Human Resource Development and Compensation Effects on Village Apparatus Performance: Evidence from Eastern Indonesia

3 Method

3.1. Research Design and Context

A quantitative, cross-sectional survey design with an explanatory purpose was employed to examine the theory-driven hypotheses regarding the relationships between HRD, compensation, and performance (Creswell & Creswell, 2018). This design aligns with organizational behavior research frameworks aimed at strictly examining variable associations (Podsakoff et al., 2012).

The study was conducted in Manulea Village, Sasitamean District, Malaka Regency, East Nusa Tenggara, from June to July 2025. The site was selected based on four purposive criteria: (a) documented performance challenges, specifically the inability to achieve full completion targets (stagnating at 95-98% for key infrastructure); (b) severe resource constraints, characterized by minimal office facilities and low training participation; (c) socio-economic representativeness of Malaka's 127 villages within a newly autonomous border region with a 13.92% poverty rate (BPS Malaka, 2024; Sutikno, 2025); and (d) leadership willingness to support data collection (Seran et al., 2021).

3.2. Sample and Procedure

The study population comprised the entire cohort of 70 formally appointed village officials (including village heads, secretaries, section heads, hamlet chiefs, and administrative staff), explicitly excluding volunteer roles such as RT/RW. Given the manageable population size, a census (saturated sampling) method was employed (Tabachnick & Fidell, 2019).

To ensure statistical adequacy, an a priori power analysis was conducted using G*Power 3.1.9.7. Parameters were set for a medium effect size (f² = 0.15), an alpha level of α = .05 and a power of 1 - β = .80 with two predictors. The analysis indicated a minimum required sample size of n = 68 (Faul et al., 2007). The final census sample (N = 70) exceeds this threshold and adheres to general guidelines suggesting 50-100 observations for regression models with few predictors (Hair et al., 2019).

Structured questionnaires were personally administered in June-July 2025 following ethics approval from University of Timor. Permissions were secured from district and village authorities, and all participants provided written informed consent. To mitigate Common Method Bias (CMB), procedural remedies included assuring respondent anonymity, separating predictor and outcome variables with filler items, randomizing item order, and using clear Bahasa Indonesia to reduce ambiguity (Podsakoff et al., 2012). This direct administration strategy resulted in the return of all 70 completed questionnaires, yielding a 100% response rate.

3.3. Measures

All constructs were measured using validated scales adopted from established studies. To ensure linguistic and conceptual equivalence, the instruments followed Brislin's (1970) back-translation protocol. A panel of four experts reviewed the translation and back-translation processes, with discrepancies resolved via consensus. Responses were recorded on a seven-point Likert scale, ranging from 1 (strongly disagree/never) to 7 (strongly agree/always), depending on the item type.

Human Resource Development (HRD). Measured using six items adapted from Swanson and Holton (2009), the scale assessed training frequency, skill development, educational support, mentoring, and competency enhancement. The measurement model demonstrated robust internal consistency and validity (α = .89, CR = .90, AVE = .61).

Compensation. This construct utilized eight items adapted from Milkovich et al. (2014), covering both financial and non-financial dimensions (e.g., salary adequacy, allowances, facilities, equipment, and work environment). The scale showed strong psychometric properties (α = .92, CR = .93, AVE = .63).

Village Apparatus Performance. Performance was evaluated using nine items adapted from Moeheriono (2012) to capture work quality, quantity, timeliness, accuracy, and responsibility. The scale exhibited high reliability and validity (α = .94, CR = .95, AVE = .67).

Control Variables. To account for potential confounding effects (Garavan et al., 2012), the study controlled for: gender (0 = female, 1 = male); age (in years); education (1 = high school, 2 = diploma, 3 = bachelor's); tenure (in years); and position level (1 = non-supervisory, 2 = hamlet chief, 3 = section head/senior).

3.4. Analytical Strategy

Data were analyzed using SPSS 26.0 (IBM). The analysis proceeded in four phases. Phase 1 involved preliminary analysis, including descriptive statistics, correlations, normality checks (skewness and kurtosis within ± 2), multicollinearity assessment (VIF < 3.0), and outlier detection (Hair et al., 2019). Phase 2 focused on measurement validation via Exploratory Factor Analysis (principal axis factoring with promax rotation), Cronbach's α, Composite Reliability (CR), Average Variance Extracted (AVE), and discriminant validity using the Fornell-Larcker criterion and HTMT ratios (Fornell & Larcker, 1981; Henseler et al., 2015). Phase 3 assessed Common Method Bias (CMB) using Harman's single-factor test and the Common Latent Factor (CLF) method with procedural remedies (Podsakoff et al., 2012). Finally, Phase 4 involved hypothesis testing via simple and multiple regression, interpreting effect sizes as small (R² = .02), medium (R² = .13), and large (R² = .26) (Cohen, 1988).

4 Results

4.1. Descriptive Statistics and Correlations

Table 2 presents the descriptive statistics, reliability coefficients, and bivariate correlations. The participants reported moderate-to-high evaluations across all constructs. Performance (M = 5.12, SD = 0.97) received the highest rating, while Compensation (M = 3.87, SD = 1.34) was rated lowest, potentially reflecting the resource constraints and facility deficiencies noted in the study setting.

Bivariate correlations supported the theoretical expectations. HRD was positively and significantly correlated with Performance (r = .64, p < .001), and Compensation showed a strong positive association with Performance (r = .69, p < .001). Additionally, a moderate positive correlation was observed between HRD and Compensation (r = .52, p < .001). Multicollinearity was not a concern, as all Variance Inflation Factor (VIF) values were below the threshold of 2.0 (Hair et al., 2019).

4.2. Measurement Validation and Common Method Bias Assessment

The measurement model was first assessed via Exploratory Factor Analysis (EFA). The results yielded a three-factor solution with eigenvalues greater than 1.0, explaining 68.4% of the total variance. Factor loadings ranged from .65 to .89 on the intended constructs, with no significant cross-loadings, thereby establishing initial construct validity (Hair et al., 2019).

Convergent validity and internal consistency were confirmed, as all constructs exhibited Cronbach's α values between .89 and .94, Composite Reliability (CR) ranging from .90 to .95, and Average Variance Extracted (AVE) values between .61 and .67 (Fornell & Larcker, 1981).

Discriminant validity was assessed using two criteria. First, the Fornell-Larcker criterion was met, as the square root of the AVE for each construct (.78, .79, .82) exceeded the highest inter-construct correlation (.69).

Second, the Heterotrait-Monotrait (HTMT) ratios ranged from .68 to .76, well below the conservative threshold of .85 (Henseler et al., 2015), providing robust evidence of discriminant validity.

To address potential Common Method Bias (CMB), Harman's single-factor test indicated that the largest unrotated factor explained 38.7% of the variance, below the 50% threshold. Furthermore, a Common Latent Factor (CLF) analysis showed that differences in standardized regression weights with and without the CLF were minimal (<.15). Combined with the procedural remedies described earlier, these results suggest that CMB is not a pervasive threat in this study (Podsakoff et al., 2012).

4.3. Hypothesis Testing

The regression analysis results, summarized in Table 6, confirm all proposed hypotheses. First, Human Resource Development (HRD) demonstrated a significant positive effect on Performance (β = .64, t(68) = 6.91, p < .001), with an R² of .413, supporting H1. Second, Compensation also showed a strong positive influence on Performance (β = .69, t(68) = 7.84, p < .001), explaining 47.5% of the variance (R² = .475) and supporting H2.

When tested simultaneously (H3), the model remained highly significant (F(2, 67) = 42.22, p < .001). The combination of HRD and Compensation explained 55.8% of the variance in Village Apparatus Performance (R² = .558). In this multiple regression model, Compensation emerged as the dominant predictor (β = .48, p < .001) compared to HRD (β = .35, p < .001), although both remained statistically significant. These effect sizes substantially exceed Cohen's (1988) threshold for large effects.

4.4. Robustness Checks

To ensure the stability of the results, a series of robustness checks were conducted. First, influence diagnostics via Cook's distance revealed no influential outliers; sensitivity analyses excluding potential high-leverage cases yielded consistent conclusions. Second, the assumption of linearity was confirmed, as the inclusion of quadratic terms proved non-significant (p > .10). Third, the Breusch-Pagan test indicated the presence of homoscedasticity, and residual analysis confirmed an approximation to normality (skewness and kurtosis within ± 2). Finally, controlling for demographic covariates (gender, age, education, tenure, position) resulted in negligible shifts in the primary coefficients (Δβ < .03), demonstrating that the main findings remain robust across model specifications (Tabachnick & Fidell, 2019; Hair et al., 2019).

5 Discussion

5.1. Summary of Findings and Theoretical Contributions

The study provides empirical evidence that Human Resource Development (HRD) and compensation are critical determinants of village apparatus performance, even within the resource-constrained context of a border regency. Consistent with human capital theory (Salas et al., 2012; Garavan et al., 2012), HRD emerged as a significant predictor, suggesting that competence-building interventions directly translate into improved bureaucratic output. Notably, compensation exhibited a slightly stronger individual effect than HRD. This finding aligns with Expectancy Theory, implying that in a setting characterized by limited infrastructure, tangible rewards and facility improvements serve as powerful "hygiene factors" and motivators that are prerequisites for high performance (Gerhart & Fang, 2014; Syamsir, 2025).

Theoretically, these findings extend the boundary conditions of Human Capital and Expectancy Theories traditionally grounded in corporate, resource-rich settings to the distinct context of public administration in an emerging market border region. The results illustrate that capability and motivation mechanisms remain robust even under severe resource scarcity. Furthermore, the synergistic effect observed in the combined model supports an integrated framework: capability-building (HRD) and motivation-enhancing (compensation) interventions function as complementary rather than substitutable forces. Neglecting one diminishes the efficacy of the other (Perry et al., 2010; Crook et al., 2011).

5.2. Practical Implications

Four actionable implications emerge for policy and practice. First, district governments should institutionalize a specific regulatory framework mandate, such as earmarking a minimum 5% of the Village Fund (Dana Desa) specifically for HRD. This allocation should fund critical competency training including financial management applications (SISKEUDES), participatory planning (Musrenbang), and regulatory compliance as well as scholarship programs for village officials seeking higher education.

Second, regarding compensation and facilities, national and district authorities must define Minimum Service Standards (SPM) for village office infrastructure. Given the strong correlation between tangible resources and performance in deprived settings, standards should ensure adequate digital tools (e.g., a target ratio of one computer per five officials) and reliable internet access to support administrative duties.

Third, policy design must shift from fragmented interventions to a holistic approach. The findings demonstrate that simultaneous investments in HRD and compensation yield synergistic returns that exceed the sum of individual interventions. Finally, strategic priority should be accorded to border and least-developed regions like Malaka. In these areas, the marginal utility of human resource investments is highest, offering the greatest potential for accelerating bureaucratic reform (Lewis, 2015; Sutikno, 2025).

5.3. Limitations and Future Research Directions

Several limitations warrant consideration when interpreting these findings. First, the cross-sectional design precludes the definitive establishment of causal inferences. Future scholarship should employ longitudinal or quasi-experimental approaches to rigorously track how HRD interventions impact performance over time (Podsakoff et al., 2012). Second, while the census method ensured high internal validity, the study's focus on a single administrative locus restricts external validity. Future research should expand to multi-site studies across Malaka's 127 villages or other border provinces to test boundary conditions and generalizability (Creswell & Creswell, 2018).

Third, reliance on self-reported measures may introduce response bias, despite the procedural controls for Common Method Variance (CMV). To enhance robustness, future studies should integrate objective performance metrics, such as budget realization rates (SISKEUDES logs) and administrative audit scores.

Finally, to refine the theoretical framework, future models should examine psychological mediators (e.g., public service motivation, self-efficacy) and contextual moderators specific to the region (e.g., leadership style, poverty levels, and proximity to cross-border checkpoints) (Perry et al., 2010; Knies et al., 2018).

6 Conclusion

This study establishes that Human Resource Development (HRD) and compensation are decisive determinants of village apparatus performance in the border region of Malaka Regency. The findings reveal that HRD investments are critical for enhancing officials' capability to navigate complex administrative demands, while compensation particularly physical work facilities serves as a fundamental motivational enabler. The substantial combined effect of these variables underscores that capability-building and motivation-enhancing mechanisms must work in tandem to achieve optimal bureaucratic output.

Theoretically, the research extends the applicability of Human Capital and Expectancy Theories beyond their traditional corporate boundaries to the resource-constrained context of rural public administration. Practically, it offers an evidence-based framework for district governments to prioritize specific HRD budgetary allocations and standardize compensation infrastructure. Ultimately, the study argues that strategic investment in the capacity and well-being of frontline officials is a prerequisite for realizing Indonesia's ambitious rural development and decentralized governance goals (Lewis, 2015; Sutiyo & Maharjan, 2017; World Bank, 2020).

D Declarations

R References