Self‐Perceived Competence and General Happiness Scores Among Acute‐Care Nurses in the United Arab Emirates: A Cross‐Sectional Study

ABSTRACT

Aim

To examine the association between scores on an investigator‐adapted self‐perceived nursing competence questionnaire and a five‐point modification of the Oxford Happiness Questionnaire among registered nurses in one acute‐care hospital in the United Arab Emirates.

Design

Cross‐sectional correlational survey.

Methods

Between 15 September and 15 November 2025, 200 direct‐care registered nurses completed an online survey. Self‐perceived competence was scored with a 21‐item investigator‐adapted questionnaire informed by Nurse Competence Scale domains; general happiness was scored with a 29‐item Oxford Happiness Questionnaire modified to use five response categories. Neither modified score is established as psychometrically equivalent to its source instrument. Exploratory factor analysis, correlation and covariate‐adjusted linear regression with HC3 robust inference were used.

Results

The adapted scores correlated modestly (r = 0.293, 95% CI 0.160–0.414); the adjusted competence‐score coefficient was B = 0.440 (95% CI 0.248–0.633). Removing a competence item about feeling valued and supported within the team yielded r = 0.290 and adjusted B = 0.447 (95% CI 0.248–0.645). Exploratory factor analyses indicated factorability but did not validate either adaptation: the competence items did not reproduce seven intended domains, and happiness items primarily separated by wording polarity.

Conclusion

Higher scores on the locally adapted self‐perceived competence questionnaire were associated with higher scores on the modified general happiness questionnaire. The modest, cross‐sectional, concurrent self‐report association does not establish construct equivalence, direction of effect or benefits from competence training.

Implications for the Profession and/or Patient Care

Support for professional capability may be considered as one component of a broader workforce strategy that addresses staffing, workload, leadership, recognition and psychological safety.

Impact

Earlier research examined competence and happiness among nurses in Iran; this study extends the question to a contemporary multicultural UAE acute‐care sample. Its modest association between adapted questionnaire scores is useful for hypothesis generation, subject to measurement and sampling limitations.

Reporting Method

The study is reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology checklist for cross‐sectional studies.

Patient or Public Contribution

No patients or members of the public contributed because this study examined registered nurses as professional participants; nurses contributed only as survey respondents.

1. Introduction

Nurses' well‐being is central to workforce sustainability and safe, high‐quality care within a profession facing a substantial global shortage (Boniol et al. 2022). Hospital nurses work in cognitively, emotionally and relationally demanding environments, and sustained well‐being depends on both personal resources and organizational conditions. Contemporary evidence links healthy work environments, supportive leadership, recognition, teamwork, mentorship and adequate staffing with more favourable nurse outcomes, while burnout is associated with poorer safety, quality and patient satisfaction (Bond et al. 2025; Lake et al. 2019; Li et al. 2024; Williams et al. 2024).

2. Background

These issues are salient in the United Arab Emirates (UAE), where hospitals depend on a culturally and linguistically diverse nursing workforce. Aqtash et al. (2017) used the original Nurse Competence Scale (NCS) to study self‐assessed competence among expatriate nurses in rural Abu Dhabi hospitals; their NCS scores should not be compared directly with scores from the locally adapted questionnaire used here. Other UAE studies have linked work engagement with job performance, critical care nurses' quality of life with safety perceptions and job satisfaction and work environment with intention to leave (Ahmed et al. 2023; Al Badi et al. 2023; Aljawarneh et al. 2025). Regional evidence also indicates that work‐life balance and organizational and leadership conditions are important correlates of nurses' happiness and well‐being (Afriana et al. 2026; Nashashiqi et al. 2026).

Happiness is related to, but not synonymous with, job satisfaction or happiness at work (Fisher 2010; Arulappan et al. 2021). The Oxford Happiness Questionnaire (OHQ) assesses general subjective happiness and psychological well‐being across life domains (Hills and Argyle 2002). Accordingly, the outcome in the present study is defined consistently as general happiness, not workplace‐specific happiness. This distinction is important because work conditions may influence general happiness but are not the only determinants of it (Wati et al. 2025).

Self‐determination theory proposes that experiences of competence, autonomy and relatedness support well‐being (Ryan and Deci 2000). In nursing, perceiving oneself as effective may strengthen confidence, meaning and professional contribution. The Job Demands‐Resources model further situates this possible association within the balance between job demands and resources (Demerouti et al. 2001; Bakker and Demerouti 2017; Broetje et al. 2020). These theories provide an interpretive framework; the present study did not directly measure psychological need satisfaction or job demands and resources.

Self‐perceived competence must be distinguished from objectively observed clinical competence and job performance. Self‐report scores reflect nurses' judgements of their own capabilities and may be influenced by confidence, mood, response style and cultural context. Madadizadeh et al. (2017) previously reported a positive association between clinical competency questionnaire responses and happiness among nurses in teaching hospitals in Kerman, Iran. The present study therefore extends an existing research question to a multicultural UAE acute‐care setting; it does not establish a previously unexamined association.

3. The Study

3.1. Aim and Hypothesis

The study aimed to examine the association between scores from a locally adapted self‐perceived nursing competence questionnaire and a modified measure of general happiness among registered nurses in one UAE acute‐care hospital. Demographic and work‐related correlates of happiness were explored secondarily. We hypothesized a positive association between the two questionnaire scores without assuming that either adapted score has the construct validity of its source instrument.

4. Methods

4.1. Design

A cross‐sectional correlational survey design was used. Reporting followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidance for cross‐sectional studies (von Elm et al. 2007).

4.2. Setting

The study was conducted between 15 September and 15 November 2025 in a large acute‐care hospital in Ajman, United Arab Emirates. The hospital provides emergency, medical‐surgical, critical care, maternity/labour, paediatric and neonatal services. The institution name is masked in this manuscript for double‐blind review; full identifying information is retained in the title‐page file.

4.3. Sampling and Recruitment

Registered nurses were eligible if they provided direct patient care, held current United Arab Emirates licensure and had at least 1 year of nursing experience in the country. Nurses working primarily in administrative, educational or research roles were excluded. Convenience sampling was used. The survey link was distributed to 300 eligible nurses; 200 submitted questionnaires, giving a response proportion of 66.7%. No submitted record was excluded from the primary analysis (Figure 1).

FIGURE 1.

Participant flow and analytic samples.

4.4. Instruments With Validity and Reliability

The outcome was the score on a five‐point modification of the Oxford Happiness Questionnaire; the principal explanatory variable was the total score on a 21‐item investigator‐adapted self‐perceived nursing competence questionnaire. Covariates were age group, gender, marital status, nationality group, years of professional experience and clinical area.

4.4.1. Self‐Perceived Nursing Competence

Competence was assessed with a 21‐item investigator‐adapted questionnaire informed by the seven conceptual domains of the original Nurse Competence Scale: Helping Role, Teaching‐Coaching, Diagnostic Functions, Managing Situations, Therapeutic Interventions, Ensuring Quality and Work Role (Meretoja et al. 2004). The original scale contains 73 items rated for competence level on a 0–100 visual analogue scale, with a separate four‐point frequency rating. The present 21 items were not a verbatim shortened subset of the 73 items. Instead, the study team used three locally worded indicators per conceptual domain to reduce respondent burden and emphasize acute‐care capability. Therefore, a removed‐item list is not applicable. All 21 items and the complete rotated exploratory loading matrix are provided in File S1, Table S1.

Each item was rated from 1 (strongly disagree) to 5 (strongly agree); total scores ranged from 21 to 105, with higher scores indicating greater agreement with self‐perceived capability and related statements. One item, ‘I feel valued and supported within my healthcare team,’ assesses perceived team support rather than nursing capability. ‘I feel prepared to manage stressful situations without compromising patient care’ mixes perceived capability with stress appraisal; ‘I take pride in upholding ethical nursing standards in my practice’ captures professional affect or attitude. These items may broaden or contaminate the intended competence construct. This instrument is neither a validated short form nor psychometrically equivalent to the original Nurse Competence Scale. The available records did not document an independent expert panel review, formal content‐validity index, cognitive interviewing or standalone pre‐study validation pilot. Internal consistency and exploratory structure in this sample are preliminary only.

4.4.2. General Happiness

General happiness was scored with a locally modified version of the 29‐item Oxford Happiness Questionnaire (Hills and Argyle 2002). The original questionnaire uses six response categories. The present survey used five categories from 1 (strongly disagree) to 5 (strongly agree), including a neutral midpoint, to harmonize the online questionnaire. This operational modification was not separately validated. The 12 negatively worded items (items 1, 5, 6, 10, 13, 14, 19, 23, 24, 27, 28 and 29) had been reverse‐coded in the analytic worksheet so that higher values represented greater happiness; scored columns were summed without a second reversal. Possible totals ranged from 29 to 145. Scores from this modified format cannot be assumed equivalent to the validated six‐point OHQ. The complete rotated exploratory loading matrix by item number and wording polarity is provided in File S1, Table S2.

4.4.3. Covariates and Data Harmonization

Age and experience were retained as categories rather than treated as equally spaced ordinal scores. Nationality was grouped as Emirati; Arab non‐Emirati (Arab League member‐state nationalities); Asian (South, Southeast or East Asian nationalities); or Western/Other (all remaining nationalities). One response containing two experience categories was assigned to the longest‐experience category. For two responses containing two clinical areas, the first listed area was retained consistently. These rules were applied before analysis.

4.5. Sample Size and Power

An a priori calculation was performed with G*Power version 3.1.9.7 (Heinrich Heine University Düsseldorf) using fixed‐model multiple regression, a medium effect size (f ² = 0.15), α = 0.05, 80% power and four core predictors; the minimum was 85 participants (Faul et al. 2009). The achieved sample of 200 also exceeded the approximately 143 participants required to detect a medium overall effect with 16 regression coefficients. Nevertheless, estimates for small nationality and clinical‐area categories remain imprecise.

4.6. Data Collection, Management and Common‐Method Safeguards

Eligible nurses were approached during non‐clinical times and invited to complete an English‐language online questionnaire. The information sheet described the study, participation requirements, voluntariness, the absence of employment consequences and the right to discontinue before questionnaire submission. Participants indicated electronic informed consent before entering the survey. Estimated completion time was 15–20 min. No direct personal identifiers were collected. A data dictionary documented variable coding and composite‐scoring conventions, and electronic data were stored on password‐protected devices accessible only to the research team. Anonymity and separation of demographic, competence and happiness sections may have reduced evaluation apprehension and response carryover. Nevertheless, both principal variables were self‐reported in the same survey at one time point, so common‐method variance cannot be excluded.

4.7. Missing and Invalid Responses

The raw dataset contained no blank demographic or scale‐item cells. However, item‐level auditing identified 11 ambiguous cells containing more than one selected response: five competence cells among five participants (5/4200 competence item cells; 0.12%) and six happiness cells among five participants (6/5800 happiness item cells; 0.10%). These invalid multiple selections were set to missing for scoring. The analytic file did not retain the survey platform's required‐response settings, so mandatory‐field configuration could not be verified. When at least 80% of scale items were valid, the total was prorated using that participant's mean across valid items; all 200 participants met this rule. Complete‐record analyses that excluded any participant with an invalid competence or happiness item (n = 192) were conducted as sensitivity checks.

4.8. Data Analysis

Data were initially managed and analysed in IBM SPSS Statistics version 28.0 (IBM Corp., Armonk, NY, USA). For this major revision, all reported analyses were independently reproduced from the item‐level Excel dataset using Python 3.12, NumPy 2.3.5, SciPy 1.17.0, pandas 2.2.3 and scikit‐learn 1.8.0. Frequencies and percentages summarized categorical variables; means, standard deviations, medians, observed ranges and 95% confidence intervals summarized scale scores. Cronbach's alpha quantified internal consistency after the same within‐person substitution used for primary scoring; percentile bootstrap 95% confidence intervals used 5000 participant resamples. Complete‐record alpha values were calculated as sensitivity estimates. Because both instruments were modified, post hoc maximum‐likelihood exploratory factor analyses with varimax rotation were conducted on complete scale records. Sampling adequacy was assessed with the Kaiser‐Meyer‐Olkin statistic and Bartlett's test of sphericity. Factor retention used the 95th‐percentile criterion from parallel analysis with 500 normally distributed random datasets. These analyses were exploratory and were not a substitute for independent validation.

Pearson's r estimated the primary association, with a Fisher r‐to‐z 95% confidence interval. Spearman's rho was a rank‐based sensitivity estimate. Multiple linear regression predicted the modified happiness score from the adapted competence score and all covariates. Age and experience were dummy‐coded. Reference categories were age 20–30 years, male, single, Asian, 1–2 years of experience and medical‐surgical area. Unstandardized B, HC3 robust standard errors, standardized β, 95% confidence intervals and p values are reported. HC3 inference was selected because residual testing indicated heteroscedasticity and mild non‐normality. Statistical significance was two‐sided p < 0.05; demographic and work‐area coefficients were exploratory and not adjusted for multiplicity. A prespecified‐for‐revision sensitivity analysis excluded the team‐support item and prorated the mean of the remaining 20 valid competence items to a 20‐item total; the same covariates and HC3 inference were retained.

Diagnostics included residual histograms and Q–Q plots, Shapiro–Wilk and Jarque‐Bera tests, residual skewness and kurtosis, the Breusch‐Pagan test, variance inflation factors (VIF), a quadratic competence term to assess linearity, leverage, studentized residuals and Cook's distance. Sensitivity models excluded observations with Cook's distance > 4/n. A one‐factor unrotated exploratory analysis across all competence and happiness items was used only as a limited statistical diagnostic of common‐method variance; it cannot prove the absence of method bias.

4.9. Ethical Considerations

The study protocol was approved by the Research Ethics Committee of RAK Medical and Health Sciences University (RAKMHSU/RES/HEC/2024‐25/9) in June 2025 and by the Ministry of Health and Prevention Research Ethics Committee (MOHAP/REC/2025/74‐2025‐PG‐N) in August 2025. Both approvals preceded the start of data collection on 15 September 2025. Participation was voluntary and unrelated to employment. All participants provided electronic informed consent before completing the online questionnaire; no direct personal identifiers were collected.

5. Results

5.1. Participants

The final sample comprised 200 registered nurses. Most were aged 30–40 years (51.0%), female (59.5%) and married (70.5%). Asian nurses formed the largest nationality group (48.5%), followed by Arab non‐Emirati nurses (38.5%). Emergency (24.0%) and medical‐surgical (22.0%) were the most frequent clinical areas (Table 1).

TABLE 1.

Participant characteristics (n = 200).

Variable	Category	n (%)
Age group (years)	20–30	55 (27.5)
	30–40	102 (51.0)
	40–50	43 (21.5)
Gender	Female	119 (59.5)
	Male	81 (40.5)
Marital status	Married	141 (70.5)
	Single	59 (29.5)
Nationality group	Emirati	16 (8.0)
	Arab (non‐Emirati)	77 (38.5)
	Asian	97 (48.5)
	Western/Other	10 (5.0)
Years of experience	1–2 years	25 (12.5)
	3–5 years	28 (14.0)
	6–10 years	83 (41.5)
	> 10 years	64 (32.0)
Clinical area	Emergency	48 (24.0)
	Medical‐surgical	44 (22.0)
	ICU/HDU	30 (15.0)
	Maternity/Labour	29 (14.5)
	Paediatric	28 (14.0)
	NICU	21 (10.5)

Abbreviations: HDU, high‐dependency unit; ICU, intensive care unit; NICU, neonatal intensive care unit.

5.2. Scale Distributions and Preliminary Psychometric Findings

The mean competence score was 87.6 (SD 8.2; 95% CI 86.4–88.7), and the mean happiness score was 99.4 (SD 9.9; 95% CI 98.0–100.8). Because both response formats were locally modified and no validated cut points exist for these versions, scores were not classified as low, moderate or high. Internal consistency after within‐person substitution was good for competence (α = 0.882, bootstrap 95% CI 0.848–0.907) and acceptable for happiness (α = 0.736, bootstrap 95% CI 0.675–0.779). Complete‐record estimates were α = 0.880 and α = 0.731, respectively (Table 2).

TABLE 2.

Descriptive and preliminary psychometric results.

Measure	n	Mean (SD)	Median (range)	α	KMO	Parallel analysis
Self‐perceived competence	200	87.6 (8.2)	86.0 (68–105)	0.882	0.850	2; 35.1%
General happiness	200	99.4 (9.9)	96.5 (76–127)	0.736	0.839	2; 29.8%

Note: Alpha values in the table use the primary within‐person substitution rule; complete‐record alpha values were 0.880 for competence and 0.731 for happiness. Kaiser‐Meyer‐Olkin and factor analyses used complete scale records (n = 195 for each scale). Bartlett's tests were significant for competence, χ ²(210) = 1446.19, p < 0.001 and happiness, χ ²(406) = 1895.99, p < 0.001. The final column shows the number of factors retained using the 95th‐percentile parallel‐analysis criterion and the variance explained by the rotated maximum‐likelihood solution.

Using the 95th‐percentile criterion, parallel analysis supported two‐factor solutions for both adapted instruments. The competence solution did not reproduce the seven intended NCS domains; items 1–8 predominantly loaded on one factor and items 9–21 on the second. In the modified happiness questionnaire, the two factors largely separated negatively and positively worded items, consistent with a wording‐polarity effect. KMO and Bartlett results indicate factorability of the observed data, not validation of either score, confirmation of an intended construct or equivalence to the original instruments. Total scores were retained for the primary exploratory association; domain‐level correlations were not interpreted. Complete loading matrices are provided in File S1.

5.3. Association Between Competence and Happiness

The locally adapted competence and modified happiness scores were positively associated (r = 0.293, 95% CI 0.160–0.414, p < 0.001), corresponding to 8.6% shared variance. The rank‐based result was similar (Spearman's ρ = 0.270, p < 0.001). Complete‐record analysis remained significant (n = 192, r = 0.252, p < 0.001).

5.4. Covariate‐Adjusted Regression

The full model had R ² = 0.253 and adjusted R ² = 0.188 (robust Wald F[16, 183] = 4.67, p < 0.001). The adapted competence score remained associated with the modified happiness score after covariate adjustment (B = 0.440, 95% CI 0.248–0.633; standardized β = 0.363; p < 0.001). This coefficient represents an estimated 0.44‐point difference in the modified happiness total per one‐point difference in the adapted competence total, conditional on model covariates. Nationality and clinical‐area contrasts were exploratory, unplanned, unadjusted for multiplicity and potentially unstable owing to small subgroups. They should not be interpreted as intrinsic group effects or used for comparative practice decisions (Table 3). Years of experience was not associated with the happiness score after categorical coding.

TABLE 3.

Multiple linear regression predicting general happiness (n = 200).

Predictor	B	HC3 SE	β	95% CI	p
Intercept	64.412	8.780		47.089–81.736	< 0.001
Self‐perceived competence score	0.440	0.098	0.363	0.248–0.633	< 0.001
Age 30–40 versus 20–30	−1.578	2.033	−0.080	−5.588 to 2.433	0.439
Age 40–50 versus 20–30	−4.218	2.792	−0.175	−9.725 to 1.290	0.133
Female versus male	2.517	1.769	0.125	−0.974 to 6.007	0.157
Married versus single	0.381	1.523	0.018	−2.624 to 3.387	0.803
Emirati versus Asian	6.962	3.110	0.191	0.826–13.097	0.026
Arab (non‐Emirati) versus Asian	−1.953	1.641	−0.096	−5.190 to 1.285	0.236
Western/Other versus Asian	−7.832	2.111	−0.173	−11.997 to −3.667	< 0.001
Experience 3–5 versus 1–2 years	−2.866	2.744	−0.101	−8.279 to 2.548	0.298
Experience 6–10 versus 1–2 years	2.339	3.002	0.117	−3.583 to 8.262	0.437
Experience > 10 versus 1–2 years	0.581	3.404	0.027	−6.135 to 7.296	0.865
Emergency versus medical‐surgical	−4.506	1.947	−0.195	−8.348 to −0.664	0.022
ICU/HDU versus medical‐surgical	−2.918	2.090	−0.105	−7.041 to 1.205	0.164
Maternity/Labour versus medical‐surgical	−4.917	2.805	−0.175	−10.451 to 0.618	0.081
Paediatric versus medical‐surgical	−8.281	2.026	−0.291	−12.278 to −4.285	< 0.001
NICU versus medical‐surgical	−4.076	2.895	−0.126	−9.789 to 1.636	0.161

Note: Reference categories: age 20–30 years, male, single, Asian, 1–2 years of experience and medical‐surgical area. B = unstandardized coefficient; β = standardized coefficient; HC3 = heteroscedasticity‐consistent standard error. R ² = 0.253; adjusted R ² = 0.188. Demographic and work‐area coefficients were exploratory and not multiplicity adjusted.

5.5. Regression Diagnostics and Sensitivity Analyses

The competence‐happiness score relation was approximately linear; adding a centred quadratic competence term was not significant (p = 0.920). Residuals showed mild positive skewness (0.30) and excess kurtosis (0.83); Shapiro–Wilk W = 0.974, p = 0.001 and Jarque‐Bera p = 0.019. The Breusch‐Pagan test indicated heteroscedasticity, χ ²(16) = 27.69, p = 0.034, supporting HC3 inference. Variance inflation factors ranged from 1.10 to 5.63. The maximum exceeded 5 for the > 10‐year experience indicator, indicating moderate overlap between age and experience, but none approached 10. Maximum Cook's distance was 0.054; 10 observations exceeded 4/n, three had leverage above 2 k/n (k = 17 model parameters) and three had absolute studentized residuals above 3. Excluding observations above the Cook's‐distance threshold yielded B = 0.440 (95% CI 0.287–0.593, p < 0.001). Complete‐record regression yielded n = 192, B = 0.375 (95% CI 0.182–0.569, p < 0.001). Removing the team‐support item from the competence score yielded a 20‐item total and left the association materially unchanged (n = 200; r = 0.290, p < 0.001; adjusted HC3 B = 0.447, 95% CI 0.248–0.645, standardized β = 0.357, p < 0.001). The corresponding complete‐record model also remained positive (n = 192; B = 0.380, 95% CI 0.181–0.579, p < 0.001). Coefficients per point across 20‐ and 21‐item totals are not strictly identical units; the similar correlations and standardized coefficients provide a scale‐independent comparison.

In the limited post hoc Harman single‐factor diagnostic, the first unrotated principal component across all 50 items accounted for 14.9% of variance after the primary within‐person substitution. This weak diagnostic did not indicate a dominant single component, but it cannot exclude common‐method bias and does not replace a longitudinal or multimethod design.

6. Discussion

6.1. Principal Findings

The principal result is a modest association between the scores on two locally modified, concurrently administered self‐report questionnaires (r = 0.293; 8.6% shared variance). It remained positive after measured covariate adjustment and after removing the team‐support item. This is consistent with, but does not test, theories relating perceived capability to well‐being. Happier nurses may rate their capabilities more favourably, and shared response tendencies or omitted organizational conditions may contribute to the association. These results should not be described as an established association between independently validated NCS and OHQ constructs.

The revised categorical analysis did not support the earlier claim that experience independently predicted happiness. This is important because treating age and experience categories as ordinal scores would have imposed equal spacing and produced an overly simple linear trend. Dummy coding showed that experience contrasts were imprecise and non‐significant after adjustment.

6.2. Interpretation in Relation to Theory and Contemporary Evidence

Self‐determination theory offers a possible interpretation: feeling effective may contribute to confidence, accomplishment and meaning (Ryan and Deci 2000). The Job Demands‐Resources model also places personal resources alongside workload and organizational demands (Bakker and Demerouti 2017). A prior Iranian study directly examined nurse clinical competency and happiness and reported a positive association (Madadizadeh et al. 2017). Its setting and questionnaires differ from the present UAE study, and neither cross‐sectional study establishes direction of effect. Aqtash et al. (2017) studied self‐assessed competence in UAE expatriate nurses using the validated original NCS, without making the present competence‐happiness comparison. Our locally adapted 21‐item score and modified OHQ score cannot be benchmarked against those instruments. Psychological need satisfaction and job demands and resources were not measured here.

The modest bivariate effect is consistent with a multifactorial view of nurse well‐being. Recent syntheses identify supportive leadership, professional development, recognition, teamwork, adequate staffing and resources, psychological support and healthy work environments as central organizational contributors to flourishing and retention (Afriana et al. 2026; Bond et al. 2025; Buckley et al. 2025; Williams et al. 2024). The present model omitted these variables. Competence‐focused initiatives therefore cannot be assumed to compensate for adverse workload, staffing or culture.

6.3. Measurement Considerations

The central methodological qualification is measurement. The 21 locally worded competence items are not a validated short form of the original 73‐item NCS, and the five‐point happiness format differs from the validated six‐point OHQ. Some competence statements capture perceived team support, stress appraisal or pride in professional conduct rather than capability alone; these could correlate with general happiness independently of competence and inflate or otherwise distort the total‐score association. Excluding the explicit team‐support item made little difference, but it does not resolve contamination from other items or establish construct validity. Internal consistency does not establish content or construct validity. The exploratory analyses showed factorability only: the competence solution did not reproduce seven intended domains, while the modified happiness solution primarily reflected item‐wording polarity. The observed association is therefore provisional and score‐specific. Independent content, cross‐cultural and structural validation is needed before benchmarking or practice use.

6.4. Exploratory Subgroup Findings and UAE Context

Some nationality and clinical‐area contrasts reached statistical significance. These findings were not hypothesized, were not adjusted for multiple testing and involved small reference or comparison groups. They may reflect unmeasured differences in work conditions, shift patterns, unit acuity, language, migration circumstances, cultural response styles or other contextual factors. They should not be interpreted as inherent differences between nationalities or specialties.

The UAE workforce context simultaneously increases the relevance and limits the transferability of the study. The multicultural sample reflects an important feature of UAE hospitals, but an English‐language self‐report measure may not function identically across linguistic and cultural groups. Findings from one Ajman hospital may not generalize to other emirates, public versus private facilities, primary care or health systems with different staffing models, licensure pathways and employment conditions.

6.5. Implications for Nursing Practice

The findings justify attention to nurses' perceived capability, not a claim that competence training will increase happiness. Nurse leaders may consider supportive onboarding, mentorship, protected learning time, formative feedback and recognition as components of a broader workforce strategy. These approaches should be evaluated prospectively and implemented alongside adequate staffing, manageable workload, psychologically safe leadership, fair recognition and access to mental‐health support. Contemporary reviews similarly favour organizational and multilevel approaches over placing responsibility for well‐being solely on individual nurses (Bond et al. 2025; Buckley et al. 2025; Williams et al. 2024).

6.6. Strengths and Limitations of the Work

Strengths include an important regional question, a sample exceeding the planned minimum, item‐level reanalysis, transparent handling of invalid multiple selections, dummy coding of categorical predictors, heteroscedasticity‐consistent regression inference, standardized and unstandardized estimates with confidence intervals, extensive diagnostics, sensitivity analyses and explicit separation of self‐perceived from objectively measured competence.

Limitations are substantial. The cross‐sectional design precludes temporal and causal inference. Both questionnaire totals were self‐reported concurrently, creating social‐desirability, mood‐congruence and common‐method risks; the post hoc single‐factor diagnostic cannot eliminate these biases. The adapted competence questionnaire lacks documented independent expert review, formal content‐validity assessment, cognitive interviewing and standalone validation, and contains wording that may capture contextual support or affect. The OHQ response format was modified, and its two‐factor pattern was predominantly wording‐related. Removing the clearest support item did not change the association materially, but does not validate the remaining total. Eleven ambiguous multiple selections required a scoring decision; complete‐record sensitivity results were similar. The single‐hospital convenience sample, 66.7% response proportion, English‐language administration, small subgroup cells, unplanned multiple comparisons and local recoding rules limit generalizability. Staffing, workload, leadership, recognition, compensation, scheduling, psychological safety and work‐life balance were not measured. Objective or multisource competence indicators were absent.

6.7. Recommendations for Further Research

Future studies should begin with documented content validation, cognitive interviewing and cross‐cultural testing of any adapted instrument, followed by confirmatory factor analysis in an independent sample and measurement‐invariance testing across major language or nationality groups. Multisite longitudinal designs should combine validated general and work‐specific well‐being measures with objective or multisource competence indicators and organizational variables. Intervention studies are required before recommending competence development as a means to improve happiness.

7. Conclusion

In one UAE acute‐care hospital, higher scores on an investigator‐adapted self‐perceived nursing competence questionnaire were modestly associated with higher scores on a modified five‐point general happiness questionnaire. The association persisted after adjustment for measured covariates and after removing an item about team support. Neither adapted score has established equivalence to its validated source instrument, and the exploratory factor analyses did not validate the intended constructs. Concurrent self‐report measurement, convenience sampling and the cross‐sectional design preclude causal or intervention conclusions. Professional development can be considered within a broader workforce strategy that addresses organizational conditions.

Author Contributions

Ibrahim Al Faouri: conceptualization, methodology, software, formal analysis, project administration, writing – original draft, supervision, visualization, validation. Rahma Nemer: data curation, visualization, investigation, validation, funding acquisition. Fatma Magdi Ibrahim: writing – original draft, writing – review and editing, formal analysis, methodology, supervision, validation.

Funding

No fund was provided for this study.

Ethics Statement

The study protocol was approved by the Research Ethics Committee of RAK Medical and Health Sciences University (RAKMHSU/RES/HEC/2024‐25/9) in June 2025 and by the Ministry of Health and Prevention Research Ethics Committee (MOHAP/REC/2025/74‐2025‐PG‐N) in August 2025. Both approvals preceded data collection, which began on 15 September 2025.

Consent

All participants provided electronic informed consent before the online survey.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

File S1: Psychometric details for the adapted instruments.

Table S1: Investigator‐adapted competence items and complete rotated loading matrix.

Maximum‐likelihood extraction with varimax rotation; complete scale records n = 195. The 95th‐percentile parallel‐analysis criterion retained two factors. Factor signs are arbitrary. h ² denotes communality.

Table S2: Modified Oxford Happiness Questionnaire exploratory loading matrix by item number.

Maximum‐likelihood extraction with varimax rotation; complete scale records n = 195. Item wording is not reproduced. Factors are oriented so that higher positive loadings represent the dominant negatively worded (F1) and positively worded (F2) item clusters. h ² denotes communality.

Data S1: STROBE checklist for cross‐sectional studies.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request and subject to institutional and ethics requirements.