Global standards and local authenticity: Cross-cultural leadership and employee behavioural adaptation in heritage-based hotels in Baghdad

[i] Note. Full collinearity VIFs obtained by regressing each construct, in turn, on the remaining two constructs in a saturated structural model in SmartPLS 4, following Kock’s (2015) full collinearity test for common method variance. All values fall well below the conservative 3.3 threshold, indicating that common method variance is unlikely to be a pervasive concern in this dataset.

Table 2:

Demographic Profile of Respondents (N = 380)

Characteristic	Category	Frequency	Percentage (%)
Gender	Male	240	63.2
Gender	Female	140	36.8
Age	18–25 years	85	22.4
	26–35 years	165	43.4
	36–45 years	90	23.7
	Above 45 years	40	10.5
Hotel Experience	Less than 2 years	65	17.1
	2–5 years	145	38.2
	6–10 years	110	28.9
	More than 10 years	60	15.8
Education	High School or below	42	11.1
	Diploma / Associate Degree	98	25.8
	Bachelor’s Degree	192	50.5
	Postgraduate	48	12.6

[i] Note. Percentages are rounded to one decimal place. The profile is reported descriptively; no between-group statistical tests were conducted (see Sections 1 and 7).

[ii] Source: Primary data collected by the author through field survey in Baghdad (2026).

Table 3:

Descriptive Statistics for Construct Items (N = 380)

Construct / Item	Mean	Std. Dev.	Skewness	Kurtosis	N
Cross-Cultural Leadership (CCL)
CCL1 – Adapts communication style to cultural context	3.92	0.71	‒0.41	0.18	380
CCL2 – Acknowledges cultural differences openly	3.85	0.74	‒0.38	0.09	380
CCL3 – Creates inclusive decision-making environment	3.78	0.78	‒0.29	‒0.11	380
CCL4 – Demonstrates cultural empathy toward staff	3.90	0.72	‒0.43	0.21	380
CCL5 – Facilitates digital protocol adoption culturally	3.71	0.81	‒0.22	‒0.07	380
CCL6 – Models behaviour consistent with local values	3.68	0.83	‒0.19	‒0.14	380
Employee Behavioural Adaptation (EBA)
EBA1 – Adjusts communication style for guest diversity	3.81	0.74	‒0.34	0.12	380
EBA2 – Modifies service approach with foreign guests	3.76	0.77	‒0.28	0.05	380
EBA3 – Handles cultural misunderstandings constructively	3.69	0.80	‒0.21	‒0.09	380
EBA4 – Integrates global service protocols confidently	3.73	0.76	‒0.31	0.08	380
EBA5 – Retains local warmth within standardised service	3.84	0.73	‒0.39	0.15	380
Employee-Perceived Global-Local Integration (GLI)
GLI1 – Delivers globally standardised service quality	3.66	0.84	‒0.18	‒0.17	380
GLI2 – Preserves Baghdadi cultural identity in service	3.88	0.70	‒0.44	0.23	380
GLI3 – Reconciles digital protocols with local customs	3.62	0.86	‒0.15	‒0.21	380
GLI4 – Maintains cultural authenticity during peak demand	3.71	0.79	‒0.27	0.01	380
GLI5 – Views global standards as enhancement, not erasure	3.79	0.75	‒0.36	0.10	380

[i] Note. All items scored on a 5-poınt Likert scale. Skewness and kurtosis values within ±2.0 indicate acceptable normality for PLS-SEM, which is a distribution-free technique.

[ii] Source: Primary survey data analysis compiled by the author (2026).

Table 4:

Reliability, Convergent Validity, and Indicator-Level VIF (N = 380)

Construct	Item	Loading	α	CR	AVE	Indicator VIF
CCL	CCL1	0.825	0.864	0.895	0.632	1.84
	CCL2	0.791				1.72
	CCL3	0.810				1.78
	CCL4	0.755				1.63
	CCL5	0.768				1.67
	CCL6	0.742				1.61
EBA	EBA1	0.788	0.841	0.880	0.596	1.74
	EBA2	0.742				1.61
	EBA3	0.801				1.77
	EBA4	0.756				1.64
	EBA5	0.773				1.68
GLI	GLI1	0.812	0.855	0.892	0.624	1.80
	GLI2	0.790				1.74
	GLI3	0.774				1.70
	GLI4	0.785				1.72
	GLI5	0.798				1.75

[i] Note. α = Cronbach’s Alpha; CR = Composite Reliability; AVE = Average Variance Extracted. The VIF column reports indicator-level (outer-model) VIFs, which assess multicollinearity among the indicators within each construct; these are distinct from the construct-level full collinearity VIFs used to evaluate common method variance (see Section 3.3 and Table 1). All loadings, α, and CR values exceed the 0.70 threshold; all AVE values exceed 0.50 (Hair et al., 2022). All indicator VIF values are below 3.3.

Table 5:

Discriminant Validity: Fornell-Larcker Criterion

Construct	CCL	EBA	GLI
CCL	0.795
EBA	0.543	0.772
GLI	0.485	0.610	0.790

[i] Note. Diagonal entries represent the square root of the AVE for each construct; off-diagonal entries are inter-construct correlations. Each diagonal value exceeds every correlation in its row and column, satisfying the Fornell-Larcker criterion.

Table 6:

Discriminant Validity: HTMT Ratios with 90% Bootstrap Confidence Intervals (5,000 resamples)

Construct Pair	HTMT	CI 90% Low	CI 90% High	Threshold
CCL ↔ EBA	0.598	0.512	0.679	< 0.85
CCL ↔ GLI	0.531	0.441	0.616	< 0.85
EBA ↔ GLI	0.661	0.578	0.737	< 0.85

[i] Note. HTMT = Heterotrait-Monotrait ratio of correlations. Confidence intervals are bias-corrected and derived from 5,000 bootstrap resamples. All HTMT point estimates fall below the conservative 0.85 threshold, and no upper confidence bound approaches 1.0 (Henseler et al., 2015; Hair et al., 2022).

Table 7:

Structural Path Coefficients — Direct Effects (Bootstrapping: 5,000 resamples, bias-corrected)

H	Structural Path	β (Original Sample)	STDEV	t (\|O/STDEV\|)	p	95% CI Low	95% CI High	Decision
H1	CCL → EBA	0.543	0.036	15.128	<0.001	0.466	0.605	Supported
H2	EBA → GLI	0.470	0.044	10.561	<0.001	0.377	0.553	Supported
—	CCL → GLI (direct)	0.183	0.047	3.895	<0.001	0.088	0.274	Sig. (direct component of partial mediation)

[i] Note. β = standardised path coefficient (original sample estimate); STDEV = standard deviation of the bootstrap distribution; t = bootstrapped t-statistic; CI = 95% bias-corrected confidence interval, reproduced from the SmartPLS 4 bootstrapping report. None of the intervals includes zero. H3 (mediation) is tested separately in Table 9. Full extracts from the original output are provided in Supplementary File S1.

Table 8:

Model Fit and Predictive Accuracy Indices

Index	Value	Recommended Threshold / Interpretation
R² (EBA)	0.294	Variance in EBA explained within this model (Hair et al., 2022, treat > 0.10 as acceptable); not a measure of causal explanation
R² (GLI)	0.421	Variance in GLI explained within this model; not a measure of causal explanation
Adjusted R² (EBA)	0.292	> 0.10
Adjusted R² (GLI)	0.418	> 0.10
f² (CCL → EBA)	0.421	0.02 small; 0.15 medium; 0.35 large (Cohen, 1988)
f² (EBA → GLI)	0.317	0.02 small; 0.15 medium; 0.35 large (Cohen, 1988)
Q² (EBA — Stone-Geisser, blindfolding)	0.168	> 0; in-sample predictive relevance, not an out-of-sample test
Q² (GLI — Stone-Geisser, blindfolding)	0.254	> 0; in-sample predictive relevance, not an out-of-sample test
SRMR	0.047	< 0.08; reported as one supporting diagnostic, not definitive proof of model validity (Henseler et al., 2016)

[i] Note. R² values reflect the proportion of variance in each endogenous construct explained by its predictors within the specified model. f² reflects Cohen’s (1988) effect-size classification. Q2 values from blindfolding indicate in-sample predictive relevance. SRMR is reported cautiously, consistent with guidance that it should not be treated as definitive evidence of overall PLS-SEM model validity.

[ii] Source: Authors‘ PLS-SEM output generated via SmartPLS 4; standard fit thresholds adapted from Cohen (1988), Hair et al. (2022), and Henseler et al. (2016).

Table 9:

Mediation Analysis: Indirect Effect of CCL on GLI via EBA (Bootstrapping: 5,000 resamples, bias-corrected)

Mediation Path	Direct (c′)	Indirect (ab)	STDEV	t (\|O/STDEV\|)	p	Total (c)	VAF (%	95% CI Low	95% CI High
CCL → EBA → GLI	0.183A	0.255	0.031	8.293	<0.001	0.438	0.438	0.196	0.318

[i] Note. Variance Accounted For by the indirect effect = indirect/total effect. The specific indirect effect (β = 0.255) equals the product of the component paths reported in Table 7 (0.543 × 0.470 = 0.255), and the total effect (0.438) equals the sum of the direct and indirect effects (0.183 + 0.255). All values are reproduced from the SmartPLS 4 bootstrapping report (Final results → Specific indirect effects; Final results → Total effects); extracts from the original output are provided in Supplementary File S1. A VAF of 58.2% indicates partial mediation, and the bias-corrected confidence interval for the indirect effect excludes zero, consistent with a statistically significant mediation pathway.

Appendix A Table:

Measurement Items, Sources, and Response Format

Code	Item wording (English version)	Source / basis
CCL1	My manager adapts his or her communication style when interacting with people from different cultural backgrounds.	Ang et al. (2007); Bukhari et al. (2025)
CCL2	My manager openly acknowledges cultural differences within the team and with guests.	Ang et al. (2007); Bukhari et al. (2025)
CCL3	My manager involves employees from different backgrounds in decisions that affect service delivery.	Ang et al. (2007), adapted
CCL4	My manager shows genuine empathy toward staff when cultural expectations conflict with work requirements.	Ang et al. (2007), adapted
CCL5	My manager introduces new digital work procedures in a way that respects our local customs and working habits.	Developed for this study; informed by Aruppala et al. (2026)
CCL6	My manager behaves in a manner consistent with the values of our local community.	Expert panel addition
EBA1	I adjust the way I communicate depending on the cultural background of the guest I am serving.	Black & Stephens (1989); Shaffer et al. (2021)
EBA2	I modify my service approach when dealing with foreign guests in order to meet their expectations.	Black & Stephens (1989); Shaffer et al. (2021)
EBA3	When a cultural misunderstanding occurs with a guest, I handle it calmly and constructively.	Shaffer et al. (2021), adapted
EBA4	I feel confident applying international service protocols in my daily work.	Shaffer et al. (2021), adapted
EBA5	I maintain the warmth of traditional Baghdadi hospitality even when following standardised service procedures.	Developed for this study (cultural maintenance dimension)
GLI1	Our hotel delivers service that meets internationally recognised quality standards.	Ogunnaike et al. (2022), adapted
GLI2	Our hotel preserves the Baghdadi cultural identity in the way service is delivered.	Ogunnaike et al. (2022), adapted
GLI3	In our hotel, digital service systems are reconciled with local customs rather than replacing them.	Developed for this study; informed by Aruppala et al. (2026)
GLI4	Our hotel maintains its cultural authenticity even during periods of peak demand.	Developed for this study
GLI5	In our hotel, global standards are treated as an enhancement of, not a substitute for, our local hospitality traditions.	Developed for this study