Evaluating Passenger Satisfaction in Electronic Customs Declarations: An Expectation-Disconfirmation Model Approach

The Expectation-Disconfirmation Model (EDM) offers a robust theoretical framework for examining how satisfaction is shaped through the interplay of expectations and perceived service outcomes. According to this framework, satisfaction increases when perceived performance exceeds what users initially anticipated, known as positive Disconfirmation, and declines when outcomes fall short of expectations, resulting in negative disconfirmation [17, 18]. When individuals experience something different from their expectations, they tend to exaggerate the difference, so service performance that exceeds expectations is perceived as exceptionally good. In contrast, service performance that falls below expectations is perceived as extremely poor. When service performance is perceived as exceptionally good (positive Disconfirmation), individuals feel more satisfied compared to when the performance is perceived as very poor (negative Disconfirmation), which leads to dissatisfaction [12].

Although the EDM has been widely applied to assess satisfaction in both public and health service contexts, empirical findings across studies remain inconsistent. For example, Serrano et al. [12] confirmed the whole structure of EDM in a telemedicine setting, showing that both expectations and Disconfirmation significantly influenced patient satisfaction and preferences. Similarly, Morgeson [17] found that EDM effectively explained satisfaction with federal government services, especially when expanded to include trust and ideological factors. Van Ryzin's earlier studies [18, 19], however, noted variability depending on how Disconfirmation was measured—perceived Disconfirmation showed weaker support compared to subtractive measures, raising concerns about the interpretability of the results. Meanwhile, Zhang et al. [20] confirmed the overall validity of EDM in their meta-analysis, but highlighted inconsistencies arising from methodological differences and service contexts. These conflicting findings underscore the need for context-specific analysis. By applying EDM in Indonesia's airport customs environment, this study helps address these empirical gaps and expands the model's application to digital government services.

In addition to the EDM, this study incorporates the UTAUT to enrich the understanding of users' pre-adoption perceptions. UTAUT identifies effort expectancy, the perceived ease of using a system, as a key determinant of behavioral intention to adopt technology [15]. In the context of the ECD system, effort expectancy reflects users’ beliefs about how intuitively and smoothly they can interact with the platform. This perception forms part of their initial expectations, which are later confirmed or disconfirmed after actual system use, aligning with the mechanisms described in EDM. By linking UTAUT to EDM, the study posits that effort expectancy serves as both an antecedent to perceived system quality and a cognitive driver of satisfaction through the process of Disconfirmation. Previous studies in digital public services, such as e-government portals and innovative travel applications, have demonstrated that effort expectancy strongly influences users' evaluations of system performance and satisfaction outcomes [21, 22]. Thus, integrating UTAUT into the EDM framework allows for a more comprehensive examination of how passengers form, experience, and ultimately evaluate their interaction with the ECD system at Soekarno-Hatta International Airport.

Effort expectancy (EE) refers to the degree of ease associated with using a technology or system [15]. In digital public services, ease of use is particularly influential in shaping perceptions of System Quality (SQ). When a system is intuitive, users are more likely to interpret its design, reliability, and overall performance positively. Prior studies on e-services and mobile platforms have consistently shown that high effort expectancy contributes to higher perceived system quality, which in turn leads to greater customer satisfaction [15, 23]. This relationship can also be understood through the EDM, which assumes a correlation between expectations and perceived performance, even though the direction of causality remains undefined. Both constructs are viewed as exogenous, existing prior to the disconfirmation process [19]. According to Morgeson [17], in alignment with the American Customer Satisfaction Index (ACSI) model, expectations are expected to have a positive influence on perceived quality due to limited user engagement or awareness, which restricts objective performance evaluation, or because individuals cognitively align their satisfaction with prior expectations to reduce dissonance [24]. Thus, when systems are perceived as easy to use (high EE), users may interpret this not only as a sign of good usability but also as a fulfillment or even exceedance of their expectations, reinforcing their perception of system quality.

H1: Effort Expectancy has a positive influence on System Quality.

According to the EDM, satisfaction arises when users' perceptions exceed their initial expectations, referred to as positive disconfirmation [17]. EE plays a key role here: when users expect difficulty but instead find a system easy to use, this gap often leads to favorable Disconfirmation.

This result aligns with a previous study on public digital platforms, which supports the notion that systems that exceed usability expectations tend to generate higher perceived disconfirmation [12, 18].

This effect occurs in government services, where users often enter with low or even skeptical expectations due to criticism of public service quality stemming from prior negative experiences [17]. In such cases, even moderately functional systems can exceed expectations, resulting in positive Disconfirmation and increased satisfaction. As Serrano et al. [12] further note, the lower the initial expectations, the greater the likelihood of positive Disconfirmation where services are perceived as "better than expected." Conversely, higher expectations increase the risk of negative Disconfirmation when performance falls short of expectations. Therefore, managing and understanding users' effort expectancy is essential in shaping perceived Disconfirmation and ultimately satisfaction with public digital services.

H2: Effort Expectancy has a positive influence on Perceived Disconfirmation.

Van Ryzin [18] emphasizes that citizens evaluate public services not purely based on objective performance, but through the lens of their subjective expectations. In other words, people assess what they receive by comparing it to what they anticipated. This result explains why objective service quality often does not directly determine satisfaction levels because individuals interpret the same service conditions differently based on their expectations.

Zhang et al. [20] further assert that both perceived performance and prior expectations independently influence satisfaction. Expectations can have a direct positive effect on satisfaction, especially when users are either unaware of or unable to assess actual service performance. In such cases, individuals may adjust their satisfaction ratings to align with their expectations to reduce discomfort or maintain cognitive consistency [19].

Morgeson [17] adds that managing public expectations poses a significant challenge for government institutions. When expectations decline due to internal inefficiencies or external criticisms, these lowered expectations can erode satisfaction and are challenging to reverse. Although the direct effect of expectations on satisfaction is generally small, it remains statistically significant. This result suggests that expectations form part of the evaluative baseline against which satisfaction is judged [17, 24].

In line with this, they suggest that expectations directly shape how citizens perceive service quality and form satisfaction judgments. Ferreira et al. [25] also support this idea, noting that satisfaction tends to occur only after customers complete their decision-making process, at which point their expectations about service quality must be met to ensure loyalty.

Beyond these indirect mechanisms, EE or the perceived ease of using a system can also have a direct influence on Customer Satisfaction. A user-friendly system reduces the effort, time, and mental strain required, thereby contributing to a more positive service experience. Empirical studies in e-government, banking, and healthcare domains consistently affirm this direct relationship [15, 25]. When users find systems accessible and intuitive, they are more likely to report satisfaction, even independent of their prior expectations.

H3: Effort Expectancy has a positive influence on Customer Satisfaction.

The relationship between EE and Customer Satisfaction (CS) is often indirect and operates through the mediating role of Perceived Disconfirmation (PD). Users may not derive satisfaction merely from the ease of use itself, but from the unexpected ease compared to their initial expectations. This result aligns with the EDM, which posits that satisfaction arises when actual performance exceeds what was anticipated in a process referred to as positive disconfirmation [17, 24].

Perceived Disconfirmation has been shown to have a significant and positive influence on satisfaction, especially in the context of public services. Citizens who meet government services often carry preconceived expectations, and when these expectations are exceeded, their satisfaction tends to increase accordingly, supporting the core assumptions of EDM. However, in public sector contexts, Disconfirmation may not always be the dominant determinant of satisfaction. Notes that declining public trust often leads to lowered expectations, which, while potentially setting the stage for positive Disconfirmation, also creates a more challenging environment for sustaining satisfaction over time [17].

Furthermore, Van Ryzin [18] argues that while performance typically assessed through evaluations of service delivery does contribute to public satisfaction, it may not fully explain how citizens form their judgments. Instead, the gap between expectations and experience, particularly in terms of ease of use, plays a crucial role. Thus, both expectations and the disconfirmation process are central mechanisms in shaping citizen satisfaction, especially when evaluating the success of digital government services.

H4: Effort Expectancy has a positive effect on Customer Satisfaction mediated by Perceived Disconfirmation.

System Quality (SQ) refers to the evaluation of an information system's performance, with a particular focus on the interaction between users and the system. It can be measured by its functional attributes, primarily usability [23]. In the context of public digital services, SQ includes aspects such as reliability, ease of use, speed, and visual design, all of which contribute to the user's perception of system effectiveness [16].

Perceived Disconfirmation (PD), as described by Van Ryzin [18], refers to the gap between what users expect and what they experience in terms of service or product quality. It reflects whether the perceived performance meets, exceeds, or falls short of initial expectations. While expectations may have a minimal direct influence on perceived Disconfirmation, the relationship between perceived quality and Disconfirmation is often the strongest among the procedural effects of satisfaction [24]. The result suggests that users may not always confirm or disconfirm their expectations per se, but rather reassess the perceived quality based on their actual interaction with the system.

According to the EDM, when a system performs better than expected, such as being highly responsive, visually appealing, or easy to navigate, users experience positive Disconfirmation, which significantly contributes to their satisfaction [17, 24]. Empirical studies in digital platforms and public service systems affirm that systems delivering smooth, efficient, and reliable interactions lead to strong disconfirmation effects [12]. Thus, the literature provides firm theoretical and empirical grounding for the following hypothesis:

H5: System Quality has a positive influence on Perceived Disconfirmation.

Perceived system quality tends to have a more substantial indirect impact on satisfaction through Perceived Disconfirmation (PD) than a direct relationship. While perceived quality is indeed important in shaping satisfaction, its effect is amplified when mediated by users' evaluation of whether their expectations were met or exceeded [24]. In this sense, users may not immediately express satisfaction based on technical or functional system performance but rather through the recognition that the system outperformed their initial expectations.

Van Ryzin [18] emphasizes that a performance-based disconfirmation mechanism often shapes satisfaction with public services. That is, when users encounter a higher-than-expected level of quality, especially in government services where expectations tend to be modest, they are more likely to experience positive Disconfirmation, which in turn elevates their satisfaction. This shift underscores the significance of structural path coefficients in understanding how citizens form their overall assessments of satisfaction.

Empirical findings confirm that perceived quality has a greater effect on satisfaction indirectly through Disconfirmation rather than through a direct relationship [24]. This relationship has been observed across various digital service platforms, including government portals, mobile applications, and self-service technologies, where users' expectations have been consistently exceeded, resulting in increased satisfaction levels. Thus, the following hypothesis is proposed:

H6: System Quality has a positive effect on Customer Satisfaction mediated by Perceived Disconfirmation.

Researcher Van Ryzin [18, 19] argues that directly measuring citizens' expectations is essential for explaining satisfaction with public services. A key reason why objective service quality does not always align with satisfaction levels is that users interpret objective conditions through their subjective expectations, ultimately judging what they receive based on what they expected to receive. Similarly, Zhang et al. [20] emphasize that both perceived performance and expectations directly influence satisfaction. This direct effect of expectations may occur when users are either unaware of actual service performance or assimilate their satisfaction evaluations to prior expectations to reduce cognitive dissonance [19].

Morgeson [17] highlights the dilemma governments face in managing public expectations. Since expectations are shaped by internal and external factors beyond administrative control, declining expectations can negatively affect satisfaction in ways that are difficult to correct. Nonetheless, expectations have been found to have a positive and significant, albeit modest influence on satisfaction. Berryman [24] also predicts that expectations directly influence satisfaction because they form the evaluative baseline.

This result is supported by Ferreira et al. [25], who state that satisfaction emerges only after service encounters are complete, meaning that meeting expectations about service quality is crucial for ensuring continued loyalty.

System Quality, as a construct, is frequently viewed as a direct antecedent of Customer Satisfaction. High-performing digital systems that are stable, intuitive, and visually appealing reduce user effort and frustration, thereby improving service evaluations. Evidence from mobile banking, e-government, and other public service platforms confirms that reliable and user-friendly systems build user confidence and enhance satisfaction [20, 26].

H7: System Quality has a positive influence on Customer Satisfaction.

The EDM posits that the comparison between initial expectations and actual performance determines satisfaction. When performance surpasses expectations, referred to as positive Disconfirmation, it typically results in higher satisfaction. Conversely, when performance falls short, negative Disconfirmation leads to dissatisfaction. This relationship underscores the dual role of expectations and perceived performance in shaping Disconfirmation, and ultimately, satisfaction [19].

Importantly, research has shown that Disconfirmation serves as a more precise predictor of satisfaction when measured explicitly. Using question formats that ask respondents to rate outcomes on a scale from "much worse than expected" to "much better than expected" provides stronger explanatory power for satisfaction outcomes [18]. Furthermore, comparisons of subtractive and direct measures of Disconfirmation suggest that subtractive Disconfirmation based on the difference between expected and perceived performance has a more substantial effect on satisfaction formation [20].

This condition is particularly evident in public service contexts, where users may have low expectations due to prior inefficiencies. For example, travelers who anticipate complex or slow customs procedures but instead encounter efficient digital processes are more likely to report high levels of satisfaction, driven not only by the service itself but also by the unexpected improvement over their expectations. Thus, we propose the following hypothesis:

H8: Perceived Disconfirmation has a positive influence on Customer Satisfaction.

The hypothesized correlation of the constructs is depicted in Figure 1.

1.png

Figure 1. Research framework

The total sample consisted of 207 respondents, all of whom were passengers arriving from abroad, submitting declarations using the ECD, and holding Indonesian citizenship. The research analyzes respondents' gender, age, latest educational background, frequency, and location of ECD, as shown in Table 1. The descriptive analysis of the survey sample reveals several key characteristics of the respondents. In terms of gender, the majority were male (56.04%), with females comprising 43.96% of the sample. Age-wise, most respondents fell into the younger adult categories, with 39.61% aged between 18 and 30 years and 38.16% in the 31-40 range. Only a small fraction was over the age of 50 (0.48%). Regarding educational background, over half of the participants (54.11%) held a DIV/bachelor's degree, followed by 34.78% with a senior high school qualification. A smaller portion had attained a master's degree (8.70%), while less than 2.5% had either junior high or doctoral education. Experience with the ECD system revealed that 88.52% of respondents were first-time users, indicating that most had limited familiarity with the platform. Only a small segment had completed the ECD multiple times. Additionally, the location where passengers filled out the ECD form was overwhelmingly concentrated at Soekarno-Hatta International Airport (89.86%), with very few completing it at their origin (6.76%) or at transit airports (3.38%).

Table 1. Respondent’s characteristic

Desc.	Criteria	Frequency (n = 207)	Percentage (%)
Gender	Male	116	56.04%
	Female	91	43.96%
Age	18-30	82	39.61%
	31-40	79	38.16%
	41-50	45	21.74%
	> 50	1	0.48%
Education	Junior High School	3	1.45%
	Senior High School	72	34.78%
	DIV/Bachelor	112	54.11%
	Master’s degree	18	8.70%
	Doctoral Degree	2	0.97%
Frequency of ECD filling when traveling from abroad	First time	184	88.52%
	2-3 times	16	8.13%
	4-8 times	3	1.44%
	> 8 times	4	1.91%
Location of ECD filling	Origin Airport	14	6.76%
	Transit Airport	7	3.38%
	SHIA Airport	186	89.86%

Prior to testing the proposed hypotheses, a two-step approach was employed, following the guidelines of PLS-SEM. The measurement model analysis is followed by structural model analysis. Measurement model analysis results (Figure 2) are a crucial aspect of research in various fields, as they provide a framework for assessing the reliability and validity of the constructs and indicators used in a study. The analysis results indicate that the model exhibits good convergent validity, internal consistency, reliability, and discriminant validity. The measurement model is important because it provides a framework for assessing the reliability and validity of the constructs in the study. It examines the relationships between the latent variables and their respective indicators. The results indicate that the model exhibits good convergent validity, internal consistency, reliability, and discriminant validity, as shown in Table 2. All the latent variable indicators have a factor loading above 0.7, with the construct's Average Variance Extracted above 0.5. The Cronbach's alpha and composite reliability, which represent internal consistency and reliability, respectively, also have values above 0.7. The discriminant validity, as measured by the HTMT, reported that the variables' confidence interval does not include one.

5a418509-459f-4e23-ab2d-cfaf7a171c5d.png

Figure 2. Measurement model analysis result

The structural model analysis (Figure 3) is vital in understanding the relationship between the latent variables in the conceptual framework. It evaluated the coefficient of determination (R²), calculated the effect size (f²), and assessed the path coefficient. In evaluating the coefficient of determination, the R² value was used to measure the exogenous effects on the endogenous variable [31]. Following Cohen [32], f² values were interpreted as small (≥ 0.02), medium (≥ 0.15), or large (≥ 0.35).

fe304b31-cb4e-491a-be18-f33d4d8eae09.png

Figure 3. Structural model analysis result

Table 2. Measurement model analysis result

LV/IND	CV		ICR		DV
	LF	AVE	CA	CR	HTMT
Customer Satisfaction (CS)
CS1	0.868	0.837	0.951	0.962	Yes
CS2	0.914
CS3	0.924
CS4	0.96
CS5	0.905
Effort Expectancy (EE)
EE1	0.886	0.822	0.946	0.958	Yes
EE2	0.914
EE3	0.913
EE4	0.917
EE5	0.904
Perceived Disconfirmation (PD)
PD1	0.874	0.812	0.923	0.945	Yes
PD2	0.895
PD3	0.914
PD4	0.92
System Quality (SQ)
SQ1	0.878	0.811	0.923	0.945	Yes
SQ2	0.919
SQ3	0.907
SQ4	0.898

Note: LV: Latent Variable, IND: Indicator, CV: Convergence Validity, ICR: Internal Consistency and Reliability, DV: Discriminant Validity, LF: Loading Factor, AVE: Average Variance Extracted, CA: Cronbach Alpha, CR: Composite Reliability, HTMT: HTMT confidence interval does not include 1

The coefficient of determination (R²), as depicted in Table 3, shows that effort expectancy (EE), system quality (SQ), and perceived disconfirmation (PD) collectively explain 86.3% of customer satisfaction (CS). These variables can justify PD by 87.3%, while EE explains SQ by 72%. The study analyzed the effect size to determine the exogenous effect on endogenous variables through the resulting f² value, as depicted in Table 4.

EE and PD have a moderate level of effect on customer satisfaction (CS), with f² values of 0.269 and 0.218, respectively, indicating a medium effect. Meanwhile, system quality (SQ) has a subdued level of influence on the CS, with an f² value of 0.002 (indicating a small effect). EE and SQ have a moderate level of influence on PD with an f² value of 0.575 and 0.467 (significant effect), respectively. SQ influenced EE with an f² value of 2.571, which also represents a significant effect according to Cohen's criteria. Subsequently, the bootstrap procedure was completed to test the path coefficient and measure the effect of exogenous variables on endogenous variables by evaluating the compiled hypothesis. Mediation testing was conducted by examining the influence of outcomes on variables. This research employs a critical t-value of 1.96 for a 5% significance level and a p-value smaller than 0.05 [31], as shown in Table 5. The SRMR estimated model has a value of 0.054, indicating that it is considered a good fit, with a threshold below 0.08 [33].

Based on the data, every external factor has a positive impact on the respective internal factors. Among the external factors, EE has the most decisive influence on SQ, with a path coefficient value of 0.849. In comparison, the correlation of SQ and CS is the weakest, with a path coefficient value of 0.012. Although some variables do not meet the relationship criteria of a t-value above 1.65 or a p-value of 0.05 or less, the measurement results suggest that the compiled variable relationships are significant, except for the relationship between SQ and CS. Based on the measurement model analysis, all hypotheses are accepted, except H7, which is rejected, as depicted in Table 6.

Table 3. Coefficient of determination (R²) measurement

Table 4. Effect size (f²) measurement

Table 5. Path coefficient and measurement analysis

Table 6. Hypothesis test result

This study examined the impact of effort expectancy and system quality on perceived Disconfirmation and customer satisfaction within the context of Indonesia's ECD system. Findings indicate that effort expectancy significantly influences both perceived Disconfirmation and satisfaction, while system quality indirectly affects satisfaction only through Disconfirmation. These results reinforce the critical role of user-perceived ease of use in digital service adoption, especially when expectations are exceeded.

Notably, system quality—despite its technical importance—was not a direct driver of satisfaction. The result suggests that passengers at Soekarno-Hatta International Airport prioritize usability over backend performance. When passengers encounter friction in navigating the system, it diminishes satisfaction, regardless of the ECD system's sophistication. Thus, effort expectancy emerged as both a cognitive and emotional precursor to satisfaction, mediated by the degree to which expectations were positively disconfirmed.

6.1 Policy and managerial implications

To enhance public satisfaction with digital customs systems, Customs authorities should implement pre-arrival tutorials, multilingual support, and human assistance counters for digital troubleshooting. Airport operators should ensure stable internet access and sufficient kiosk infrastructure to reduce technical complaints. System developers should prioritize intuitive interfaces, regular user testing, and responsive design to accommodate passengers with varying levels of digital literacy. These actionable measures are expected to minimize the usability gap and improve the alignment between user expectations and system performance.

6.2 Limitations and future research

This study's primary limitation is its single-site design, which focuses solely on SHIA and is limited to Indonesian passengers. While this sample reflects the most significant international hub in the country, it may not capture the heterogeneity of user experiences across other Indonesian airports or among foreign passengers. Future studies should extend this research to multiple airports and diverse traveler demographics to increase generalizability.

Furthermore, while this study controlled key structural relationships, it did not account for common method bias or potential measurement error inherent in self-reported data. Future research should consider incorporating objective performance metrics or experimental designs to enhance the accuracy of the findings. Expanding the theoretical scope by integrating constructs such as trust, perceived risk, or loyalty could also deepen the understanding of satisfaction in digital public service contexts.

In summary, this research offers both theoretical and practical insights into the digitalization of public services, demonstrating that managing expectations and simplifying user interaction are crucial levers for enhancing satisfaction with government technologies, such as the ECD.