Assessing professional values in nursing education: adaptation and validation of the Hong Kong nurses professional values scale – revised (HK-NPVS-R)

Abstract

Background

Professional values are central to nursing practice and education, yet their measurement requires culturally adapted instruments. Despite the importance of these values, the NPVS-R has not been systematically validated in the Hong Kong context. This study was to evaluate the reliability and validity of a Hong Kong version of the Nurses Professional Values Scale-Revised (HK-NPVS-R) among nursing students.

Methods

A methodological cross-sectional design was used, with a total of 1,623 nursing students recruited from eight institutions representing all major categories of nursing education providers in Hong Kong. The NPVS-R was translated into traditional Chinese using forward-backward translation. Content validity was evaluated by eight nursing experts. Construct validity was examined through principal component analysis (PCA) and confirmatory factor analysis (CFA). Internal consistency was assessed using Cronbach’s alpha. Measurement invariance was tested across gender, age groups, program types, and academic years.

Results

PCA revealed a three-factor structure — Caring, Professionalism, and Activism — explaining 63.17% of total variance. CFA indicated good model fit (CFI = 0.940; TLI = 0.934; RMSEA = 0.042 [90% CI: 0.040, 0.045]). The HK-NPVS-R demonstrated excellent reliability (Cronbach’s α = 0.96). Measurement invariance analyses confirmed stability of the factor structure across all demographic subgroups.

Conclusions

The HK-NPVS-R exhibits strong psychometric properties and cultural relevance. It is a valid and reliable instrument for assessing professional values among Hong Kong nursing students, supporting both research and educational applications.

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12912-026-04661-5.

Background

Professional values are central to practice and education worldwide. Commonly defined as the beliefs and principles that guide ethical behavior and decision-making [10], these values provide a moral framework for navigating dilemmas and maintaining professional standards [8, 54]. They are embedded in codes of ethics and standards of practice, such as the American Nurses Association (ANA) Code of Ethics, which emphasizes activism, caring, justice, professionalism, and trust [5], and the American Association of Colleges of Nursing framework, which highlights domains including nursing knowledge, person-centered care, and professionalism [4]. While broadly recognized, the interpretation and application of professional values vary across sociocultural and healthcare contexts [6, 20, 72].

Beyond their ethical significance, professional values underpin healthy work environments, patient satisfaction, and quality of care [12]. They foster resilience and guide nurses in aligning personal, professional and organization values — a congruence linked to improved wellbeing, reduced burnout, and enhanced patient safety [28]. Nursing students’ internalization of professional values critically shapes their ethical conduct, clinical decisions, and professional identities, particularly during the challenging transition from student to practicing nurse [25, 27, 44, 45]. This transitional period is often marked by role ambiguity, increased responsibility, and heightened stress, making professional values an essential anchor for navigating uncertainty and building conference. These values enhance decision-making capacities during the transitional periods [76] while also enhancing self-efficacy [77], shaping professional identity and job satisfaction [41].

Instruments have been developed to measure nursing-specific professional values. Early assessments adapted general measures like Rokeach’s Value Survey [65], while later tools, such as the Salford-Scott Nursing Values Questionnaire [39] and Career Calling Scale [26], offered context-specific evaluations. Weis and Schank, [73] introduced the Nurses Professional Values Scale (NPVS) grounded in the ANA Code of Ethics. The NPVS provided a 44-item inventory of behaviors exemplifying core nursing values, and initial testing demonstrated strong reliability and validity [73]. Following updates to the Code of Ethics and subsequent research, the instrument was refined into the Nurses Professional Values Scale–Revised (NPVS-R) [74].

In this study, professional values are understood as ethical commitments and dimensions of professional identity, consistent with the professional values model [41] and the ANA Code of Ethics. The NPVS-R operationalizes these values in a multidimensional model, reflecting the structured domains of nursing ethics and professional identity formation. International validation of the NPVS-R among nurses [34, 59, 68] and nursing students [1, 9, 35, 37, 47, 54] has confirmed its robustness, though region-specific variations in factor structures and item loadings have been reported, including within Chinese-speaking contexts such as Mainland China [35] and Taiwan [47].

In Hong Kong, the development of nursing education is uniquely influenced by a blend of Western and Chinese traditions [15, 16]. Although Mainland China, Hong Kong and Taiwan share similar ethnic origins, they differ in political regimes, economic status, healthcare systems, and educational structures [32, 46, 48, 78]. Hong Kong’s nursing training and professional regulation were historically modeled on the British system [16], creating fundamental differences in educational philosophy and clinical practice frameworks compared to other Chinese-speaking regions. The healthcare system itself operates as a tax-financed model with small co-payments, while Mainland China and Taiwan developed distinct financing and delivery models. These systemic differences influence how professional values are conceptualized, taught, and prioritized within nursing education and practice.

Previous studies among students and practicing nurses highlight strong endorsement of values such as patient welfare and accountability [18, 51], yet linguistic barriers and abstract ethical guidelines pose challenges in clinical practice [18]. Despite the recognized significance of professional values, local research has often relied on self-developed surveys with limited psychometric rigor. The use of validated instruments like the NPVS-R in Hong Kong remains very limited, likely due to the absence of locally validated tools. Meanwhile, the expansion of higher education has diversified nursing programs across diploma, baccalaureate, and postgraduate levels in both public and private institutions. While researchers might assume the appropriateness of existing Chinese versions from Mainland China [35] or Taiwan [47], the distinct sociolinguistic contexts (for instance, simplified vs. traditional Chinese, Mandarin vs. Cantonese), educational structures, and healthcare delivery models warrant independent validation. Such need for cross-cultural validation of psychometric instruments within the Chinese-speaking context was emphasized in previous studies [49, 55, 81]. A validated NPVS-R tailored to Hong Kong’s sociocultural and educational contexts is therefore essential to systematically assess nursing professional values. This validation provides a reliable foundation for educators and researchers to assess professional values development, inform curricular improvements, and facilitate meaningful international comparisons.

Methods

Design and participants

This study employed a cross-sectional survey design. Nursing students were recruited from eight institutions in Hong Kong, representing all four major categories of nursing education providers accredited by the Nursing Council of Hong Kong: public universities, private universities, private post-secondary colleges, and hospital-affiliated nursing schools. These institutions collectively account for approximately 95% of the annual registered nurse student intake in Hong Kong and offer diverse pre-registration pathways, including five-year bachelor’s degrees, three-year articulation bachelor’s degrees (for diploma holders), applied degrees (approximately 3.5 years), and professional diplomas in nursing.

Eligible participants were Year 1 and Year 2 nursing students enrolled in pre-registration programs leading to registered nurse licensure across the eight institutions during the 2022–2023 academic year. The total eligible population comprised approximately 4,500 students (2,250 per academic year across these eight institutions, based on enrollment data of the respective institutions). Based on a 99% confidence level and 3% margin of error, the required sample was estimated at 1,307 students, representing approximately 29% of the eligible population.

A stratified purposive sampling approach was employed to ensure representativeness across Hong Kong’s nursing education landscape while maintaining data collection feasibility. Recruitment targets were set to approximate the proportional distribution of students across the four institutional categories based on their annual intake figures. Within-institution recruitment used convenience sampling procedures due to classroom-based access and scheduling constraints. Research team members attended scheduled classes at each institution and invited all present students meeting eligibility criteria to participate voluntarily. Overall, 1,866 responses were collected from the 4,500 eligible students, yielding a total response rate of 41.5% — above the required sample size. The achieved sample distribution approximately matched target proportions, and the respective response rates varied by institution type ranging from 21.6% in post-secondary colleges to 63.2% in hospital-affiliated schools.

The decision to focus on early-year students was methodologically intentional. Students at this stage are in the formative stage of their professional development and generally display relatively homogeneous professional values, with limited variation arising from clinical exposure. Such homogeneity is advantageous in initial scale validation, as it reduces extraneous variability and supports clearer assessment of the instrument’s underlying structure [23]. In addition, access to senior students was logistically constrained because of their dispersed clinical placements and irregular schedules, making systematic recruitment substantially more challenging.

Data collection was conducted between March and June 2023. Each institution granted approval for on-site data collection, and informed consent was obtained from all participants. Ethical approval was secured from the Research Ethics Committee of the corresponding author’s host university (reference number: REC/22–23/0334).

Measures

The survey comprised two sections relevant to this study: (a) demographics; (b) perception of nursing values. Additional sections (e.g. gender-role attitudes, reasons for study, and future aspirations) were included in the broader survey but are not reported here, as they were not analyzed in this paper.

The demographic section collected information on participants’ age, gender, nursing institution, and year of study. Additional socioeconomic indicators (e.g. housing type, family composition, household income, and parents’ education and employment) were also recorded but are not included in the present analyses.

Perception of nursing values was assessed using the Nurses Professional Values Scale-Revised (NPVS-R). The NPVS-R consists of 26 items rated on a five-point Likert scale, ranging from 1 (“Not important”) to 5 (“Most important”). Content validity of the traditional Chinese version was established through expert review by nursing educators and clinical practitioners in Hong Kong.

Translation and cross-cultural adaptation

Permission to use and adapt the NPVS-R was obtained from the original authors (Professors Darlene Weis and Mary Jane Schank) through the College of Nursing of Marquette University. To ensure linguistic and cultural equivalence, the NPVS-R was translated into traditional Chinese following a forward-backward translation procedure [13]:

1. Forward translation: Two bilingual translators independently translated the NPVS-R into Chinese, producing two versions (A1 and A2). The first translator was a doctoral researcher in nursing (native Chinese speaker with professional English fluency and > 10 years of experience in graduate studies and international collaborations). The second translator was a senior research assistant with a Bachelor’s degree in bilingual studies, and Master’s degree in social sciences (native Chinese speaker with > 8 years of experience in bilingual research environments).

2. Reconciliation: The research team met with both translators to discuss language expression and cultural adaptation. By systematically comparing each item of the original English version with A1 and A2, discrepancies related to cultural nuances, terminology, and semantic meaning were identified and resolved. This process resulted in a synthesized initial draft (Version A).

3. Back-translation: Version A was independently back-translated into English by two bilingual Ph.D. candidates (both native Chinese speakers with overseas study experience) who had no prior familiarity with the scale and no medical background, producing versions B1 and B2. The research team then compared the back-translated versions with the original English scale to identify discrepancies and refine the Chinese translation accordingly.

4. Expert validation: While we did not consult the original scale authors on the back-translation, the revised version underwent rigorous expert validation. A panel of eight experts in nursing education, professional practice, and ethics (one clinical nurse, two advanced practice nurses, and five assistant/associate professors) evaluated each item for relevance, clarity, and brevity. Minor wording adjustments were made based on expert feedback to ensure conceptual equivalence within Hong Kong’s nursing context. The Chinese translated versions from Mainland China, Taiwan and our Hong Kong version are shown in comparison in Supplementary Table 3. All items demonstrate nuanced but meaningful lexical differences among these versions.

5. Pilot testing: Finally, the HK-NPVS-R was pre-tested with 10 undergraduate nursing students to confirm clarity and estimate completion time.

Data collection

The survey was administered via the online platform Qualtrics (https://www.qualtrics.com/) to ensure anonymity, confidentiality and comfort. Following a routine nursing lecture, our research staff conducted a briefing session to explain the study purpose and participation procedures. Students completed the survey on their own electronic devices during class time, which allowed standardized instructions and minimized missing data, while voluntary participation was emphasized.

To safeguard data integrity and reduce technical issues, each institution was provided with a unique survey link that was activated only during the scheduled research session. The average completion time was approximately 12 min.

Data analysis

1. Descriptive statistics: Data were screened and imported into IBM SPSS (version 31.0; IBM Corp., Armonk, NY, USA) to compute frequencies, percentages, means, and standard deviations of the demographic characteristics of participants. Comparisons of the composite scores of HK-NPVS-R were conducted across the sub-groups of the demographic variables with the independent t-test and one-way ANOVA. The test statistics and their probability values were adjusted with the appropriate methods if the test assumptions were not met.

2. Content validity was evaluated using the Content Validity Index (CVI). Eight nursing experts evaluated each item across three dimensions — relevance, clarity, and brevity — using a 4-point scale (1 = not relevant/clear/brief, 2 = somewhat relevant/clear/brief, 3 = quite relevant/clear/brief, 4 = highly relevant/clear/brief). Item-level CVI (I-CVI) was calculated as the proportion of experts rating an item 3 or 4, with I-CVI ≥ 0.78 considered acceptable for eight experts [52]. Modified kappa (κ*) was calculated to adjust I-CVI for chance agreement using the formula κ* = (I-CVI − pc) / (1 − pc), where pc represents the probability of chance agreement [61]. Values of κ* >0.74 were considered excellent. Scale-level CVI using universal agreement (S-CVI/UA) was computed as the proportion of items achieving I-CVI = 1.00, with S-CVI/UA ≥ 0.80 indicating adequate scale-level content validity [60].

3. Sample splitting: The sample was divided into two equivalent subsamples using the SOLOMON method [50] for principal component analysis (PCA) and confirmatory factor analysis (CFA).

4. Exploratory step: PCA was conducted on the first subsample to determine the factor structure of the HK-NPVS-R. Orthogonal rotation (Varimax) was applied to extract the maximum variance, under the assumption that the underlying factors were uncorrelated. As a sensitivity analysis, an exploratory factor analysis (EFA) with oblique rotation (Promax) was conducted under the assumption of factor correlations. The factor structures resulted from PCA and EFA were further examined in terms of theoretical interpretability and model fit in CFA.

5. Confirmatory step: CFA was performed on the second subsample using the R package lavaan [66] to test the factor structures derived from the exploratory step, as well as alternative models reported in previous studies. Model fit was evaluated by multiple indices: chi-square to degrees of freedom ratio (χ²/df < 5.0), root mean square error of approximation (RMSEA < 0.080), comparative fit index (CFI > 0.900), Tucker-Lewis Index (TLI > 0.900) and standardized root mean squared residual (SRMR < 0.080) [38]. The Akaike information criterion (AIC) was used to compare competing models, with lower values indicating superior model fit [38].

6. Test of reliability: Cronbach’s alpha coefficient was used to assess internal consistency, with values ≥ 0.70 indicating acceptable reliability [56]. The test-retest reliability was evaluated using the intra-class correlation (ICC) coefficient. ICC values larger than 0.75 are indicative of good reliability [43]. The Spearman-Brown coefficient was used for split-half reliability, with coefficients ≥ 0.70 indicating acceptable internal consistency [29].

7. Convergent validity was assessed using composite reliability and average variance extracted (AVE). Discriminant validity was assessed with the Heterotrait-Monotrait (HTMT) ratio [36]. AVE was calculated as the mean of squared standardized factor loadings from the CFA model, with AVE > 0.50 supporting the convergent validity of the construct [33]. The HTMT criterion was used to assess discriminant validity, with HTMT < 0.90 indicating adequate discriminant validity between factors.

8. Measurement invariance: Multi-group CFA was conducted to assess invariance across demographic subgroups (gender, age groups, program types and academic years). Configural invariance was first established, followed by progressive equality constraints on factor loadings (metric invariance) and item intercepts (scalar invariance). Because the chi-square difference test is highly sensitive to sample size, changes in CFI (ΔCFI < 0.01) were adopted as the criterion to establish invariance [19].

Results

Demographic characteristics

A total of 1,866 responses were collected from the nursing students. After data inspection to clear ineligible cases with incomplete response on mandatory questions (including demographics and measures of the scale), a total of 1,623 cases were included for analysis. The sample comprises 422 males (26.0%) and 1201 females (74.0%). Participants’ age ranged from 17 to 50 years (M = 21.59, SD = 4.24). Students were recruited from private universities (41.5%), public universities (25.7%), post-secondary colleges (24.9%) and hospital-affiliated nursing schools (7.9%). For group comparisons and factor analysis, age was recoded into three categories: below 20 years, 20–24 years, and ≥ 25 years. These three age categories reflect distinct educational pathways in Hong Kong’s nursing education system: individuals under 20 years represent traditional students entering five-year bachelor programs directly after secondary education; those aged 20–24 years typically include students who completed community college studies (e.g., associate degree, higher diploma) before entering articulation degree programs; and those aged 25 years or above represent mature students with potential prior work or life experiences. The mean composite score of the HK-NPVS-R was 104.37 (SD = 12.93) (Table 1). All subgroups of these demographic characteristics did not show significant difference in the HK-NPVS-R scores (all ps > 0.05).

Table 1.

Summary of demographic characteristics, and HK-NPVS-R mean score (N = 1,623)

Variable	Frequency (%)	HK-NPVS-R
		Mean score (SD)	Test statistic	p value
Age (years)	Age range: 17–50, M = 21.59, SD = 4.24	104.37 (12.93)
Age group			F = 2.026	0.132
Below 20	667 (41.1)	103.64 (13.02)
20–24	625 (38.5)	104.65 (12.29)
25 or above	331 (20.4)	105.28 (13.84)
Gender			t = -1.457	0.146
Female	1201 (74.0)	104.68 (11.98)
Male	422 (26.0)	103.48 (15.27)
Institution type			F = 2.117	0.096
Public university	417 (25.7)	103.71 (12.45)
Private university	673 (41.5)	104.34 (13.00)
Post-secondary college	404 (24.9)	105.57 (13.32)
Hospital nursing school	129 (7.9)	102.85 (12.60)
Program type			F = 0.135	0.874
Diploma	368 (22.7)	104.48
Bachelor	1178 (72.6)	104.29
Master	77 (4.7)	105.03
Academic Year			t = 0.715	0.474
Year 1	1058 (65.2)	104.54
Year 2	565 (34.8)	104.05

Content validity

All items demonstrated excellent content validity. I-CVI values ranged from 0.88 to 1.00 across the three dimensions (relevance, clarity, brevity), with modified kappa values ranging from 0.87 to 1.00, indicating excellent agreement beyond chance [61]. S-CVI/UA values were 1.00 for relevance, 0.85 for clarity, and 0.88 for brevity, all exceeding the 0.80 threshold. Detailed CVI and modified kappa values are provided in Supplementary Table 1.

Construct validity

Descriptive statistics of the 26 items of the HK-NPVS-R are presented in Table 2. The highest-rated item was Item 25 (“Maintain confidentiality of patient”, M = 4.42, SD = 0.69), Item 21 (“Safeguard patient’s right to privacy”, M = 4.29, SD = 0.70) and Item 15 (“Maintain competence in area of practice”, M = 4.23, SD = 0.66). The lowest-rated items were Item 26 (“Participate in activities of professional nursing association”, M = 3.48, SD = 0.89), Item 5 (“Participate in peer review”, M = 3.56, SD = 0.83), and Item 4 (“Participate in public policy decisions affecting distribution of resources”, M = 3.72, SD = 0.80).

Table 2.

Items means, standard deviation, corrected item-total correlation, squared multiple correlation, and alpha if item deleted for HK-NPVS-R (N = 1,623)

Item		Mean	SD	CITC	SMC	AID
1	Engage in on-going self-evaluation	4.10	0.67	0.68	0.52	0.96
2	Request consultation/collaboration when unable to meet patient needs	4.13	0.65	0.64	0.50	0.96
3	Protect health and safety of the public	4.17	0.73	0.69	0.56	0.96
4	Participate in public policy decisions affecting distribution of resources	3.72	0.80	0.54	0.45	0.96
5	Participate in peer review	3.56	0.83	0.44	0.34	0.96
6	Establish standards as a guide for nursing practice	3.93	0.74	0.67	0.57	0.96
7	Promote and maintain standards where planned learning activities for students take place	4.03	0.68	0.70	0.58	0.96
8	Initiate actions to improve environments of practice	3.96	0.70	0.68	0.52	0.96
9	Seek additional education to update knowledge and skills	4.08	0.70	0.69	0.56	0.96
10	Advance the profession through active involvement in health-related activities	3.92	0.73	0.68	0.57	0.96
11	Recognize role of professional nursing associations in shaping health care policy	3.91	0.72	0.69	0.56	0.96
12	Promote equitable access to nursing and health care	4.00	0.72	0.70	0.56	0.96
13	Assume responsibility for meeting health needs of the culturally diverse population	4.05	0.69	0.73	0.61	0.96
14	Accept responsibility and accountability for own practice	4.11	0.69	0.73	0.61	0.96
15	Maintain competency in area of practice	4.23	0.66	0.74	0.67	0.95
16	Protect moral and legal rights of patients	4.19	0.69	0.75	0.67	0.95
17	Refuse to participate in care if in ethical opposition to own professional values	4.13	0.70	0.68	0.53	0.96
18	Act as a patient advocate	3.95	0.73	0.67	0.51	0.96
19	Participate in nursing research and/or implement research findings appropriate to practice	3.83	0.73	0.63	0.50	0.96
20	Provide care without prejudice to patients of varying lifestyles	4.10	0.69	0.74	0.63	0.95
21	Safeguard patient’s right to privacy	4.29	0.70	0.71	0.67	0.96
22	Confront practitioners with questionable or inappropriate practice	3.89	0.74	0.61	0.44	0.96
23	Protect rights of participants in research	4.01	0.70	0.73	0.61	0.96
24	Practice guided by principles of fidelity and respect for person	4.17	0.67	0.73	0.65	0.96
25	Maintain confidentiality of patient	4.42	0.69	0.62	0.57	0.96
26	Participate in activities of professional nursing associations	3.48	0.89	0.42	0.31	0.96
	Overall Cronbach’s alpha = 0.96

Note. CITC: corrected item-total correlation; SMC: squared multiple correlation; AID: alpha if item deleted

Item retention criteria included corrected item-total correlations > 0.30 [31] and no substantial improvement in alpha coefficients upon item deletion [56]. All items demonstrated acceptable corrected item-total correlations (0.42 to 0.75); therefore, no items were removed from this scale.

Split-half sample

The sample of 1,623 respondents was divided into two equivalent subsamples using the SOLOMON method for factor analysis [50]. One subsample (n = 812) was used for principal component analysis (PCA), and the other (n = 811) for confirmatory factor analysis (CFA). The two subsamples demonstrated high equivalence, indicated by their Kaiser-Meyer-Olkin (KMO) values and communality ratio (S): KMO_PCA = 0.968, KMO_CFA = 0.973, S = 0.995.

Principal component analysis

PCA with Varimax rotation and Kaiser normalization was conducted on the subsample of 812 nursing students. The data was well-suited for factor analysis (KMO = 0.968, Bartlett’s Test of Sphericity: χ² (325) = 14911.124; p < .001). Factor extraction was guided by the Kaiser criterion of eigenvalue > 1, scree plot inspection, factor loadings > 0.40, and percentage of variance explained.

The results of PCA are presented in Table 3, and the scree plot is shown in Fig. 1. A three-factor structure was derived for the 26-item scale, accounting for 63.17% of total variance. Eigenvalues ranged from 13.30 to 1.30, with all 26 items showing moderate to strong loadings (0.52–0.82) and communalities between 0.54 and 0.75.

Table 3.

Principal component analysis of the HK-NPVS-R (n = 812)

		Factor and loading			Communality
Item		1	2	3
Eigenvalue		13.30	1.83	1.30
1	Engage in on-going self-evaluation	0.47	0.56	0.19	0.58
2	Request consultation/collaboration when unable to meet patient needs	0.47	0.61	0.07	0.60
3	Protect health and safety of the public	0.50	0.56	0.09	0.57
4	Participate in public policy decisions affecting distribution of resources	0.06	0.65	0.36	0.55
5	Participate in peer review	0.09	0.56	0.47	0.54
6	Establish standards as a guide for nursing practice	0.30	0.75	0.20	0.69
7	Promote and maintain standards where planned learning activities for students take place	0.40	0.67	0.14	0.63
8	Initiate actions to improve environments of practice	0.36	0.65	0.19	0.59
9	Seek additional education to update knowledge and skills	0.41	0.59	0.25	0.58
10	Advance the profession through active involvement in health-related activities	0.30	0.58	0.37	0.57
11	Recognize role of professional nursing associations in shaping health care policy	0.34	0.64	0.36	0.65
12	Promote equitable access to nursing and health care	0.52	0.44	0.29	0.55
13	Assume responsibility for meeting health needs of the culturally diverse population	0.59	0.49	0.20	0.64
14	Accept responsibility and accountability for own practice	0.69	0.45	0.09	0.69
15	Maintain competency in area of practice	0.73	0.45	0.04	0.74
16	Protect moral and legal rights of patients	0.75	0.37	0.09	0.71
17	Refuse to participate in care if in ethical opposition to own professional values	0.68	0.26	0.20	0.57
18	Act as a patient advocate	0.52	0.19	0.55	0.61
19	Participate in nursing research and/or implement research findings appropriate to practice	0.32	0.29	0.68	0.64
20	Provide care without prejudice to patients of varying lifestyles	0.72	0.25	0.34	0.69
21	Safeguard patient’s right to privacy	0.82	0.20	0.19	0.75
22	Confront practitioners with questionable or inappropriate practice	0.45	0.14	0.63	0.62
23	Protect rights of participants in research	0.61	0.22	0.54	0.71
24	Practice guided by principles of fidelity and respect for person	0.77	0.21	0.31	0.73
25	Maintain confidentiality of patient	0.78	0.17	0.14	0.66
26	Participate in activities of professional nursing associations	0.02	0.26	0.71	0.58
	% of variance; total: 63.17%	28.57	22.16	12.44

Fig. 1.

Scree plot of PCA

The first factor, labelled Caring, consisted of 11 items (Item 12–17, 20, 21, 23, 24, 25) and accounted for 28.57% of the total variance. These items primarily reflected respect for patient dignity, safeguarding patient rights, maintaining confidentiality, and providing unbiased and culturally sensitive care.

The second factor, labelled Professionalism, also consisted of 11 items (Item 1–11), explaining 22.16% of the variance. Items in this factor emphasized self-assessment, collaborative practice, public health protection, professional development, and maintenance of nursing standards.

The third factor, labelled Activism, included Item 18, 19, 22, 26 (12.44% variance) captured advocacy for patients, engagement in research, confronting unethical practices, and participation in professional nursing associations.

Although PCA with Varimax rotation assumes uncorrelated factors, we are cautious that the underlying domains of the scale may be interconnected in reality. As a sensitivity analysis, an EFA with oblique rotation (Promax) was conducted under the assumption of factor correlations. The EFA produced a single-factor solution dominated by a large general factor (eigenvalue = 13.30, 49.48% variance), with the third factor falling below the Kaiser criterion (eigenvalue = 0.85). No items showed primary loadings on Factors 2 or 3, pointing to a unidimensional factor solution. We subsequently examined these two factorial structures using CFA on the fit indices along with other competing models.

Confirmatory factor analysis

Tests of multivariate skewness and kurtosis indicated non-normality (ps < 0.001) [53]. Robust maximum likelihood estimation with Satorra-Bentler scaling corrections was therefore applied [67].

As shown in Table 4, we first tested two nested models: (1) Model 1 – orthogonal three-factor model consistent with PCA Varimax assumptions that factors are uncorrelated, with factor covariances constrained to zero, and (2) Model 2 – correlated three-factor model allowing factors to covary freely. The decision to allow factor correlations in Model 2 was based on: (a) modification indices from Model 1 indicating substantial factor covariances, (b) theoretical expectation that professional values dimensions are conceptually interconnected (nursing caring requires professional competence; activism involves caring principles), and (c) consistent findings of moderate to high factor correlations across prior studies of nursing professional values especially in the Chinese context [82, 83]. Although modification indices from Model 2 suggested covariances between items, the model fit indices did not improve significantly after attempting these modifications (ΔCFI < 0.01). To avoid data-driven manipulations that risk over-fitting and compromise replicability, we decided to retain Model 2 as the final model in CFA.

Table 4.

Confirmatory factor analyses for model comparison (n = 811)

Model	S-Bχ2	df	RMSEA (90% CI)	SRMR	CFI	TLI	AIC
Model 1: Uncorrelated 3-factor	1598.869***	299	0.073 [0.071, 0.076]	0.332	0.818	0.802	36069.088
Model 2: Correlated 3-factor	727.772***	296	0.042 [0.040, 0.045]	0.055	0.940	0.934	34629.764
Model 3: 1-factor	1237.886***	299	0.062 [0.060, 0.065]	0.061	0.854	0.841	35219.252
Model 4: 3-factor (Taiwan)	790.047***	296	0.045 [0.043, 0.048]	0.057	0.931	0.924	34803.209
Model 5: 5-factor (U.S.)	829.386***	289	0.048 [0.045, 0.051]	0.050	0.916	0.906	34396.030

Note. 1-factor model = all items loaded onto one single construct; 3-factor model (Taiwan) = factor structure from a Taiwanese version [47]; 5-factor model = factor structure from the original English version [74]; ^***p < .001

We also examined the unidimensional factorial structure generated from EFA with oblique rotation (Model 3), and it failed to achieve acceptable fit. Compared to Model 4 from a Taiwan’s study with a 3-factor structure [47] and Model 5 the original authors’ 5-factor structure [74], Model 2 maintained better fit than these models: S-Bχ² (299, n = 811) = 727.772, p < .001; χ²/df = 2.459; RMSEA = 0.042 [90% CI: 0.040, 0.045]; CFI = 0.940; TLI = 0.934; AIC = 34629.764. The path diagram of Model 2 is illustrated in Fig. 2.

Fig. 2.

Path diagram and CFA factor loadings

Reliability analysis

Internal consistency reliability was excellent for the total scale (Cronbach’s α = 0.96, McDonald’s ω = 0.95) and all factors: Caring (α = 0.94, ω = 0.93), Professionalism (α = 0.91, ω = 0.88), and Activism (α = 0.77, ω = 0.75). All values exceeded the 0.70 acceptability threshold [56, 69]. The split-half reliability coefficient for the scale was 0.90, and the test-retest reliability was 0.96. The reliability coefficients are shown in Table 5.

Table 5.

Descriptives, bivariate correlations and reliability coefficients of the HK-NPVS-R

	1	2	3	4
1. Caring
2. Professionalism	0.78***
3. Activism	0.67***	0.68***
4. Overall HK-NPVS-R	0.94***	0.93***	0.80***
M	4.16	3.96	3.79	4.02
SD	0.54	0.52	0.59	0.49
Cronbach’s α	0.94	0.91	0.77	0.96
McDonald’s ω	0.93	0.88	0.75	0.95
ICC	0.94	0.91	0.77	0.96
Split-half reliability	0.90	0.87	0.76	0.90

Note. ^***p < .001, Pearson correlation coefficient. ICC = Intraclass correlation coefficient. Spearman-Brown coefficient for the split-half reliability

Convergent and discriminant validity

Despite two factors falling marginally below the conventional AVE ≥ 0.50 threshold (Caring: AVE = 0.57; Professionalism: AVE = 0.47; Activism: AVE = 0.45), convergent validity was considered adequate: all standardized factor loadings were significant (p < .001) and substantial (range: 0.49–0.81), composite reliability values substantially exceeded 0.70. As Fornell and Larcker ([33], p. 46) stated, convergent validity should be considered “adequate even though more than 50% of the variance is due to error” when CR > 0.70. Discriminant validity was assessed using the Heterotrait-Monotrait (HTMT) ratio, which outperforms traditional criteria with correlated factors [36]. All HTMT values ranged from 0.79 to 0.84, below the 0.90 threshold, supporting discriminant validity despite moderate factor correlations (r = .67–0.78).

Measurement invariance

Multi-group CFA was conducted to test invariance across gender, age groups, program types and academic years. Results are summarized in Table 6.

Table 6.

Measurement invariance across gender, age groups, nursing program types, and years of study (n = 811)

Model	S-Bχ2	df	RMSEA [90% CI]	SRMR	CFI	TLI	Comparison (Δ CFI)
Gender invariance
1. Configural invariance	909.181***	592	0.036 [0.033, 0.039]	0.060	0.986	0.985	/
2. Metric invariance	923.919***	615	0.035 [0.032, 0.038]	0.064	0.986	0.986	2 vs. 1 (0.000)
3. Scalar invariance	958.764***	638	0.035 [0.032, 0.038]	0.062	0.986	0.986	3 vs. 2 (0.000)
Age group invariance
1. Configural invariance	918.385	888	0.011 [0.001, 0.017]	0.065	0.997	0.997	/
2. Metric invariance	940.083	934	0.005 [0.001, 0.014]	0.070	0.999	0.999	2 vs. 1 (0.002)
3. Scalar invariance	1000.628	980	0.009 [0.001, 0.016]	0.068	0.998	0.998	3 vs. 2 (-0.001)
Program type invariance
1. Configural invariance	964.688*	888	0.018 [0.012, 0.022]	0.063	0.989	0.988	/
2. Metric invariance	960.375	934	0.010 [0.001, 0.016]	0.067	0.996	0.996	2 vs. 1 (0.007)
3. Scalar invariance	1034.781	980	0.014 [0.007, 0.019]	0.065	0.992	0.992	3 vs. 2 (-0.004)
Years of study invariance
1. Configural invariance	890.437***	592	0.035 [0.032, 0.038]	0.060	0.965	0.961	/
2. Metric invariance	901.893***	615	0.034 [0.031, 0.037]	0.066	0.966	0.964	2 vs. 1 (0.001)
3. Scalar invariance	945.950***	638	0.035 [0.032, 0.037]	0.064	0.963	0.963	3 vs. 2(-0.003)

^*p < .05, ^***p < .001

1. Gender: Baseline model fit was good (S-Bχ² (592, n = 811) = 909.181, χ²/df = 1.536, RMSEA = 0.036, CFI = 0.986, TLI = 0.985). Constraining factor loadings and intercepts (ΔCFI = 0.000) did not degrade fit, supporting configural, metric, and scalar invariance.

2. Age groups: Baseline fit was excellent (S-Bχ² (888, n = 811) = 918.385, p = .233, χ²/df = 1.034, RMSEA = 0.011, CFI = 0.997, TLI = 0.997). Equality constraints confirmed invariance across age categories (ΔCFI < 0.01).

3. Program types: Baseline fit was strong (S-Bχ² (888, n = 811) = 964.688, p = .037, χ²/df = 1.086, RMSEA = 0.018, CFI = 0.989, TLI = 0.988). Constrained models retained good fit, confirming invariance across program types (ΔCFI < 0.01).

4. Academic years: Baseline fit was acceptable (S-Bχ² (592, n = 811) = 890.437, χ²/df = 1.504, RMSEA = 0.035, CFI = 0.960, TLI = 0.961). Equality constraints maintained fit (ΔCFI < 0.01), supporting invariance across Year 1 and Year 2 students.

Discussion

This is the first study to validate the NPVS-R in Hong Kong, benefiting from a large, representative sample of nursing students from all types of accredited nursing programs. Findings demonstrate that the HK-NPVS-R possesses strong psychometric properties with meaningful practical applications for nursing education and research in the region. The mean composite score in our sample (104.37, SD = 12.39) falls within the range reported across international NPVS-R validation studies, though notable variation exists across cultural contexts. Asian samples have yielded mean scores ranging from 95.80 to 104.27: Indonesia (95.80; [1]), Taiwan (99.10–104.27; [2, 47, 48]), Mainland China (100.47; [48]), South Korea (101.86; [54]), and our Hong Kong sample (104.37). North American and European samples have generally reported scores in the 106–118 range: United States (106.16 in [2]; 106.86 in [63]), Spain (113.35; [11]), and Canada (118.30; [7]). Middle Eastern samples have shown comparable patterns to Asian contexts: Iran (101.79; [62]), and Jordan (99.68; [3]).

Study design and sample size

The structural validity of the scale was evaluated in two consecutive steps, beginning with principal component analysis (PCA) to identify how the items would cluster into domains, followed by confirmatory factor analysis (CFA) to validate the proposed model. In addition, an EFA with oblique rotation was performed as a sensitivity analysis, which resulted in a less interpretable solution dominated by a single general factor. While prior studies have provided valuable insight in the regional variation of professional values within the Chinese context (e.g., [35, 47]), our approach further combined CFA to provide robust tests of the models from the exploratory step of this study and the models from other validation studies [14, 42]. As a result, CFA supported the three-factor structure derived from PCA, with fit indices (e.g., CFI, TLI, RMSEA) surpassing recommended thresholds. Importantly, multi-group CFA established measurement invariance across gender, age group, program type, and academic year, confirming that the scale assesses professional values stably across these subgroups [64].

The sample size (N = 1,623) was among the largest reported in NPVS-R validation studies [6, 21], representing approximately one-third of the total student intake across Hong Kong’s accredited nursing programs over two consecutive years. Students were drawn from a wide range of program types (diploma to master) and institutional settings (public, private, and hospital-based), enhancing representativeness and generalizability.

Content validity was further ensured through expert review, which identified minor but important lexical and contextual differences compared to Mainland Chinese [35] and Taiwanese [47] versions, resulting in a version better aligned with Hong Kong’s linguistic and professional norms. Internal consistency reliability was high across the overall scale and subscales, affirming the HK-NPVS-R as a psychometrically sound instrument.

Three-factor structure of HK-NPVS-R

Principal component analysis identified a three-factor structure — Caring, Professionalism, and Activism — accounting for 63.17% of the total variance. This structure was further supported by CFA, yielding satisfactory model fit indices. This structure diverges notably from the original five-factor U.S. model [74], and demonstrates variability compared to other international validations: five factors in Korea [54], Canada [7], and Iran [37]; four factors in Mainland China [35], Turkey [59]; three factors in Taiwan [47] and Spain [11]; and one-factor structure in Italy [34]. This cross-cultural heterogeneity in factor structures supports theoretical frameworks positing that professional values, while rooted in universal ethical principles, are culturally constructed through context-specific professional socialization processes [22, 40].

Three points should be particularly noted when interpreting the factor structure in this study. First, the Hong Kong three-factor solution reflects specific features of the region’s professional and cultural context. The Caring factor emerged as the dominant dimension (28.57% variance), incorporating items originally distributed across the U.S. model’s Caring, Trust, and Justice factors. Notably, items addressing equitable access and culturally diverse populations (Items 12–13, originally Justice) and professional accountability items (Items 14–15, originally Trust) clustered with core caring behaviors in the Hong Kong adaptation. This consolidation aligns with Hong Kong’s demographic homogeneity (over 90% ethnic Chinese) and Confucian cultural values emphasizing relational ethics and collective responsibility [17, 70]. In contexts with high ethnic homogeneity, equity and diversity may not be perceived as separate justice concerns but as intrinsic elements of patient-centered care. Hong Kong nurses perceived trusting nurse-patient relationship and being accountable as important caring behaviors [80]. The incorporation of accountability and competence items (originally Trust) into Caring suggests that within Hong Kong’s regulatory environment, professional trustworthiness is inseparable from caring competence. This reflects the integration of professional competence with ethical sensitivity in the NCHK’s Code of Ethics and Professional Conduct [58] and our nurses’ perception of caring behaviors [80].

Second, the Professionalism factor in HK-NPVS-R absorbed items distributed across the original Professionalism, Trust, and Activism factors, including ongoing self-evaluation (Item 1), collaboration (Item 2), and peer review (Item 5). This reconfiguration likely reflects NCHK’s Hong Kong’s Core-competencies [57] in emphasizing individual accountability, reflective practice, and peer governance. These behaviors are conceptualized not primarily as caring attributes (as in more collectivist frameworks) but as professional obligations transcending nursing to characterize learned professions in general. This aligns with contemporary nursing discourse in Hong Kong positioning nursing as a profession rather than vocation [75, 79], advocating autonomous judgment, evidence-based practice, and professional accountability.

Third, the Activism factor in our study is particularly revealing of cultural-professional priorities. Items involving patient advocacy (Item 18) and confronting unsafe practice (Item 22) — originally coded as Caring in the U.S. model — emerged under the Activism dimension in Hong Kong. This suggests that activism and advocacy are conceptualized as active professional engagement requiring agency and assertiveness, rather than as an intrinsic caring behavior [71]. In the hierarchical healthcare cultures [22, 84], speaking up for patients or challenging practitioners represents a deliberate departure from normative practices, constituting professional activism distinct from routine caring or technical competence. This cultural reframing of advocacy behaviors aligns with the broader recognition that professional values are operationalized through specific nurse-patient interaction competencies, wherein therapeutic communication itself constitutes an ethical imperative [30]. The recent development of culturally sensitive instruments such as the Nurse-Patient Therapeutic Interaction Scale (NuPTIS) [30] underscores the ongoing need for contextually valid measures that capture how nurses enact professional values, including patient advocacy, within diverse clinical and cultural settings. Our findings reinforce this imperative, demonstrating that what constitutes ‘caring’ and ‘activism’ is not universally fixed but shaped by cultural expectations regarding professional agency and hierarchical norms. Furthermore, it is worth noting that the three items with lowest scores (Item 4, 5 and 26) also have relatively lower factor loadings and item-total correlations. While these activities are ‘participatory’ in nature and require higher professional agency, it is consistent with the findings that nurses are good rule followers and tend to conform to the social identity paradigm but may lack development in character and application of situational thinking and practice [22, 24]. The relatively lower reliability coefficient of the Activism factor (Cronbach’s alpha = 0.77) also likely reflects the early stage of professional socialization as the characteristics of our junior-year cohort, where students may prioritize clinical tasks over social advocacy.

Overall, the three domains identified in our study align closely with the NCHK’s Code of Ethics and Professional Conduct [58]. Caring resonates with the domains of “Nurses and People” and “Nurses and Practice”, emphasizing respect for patient dignity, confidentiality, and competent care. Professionalism parallels “Nurses and the Profession”, underscoring lifelong learning, maintenance of standards, and personal conduct. Activism reflects “Nurses and Society”, particularly in promoting community health, equitable resource allocation, and professional association involvement. These overlaps reinforce the contextual validity of the HK-NPVS-R and underscore its relevance to Hong Kong’s nursing education and practice.

Additionally, several items exhibited cross-loadings (> 0.40) on dual factors, such as Items 1–3, 12–15 on Caring and Professionalism. Assignment to factors was determined by considering: (1) primary factor loading strength; (2) contribution to the factor’s internal consistency (Cronbach’s α); (3) semantic coherence with other factor items; and (4) alignment with the original NPVS-R structure. These cross-loadings are likely to point to a conceptual overlap between multiple domains, and reflect the dual role of professional codes of conduct, which articulate expected behaviors while simultaneously guiding ethical decision-making [20]. In nursing, attributes such as accountability, advocacy, and competence are not only professional standards but also ethical imperatives. Public trust in the profession is built on this integration, with ethical judgment and professional integrity often operationalized jointly in practice. It is important to note that cross-loadings are not necessarily problematic, for example factor loadings of Items 10–12 were high on two factors in the Taiwan study [47], and similarly on items (9, 10, 13, 15) in the South Korea study [54]. Unlike conventional psychological latent constructs that can be highly distinguishable from each other, perhaps this issue reflects the inherent breadth and diversity of professional values as operationalized in the NPVS-R. Each factor encompasses multiple behavioral manifestations of professional ethics. For future research, the interpretation and application of the factor structure of any NPVS-R adaptation, including the one from the present study, should be treated with flexibility and focus more on the overall psychometric properties. Cross-region comparisons are more appropriately conducted using total scale scoring or by examining individual items rather than assuming strict factorial invariance.

Implications

The findings of this study carry important implications for nursing education and professional development in Hong Kong and similar contexts. By validating a culturally adapted version of the NPVS-R, educators and researchers now have a reliable and contextually appropriate tool to assess the development of professional values among nursing students. This enables curriculum coordinators to systematically evaluate how well educational programs instill professional values such as caring, professionalism and activism. Since these values are fundamental to ethical nursing practice, their integration into the curriculum — through ethics courses, reflective practice, clinical mentorship, and service learning — should be intentional and measurable. The ability to assess students’ values at different stages of training offers insights into the effectiveness of pedagogical approaches and helps identify areas where further support or curriculum revision may be needed.

Although measurement invariance was established across Year 1 and 2 cohorts, it is likely that students in senior years may interpret or prioritize certain values differently. Value development is a dynamic and evolving process throughout nursing education [6], and educators should be mindful of this progression when tailoring interventions and learning experiences. Reinforcing core values progressively across the curriculum can help ensure that students develop a strong ethical foundation as they advance in their training.

The three-factor structure identified in this study also provide a framework for aligning teaching strategies with these domains. For example, content related to patient advocacy, policy engagement, and community health may require greater emphasis, particularly since items under Activism were rated relatively lower in importance by students. This reflects a potential gap between the ideal and perceived relevance of certain professional roles, highlighting the need for role modelling, clinical exposure, and ethical reasoning exercises to bridge that gap.

From a broader perspective, this validated HK-NPVS-R aims to benefit cross-institutional comparison and tracking of student development across different types of programs and settings. This can further contribute to the process of informing policy decisions in nursing education and supporting standardization of professional training outcomes, ensuring that graduates across diverse institutions share a common foundation of professional values.

Limitations

Several limitations should be acknowledged. First, the cross-sectional design precludes assessment of changes over time. Future research should adopt longitudinal design to examine test-retest reliability, longitudinal invariance, and predictive validity. Second, this study focused on nursing students. While this population is critical for understanding value formation during education, further validation among practicing nurses is needed to assess the generalizability of the HK-NPVS-R to the broader workforce. Professional values may evolve with clinical experience, organizational culture, and exposure to ethical dilemmas, and these influences warrant systematic investigation. Third, the reliance on self-report measures introduces potential biases, including social desirability and acquiescence effects. Although the NPVS-R has been widely used internationally, responses may be influenced by students’ perceptions of what constitutes “ideal” professional behavior rather than their actual attitudes. Future adaptations could incorporate reverse-worded items or triangulate findings with observational methods to enhance response validity.

Finally, while measurement invariance was established across Year 1 and Year 2 cohorts, caution is warranted in interpreting these results. Our sample was concentrated in the early years of training, whereas courses on nursing ethics and professional values are typically introduced later in the curriculum. As senior students accumulate more clinical exposure and develop more complex ethical reasoning skills, their professional values may be different from those of early-year students, as suggested by previous work on the effects of education duration on professional values development [6]. This study intentionally focused on early-stage students — both for methodological reasons of achieving a relatively homogeneous baseline for initial scale validation and due to practical constraints associated with accessing senior students during clinical placements — the current findings primarily reflect the early educational stage. Therefore, generalizing the factor structure to senior cohorts should be treated with caution. Future studies may employ a longitudinal design or include students across all academic years, including those in clinical practicum, to determine whether the factor structure remains stable throughout the full trajectory of nursing education.

Conclusion

This study provides the first comprehensive validation of the Nurses Professional Values Scale – Revised (NPVS-R) in traditional Chinese among nursing students in Hong Kong. Using a large and diverse sample, rigorous statistical methods, and expert input, the HK-NPVS-R demonstrated strong psychometric properties, cultural relevance, and measurement invariance across key demographic subgroups. The three-factor structure —Caring, Professionalism, and Activism — reflects both international nursing values and the local ethical framework articulated by the Nursing Council of Hong Kong.

By establishing the reliability and validity of this instrument, the study offers educators, researchers, and policymakers a robust tool for assessing professional values in nursing education. The findings highlight the importance of integrating caring, professionalism, and activism into curricula and professional development, while also enabling meaningful cross-institutional and cross-cultural comparisons.

Besides affirming the instrument’s psychometric soundness, this study provides a culturally grounded perspective on how professional values are interpreted within Hong Kong’s educational and sociocultural context. It underscores the need to continuously re-examine these values in light of evolving healthcare challenges, generational shifts, and broader sociocultural influences. Future research should investigate how values develop longitudinally, manifest in clinical decision-making, and shape early-career transitions. As the nursing profession increasingly emphasizes global competence and ethical responsiveness, locally validated tools such as HK-NPVS-R will remain essential for guiding education, supporting professional identity formation, and promoting ethical practice in Hong Kong and beyond.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AIC

Akaike information criterion

ANA

American Nurses Association

CFA

Confirmatory factor analysis

CFI

Comparative fit index

CVI

Content validity index

KMO

Kaiser-Meyer-Olkin

NCHK

Nursing Council of Hong Kong

NPVS-R

Nurses Professional Values Scale – Revised

PCA

Principal component analysis

RMSEA

Root mean square error of approximation

SRMR

Standardized root mean squared residual

TLI

Tucker-Lewis index

Author contributions

KC conceptualized the study and provided methodological support, resources and supervision, and participated in writing, review and editing. CC collected and analyzed the data, and prepared the original draft. AK acquired project funding, conceptualized the study and provided methodological support, supervision, resources, and was the major contributor in writing the manuscript. All authors read and approved the final manuscript.

Funding

This work was supported by the General Research Fund – Research Grants Council of Hong Kong (project number 15605522; project title: Vocational/Professional Training against a Changing Landscape of Higher Education in Hong Kong: Choices and Experiences of Students in Accredited Nursing Programs).

Data availability

The datasets generated and analyzed during the current study are not publicly available because this study is part of a larger study, but are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Helsinki Declaration. All methods were performed in compliance with relevant guidelines and regulations. The study was approved by the research ethics committee of Hong Kong Baptist University (reference number: REC/22–23/0334). Written informed consent was obtained from all participants prior to the data collection. The participants were ensured about voluntary participation in the study, as well as anonymity and confidentiality of the data.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.