Preliminary cross-cultural adaptation and validation of the organizational commitment questionnaire (OCQ-18) in a Tunisian nursing context

Abstract

A few validated instruments exist in Arabic to assess Organizational commitment, which limits research and practical applications in Arabic-speaking contexts. This study aimed to translate, culturally adapt and validate the Organizational Commitment Questionnaire developed by Allen, Meyer and Smith (1993) for use in Arabic-speaking nurses. A cross-sectional study was conducted among nurses at Sahloul University Hospital in Sousse, Tunisia using a convenience sampling method. The main inclusion criterion was primarily having more than 6 months of work experience. The Organizational Commitment Questionnaire was translated following the guidelines of Vallerand (1989). Content validity was ensured by expert review, construct validity was examined through exploratory factor analysis, and reliability was assessed using Cronbach’s alpha coefficients and intraclass correlation coefficients to evaluate internal consistency and stability. A total of 180 participants completed the Arabic OCQ. The Arabic scale showed a strong Pearson correlation with the original English version (r = 0.696, p < 0.01). EFA supported the three-factor structure (affective, continuance, and normative commitment) consistent with Allen and Meyer’s model. The Kaiser–Meyer–Olkin index was 0.806, and Bartlett’s test was significant (p < 0.001). Internal consistency was acceptable across subscales (Cronbach’s α = 0.63, 0.67, and 0.74), and the intraclass correlation coefficients indicated excellent reliability of a single administration (ICC mean = 0.941). The Arabic version showed adequate psychometric properties, supporting its validity and reliability for assessing organizational commitment among Arabic-speaking nurses and enabling cross-cultural research in occupational and health psychology.

1. Introduction

Organizational commitment (OC) reflects the psychological attachment between employees and their organization and is consistently associated with job satisfaction, performance, absenteeism, turnover intention, and employee well-being. Measuring OC accurately is therefore essential for both research and organizational practice [1].

In the human, social, and health sciences, researchers frequently rely on standardized measurement instruments, such as self-administered questionnaires. These tools allow researchers to objectify complex phenomena, such as attitudes, perceptions, and psychological states, particularly to enable comparison across cultural contexts [2].

In 1990, Meyer and Allen’s three-component model conceptualized OC as comprising affective commitment (AC) is the emotional attachment to the organization, continuance commitment (CC), which is the perceived costs associated with leaving the organization, and normative commitment (NC) is the sense of obligation to remain. The Organizational Commitment Questionnaire-18 Items (OCQ-18) has since been translated into several cultural contexts and occupational sectors, allowed staff to better understand how retention, job satisfaction, and work performance [3,4].

The development of a questionnaire is a lengthy and strict procedure, often requiring several months. Therefore, using previously validated instruments is recommended. Despite its widespread use, the availability of validated Arabic instruments measuring OC remains limited. In many Arab countries, measurement scales are often administered in English or French., which may introduce linguistic, cultural, and conceptual biases. Such practices can compromise the validity of findings and restrict participation to individuals proficient in foreign languages. Cross-cultural validation should thus be viewed as a rigorous process in its own right [5,6].

Cross-cultural adaptation of psychometric instruments requires more than literal translation. It involves ensuring semantic, conceptual, and functional equivalence between the source and target versions. International Test Commission Guidelines for Translating and Adapting Tests emphasize the importance of systematic translation, expert review, and empirical validation to maintain measurement validity across cultures [7].

Moreover, researchers frequently avoid translation into Arabic, because of its linguistic complexity and variations across countries. Although, some authors were tried a cross-cultural validation in Arabic. Their number remains extremely limited. This scarcity affected the field of work psychology and, particularly OC [8–11].

Given the absence of a rigorously validated Arabic version of the OCQ-18, this study seeks to address this gap by translating, adapting, and validating the questionnaire among healthcare professionals in Tunisia. Establishing the psychometric properties of the Arabic OCQ-18 will provide researchers and practitioners with a reliable tool for assessing OC in Arabic-speaking contexts [12]. Thus, the aim of the present study is to translate Allen and Meyer’s (OCQ-18) from English into Arabic and to evaluate its validity and internal consistency.

2. Materials and methods

2.1. Study design and setting

A cross-cultural validation study was conducted at Sahloul University Hospital, a tertiary public hospital in Sousse, Tunisia, and first one accredited, on August 26, 2023 [13].

2.1.1. Participants and sampling

This study included healthcare professionals employed at the hospital. The inclusion criteria for the study participants were Healthcare providers who had been employed at the hospital for more than six months and who expressed willingness to participate. A conveniently sampling method was used.

2.1.2. Instrument

2.1.2.1. Organizational commitment questionnaire (OCQ) for 18 items.

The OCQ-18, composed of 18 items, divided equally into three subscales: affective, continuance, and normative commitment. Responses are rated on a 7-point Likert scale ranging from 1 (strongly disagree) to 7 (strongly agree). The OCQ-18 was selected as the most widely used instrument, with 6-item subscales demonstrating strong correlations [14]. It was administered to healthcare providers and took approximately seven minutes to complete. Written consent was obtained from all participants prior to questionnaire administration. Negatively worded items were reverse-coded, with higher scores indicating higher levels of OC. Domain Scoring: (Low level: 1 to 3; Moderate level: 3 to 5; High level: 5 to 7).

2.2. Translation and cultural adaptation

2.2.1. Translation of the scale

1. Authorization from the Original Author

   The original author of the OCQ-18 Allen, J. was contacted and informed about the study. Permission was obtained.

2. Translation Process

   The translation followed Vallerand’s cross-cultural validation methodology [15].

   • a) Forward Translation

     A rigorous translation–back-translation procedure was implemented to ensure the conceptual equivalence of the scale. An initial Arabic translation was produced by a certified bilingual translator whose native language is Arabic.

   • b) Back-Translation

     The first Arabic version was then back-translated into English by a certified second traductor. The version back-translated was assessed with the original instrument and to detect any remaining inconsistencies by the comparison committee (the research team and linguistic experts). This iterative process resulted in a preliminary version of the scale demonstrating satisfactory equivalence with the original measure. This finding was carried out as follows. 67 participants were selected based on two criteria: being a caregiver and having a basic knowledge of English. The participants completed the scale in original version and then in Arabic preliminary version one month later.

   • c) Comparison of the Source and Back-Translated Versions

     Discrepancies were discussed by the research team and linguistic experts to ensure semantic, idiomatic, technical, experiential, and conceptual equivalence.

     • Semantic equivalence: ensuring that the target language words convey the same core meaning and nuances as the source language.

     • Idiomatic equivalence: replacing culture-specific expressions with semantics and stylistics appropriate equivalents.

     • Technical equivalence: where possible, retaining the same grammatical frameworks across the two languages.

     • Experiential equivalence: adapting culturally with locally meaningful ones.

     • Conceptual equivalence: ensuring that the underlying concepts are perceived similarly in both cultures.

After these steps, the back-translated and verified version was reviewed by 3 researchers in health sciences and a linguistic expert from faculty of literature. The result is an experimental Arabic literature version of the OCQ-18, the subject of linguistic validation.

2.3. Cultural adaptation

1. Content Validity

   Content validity ensures the clarity, comprehensibility, and representativeness of the construct being measured. It also examines the ability of items to discriminate between responses and the logical flow of questions [16]. Content validity was assessed through a pre-test and a Content Validity Index (CVI).

   • a) Pre-test

     The pre-test assessed whether the items in the experimental Modern Standard Arabic version were explicit, comprehensible to the target population, and clearly formulated. Also, it evaluated the response rate, participant reactions and confusion caused by items. It noted the difficulties related to questionnaire design and administration time [17].

     From two departments, 67 healthcare providers were involved. They refined the final version based on their feedback and identified comprehension issues. The pretest participants' comments and observations were taken into consideration during the debriefing meetings. These discussions allowed us to identify potential difficulties in understanding the items and to make the necessary linguistic adjustments to improve clarity and comprehensibility while preserving the original meaning of the items, which were subsequently reviewed by an expert panel for further assessment.

   • b) Expert review

     The revised experimental version was submitted to experts along with an evaluation. The panel included four experts: two medical specialists, and two doctoral students in health sciences. Each item was rated on a 4-point scale: not relevant (1), slightly relevant (2), quite relevant (3), relevant (4). A CVI was then calculated by dividing the number of items rated 3 or 4 by the total number of items. A CVI ≥ 0.9 was considered satisfactory [16]. Expert comments and suggestions were incorporated into the revision.

2. Reliability (Internal Consistency)

   Reliability refers to the reproducibility of a measurement. the original English version was administered first to 73 participants. A coding procedure was implemented to identify and match participants across the two administrations while preserving the confidentiality of their responses. Thus, the correlation between the English and Arabic versions was calculated using the paired observations from the same participants.

   Cronbach’s alpha coefficient was calculated twice, with a one-month interval, both for the overall questionnaire and its subdomains. Regarding the one-month interval between the two administrations, this interval was deliberately selected to reduce the potential recall or memory effect. This decision caused us to lose 6 participants, who left the hospital to go abroad. No relevant organizational changes were identified during this period. Cronbach’s alpha assesses internal consistency, the degree to which items measuring the same construct yield similar responses. A value over 0.7 is generally considered acceptable [18,19].

3. Construct Validity

   Construct validity was assessed using multiple methods to confirm that the questionnaire measured the intended construct [19]. Factor analysis was conducted to evaluate construct validity.

2.4. Data collection procedure

After receiving ethical approval from the Ethics Committee of Sahloul University Hospital. Participants received written information explaining the study objectives, voluntary participation, confidentiality, and anonymity. Written informed consent was obtained from all participants prior to questionnaire administration. Participants were informed of their rights: to decline participation, and to skip any questions. The questionnaires were collected on the same day as administration.

2.5. Data analysis

Data were analyzed using SPSS version 27. Descriptive statistics were computed for sociodemographic variables. Construct validity was assessed using exploratory factor analysis (principal axis factoring with an oblique promax rotation). Reliability was evaluated using Cronbach’s alpha and intraclass correlation coefficients (ICC). Statistical significance was set at p < .05.

2.5.1. Ethical considerations

Ethical approval was obtained from the Ethics Committee of Sahloul University Hospital. Authorization was granted by department heads. Participants received detailed study information in advance and reminded of their right: voluntary participation, confidentiality, anonymity and to withdraw or skip any question. Written consent was obtained from all participants prior to questionnaire administration. All data were anonymized and stored on secure, in accordance with institutional confidentiality requirements.

3. Results

3.1. Sociodemographic

As shown in Table 1, 60.6% of our population, is between [30-40] years old, with a female predominance of 78.9%, corresponding to a sex ratio (M/F) of 0.26.

Table 1.

Socio-demographic representation.

		Frequency	Percentage
Age	<30	48	26,7
	[30-40]	109	60,6
	>40	23	12,8
Gender	Male	38	21.1
	Female	142	78,9
Profession	Nurse	180	100
Years of experience	6 month_5 years	104	57,8
	5,1_ 10 years	36	20,0
	>10,1	40	22,2

3.2. Content validity

3.2.1. Test retest

The correlation demonstrated a strong positive association (Pearson r = .696; p < .01) between the two versions, are presented in Table 2. This result shows that responses in English were highly consistent with those in Arabic, supporting the validity of the linguistic adaptation. With comparable sample sizes (N = 73 for the English version; N = 67 for the Arabic version), the results suggest that the translated items measure the same constructs across both languages.

Table 2.

Concurrent validity.

		Score _ Arabic	Score _ English
Score _ Arabic	Pearson Correlation	1	,696**
	Sig. (bilateral)		,000
	N	67	67
Score _ English	Pearson Correlation	,696**	1
	Sig. (bilateral)	,000
	N	67	73

**. P< 0.01.

3.2.2. Expert review

The CVI was calculated and yielded a value of .91, indicating satisfactory content validity. The expert panel aimed to maintain the greatest possible similarity to the original questionnaire. Each item was carefully reworded to align with its corresponding item in the original instrument. Minor modifications were applied, primarily to items containing terms considered inappropriate in terms of semantics, phrasing, or relevance.

3.2.3. Reliability (internal consistency)

3.2.3.1. Cross cultural comparisons.

Table 3 presents descriptive statistics, internal consistency coefficients (Cronbach’s α), and intercorrelations among the three scales (AC, CC, and NC) across four cultural samples.

Table 3.

Cross cultural comparisons.

Sample	n	AC			CC			NC			Inter-correlations among three scales
		M	SD	α	M	SD	α	M	SD	α	rac	rNC	raN
Chinese	299	4.69	1.2	.78	3.71	0.89	.52	4.24	1.07	.75	.36	.48	.64
Canadian	603	3.91	1.47	.82	4.03	1.39	.74	3.04	1.41	.83	−.05	.11	.63
South Korea	227	4.21	1.28	.86	4.15	0.94	−.61	3.81	1.0	.74	−.03	.17	.58
Arabic	180	4.07	.49	.63	2.52	1.18	.67	2.99	.87	.74	−.14	.61	.16

Note: P < 0.01.

In the Arabic sample, all three scales demonstrated acceptable internal consistency (α = .63 to .74). The intercorrelations were lower with AC correlated with NC (r = .16), and negative observed between AC and CC (r = −.14) and acceptable between NC and CC (r = .61). Compared with the Arabic sample, the Chinese sample exhibited stronger associations among the scales. Specifically, the correlation between ACS and NC was higher in the Chinese sample (r = .64) than in the Arabic sample (r = .16). In contrast, the South Korea sample showed weaker associations involving CC relative to the Arabic sample. While the AC-NC correlation remained strong and comparable (r = .58 vs. r = .16), between CC - NC (r = .17 vs. r = .61) and the correlations between CC -AC (r = −.03 vs. r = −.14) were substantially lower, indicating minimal overlap between CC and the other constructs in the South Korean context.

3.2.3.2. Intra-class correlation coefficient.

The ICC reported in Table 4 provide a comprehensive assessment of the reliability of the measurement procedure. For single measurements, the ICC was estimated at 0728, with a (95%.CI [0.679–0.775], p < .001) suggesting a moderate level of agreement across observations. For the average of measurements, the ICC increased to .941 (95% CI [0.927–0.954], p < .001) indicating a clear improvement in measurement consistency when multiple observations were combined.

Table 4.

Intraclass Correlation Coefficient.

	Intraclass Correlationb	95% Confidence Interval		F Test with True Value 0
		Lower Bound	Upper Bound	Value	df1	df2	Sig
Single Measures	,728a	,679	,775	17,461	179	895	,000
Average Measures	,941c	,927	,954	17,461	179	895	,000

Two-way mixed-effects model where people effects are random and measures effects are fixed.

^aThe estimator is the same whether the interaction effect is present or not.

^bType A Intraclass correlation coefficients using an absolute agreement definition.

^cThis estimate is computed assuming the interaction effect is absent, because it is not estimable otherwise.

3.3. Construct validity

3.3.1. The Kaiser–Meyer–Olkin index (KMO) and Bartlett’s test

Prior to conducting the principal component analysis, the suitability of the data for factor analysis was assessed. (KMO) measure of sampling adequacy was .806, indicating a good level of factorability. In addition, Bartlett’s test of sphericity was statistically significant, χ² (153) = 1495.65, p < .001. See Table 5 for an overview.

Table 5.

KMO and Bartlett’s test.

Kaiser–meyer–olkin measure of sampling adequacy		,806
Bartlett’s test of sphericity	Approx. Chi-Square	1495,650
	df	153
	Significance	,000

3.3.2. Exploratory factor analysis

The suitability of the data for factor analysis was confirmed by a KMO measure of 0.806 and a significant Bartlett’s test of sphericity (χ², p < .001). Exploratory factor analysis using principal axis factoring with an oblique Promax rotation revealed a three-factor solution consistent with the theoretical model.

The three factors had eigenvalues greater than 1 and jointly explained a substantial proportion of the total variance. Factor 1 AC included items 1–6, with loadings ranging from 0,647 to 0.984. Factor 2 CC included items 7–12, with loadings between 0.533 and 0.922. Factor 3 NC included items 13–18, with loadings ranging from 0.533 to 0.827. No items were removed, as all factor loadings exceeded the recommended threshold of 0.40. Show Table 6.1.

Table 6.1.

Exploratory Factorial Analysis.

	Factor
	1	2	3	Uniquity
EE_1		0.768		0.344
EE_2		0.859		0.254
EE_3		0.723		0.450
EE_4		0.854		0.217
EE_5		0.984		0.134
EE_6		0.647		0.381
EC_7			0.533	0.576
EC_8			0.551	0.432
EC_9			0.743	0.398
EC_10			0.827	0.287
EC_11			0.690	0.395
EC_12			0.660	0.521
EN_13	0.531			0.542
EN_14	0.920			0.164
EN_15	0.922			0.241
EN_16	0.761			0.323
EN_17	0.424			0.478
EN_18	0.860			0.228

The exploratory factor analysis supported a three-factor solution, which explained 61.9% of the total variance. Factor 1 accounted for 21.27%, Factor 2 for 18.17%, and Factor 3 for 22.35% of the variance. The cumulative proportion of explained variance increased from 21.3% after the first factor to 39.5% after the second factor, reaching 61.9% with the three-factor solution. Overall, these results indicate that the retained factors captured a substantial proportion of the variance in the 18-item scale, supporting the adequacy of the three-dimensional structure. Show Table 6.2.

Table 6.2.

Factor Statics.

Factor	Sum of Squared Laodings	% of variance	Cumulative %
1	3.83	21.27	21.3
2	3.27	18.17	39.5
3	4.03	22.35	61.9

3.3.3. Confirmatory factor analysis

The confirmatory factor analysis (CFA) was conducted to examine the factorial structure of the Arabic version of the OCQ-18 and to assess the fit of the hypothesized three-factor model, comprising affective, continuance, and normative commitment.

As shown in Table 7.1, all 18 items contributed significantly to their respective latent factors (all p < .001). The parameter estimates ranged from 0.798 to 1.633 for the affective commitment dimension, from 0.560 to 1.693 for continuance commitment, and from 0.751 to 1.672 for normative commitment. All corresponding Z-values were statistically significant, ranging from 5.95 to 13.77 for affective commitment, from 3.84 to 12.44 for continuance commitment, and from 4.43 to 13.56 for normative commitment. Moreover, the 95% confidence intervals for all parameter estimates did not include zero, providing further evidence that each item made a statistically significant contribution to its respective latent construct.

Table 7.1.

Confirmatory analysis factor.

Factors Contributions
				95% Confidence Interval
Factor	Indicator	Estimate	Standard error	Lower Bound	Upper Bound	Z	p
Affective	EE_1	0.857	0.144	0.575	1.139	5.95	<.001
	EE_2	0.798	0.131	0.542	1.055	6.10	<.001
	EE_3	1.392	0.130	1.137	1.647	10.69	<.001
	EE_4	1.633	0.119	1.401	1.866	13.77	<.001
	EE_5	1.481	0.123	1.241	1.722	12.08	<.001
	EE_6	1.055	0.150	0.761	1.350	7.02	<.001
Continuance	EC_7	1.137	0.149	0.845	1.428	7.65	<.001
	EC_8	1.530	0.141	1.254	1.807	10.84	<.001
	EC_9	1.693	0.136	1.427	1.960	12.44	<.001
	EC_10	1.477	0.130	1.222	1.731	11.38	<.001
	EC_11	0.560	0.146	0.274	0.846	3.84	<.001
	EC_12	0.758	0.148	0.468	1.047	5.13	<.001
Normative	EN_13	0.751	0.169	0.419	1.083	4.43	<.001
	EN_14	1.605	0.127	1.355	1.854	12.59	<.001
	EN_15	1.595	0.127	1.345	1.845	12.51	<.001
	EN_16	1.440	0.130	1.186	1.695	11.10	<.001
	EN_17	1.672	0.123	1.430	1.914	13.56	<.001
	EN_18	1.281	0.129	1.028	1.533	9.96	<.001

Note: All indicators show positive and statistically significant estimates, with p < .001.

Regarding the overall model fit (Table 7.2), the chi-square test was statistically significant, χ²(132) = 311, p < .001. However, given the sensitivity of the chi-square statistic to sample size and model complexity, additional fit indices were considered. The Comparative Fit Index (CFI = 0.954) and Tucker–Lewis Index (TLI = 0.948) indicated a good fit of the three-factor model. The RMSEA was 0.087, with a 90% confidence interval ranging from 0.075 to 0.100, suggesting an acceptable but relatively borderline level of approximation error.

Table 7.2.

Goodness-of-fit indices for the three-factor measurement model.

			Fit Measures
Exact fit test						IC90% of RMSEA
χ2	df	p	CFI	TLI	RMSE	Lower Borne	Upper Borne
311	132	<.001	0.954	0.948	0.087	0.075	0.100

Overall, the results provide support for the hypothesized three-factor structure of the OCQ-18, with all items showing statistically significant relationships with their respective dimensions. Although the RMSEA indicates that the model fit is not optimal, the high CFI and TLI values suggest that the proposed three-factor model demonstrates an overall satisfactory fit to the data.

			Fit Measures
Exact fit test						IC90% of RMSEA
χ2	df	p	CFI	TLI	RMSE	Lower Borne	Upper Borne
311	132	<.001	0.954	0.948	0.087	0.075	0.100

4. Discussion

This study provides psychometric evidence supporting the validity and reliability of the OCQ-18 among Tunisian healthcare providers. The findings support the applicability of the three-component model of organizational commitment (OC), comprising affective, continuance, and normative commitment, in this specific healthcare context. By examining the instrument in a North African and Arabic-speaking population, the present study contributes to the growing body of research on the cross-cultural applicability of organizational commitment measures [20].

The exploratory and confirmatory factor analytic results support the multidimensional nature of OC. The use of an oblique Promax rotation was theoretically and methodologically appropriate because the three commitment dimensions are conceptually related while remaining empirically distinct. This finding reinforces the theoretical proposition that organizational commitment reflects different, yet related, psychological bonds linking individuals to their organization [21–23].

The present factorial structure differs from that reported in some previous adaptations. For example, the French adaptation identified four subdimensions, distinguishing affective and normative commitment from two components related to perceived sacrifice and limited alternatives. Such differences may reflect variations in the adaptation procedures, sample characteristics, and cultural or occupational contexts. Therefore, cross-cultural differences in factor structures should not necessarily be interpreted as evidence that one version is inherently superior to another; rather, they highlight the importance of conducting comprehensive psychometric validation within each target population.

The CFA provided additional support for the hypothesized three-factor structure. All 18 items were significantly associated with their respective latent factors (p < .001), and their confidence intervals did not include zero. However, some variability was observed in the parameter estimates, particularly within the continuance and normative dimensions, suggesting that items may not contribute equally to their respective constructs. This finding is consistent with previous research indicating that the psychometric performance of organizational commitment dimensions may vary across healthcare and cultural contexts [24,25].

The overall model fit showed a mixed but generally satisfactory pattern. Although the chi-square statistic was significant, χ²(132) = 311, p < .001, the CFI (0.954) and TLI (0.948) indicated good comparative fit. In contrast, the RMSEA of 0.087 (90% CI [0.075, 0.100]) suggested that some model misfit remained. Thus, the three-factor model should be interpreted as providing satisfactory rather than excellent fit. These results are more favorable than those reported by Agegnehu et al. (2022) among Ethiopian healthcare professionals, where the original three-factor model did not achieve acceptable fit and required modification. Together, these findings suggest that cultural, linguistic, and occupational contexts may influence the empirical expression of organizational commitment dimensions [24,25].

From a theoretical perspective, the confirmation of the three dimensions in this sample supports the view that organizational attachment cannot be reduced to a single motivational process. Instead, healthcare providers' relationships with their organizations appear to involve emotional attachment, a sense of obligation, and perceived costs associated with leaving. However, the relative importance of these components may vary according to contextual conditions.

Beyond the factorial structure, the present findings highlight the potential importance of continuance commitment in economically and structurally constrained work environments. In contexts characterized by limited employment opportunities or restricted professional mobility, employees may remain in an organization partly because of the perceived costs associated with leaving. Similar patterns have been reported in non-Western and resource-constrained contexts [26,27]. Nevertheless, continuance commitment should not be equated with positive engagement, as it may differ substantially from affective commitment in its associations with well-being, job satisfaction, and organizational outcomes [23,28].

The reliability analyses provide further support for the use of the OCQ-18 in this setting. The intraclass correlation coefficients indicated moderate reliability for single measurements and higher reliability for aggregated measurements, consistent with measurement theory indicating that aggregation may reduce random measurement error and improve score stability [23,28]. These findings support the use of the instrument for assessing organizational commitment while also suggesting that repeated measurements may be useful in longitudinal healthcare research.

Finally, the predominance of female participants reflects the composition of the healthcare workforce in the study setting. Although this may limit the generalizability of the findings to populations with different gender distributions, it enhances the ecological relevance of the results for the target healthcare context. Future studies should nevertheless examine whether the factorial structure of the OCQ-18 remains stable across gender, professional categories, and different healthcare institutions.

Overall, the present study provides substantial, although qualified, support for the three-factor structure and reliability of the OCQ-18 among Tunisian healthcare providers. The significant item–factor relationships and strong CFI and TLI values support the proposed structure, whereas the RMSEA indicates that further replication in larger and independent samples is warranted. Future research should also examine convergent and discriminant validity, composite reliability, and measurement invariance across relevant demographic and professional groups.

Taken together, these findings support the use of the OCQ-18 as a relevant instrument for assessing organizational commitment in Tunisian healthcare settings. More broadly, the results emphasize the importance of evaluating psychometric instruments within their specific linguistic, cultural, and occupational contexts rather than assuming that a factorial structure established in one population will necessarily be reproduced identically in another.

5. Limitations and future research

Several limitations should be acknowledged when interpreting the present findings. First, the use of a convenience sample drawn from a single public may limit the external validity of the results. Previous research has emphasized that organizational commitment may vary according to institutional characteristics, employment conditions, and healthcare system organization, suggesting the need for multi-site and multi-sector samples to enhance generalizability [25,29]. Future studies should therefore replicate these findings in different healthcare settings, including private institutions and home health care services.

Then, the cross-sectional design prevents conclusions regarding causal relationships or the temporal stability of organizational commitment. Prior longitudinal studies have shown that commitment levels may fluctuate over time in response to organizational changes, workload, and leadership practices [30]. Longitudinal designs would allow examination of the predictive validity of the Arabic OCQ-18 with respect to outcomes such as turnover intention, burnout, and quality of care. We also acknowledged the absence of several complementary psychometric analyses as a limitation and emphasized the need for future studies to examine additional aspects of validity and measurement invariance using larger and independent samples.

6. Implications for policy and practice

The present findings have important implications for healthcare policy and management practices. Research consistently indicates that not all forms of organizational commitment are equally beneficial. Affective commitment has been shown to be the strongest predictor of desirable outcomes, including employee well-being, job performance, retention, and quality of care, whereas continuance commitment is often weakly associated with these outcomes [22,31] In contexts characterized by limited labor mobility, high levels of continuance commitment may therefore reflect structural constraints rather than genuine engagement.

Healthcare managers and policymakers should consequently prioritize strategies that foster affective commitment, such as supportive leadership, recognition practices, opportunities for professional development, and fair organizational procedures. Empirical studies in healthcare settings demonstrate that such practices are associated with higher affective commitment and improved staff outcomes [32–34] The validated Arabic version of the OCQ-18 provides a reliable instrument for identifying distinct commitment profiles and monitoring workforce engagement over time, thereby supporting evidence-based human resource management.

Moreover, the reliability findings underscore the value of repeated assessment strategies. Measurement research suggests that aggregated or longitudinal assessments enhance reliability and improve the accuracy of organizational diagnostics, particularly in high-stress environments such as healthcare [24,25]. Regular monitoring of organizational commitment can thus inform organizational reforms and workforce policies aimed at sustainability and care quality.

Finally, given the predominantly female composition of the healthcare workforce, policies promoting work–life balance, flexible scheduling, and equitable career advancement are particularly relevant. Prior research indicates that such policies contribute to stronger affective commitment and lower turnover among healthcare professionals [35,36]. Integrating organizational commitment assessment into healthcare management frameworks may therefore serve as a strategic lever for improving both workforce stability and patient care outcomes.

7. Conclusion

This study provides acceptable evidence supporting the validity and reliability of the Arabic version of the OCQ-18 among nurses in Tunisia. The findings confirm the three-component structure of OC and demonstrate that the instrument is psychometrically adequate and culturally appropriate for Arabic-speaking nurses’ contexts. By addressing a critical gap in available measurement tools, this study enhances methodological resources for occupational and health psychology research in the Arab world and facilitates meaningful cross-cultural comparisons. Beyond its psychometric contribution, the study highlights the importance of interpreting OC within its cultural and structural context, particularly the distinction between affective and continuance forms of attachment. The Arabic OCQ-18 thus represents a valuable instrument for both research and practice, supporting a more nuanced understanding of workforce engagement in healthcare organizations.

Supplementary Material

Supplementary Material

Arabic OCQ.docx

Disclosure statement

No potential conflict of interest was reported by the author(s).

Funding

No financial support was provided to the authors for the research, writing, or publication of this work.

Supplementary material

Supplemental data for this article can be accessed at https://doi.org/10.1080/19932820.2026.2743457.