When Clarity Outweighs Capacity: PLS-SEM Evidence on Rerun Election Implementation in Tasikmalaya, Indonesia

Abstract

Rerun elections (Pemungutan Suara Ulang/PSU) are corrective electoral policies implemented under strict judicial deadlines and intense public scrutiny. However, empirical evidence on the factors associated with their operational effectiveness at the regency level remains limited. This study examines the associations of policy quality, institutional capacity, external support, and administrator commitment with perceived PSU effectiveness. Cross-sectional questionnaire data were obtained from 240 election administrators across 39 sub-districts in Tasikmalaya Regency and analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM). Policy quality showed the strongest association with perceived effectiveness (β = .466, p < .001), followed by administrator commitment (β = .241, p < .001) and external support (β = .170, p = .004), whereas institutional capacity was not significant. The model explained 53.2% of the variance in perceived effectiveness (R² = .532). Importance-Performance Map Analysis identified regulatory misinterpretation and low public enthusiasm as priority areas for further attention. These results support a context-bound proposition that regulatory clarity may function as an enabling condition in time-constrained PSU implementation. The findings are based on administrator perceptions in one regency and should be validated using multi-source and cross-regency evidence.

Keywords: policy implementation; electoral governance; rerun election; electoral effectiveness; street-level bureaucracy; PLS-SEM.

Introduction

The implementation of rerun elections (PSU) following Constitutional Court rulings differs from ordinary election administration because it requires a complete corrective electoral process within a fixed judicial deadline, after political contestation has already produced dispute, polarization, and heightened public scrutiny. In Tasikmalaya Regency, Constitutional Court Decision No. 132/PHPU.BUP-XXIII/2025 required administrators to coordinate a PSU within 60 days following a change in the contestant configuration. Tasikmalaya is analytically relevant because it was the only district in West Java mandated to conduct a PSU in this cycle, allowing examination of policy implementation under simultaneous legal urgency, local political pressure, and compressed cross- institutional coordination (Indonesia at Melbourne, 2024). The Constitutional Court ordered PSUs in 24 regions following the 2024 Local Elections, with deadlines ranging from 30 to 60 days (Komisi Pemilihan Umum Republik Indonesia, 2025), making this corrective mechanism an increasingly important electoral governance issue. Institutional readiness in Tasikmalaya Regency was rated relatively highly, yet administrators reported variation in operational effectiveness that could not be understood solely through budget and logistics. This pattern suggests that perceived PSU effectiveness may also depend on how regulations are interpreted, how external actors support implementation, and how administrators respond to pressure. Indonesian electoral governance studies remain largely juridical-normative and focused on formal Rerun elections (Pemungutan Suara Ulang/PSU) are corrective electoral policies implemented under strict judicial deadlines and intense public scrutiny. However, empirical evidence on the factors associated with their operational effectiveness at the regency level remains limited. This study examines the associations of policy quality, institutional capacity, external support, and administrator commitment with perceived PSU effectiveness. Cross-sectional questionnaire data were obtained from 240 election administrators across 39 sub-districts in Tasikmalaya Regency and analyzed using Partial Least Squares Structural Equation Modeling (PLS-SEM). Policy quality showed the strongest association with perceived effectiveness (β = .466, p < .001), followed by administrator commitment (β = .241, p < .001) and external support (β = .170, p = .004), whereas institutional capacity was not significant. The model explained 53.2% of the variance in perceived effectiveness (R² = .532). Importance-Performance Map Analysis identified regulatory misinterpretation and low public enthusiasm as priority areas for further attention. These results support a context-bound proposition that regulatory clarity may function as an enabling condition in time-constrained PSU implementation. The findings are based on administrator perceptions in one regency and should be validated using multi-source and cross-regency evidence. procedures and compliance (Khaira et al., 2025). Empirical studies examining associations among implementation variables at the regency level remain limited, particularly for PSUs that are strictly time-bound, politically sensitive following candidate disqualification, and dependent on rapid adaptive responses without compromising electoral integrity. This empirical and methodological gap forms the starting point of the study. The research question is: To what extent are policy quality, institutional capacity, external support, and administrator commitment associated with perceived PSU effectiveness, and which indicators warrant prioritized attention? The study examines predictive associations among these constructs through a cross-sectional quantitative design using PLS-SEM and data from 240 election administrators across 39 sub-districts. The study offers three context-specific contributions. Empirically, it examines regency-level PSU implementation under a judicial deadline, a setting that has received limited quantitative attention in Indonesian electoral governance research. Methodologically, it estimates a multi-construct predictive model using PLS-SEM and transparently reports post hoc measurement refinement, while IPMA is used to identify indicator-level priorities rather than merely ranking structural coefficients. Theoretically, it evaluates a context- bound proposition that statutory clarity, organizational capacity, external support, and street-level commitment may occupy different functional positions in time-constrained implementation. Accordingly, the study does not claim a universal hierarchy; it tests whether regulatory clarity appears more strongly associated with perceived effectiveness than resource adequacy in the Tasikmalaya PSU context. Public policy implementation under emergency conditions demands bureaucratic adaptation speeds far exceeding normal procedures. The classical model developed by Mazmanian & Sabatier (1983) identifies three clusters of determinants of implementation success: statutory variables (goal clarity, regulatory consistency, technical guidance availability), non-statutory variables (bureaucratic capacity, environmental support), and socioeconomic conditions and target group characteristics. In the PSU context, Constitutional Court decisions and KPU SOPs function as statutory variables that must be transformed into operational actions within a highly constrained timeframe. This pressure creates implementation conditions structurally distinct from conventional public policies: administrators lack the luxury of learning by doing, and any procedural ambiguity can directly compromise electoral outcomes. Edwards III (1980) emphasizes that clarity of objectives provides an important foundation for field implementers to make decisions under time pressure. Mubarok et al. (2020), in a comparative analysis of policy implementation models in Indonesia, support the relevance of the Mazmanian and Sabatier framework in the local context while noting that its application remains largely descriptive-conceptual and that simultaneous quantitative examination of implementation variables is limited. Accordingly, within the literature reviewed for this manuscript, quantitative testing of classical implementation models in Indonesia’s emergency electoral context appears limited. This corpus-bounded observation should not be interpreted as an exhaustive claim about all published or unpublished research. The gap is particularly relevant to contexts involving incumbent disqualification, political polarization, a 60-day implementation window, and procedural uncertainty arising from changes in contestant configurations. The framework’s relevance to PSU rests on the proposition that regulatory ambiguity may hinder consistent and timely implementational decision-making, even when logistical and budgetary resources are adequate. In Tasikmalaya’s PSU context, the indicator ‘absence of regulatory misinterpretation’ recorded the lowest mean (M = 3.692). This descriptive result suggests that regulatory interpretation was perceived as a practical challenge by administrators, but it does not independently establish that ambiguity caused implementation delays. Direct process evidence, longitudinal observation, or qualitative accounts are needed to test the proposed mechanism.

Institutional Theory and Normative Pressures in Electoral Governance

Institutional theory provides a productive lens for understanding why election management organizations behave as they do, particularly under high-pressure conditions. DiMaggio & Powell (1983) introduced institutional isomorphism mechanisms to explain organizations’ tendencies to adopt similar structures and practices within a field, driven by coercive, normative, and mimetic pressures. Meyer & Rowan (1977) earlier demonstrated that formal organizational structures often reflect institutional myths that sustain legitimacy rather than technical needs. In electoral governance, this proposition has relevant implications: KPU and Bawaslu, operating under intense public legitimacy pressure, tend to adopt procedures and structures deemed institutionally appropriate, even if actual field practices are far more complex. Scott (2008) refined this framework by distinguishing three mutually reinforcing institutional pillars. The regulative pillar refers to legal mandates and sanction mechanisms enforcing compliance; the normative pillar concerns professional expectations, values, and agreed behavioral standards; the cognitive pillar encompasses how administrators internalize electoral procedures into daily action schemas. In the PSU context, all three pillars operate simultaneously but with varying intensity. The regulative pillar manifests through Constitutional Court rulings and KPU SOPs; the normative pillar emerges through Forkopimda support, political parties, and professional administrator communities; the cognitive pillar materializes in how PPK and Panwascam officers internalize PSU as a democratic restoration mission rather than a mere technical voting repetition. James et al. (2019), in a cross-national study examining organizational determinants of electoral integrity, found that EMB institutional design, performance capacity, and electoral integrity interrelate in non-linear relationships. Garnett (2019) adds that resource and personnel data only partially capture EMB capacity; what matters more is how capacity translates into communication, coordination, and field responsiveness. This helps explain why institutional capacity, rated very highly by respondents (M = 4.243), does not automatically correlate with PSU effectiveness variation. Recent literature also cautions that external support does not always function as a linear predictor of implementation effectiveness. When regional government or security apparatus support is ceremonial and lacks structured technical coordination, the variable tends to act as an enabler, facilitating working conditions rather than directly determining outcome quality. This suggests that in emergency electoral contexts, the regulative pillar (SOP clarity) and cognitive pillar (administrator commitment and value internalization) are more determinant than mere normative pressure from external actors.

Street-Level Bureaucracy and Administrator Discretion

The street-level bureaucracy (SLB) framework developed by Lipsky (2010) positions field implementers not as passive regulation executors, but as actors with substantive discretion in translating policies into real actions. In the PSU context, sub- district level administrators (PPK, Panwascam) face classic SLB conditions: high workload, resources limited relative to task complexity, and regulations that rarely cover all field scenarios. Their exercised discretion—whether handling procedural ambiguity or responding to local political pressure—directly shapes implementation quality not fully captured by top-down policy design. Hupe & Hill (2007) reinforce this by arguing that public accountability in policy implementation can only be adequately understood if field discretion is treated as a substantive variable rather than a deviation to be fully controlled. In PSU, this means administrator integrity and moral resilience are not merely individual psychological variables; they constitute invisible infrastructure sustaining process credibility, especially under political polarization that heightens partisan temptations. However, Lotta et al. (2022) caution that SLB propositions require re-examination when applied to developing countries, where institutional conditions fundamentally differ from the theory’s origins. Peeters & Campos (2023) strengthen this warning through a systematic review identifying four institutional differentiators in weak institutional contexts: public trust levels in the state, bureaucratic politicization, precarious working conditions, and limited state capacity, forcing frontline workers to improvise. At least two are relevant to Tasikmalaya’s PSU: shaken public trust post- incumbent disqualification and informal politicization pressure from local elites with stakes in PSU outcomes. Notably, the SLB perspective reveals that discretion does not automatically produce deviation. In high-risk electoral situations, as shown by empirical studies across developing nations, administrators with strong moral commitment actually use discretion as an integrity-preserving mechanism when regulations fail to anticipate all field scenarios (Peeters & Campos, 2023). Thus, administrator commitment is not separate from institutional capacity but complementary, determining whether available capacity truly translates into effective and legitimate action.

Literature Gaps and Research Positioning

The literature indicates an interconnected empirical, methodological, and theoretical gap. Empirically, Indonesian PSU studies remain dominated by juridical-normative analyses of legal mechanisms and formal compliance (Khaira et al., 2025; Mubarok et al., 2020), providing limited evidence on why operational effectiveness may vary among administrators working under the same formal mandate. Methodologically, few studies estimate multiple implementation factors simultaneously using administrator- level data. Theoretically, classical implementation models identify relevant variables but do not specify how their relative importance may change when implementation is compressed by judicial deadlines and political polarization. International electoral governance research has advanced understanding of electoral management body design and capacity (Garnett, 2019; James et al., 2019), but it is concentrated at national or cross-country levels. Subnational corrective elections remain underexamined even though they differ from regular elections in three respects: they operate under rigid judicial deadlines, begin within an already polarized environment, and require accelerated coordination among election bodies, regional government, security institutions, political parties, and the public. These conditions make it necessary to examine not only whether resources are available, but also whether regulations are sufficiently clear and whether frontline administrators can translate them consistently into action. PLS-SEM has also been used only sparingly in Indonesian electoral administration research. Its predictive orientation and indicator-level extension through IPMA are useful for exploring which perceived implementation dimensions are most strongly associated with the outcome and which indicators may deserve further managerial attention (Hair et al., 2022; Shmueli et al., 2019). Nevertheless, IPMA priorities are treated here as evidence-informed diagnostic signals rather than definitive policy prescriptions. The proposed model integrates the three perspectives through a division of explanatory roles. Policy implementation theory identifies statutory clarity and organizational capacity as core implementation conditions; institutional theory explains the external regulative and normative environment surrounding election management bodies; and street-level bureaucracy explains how administrator commitment and discretion shape the translation of rules into field practice. These perspectives are combined because PSU implementation simultaneously involves formal rules, organizational resources, external legitimacy, and frontline judgment. The study therefore examines four constructs in one predictive model while maintaining their distinct theoretical origins. Synthesizing these frameworks yields four directional hypotheses concerning predictive associations with perceived PSU effectiveness. The hypotheses are not treated as proof of causality because the data are cross-sectional and obtained from a single respondent source. Policy quality, operationalized as PSU regulation clarity, consistency, and ease of application, is the most fundamental statutory variable in Mazmanian & Sabatier (1983) framework. In strictly time-bound electoral policy contexts, regulatory clarity functions as a cognitive framework enabling field administrators to make rapid operational decisions without ambiguity, as emphasized by Edwards III (1980). Thus: H1: Policy quality is a positive and significant predictor of PSU effectiveness. Institutional capacity reflects organizational resource adequacy, including human resources availability, budget, logistics, and technical training. Mazmanian & Sabatier (1983) position resource adequacy as a key determinant of policy tractability, so capacity sufficiency should theoretically correlate positively with implementation success. Garnett (2019) also identifies EMB capacity as a vital electoral governance variable, albeit multidimensionally measured. Based on this: H2: Institutional capacity is a positive and significant predictor of PSU effectiveness. External support encompasses normative and regulative support from actors outside election management organizations, including regional governments, security apparatus, political parties, and the public. Institutional theory asserts that normative pressure from external environments strengthens organizational compliance with formal procedures and enhances implementation legitimacy (DiMaggio & Powell, 1983; Meyer & Rowan, 1977; Scott, 2008). In PSU operating under public scrutiny and potential local conflict, external ecosystem support becomes a prerequisite for conducive execution. Therefore, H3: External support is a positive and significant predictor of PSU effectiveness. Administrator commitment refers to the integrity, professionalism, and moral resilience of implementers conducting PSU without partisanship. Lipsky (2010) emphasizes that street-level bureaucrats’ discretion is a vital adjustment mechanism when regulations fail to anticipate all field scenarios. In post-disqualification political polarization, administrator moral commitment functions as a process credibility stabilizer irreplaceable by any formal infrastructure (Hupe & Hill, 2007; Peeters & Campos, 2023). Lotta et al. (2022) caution that this variable requires re-testing in Global South institutional contexts where political pressure on field implementers operates differently from conventional SLB assumptions. Thus: H4: Administrator commitment is a positive and significant predictor of PSU effectiveness.

Methods

This study uses a cross-sectional quantitative explanatory design to estimate predictive associations among latent

Construct Theoretical Dimension Example Indicators Adaptatio n Source Policy Quality (KUKE) Goal clarity, regulatory consistency , and technical guidance availability “Constitutional Court ruling easily understood”, “PSU SOP clear”, “No regulatory misinterpretation .”

Mazmani an & Sabatier (1983); Edwards III (1980) Institution al Capacity (SITAS) Human resources availability, budget, logistics, and technical training “Budget available on time”, “Logistics complete”, “Technical training provided.”

Mazmani an & Sabatier (1983); Garnett (2019) External Support Normative & regulative pressure from actors outside administrat ors “Regional government supports budget”, “Security apparatus active”,” Public enthusiastic.”

DiMaggio & Powell (1983); Scott (2008) Administr ator Commitm ent (KOM) Integrity, neutrality, moral resilience, and code of ethics compliance “Non-partisan”, “Able to withstand pressure”, “Complies with code of ethics.”

Lipsky (2010); Hupe & Hill (2007) variables. Partial Least Squares Structural Equation Modeling (PLS-SEM) was performed using SmartPLS 4. PLS-SEM was selected because the study emphasizes prediction and indicator-level diagnosis, includes reflective constructs, and uses perceptual data that may not satisfy multivariate normality assumptions. It also permits Importance- Performance Map Analysis (IPMA). These advantages were considered more aligned with the study objective than covariance-based SEM, although the design cannot establish temporal or causal ordering (Hair et al., 2019; Ringle & Sarstedt, 2016). The target population comprised all 240 PSU administrators in Tasikmalaya Regency: Kabupaten KPU (31), Kabupaten Bawaslu (34), PPK across 39 sub-districts (109), and Panwascam (66). Because the analyzed dataset included the full stated population, the study used a census approach rather than probability sampling. The analysis therefore describes predictive relationships within this administrative population. Its statistical adequacy exceeds the minimum PLS-SEM requirements (Kock & Hadaya, 2018), but census coverage within one regency does not eliminate the limitation on generalization to other PSU settings. Data were collected using a structured 1-5 Likert-scale questionnaire adapted from policy implementation theory (Edwards III, 1980; Mazmanian & Sabatier, 1983; Rulandari et al., 2022), institutional theory (DiMaggio & Powell, 1983; Scott, 2008), street-level bureaucracy (Lipsky, 2010), and program evaluation (Rossi et al., 2004). The initial instrument contained indicators covering policy quality, institutional capacity, external support, administrator commitment, and PSU effectiveness, as summarized in Table 1. A pilot test involving 30 respondents outside the analytical census assessed item clarity and preliminary internal consistency. However, the retained documentation does not include detailed pilot reliability coefficients, the expert-review procedure, item-revision records, or a formal content-validity index. The instrument and resulting measurement model should therefore be regarded as preliminary and exploratory- predictive for the PSU context. This limitation reduces confidence in construct-content validity and requires independent validation using documented expert assessment, pilot reliability results, and item-revision procedures. Data analysis proceeded in three standard stages following Hair et al. (2022) and Shmueli et al. (2019): (1) outer model evaluation (convergent validity, reliability, discriminant validity via HTMT and Fornell-Larcker); (2) inner model evaluation (determination coefficient R², effect size f², predictive relevance Q²predict); and (3) hypothesis testing via bootstrapping with 5,000 subsamples at α = .05. Before estimating the final model, a theory-informed post hoc refinement procedure was undertaken. Administrator experience as a moderator and education as a control were removed after non-significant estimates (p > .05); the limited variation in experience, with 90.83% handling a PSU for the first time, reduced the capacity to detect moderation. Four indicators were then removed after joint statistical and conceptual review. EFE1 and EFE5 showed high cross-loadings and appeared to represent implementation inputs rather than substantive effectiveness outcomes; EFE3 showed near-zero variance and could not discriminate respondents; and KOM1 was highly redundant with KOM2 and KOM3. The retained indicators continued to represent the central dimensions of each construct, but these changes may have narrowed construct coverage. The initial model was specified from the theoretical framework and hypotheses before analysis, whereas the indicator deletions and exclusion of moderation and control variables were data-informed modifications. The final estimates should therefore be interpreted as results of a theory-informed post hoc refinement rather than a fully pre- specified confirmatory model. This procedure introduces a risk of capitalizing on sample-specific patterns and requires replication in independent PSU samples. The refinements were intended to improve parsimony, interpretability, and measurement performance while retaining theoretical coherence. Initial-model results and deleted-item diagnostics should accompany the manuscript as supplementary analytical documentation where available. In the absence of independent validation, the final model is treated as exploratory-predictive evidence and not as a definitive measurement structure for PSU implementation. Research ethics procedures included written informed consent, assurances of respondent confidentiality, and an explanation that the data would be used solely for academic purposes and public policy recommendation formulation. Participation was presented as voluntary. The study was conducted under research permission from Garut University. Participants were informed about the study purpose and procedures, and participation was voluntary. The informed consent procedure covered audio recording, confidentiality and anonymity, academic use of the information, secure data storage, the right to withdraw, and the right to request correction or clarification of interview transcripts. The absence of these details limits independent verification of ethical oversight and should be rectified in future reporting through complete institutional approval and data-governance documentation.

Construct Operationalization and Indicators

Table 1. Construct Operationalization and Indicators

Result and Discussion

Measurement Model Evaluation (Outer Model)

Outer model evaluation ensures the instrument measures intended constructs before hypothesis testing. Convergent Outer Loading Cronbach's α CR AVE √AVE Status Policy Quality (KUKE) KUKE1 – Constitutional Court ruling easily understood .809 .806 .872 .631 .794 Valid KUKE2 – PSU SOP clear .840 KUKE3 – No regulatory misinterpretation .811 KUKE4 – Technical guidelines complete .713 Institutional Capacity (SITAS) SITAS1 – Budget available on time .774 .820 .881 .649 .805 Valid SITAS2 – Human resources sufficient .802 SITAS3 – Logistics complete .827 SITAS4 – Technical training provided .817 External Support (EKS) EKS1 – Regional government supports the budget* .516 .677 ** .801 .509 .713 Marginal EKS2 – Security apparatus active .725 EKS3 – Political parties cooperative .816 EKS4 – Public enthusiastic .759 Administrator Commitment (KOM) KOM2 – Non partisan .903 .855 .912 .776 .881 Valid KOM3 – Able to withstand pressure .849 KOM4 – Complies with the code of ethics .889 PSU Effectiveness (EFE) EFE2 – No further lawsuits .800 .595 ** .788 .555 .745 Marginal EFE4 – Voter participation stable* .651 EFE6 – PSU assessed fair & transparent .775 Minimum Threshold ≥ .50 ≥ .70 ≥ .70 ≥ .50 — theoretical domains and preserve content validity despite exhibiting marginal outer loadings (Hair et al., 2022). Shaded cells indicate marginal values requiring careful interpretation—source: SmartPLS 4 output, Primary Data Processed (2026).

conducted using Fornell-Larcker. Source: SmartPLS 4 output, Primary Data Processed (2026). Criteria follow Henseler et al. (2015). validity and reliability results are presented in Table 2. Substantive Interpretation: Most constructs meet the conventional PLS-SEM criteria for convergent validity and composite reliability. Policy Quality, Institutional Capacity, and Administrator Commitment show satisfactory internal consistency. External Support and PSU Effectiveness require cautious interpretation because their Cronbach's alpha values are below .70, despite composite reliability above .70 and AVE above .50. This pattern may reflect the conceptual breadth and heterogeneity of their indicators rather than uniformly strong internal consistency. The measurement evidence is therefore acceptable for exploratory-predictive analysis but not uniformly strong across all constructs. Discriminant validity was primarily assessed using the HTMT criterion (Table 3), while the Fornell-Larcker criterion (Table 4) was used as complementary evidence to support construct distinctiveness. The discriminant validity assessment reveals a substantive limitation. Although the Fornell-Larcker criterion and cross- loading inspection suggest separation among constructs, the HTMT value between Policy Quality and PSU Effectiveness is .946, and its bias-corrected confidence interval (.839-1.046) includes 1.00. These results do not establish discriminant validity between the two constructs with sufficient confidence. Fornell-Larcker evidence cannot be treated as fully resolving this concern because HTMT is generally more sensitive to conceptual overlap. One plausible explanation is that administrators interpreted regulatory clarity as part of effective PSU implementation, creating overlap between an explanatory construct and the perceptual outcome. This interpretation is consistent with the time-constrained context, but it remains an inference rather than a confirmed mechanism. Alternative specifications, redundancy analysis, revised outcome indicators, and multi- source effectiveness measures are needed to determine whether the constructs can be empirically separated more clearly. Accordingly, the Policy Quality-PSU Effectiveness coefficient must be interpreted cautiously. Its magnitude may be partly inflated by conceptual overlap, even though the constructs remain theoretically distinguishable. Subsequent findings are therefore reported as associations with perceived PSU effectiveness, and the dominance of policy quality is treated as suggestive rather than definitive. Triangulation confirms that the square root of AVE for Policy Quality (.794) and PSU Effectiveness (.745) both exceed their Triangulation Construct EKS EFE SITAS KOM KUKE EKS .713 — — — — EFE .561 .745 — — — SITAS .531 .462 .805 — — KOM .613 .597 .433 .881 — KUKE .575 .673 * .674 .584 .794 diagonal = inter-construct correlations. Discriminant validity is met if √AVE > all correlations in its row/column (Fornell & Larcker, 1981). * KUKE–EFE correlation (.673) is lower than √AVE KUKE (.794) and √AVE EFE (.745), suggesting that complete construct redundancy is unlikely, although conceptual overlap cannot be ruled out. Source: SmartPLS 4 output, Primary Data Processed (2026). VIF EKS ↔ EFE 1.912 SITAS ↔ EFE 1.943 KOM ↔ EFE 1.842 KUKE ↔ EFE 2.330 problematic collinearity and suggest that common method bias is unlikely (Kock, 2015). Source: SmartPLS 4 output, Primary Data Processed (2026). Relevance Variable R² R² Adjusted Q²predi ct Interpretat ion PSU Effective ness PSU (EFE) .532 .524 .507 Moderate / Substantia l .50 = strong (Hair et al., 2022). Q²predict interpretation: .25– .50 = moderate; > .50 = substantial (Shmueli et al., 2019). Processed (2026). inter-construct correlation (.673). Cross-loading analysis also shows all indicators load highest on their originating constructs. The high HTMT reflects the statutory-operational linkage inherent in time-bound electoral policy implementation, where administrators perceive regulatory clarity as a direct prerequisite fused with effectiveness achievement (Edwards III, 1980; Mazmanian & Sabatier, 1983; Purwanto & others, 2015). Following Hair et al. (2022) both constructs were retained in the final model. Although the HTMT value exceeded the conservative threshold of .90, the Fornell–Larcker criterion and cross-loading analysis provide complementary evidence regarding construct separation but do not resolve the discriminant validity concern. Accordingly, Policy Quality and PSU Effectiveness were retained on theoretical grounds, while their empirical distinctiveness remains unresolved. Nevertheless, the HTMT confidence interval results indicate that discriminant validity remains a potential limitation of the measurement model and should be considered when interpreting the magnitude of the structural relationship between both constructs. Before evaluating the structural relationships, collinearity diagnostics were assessed to ensure that the predictor constructs did not exhibit problematic overlap that could distort parameter estimates. Variance Inflation Factor (VIF) values are presented in Table 5. To assess potential collinearity and provide a diagnostic check for common method bias, variance inflation factor (VIF) values were examined. All predictor constructs recorded VIF values ranging from 1.842 to 2.330, below the conservative threshold of 3.3 suggested by Kock (2015), indicating that problematic collinearity was not detected. However, full- collinearity VIF values alone cannot establish the absence of common method bias in a single-source perceptual survey. The results should therefore be interpreted as reducing, rather than eliminating, this concern. Future studies should combine procedural remedies, such as temporal or source separation and anonymity assurances, with additional statistical assessments or a marker-variable approach.

Table 2. Measurement Model Evaluation — Convergent Validity & Reliability

Outer loadings between .50 and .70 were retained because they represent theoretically important construct content. Source: SmartPLS 4 output, Primary Data Processed (2026).

Table 3. Discriminant Validity — HTMT Matrix

Table 4. Fornell–Larcker Criterion — Discriminant Validity Triangulation

Table 5. Collinearity Assessment

Structural Model Evaluation (Inner Model) & Hypothesis Testing

Before hypothesis testing, the structural model's predictive power was evaluated via three primary measures presented in The model records an R² of .532 (adjusted R² = .524), meaning 53.2% of PSU effectiveness variation is simultaneously explained by the four independent variables. This is categorized as moderate-to-strong per Hair et al. (2022). Predictive relevance was supported by the PLSpredict with a Q²predict of .507, exceeding the substantial threshold of .35 per Shmueli et al. (2019). The combination of substantial R² and Q²predict confirms the model possesses adequate explanatory power and consistently predicts outcome variation in out-of-sample data. Hypothesis testing was conducted via bootstrapping with 5,000 subsamples, two-tailed, at α = .05. Hypotheses are accepted if the t-statistic > 1.96 and p < .05 (Hair et al., 2022). Results alongside effect sizes (f²) are presented in Table 7. H1 accepted: Policy Quality is positively and significantly associated with perceived PSU effectiveness (β = .466, t = 6.711, p < .001). With a medium effect size (f² = .199), it has the largest standardized path coefficient in the model. The coefficient indicates that higher standardized levels of perceived policy quality are associated with higher standardized levels of perceived PSU effectiveness, conditional on the other predictors. Because the data are cross-sectional and single-source, the estimate does not establish temporal ordering or causality. The strong association observed between Policy Quality and PSU Effectiveness (β = .466; f² = .199) must be interpreted in light of unresolved discriminant validity. The HTMT value was .946, exceeding the recommended threshold, and its bias- corrected confidence interval included 1.00. Thus, the data do not demonstrate sufficient empirical separation between the two constructs, and part of the estimated relationship may reflect overlapping item content or respondent perceptions. Although the constructs remain theoretically distinguishable, the coefficient should not be treated as evidence that policy quality is an independently dominant predictor. Future research should revise item wording and construct boundaries, report complete cross-loadings, test alternative specifications, and use multi-source effectiveness measures before drawing stronger substantive conclusions. Consequently, part of the observed association may reflect the close conceptual and operational relationship between the two constructs rather than entirely independent explanatory effects. This does not invalidate the substantive importance of Policy Quality; however, it suggests that the magnitude of the relationship should be interpreted cautiously. Future studies are encouraged to employ alternative measurement specifications, multi-source data, or longitudinal designs to further examine the empirical separation between regulatory quality and implementation effectiveness. H2 rejected (Institutional Capacity) constitutes a crucial finding requiring analytical emphasis. Descriptively, SITAS records the highest mean (4.243), yet structurally shows no significance (β = −.047, p = .411). This pattern suggests a Β t-statistic p-value f² Effect Size Decision KUKE → EFE .466 6.711 < .001 .199 Medium H1 Accepted SITAS → EFE −.047 .823 .411 .002 Negligible H2 Rejected EKS → EFE .170 2.876 .004 .032 Small H3 Accepted KOM → EFE .241 4.039 < .001 .067 Small–Medium H4 Accepted medium; > .35 = large (Cohen, 1988 in (Hair et al., 2022). Source: SmartPLS 4 Bootstrapping output, Primary Data Processed (2026).

high performance; ‘Monitor’ = moderate importance; ‘Saturation Point’ = negative importance (diminishing returns). Source: SmartPLS 4 IPMA output, Primary Data Processed (2026).

requiring specific intervention. Source: SmartPLS 4 IPMA output, Primary Data Processed (2026). ceiling effect and perceptual homogeneity: once minimum logistics, human resources, and budget readiness standards are uniformly met across sub-districts, capacity variation no longer explains effectiveness differences. SITAS indicators focusing on stock availability (budget, logistics, training) also tend to be binary in emergency contexts (sufficient/insufficient), reducing quantitative discriminative power. Theoretically, this confirms Garnett (2019) proposition that resource data only partially capture capacity; what determines outcomes is capacity translation into adaptive response, not initial input volume. The policy implication is clear: additional budget allocation for logistics no longer yields significant marginal gains. Funds should be redirected to scenario simulation training, stress management, and cross- functional coordination strengthening. H3 accepted: External Support is a positive and significant predictor of perceived effectiveness (β = .170, t = 2.876, p = .004) with a small effect (f² = .032). This aligns with institutional theory propositions (DiMaggio & Powell, 1983) that external environmental legitimacy and normative support strengthen administrator procedural compliance. H4 accepted: Administrator Commitment is positively and significantly associated with perceived effectiveness (β = .241, t = 4.039, p < .001), with a small-to-medium effect size (f² = .067). This result is consistent with the street-level bureaucracy perspective (Lipsky, 2010), which emphasizes the relevance of frontline discretion and integrity under pressure. However, the cross-sectional evidence does not confirm the theory or establish that commitment determines implementation success. Reciprocal relationships and omitted contextual factors, including organizational support, political pressure, and prior administrator experience, may also contribute to the observed association.

Table 6. Model Determination Coefficient & Predictive Relevance

Table 7. Hypothesis Testing Results (Bootstrapping) & Effect Sizes (f²)

Importance-Performance Map Analysis (IPMA)

Importance-Performance Map Analysis (IPMA) was conducted to translate statistical findings into measurable, priority-based intervention recommendations (Ringle & Sarstedt, 2016). IPMA maps constructs and indicators along two dimensions: (1) importance, shown by total effect on PSU effectiveness, and (2) performance, shown by latent scores transformed to a 0–100 scale. Construct-level results are presented in Table 8. Policy Quality occupies a priority position with the highest importance (total effect = .466) yet moderate performance (65.045), indicating significant improvement space with substantial effectiveness impact. Administrator Commitment lies in a maintain position: very high performance (72.920) with substantial importance (.241), reflecting administrator integrity as PSU’s primary success capital. External Support occupies a monitor position with moderate importance (.170) and performance (68.064). Institutional Capacity sits at a saturation point: negative importance (−.047) despite adequate performance (67.505), confirming that additional resource investment no longer correlates with effectiveness improvement. Indicator-level IPMA analysis provides sharper granularity, presented in Table 9.

Table 8. IPMA — Construct Level

Table 9. IPMA — Indicator Level

Operational Recommendations Based on IPMA

1. KUKE3 (No regulatory misinterpretation) exhibits the largest importance–performance gap (56.389). Kabupaten KPU and Provinsi Bawaslu must develop scenario-based guidelines covering decision matrices for logistics redistribution, ballot adjustment, and stage scheduling within the 60-day window, tested via participatory simulations before execution. 2. EKS4 (Public enthusiasm) records the lowest performance (50.833). KPU and Bawaslu should collaborate with the Regional Communication and Information Office and civil society forums to develop restorative communication campaigns positioning PSU as a democratic trust recovery mechanism, not merely attendance target fulfillment. Narratives must include process data transparency and dialogue with polarization- affected voter groups. 3. SITAS has reached saturation. Budget reallocation from additional logistics procurement to resilience training for PPK/Panwascam will yield more measurable effectiveness impacts.

Regulatory Clarity as the Cognitive Foundation of Implementation

The finding that policy quality is the strongest predictor (β = .466, f² = .199) confirms that in 60-day time-bound electoral policies, regulatory clarity functions as a cognitive framework accelerating operational decision-making. the lowest performance (56.389) yet highest importance (.179), revealing a policy paradox: the most crucial aspect is the weakest. This confirms that post-Constitutional Court ruling socialization has not fully translated judicial language into uniform operational guidelines, particularly regarding time limits and corrective procedures. This strengthens Mazmanian & Sabatier (1983) proposition while extending it: in time-constrained contexts, SOP clarity can function as a partial substitute for limited experience, as there is no room for gradual clarification.

Institutional Capacity: When Resource Sufficiency Ceases to Differentiate

The non-significant association for institutional capacity (β = -.047, p = .411) is consistent with, but does not prove, a threshold or ceiling explanation. Capacity received high descriptive ratings, suggesting that limited variance may have reduced its ability to differentiate perceived effectiveness. Without additional distributional statistics, objective resource measures, or robustness tests, institutional capacity should be interpreted as a possible baseline condition rather than a confirmed threshold mechanism. This interpretation aligns cautiously with Garnett (2019) and James & Jervier (2017), who distinguish resource availability from the organizational processes through which resources affect performance.

Administrator Commitment as Irreplaceable Electoral Social Capital

Administrator commitment is significantly associated with perceived effectiveness (β = .241, f² = .067), confirming that integrity and neutrality constitute invisible infrastructure sustaining process credibility amid polarization. Discretion directed toward accountability principles enhances procedural legitimacy, consistent with Norris (2014). However, this commitment is vulnerable to erosion without professional recognition, stress management systems, and organizational ecosystems protecting implementer neutrality (James, 2019; Lotta et al., 2022).

External Support, Voter Fatigue, and Substantive Legitimacy Crisis

External support was positively associated with perceived effectiveness (β = .170, f² = .032), while EKS4 recorded the lowest performance score (50.833). Separate KPU administrative data reported a participation decline from 68.06% to 63.44%; this contextual evidence should not be interpreted as an outcome directly measured by the structural model. Taken together, the survey indicator and participation data suggest that public engagement warrants further investigation. Restorative communication, data transparency, and dialogue with affected groups may be considered as evidence-informed responses, but their effectiveness requires practitioner validation and evaluation.

Theoretical Implications: Hierarchy, Not Linearity, in Emergency Electoral Implementation

The findings support a context-bound proposition that implementation variables may occupy different functional positions in a time-constrained PSU. Regulatory clarity appears to operate as an enabling condition, institutional capacity as a possible baseline, and administrator commitment and external support as complementary factors. This interpretation is derived from cross-sectional data in Tasikmalaya and should not be generalized to all emergency policies or electoral settings without comparative replication. Within this case, clearer regulations were associated with more consistent perceived implementation outcomes, whereas highly rated capacity did not differentiate those outcomes. This pattern is compatible with a conditional model, but the data do not establish temporal sequencing or causal hierarchy. The proposed ordering should therefore be treated as a theoretical interpretation for future testing rather than a confirmed universal structure. Findings also carry implications for SLB frameworks in Global South contexts. As argued by Lotta et al. (2022) and Peeters & Campos (2023), SLB assumptions developed in established liberal democracies require contextualization when applied to developing countries with differing institutional characteristics. In Tasikmalaya’s PSU, administrator discretion operates not in a space relatively safe from political pressure, but amid intense polarization and informal intervention pressure. That commitment remains significant under these conditions, actually strengthens, rather than weakens, the framework’s relevance in this context, with the modification that commitment must be understood as a collective construct dependent on organizational ecosystems, not merely individual disposition.

Policy Recommendations: From IPMA Map to Strategic Action

The IPMA findings suggest three evidence-informed directions for practitioner consideration. First, KUKE3 shows the largest importance-performance gap, indicating that clearer operational guidance may deserve priority. KPU could test plain-language PSU guidelines, scenario matrices for post- disqualification procedures, cross-institutional coordination protocols, and stage-specific decision pathways within the 60- day window. Participatory simulations may also be piloted to assess whether they reduce regulatory misinterpretation. These recommendations follow the KUKE3 diagnostic result but require validation through implementation trials and practitioner feedback. Second, the non-significant capacity path and high ratings suggest examining the balance between stock resources and adaptive capabilities. Rather than assuming that additional logistics or budget will automatically improve outcomes, future programs could assess training in crisis communication, procedural conflict management, and administrator well-being. This is a cautious implication of the observed pattern, not evidence that resource investment is unnecessary. Objective capacity audits should precede any reallocation decision. Third, the low EKS4 performance score suggests that public engagement should be examined beyond the frequency of technical socialization. KPU and Bawaslu could pilot communication approaches that explain the corrective purpose of PSU, disclose process data, and provide structured dialogue with affected communities. Because the model measured administrator perceptions rather than voter motivation, these approaches should be evaluated using voter surveys, participation data, and independent observation before being institutionalized.

Conclusion

This study examined factors associated with perceived rerun election (PSU) effectiveness under a strict judicial deadline. PLS-SEM analysis of 240 administrators in Tasikmalaya Regency showed that policy quality had the strongest positive association with perceived effectiveness, followed by administrator commitment and external support, whereas institutional capacity was not significant. The capacity result is compatible with a possible ceiling or baseline explanation but does not confirm a threshold mechanism. IPMA identified regulatory misinterpretation and low public enthusiasm as areas warranting further attention. The study contributes a context-bound predictive model suggesting that regulatory clarity, organizational capacity, administrator commitment, and external support may play different functional roles in time-constrained PSU implementation. However, the proposed hierarchy remains a theoretical interpretation that requires comparative testing. The practical implications - scenario-based guidance, adaptive administrator support, and restorative public communication - should be treated as evidence-informed options for piloting rather than definitive prescriptions. Interpretation is limited by the cross-sectional single- regency design, reliance on administrator perceptions, post hoc model refinement, marginal internal consistency for two constructs, and unresolved discriminant validity between Policy Quality and PSU Effectiveness. The policy-quality coefficient may therefore be partly affected by conceptual overlap. Future research should use multi-source objective indicators, voter and observer perspectives, cross-regency replication, alternative measurement specifications, and longitudinal designs before broader conclusions are drawn.

Research Limitations

Several limitations require explicit consideration. First, the cross-sectional design and single-source perceptual data do not establish causal or temporal ordering and may introduce residual common method bias (Kock, 2015, 2015; Spector, 1987). Second, discriminant validity between Policy Quality and PSU Effectiveness remains unresolved: HTMT was .946 and the bias-corrected confidence interval (.839-1.046) included 1.00. Consequently, the reported dominance of policy quality and the proposed hierarchical interpretation are suggestive and may partly reflect conceptual overlap in administrator perceptions. These findings should not be treated as definitive evidence that clarity causally outweighs capacity. The second limitation is the exclusive reliance on administrator perceptions. Voter perspectives, independent observer assessments, and objective administrative data - such as sub-district participation rates, recorded violations, stage-completion times, and resource utilization - were not integrated into the structural model. The dependent construct therefore represents implementer-perceived effectiveness rather than recipient-experienced or independently observed effectiveness. Third, the study is a single-case census within Tasikmalaya Regency. Census coverage strengthens description of the stated administrative population but does not support automatic generalization to other regions. Tasikmalaya's polarization following incumbent candidate disqualification and its local institutional configuration may differ from those of the other regions conducting PSUs. Cross-regency replication is required before the proposed hierarchy can be considered transferable. Fourth, potentially relevant contextual variables were not modeled, including political conflict intensity, prior public trust in KPU and Bawaslu, administrative experience, voter demographics, and objective PSU performance. These omitted factors may account for part of the 46.8% variance not explained by R² = .532. The unexplained variance reinforces the need to interpret the model as partial predictive evidence rather than a comprehensive account of PSU effectiveness.

Future Research Agenda These limitations open several productive research directions that can strategically build upon this study’s findings. Priority is cross-regency comparative replication. With 24 regions conducting PSU in one 2024 Local Election cycle, Indonesia provides rare, valuable empirical variation: relatively controlled PSU cases within a single national regulatory framework, yet with significant local context variation. Simultaneous structural model testing across several regencies enables measurement invariance testing and comparative analysis of what moderating conditions make regulatory clarity more or less determinant. Specific answerable questions: Does the variable hierarchy found in Tasikmalaya hold in regions with lower polarization or more heterogeneous institutional capacity? The second priority is a mixed-methods design integrating voter perspectives. Findings on low public enthusiasm (EKS4, performance = 50.833) and actual participation decline require deeper explanation beyond statistical correlation. Post-PSU voter surveys measuring legitimacy perception, procedural satisfaction, and participation motivation can reveal mechanisms linking implementation quality with public engagement, invisible from the administrator side alone. This approach aligns with Norris (2014) recommendation that electoral integrity must be measured from citizen perspectives, not only administrator perspectives. The third priority is longitudinal design tracking perception and capacity changes before, during, and after PSU. This design enables testing whether regulatory clarity investments (e.g., via SOP revision or scenario-based training) measurably improve effectiveness in subsequent implementation cycles. It also allows researchers to trace whether high administrator commitment in one PSU episode erodes or strengthens in subsequent regular election cycles, providing insights into built electoral social capital sustainability. Fourth priority, more methodologically ambitious, is developing a multi-source, cross-context validated PSU implementation capacity index. This study shows that perceptual capacity measurement has diagnostic limitations when variables reach saturation. Garnett (2019) demonstrated that content analysis based on public EMB documents can

provide more objective, comparable capacity measures. Integrating these with perceptual data in a single multi-level hierarchical model could yield a far richer understanding of emergency electoral implementation effectiveness determinants in Indonesia..

References

DiMaggio, P. J., & Powell, W. W. (1983). The Iron Cage Revisited: Institutional Isomorphism and Collective Rationality in Organizational Fields. American Sociological Review, 48(2), 147–160. https://doi.org/10.2307/2095101

Edwards III, G. C. (1980). Implementing Public Policy. Congressional Quarterly Press.

Fornell, C., & Larcker, D. F. (1981). Evaluating Structural Equation Models with Unobservable Variables and Measurement Error. Journal of Marketing Research, 18(1), 39–50. https://doi.org/10.2307/3151312

Garnett, H. A. (2019). Evaluating Electoral Management Body Capacity. International Political Science Review, 40(3), 335–353. https://doi.org/10.1177/0192512119832924

Hair, J. F., Hult, G. T. M., Ringle, C. M., & Sarstedt, M. (2022). A Primer on Partial Least Squares Structural Equation Modeling (PLS-SEM) (3rd ed.). Sage Publications. https://doi.org/10.1007/978-3-030-80519-7

Hair, J. F., Risher, J. J., Sarstedt, M., & Ringle, C. M. (2019). When to Use and How to Report the Results of PLS-SEM. European Business Review, 31(1), 2–24. https://doi.org/10.1108/EBR-11-2018-0203

Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A New Criterion for Assessing Discriminant Validity in Variance-Based Structural Equation Modeling. Journal of the Academy of Marketing Science, 43(1), 115–135. https://doi.org/10.1007/s11747-014-0403-8

Hupe, P., & Hill, M. (2007). Street-Level Bureaucracy and Public Accountability. Public Administration, 85(2), 279–299. https://doi.org/10.1111/j.1467-9299.2007.00650.x Indonesia at Melbourne. (2024). Election Fatigue or Elite Shenanigans? Why Fewer Indonesians Went to Polls to Elect Their Local Leaders.

James, T. S. (2019). Better Workers, Better Elections? Electoral Management Bodies and Electoral Integrity Worldwide. International Political Science Review, 40(3), 370–390. https://doi.org/10.1177/0192512119829516

James, T. S., Garnett, H. A., Loeber, L., & van Ham, C. (2019). Electoral Management and the Organisational Determinants of Electoral Integrity. International Political Science Review, 40(3), 295–312. https://doi.org/10.1177/0192512119828206

James, T. S., & Jervier, T. (2017). The Cost of Elections: The Effects of Public Sector Austerity on Electoral Integrity and Voter Engagement. Public Money and Management, 37(7), 461–468. https://doi.org/10.1080/09540962.2017.1351834

Khaira, A. U., Hermita, N., & Alim, J. A. (2025). Politik Hukum Pemungutan Suara Ulang (PSU) Pilkada 2025. Anterior Journal, 24(3). https://doi.org/10.33084/anterior.v24i3.10827

Kock, N. (2015). Common Method Bias in PLS-SEM: A Full Collinearity Assessment Approach. International Journal of E-Collaboration, 11(4), 1–10. https://doi.org/10.4018/ijec.2015100101

Kock, N., & Hadaya, P. (2018). Minimum Sample Size Estimation in PLS-SEM: The Inverse Square Root and Gamma-Exponential Methods. Information Systems Journal, 28(1), 227–261. https://doi.org/10.1111/isj.12131 Komisi Pemilihan Umum Republik Indonesia. (2025). Rilis KPU Perkembangan Pemungutan Suara Ulang (PSU) di 3 Daerah dan Pilkada Ulang di 2 Daerah.

Lipsky, M. (2010). Street-Level Bureaucracy: Dilemmas of the Individual in Public Services (30th anniversary). Russell Sage Foundation.

Lotta, G., Pires, R., Hill, M., & Møller, M. O. (2022). Recontextualizing Street-Level Bureaucracy in the Developing World. Public Administration and Development, 42(1), 3–10. https://doi.org/10.1002/pad.1968

Mazmanian, D. A., & Sabatier, P. A. (1983). Implementation and Public Policy. Scott, Foresman and Company.

Meyer, J. W., & Rowan, B. (1977). Institutionalized Organizations: Formal Structure as Myth and Ceremony. American Journal of Sociology, 83(2), 340–363. https://doi.org/10.1086/226550

Mubarok, S., Zauhar, S., Setyowati, E., & Suryadi. (2020). Policy Implementation Analysis: Exploration of George Edward III, Marilee S Grindle, and Mazmanian and Sabatier Theories in the Policy Analysis Triangle Framework. Journal of Public Administration Studies, 5(1), 1–7. https://doi.org/10.21776/ub.jpas.2020.005.01.1

Norris, P. (2014). Why Electoral Integrity Matters. Cambridge University Press. https://doi.org/10.1017/CBO9781107280861

Peeters, R., & Campos, S. A. (2023). Street-Level Bureaucracy in Weak State Institutions: A Systematic Review of the Literature. International Review of Administrative Sciences, 89(4), 977–995. https://doi.org/10.1177/00208523221103196

Purwanto, E. A., & others. (2015). Public Policy Implementation: Ideas and Practices in Indonesia. Gava Media.

Ringle, C. M., & Sarstedt, M. (2016). Gain More Insight from Your PLS-SEM Results: The Importance-Performance Map Analysis. Industrial Management & Data Systems, 116(9), 1865–1886. https://doi.org/10.1108/IMDS-10-2015-0449

Rossi, P. H., Lipsey, M. W., & Freeman, H. E. (2004). Evaluation: A Systematic Approach (7th ed.). Sage Publications.

Rulandari, N., Natision, A., Esien, E. B., & Kesmawan, A. P. (2022). The policy implementation of Social Ministry’s cash assistance program during the Covid-19 pandemic in Jakarta. Journal of Governance and Public Policy, 9(1), 48–61.

Scott, W. R. (2008). Institutions and Organizations: Ideas and Interests (3rd ed.). Sage Publications. Shmueli, G., Sarstedt, M., Hair, J. F., Cheah, J. H., Ting, H., Vaithilingam, S., &

Ringle, C. M. (2019). Predictive Model Assessment in PLS-SEM: Guidelines for Using PLSpredict. European Journal of Marketing, 53(11), 2322–2347. https://doi.org/10.1108/EJM-02-2019-0189

Spector, P. E. (1987). Method Variance as an Artifact in Self-Reported Affect and Perceptions at Work: Myth or Significant Problem? Journal of Applied Psychology, 72(3), 438–443. https://doi.org/10.1037/0021- 9010.72.3.438