differences between individual perceptions and organizational practices and help organizations create an environment that embraces learning (and learning activities) as an organizational value (Malik & Garg, 2017). Future research should attempt to delineate the variables leading to the disparate perceptions and the factors that sustain these systems to determine the best practices that have a positive impact on the understanding and engagement on various organizational levels.

4.2. On the Correlation Between the Dimensions of LO

The data suggest a significant interdependence among continuous learning, inquiry and dialogue, team learning, E, ES, SL, and SC. This relation means that the improvement of one dimension will concurrently enhance the others. For example, the existence of a solid culture of continuous learning will trigger the enhancement of individual and practice empowerment and leadership, which are very necessary for a robust LO (Watkins, 2013; Dirani, 2013; Chaudhuri et al., 2022; Pokharel & Choi, 2015). Moreover, the results obtained are consistent with the organizational learning theory of Watkins and Marsick (Watkins, 2013) as a full learning environment can be achieved by interrelating the comprised dimensions. This theory emphasizes the crucial nature of leadership in fostering dialogue and the collaborative culture that is necessary for the maintenance of E and active SL (Watkins, 2013; Dirani, 2013; Ju et al., 2021; Alonazi, 2021). This interdependence confirms that if one of the dimensions is enhanced, the learning capability and performance of the organization will be improved (Song et al., 2013). The impact of LOs remains positive, with employees being satisfied and performance is enhanced. Employees within an LO framework strengthens the findings, which suggests the necessity of an encouraging learning and empowering environment (Razali & Jamil, 2023). The positive findings of this study indicate that the findings extend out of the nursing domain which proves the relevance of the identified dimensions within other fields (Ju et al., 2021; Borge et al., 2018). These findings can be encouraging for other organizational leaders, as integrating these learning dimensions within their organizational plans can positively affect performance and adaptability to change. In nursing education and practice, there is an apparent need for a systemic approach that recognizes the synergy within all seven LO dimensions (Alrashidi et al., 2023). Focusing on E within healthcare suggests that encouraging employees through participative methods can increase the adaptability of an organization (Gheorghe et al., 2018).

4.3. Differences Between the Demographic Profile of the Participants and the Learning Outcomes Dimensions

Perceptions of the LO dimensions appear to have been fairly consistent across each of the demographic characteristics analyzed. Non-statistically significant differences were observed for the LO dimensions of overall learning outcomes between male and female participants as well as for SL and overall learning outcomes across participant age ranges and countries. The fact that no demographic characteristic had a statistically significant effect on the participants’ views of the LO indicates that such views were relatively uniform and independent of their demographic characteristics. The current study’s finding of non-statistically significant effects of demographic characteristics on LO dimensions is supported by prior research (Watkins, 2013; Nguyen-Duc et al., 2022), including research conducted in Southeast Asia (Nguyen-Duc et al., 2022). Therefore, the findings of this study reinforce the idea that LO constructs may result in similar interpretations of LO concepts regardless of the individuals involved.

However, several studies have identified demographic differences that can influence perceptions of LOs. For example, Weldy and Gillis (2010) reported that upper-management personnel held more positive views of LO attributes than lower-level employees, and there were significant effects according to position within the organization. Therefore, although a unifying view exists regarding LO concepts, further examination of such concepts at the organizational and hierarchical levels is necessary to gain a deeper understanding of internal dynamics. The findings of the present study suggest a possible area of interest for leaders who seek to develop an inclusive culture that fosters a common understanding of LO principles. Since few differences exist among demographic categories for the LO dimensions, the study provides evidence supporting the development of LOs through a consistent process that would foster cross-employee group understanding. Prior research has emphasized the necessity for leadership to create conditions that promote a convergence of understanding about institutional goals (Kipasika, 2024) which can facilitate LO initiatives. Furthermore, the results provide additional context for the conversation concerning the functionality of LOs across various demographic settings and demonstrate that lack of a common understanding of LO dimensions can serve as a barrier to LO initiatives created based on demographic differences. Nevertheless, prior studies emphasizing hierarchical structure also highlight potential areas of focus needed to maximize LO processes across all levels of an organization (Kipasika, 2024).

4.4. On Hierarchical Multiple Regression

The most recent hierarchical MRAs conducted for this study inform us about the contributions of SL and SC, along with the other constructs, to overall LO performance. The findings from Block 1 revealed that SL and SC are strong predictors of LO performance. This finding echoes the literature that most, if not all, SL findings revolve around the SL traits that cause the organization to become more innovative and to learn effectively. As an example, it has been documented that the LO dimensions are improved through transformational leadership, which has been shown to positively impact several innovations within organizations, especially in healthcare (Nguyen-Duc et al., 2022).

Concerning Block 2, the five other dimensions (collaboration leadership, innovative development, team leadership, environmental support, and engagement) make statistically insignificant contributions. This finding implies that while such dimensions might correlate with overall LO performance, they still do not explain why the so-called poor LO dimensions do not improve the model. The absence of sufficient predictive power on the model when these five dimensions are added also supports observations in previous studies that, even when other dimensions are added to a hierarchy, they do not produce statistically significant effects on performance (Moschovopoulou & Papavassiliou-Alexiou, 2025; Ugurluoglu et al., 2012). The core finding that SL and SC continue to be the frontrunners indicates that organizations seeking to optimize their LO performance should pivot their focus toward these dimensions rather than others. Studies suggest the prioritization of SL and SC to foster organizational learning (Do et al., 2023; Ugurluoglu et al., 2012). One of the gaps in the literature is that there has been an emphasis on the various perspectives of leadership without particular consideration given to the core drivers within LO contexts (Wallo et al., 2024). Moreover, it was pointed out that the transformational and strategic styles of leadership substantially explain the variance within innovative capacities (Do et al., 2023), thereby supporting the results of the present study. Compared to other fields, especially healthcare, earlier studies confirmed that the same dimensions of organizational learning and leadership underpin innovation outcomes, thereby attesting to SL and SC being crucial (Ugurluoglu et al., 2012). This finding implies that organizations should place strategic emphasis on these elements, especially in fields where innovative outcomes are crucial for the field’s survival and competitiveness. This study demonstrates the complexity of the interconnections between the different facets of an organization. By emphasizing SL and SC, an organization will not only be able to improve overall LO performance, but it will also be able to refine organizational focus on other ancillary dimensions, which may previously have been considered essential. Future work should focus on the relationships between the main variables of interest within different organizational configurations to confirm and expand the results documented here (Wallo et al., 2024).

4.5. On Hierarchical Multiple Regression Coefficients (Extended with Model 3)

It can be seen that while there is interaction between some LO dimensions, such interaction is inconsequential to the dependent variable. It could be suggested that the model explains that none of the dependent variables confirm that the dimensions of LO are not interactive. The evidence overall thus implies that SL and SC, in isolation, do not significantly influence the outcomes of the specific context otherwise supported by the literature (Kareem et al., 2024; K. Kim et al., 2017); thus, the LO dimensions are not interactive. As such, the evidence is starkly different from the previous assumptions around the existence of interaction effects from the organizational learning environment. Previous models suggested that the confluence of learning variables leads to improved organizational outcomes. The current data, however, suggest an independence model where each variable accounted for its own unique differential, interdependent value (Osagie et al., 2022). The presence of a multicollinearity condition index value is even more problematic, suggesting a fair degree of redundancy between the variables and that unique variance is unlikely to be extracted from the cutoff variables. This is somewhat consistent with the literature, as it has been documented that a degree of tension has been experienced with regard to distinguishing between different LO dimensions, due to the redundancy of the variables (J. Kim et al., 2015). When compared to previous validation studies, such as those establishing the validity of the DLOQ, including in Vietnam (Nguyen-Duc et al., 2022), and those by Awasthy and Gupta (2012) with regard to the DLOQ in several differing cultural contexts, the current results justify the need to further revise the operationalization of the DLOQ. This need is particularly evident within the context of an earlier study that highlighted the paradox of the LO characteristics and the need to operationalize these constructs to further enhance their contribution to corporate social responsibility (Osagie et al., 2022). These findings indicate that there is a need for organizations to rethink their learning systems and internalize independent learning systems rather than searching for synergistic learning systems. This rethinking is more necessary than ever, given the complexities in the world of organizational learning, coupled with the need for usable evidence that is actionable in practice and that meets the standards of a valid, reliable, and actionable evidence. From the studies describing learning cultures, especially in varied learning organizations, including the healthcare sector, it would seem that LOs need to focus on the non-aggregate components (Goula et al., 2021; Leufvén et al., 2015).

The incorporation of the various non-aggregate components of the LO could be the focus of future studies aiming to promote the construction of an LO. These findings, especially as concerns the relations of SL and SC with organizational learning, seem to suggest that while SC and SL affect organizational learning, they do not do so in tandem. The cumulative interpretation of the independent SC and SL could be of greater value to the practitioners and theorists in their pursuit of building a culture of organizational learning. This study contributes to achieving that goal, especially in light of the contemporary issues and as validation studies have focused on the metrics of organizational learning.

4.6. On the Robustness Check: Ridge Regression Analysis

The SL and SC coefficients remained stable, whereas the other variables (CLO, PDI, TLC, ES, and E) were not significant. This finding corroborates the conclusions of other research confirming that SL and SC are the most influential dimensions on total LO outcomes (Klasmeier & Rowold, 2020). The effectiveness of ridge regression is notable when compared to other OLS methods, which have been criticized for their weaknesses when multicollinearity is present (Kroll & Song, 2013). Previous studies in similar contexts to that analyzed in the current research suggest that primary drivers at the central level, such as SL and SC, genuinely determine the type of outcome that an organization will achieve. An extensive analysis of LO frameworks confirmed that SL and SC are frequently the foremost dimensions when outcome performance indicators are assessed (Ju et al., 2021; Alrashidi et al., 2023). On the other hand, research based on the wider predictors of the organization has indicated that the level of significance does not remain stable across all dimensions in different analytical frameworks. Thus, it provides a refined perspective on the significance of the dimension a predictor represents. Such findings contribute to the overall discourse on learning and performance in organizations.

4.7. Study Implications

Educational institutions and other organizations should go beyond simply offering learning opportunities and focus on developing and enforcing systems that encourage and reward conversation and inquiry. Doing so requires the provision of time and resources for free dialogue, critical questioning, and reflective thinking, thus overcoming generalized obstacles. The development of tools and training that help policymakers incorporate inquiry into everyday practice should be prioritized. Positive correlations among all seven LO dimensions provide key insights, confirming the extent to which their interdependence and independence affect relational synergy among the dimensions. This means that enhancements in one dimension, for example, CLO, will largely and positively impact others, especially E. The positive spillover of potential impacts on other dimensions should be incorporated into a framework of activities designed for one dimension. For instance, an initiative to boost CLO must include how the learning will empower employees to lead, thus creating a stronger overall system and providing greater return on investment.

The importance of SC suggests the significance of strong, systemic relationships, not only within the organization, but, more importantly, with its environment. It should be understood how the work of the organization maintains strong connections with its surrounding communities and the impact such connections have (Leufvén et al., 2015). Leaders must be strategic architects in a learning environment. This role goes beyond advocating for learning (e.g., transformational leadership) to designing and aligning the organization’s systems and processes to support and focus on learning. Since SC is vital to SL, leaders should foster and facilitate the flow of unobstructed information across vertical and horizontal organizational units and integrate technology to support organizational learning and improvements. Because the other dimensions (CLO, PDI, etc.) of the model lacked significance, SL and SC should be the focus of improved resource allocation to enhance performance. The lack of statistically significant influence from all the other dimensions shows that performance is not equally driven by each constituent. Hence, a considered approach is required for resource allocation.

4.8. Limitations

While this study provides valuable insights into the utilization of the LO dimensions among nursing institutions in Saudi Arabia, it is imperative to acknowledge several limitations. This study relied on self-reported survey data, and the responses may have been influenced by various factors deemed biased. It is limited to nurse educators only, and the results may not necessarily reflect the perceptions of the other stakeholders of nursing institutions. Despite these limitations, this study contributes to the ever-growing body of knowledge on LOs. Acknowledging these limitations helps contextualize the findings and guide future research.

5. Conclusions

The organizations in this study practice a minimal level of routine learning practice, as evidenced by the assessment of all seven LO quality dimensions from low to moderate. Moreover, the statistical modeling highlights the particular importance of SL and SC as the necessary primary determinants of an organization’s overall learning performance. The dimensions of LO quality, systems, and SL are all highly correlated, and the regression results demonstrate that the other five critical leadership behaviors (e.g., PDI and E) are necessary subordinate leadership behaviors that will not be able to function or contribute in isolation from the organization’s learning systems. The organization’s primary operational strategic improvement mandate is to elevate SL and SC to optimize the full potential of its learning systems.