This model was initially proposed by Oliver (1980) to explain consumer satisfaction and repeat purchase behavior. Later, Bhattacherjee (2001) introduced it into the field of information systems, developing it into the expectation-confirmation Model (ECM) to interpret users’ willingness to continue using information systems (Bhattacherjee, 2001; Oliver, 1980). Its fundamental logic posits that users form certain expectations regarding the performance and effectiveness of a product or service prior to use. After use, they compare their actual experience with these initial expectations, leading to confirmation The level of expectation confirmation affects how useful and satisfying users think the product or service is, which in turn affects whether they want to keep using it.

Recently, ECM has been widely applied in the field of education, such as in researching students’ willingness to continue using online learning platforms, blended learning, or other educational technologies (Dou et al., 2025; Obeid et al., 2024; She et al., 2024). Expectation describes how a user anticipates the performance, functionality, or effectiveness of a product or service before using it (Bhattacherjee, 2001). In the context of esthetic education courses, non-art majors may expect the course to enhance their esthetic abilities, enrich their extracurricular lives, or provide enjoyable artistic experiences. Course promotions, instructor introductions, or peer reviews can shape these expectations. Confirmation refers to the degree to which users’ actual experiences align with their initial expectations after using a product or service (Bhattacherjee, 2001). If students observe that the esthetic education course meets or exceeds their initial expectations, their level of expectation-confirmation will be higher. Perceived usefulness refers to the extent to which users believe a product or service helps them achieve their goals (Bhattacherjee, 2001).

Students in esthetic education classes might think that the curriculum is helping them become more esthetically literate, opening their minds, or improving their artistic skills. This perception strengthens their perceived usefulness. Satisfaction refers to users’ overall emotional evaluation of a product or service based on actual experience (Bhattacherjee, 2001). If students are satisfied with the teaching quality, content design, and learning experience of esthetic education courses, they are more likely to develop a willingness to continue selecting. Continuance intention refers to a user’s willingness to continue using a product or service after initial exposure (Bhattacherjee, 2001). As the core dependent variable in ECM, it reflects whether users intend to engage in similar behaviors in the future. In this study, continuance intention manifests as non-art majors’ willingness to selectl in additional esthetic education courses after completing one. Zhang et al. (2016) developed an enhanced model based on ECM when examining MOOC continuance learning intentions. They found that the degree of expectation-confirmation positively influenced learning satisfaction and perceived usefulness experiences. Conversely, satisfaction and perceived usefulness further positively impacted continuance learning intentions (Zhang et al., 2016).

Yan and Yan (2013) also validated this logic in their study on the intention to continue using electronic resources in university libraries. They found that expectation-confirmation is a key predictor influencing perceived usefulness and user satisfaction, which in turn affects the intention to continue using (Yan and Yan, 2013). Additionally, Davis proposed that perceived ease of use and perceived enjoyment are also important factors influencing confirmation expectancy. Perceived ease of use refers to the degree to which an individual believes using a specific system will not be effortful (Davis, 1989). This aligns with students’ perception that selecting and completing esthetic education courses requires minimal additional effort. Perceived enjoyment denotes the degree to which the activity itself is considered pleasurable (Davis et al., 1992). Students experience visual enjoyment and esthetic pleasure during esthetic education courses. Based on this, the study proposes the following hypotheses:

Although ECM excels at explaining sustained usage behavior, it also has certain limitations. First, ECM primarily focuses on post-experience satisfaction and perceived usefulness. This type of analysis may overlook the complexity of initial behavioral decisions, such as factors like social pressure or personal attitudes.

Secondly, ECM primarily concentrates its applicability on continuous usage behaviors related to technology. This may require further adjustment for non-technical behaviors. Wandira et al. (2024), in their study of cloud-based academic information systems, integrated the technology acceptance model (TAM) with ECM. They found that satisfaction, perceived usefulness, perceived ease of use, and enabling conditions all significantly influence continuous usage intention. This demonstrates that combining ECM with other theories can enhance its explanatory power (Wandira et al., 2024).

Furthermore, most ECM research relies on cross-sectional data from a single point, with limited attention paid to the dynamic evolution of behaviors. Studies by Ye et al. (2022) indicate that the effects of expectation confirmation may vary depending on course type or learning context (Ye et al., 2022). This suggests that when examining esthetic education courses, we must consider the diversity of course content and its impact on student expectations.

In summary, ECM offers a dynamic perspective for understanding non-art majors’ willingness and behavior in continuously selecting in esthetic education courses. It emphasizes the role of post-course validation, satisfaction, perceived usefulness, perceived ease of use, and perceived enjoyment in sustaining behavior. However, its limitations lie in insufficient explanation of initial behavioral decisions and the need for further validation regarding its applicability to non-technical behaviors. Therefore, integrating ECM with TPB can compensate for their respective shortcomings, forming a more comprehensive theoretical framework.

The theory of planned behavior (TPB) is one of the most renowned and widely applied theories in social psychology concerning the relationship between attitudes and behavior. It was developed by Ajzen based on the Theory of Reasoned Action (TRA). Ajzen (2020) further elaborated on the distinctions between TPB and TRA in his work, emphasizing that the introduction of perceived behavioral control (PBC) represents a key extension of TPB relative to TRA. This extension enables TPB to better explain behaviors that are not fully under volitional control (Ajzen, 2020).

TPB has gained widespread application in the field of educational and learning behavior research due to its robust predictive power. It provides a powerful theoretical tool for understanding students’ learning intentions and behaviors across diverse contexts. These applications validate the universality of TPB and offer significant insights for this study in understanding non-art majors’ willingness to continuously selectl in esthetic education courses.

Specifically, TPB comprises three core components: attitude, subjective norms, and perceived behavioral control. Attitude refers to an individual’s positive or negative evaluation of performing a specific behavior (Ajzen, 1991). Research by Davis et al. (1992) and Lee (2010) explains that perceived usefulness, perceived ease of use, and perceived enjoyment significantly influence attitudes toward technology acceptance (Davis et al., 1992; Lee, 2010). When a technology makes users feel they benefit without excessive effort and brings them enjoyment and pleasure, they develop an intrinsic motivation to adopt it.

As mentioned above, esthetic education courses typically require only sensory and emotional engagement, and students often derive significant visual enjoyment and esthetic experiences from them. Therefore, we can anticipate that perceived usefulness, perceived ease of use, and perceived enjoyment will enhance students’ emotional attitudes toward esthetic education courses and increase their willingness to continue selecting in them. Based on this, the study proposes the following hypotheses:

Subjective norm refers to the perceived social pressure an individual experiences, specifically whether significant others (such as family members, teachers, friends, social groups, etc.) support or oppose their engagement in a particular behavior (Ye et al., 2022). Hsu et al. proposed that subjective norms are formed by interpersonal influences and external sources of influence. Both exert significant effects on sustained usage intentions (Hsu et al., 2004). Interpersonal influence refers to the impact of family, friends, and other relevant individuals within social networks. External source influence pertains to the impact of mass media, such as television, newspapers, and the internet (Brown and Venkatesh, 2005). Regarding students’ enrollment in esthetic education courses, support from family, teachers, or friends constitutes the interpersonal influence component. Course promotion and previous students’ evaluations of instructors and courses shape external source influence. The research model proposes the following hypotheses to account for these two potential external variables:

Perceived behavioral control refers to an individual’s perception of the ease or difficulty of performing a specific behavior and their sense of control over it (Ajzen, 1991). Su et al. (2021) integrated Perceived University Support with TPB to explore factors influencing entrepreneurial intentions among Chinese university students. Their finresearch indicated that perceived university support had a substantial impact on students’ behavioral control regardingtrepreneurship (Su et al., 2021). This result provides important implications for the present study: “university support” in esthetic education courses (e.g., course resources, faculty expertise, campus cultural atmosphere) may also significantly influence non-arts majors’ perceived behavioral control toward esthetic education courses. This, in turn, affects their willingness to continue selecting in such courses. Therefore, the study proposes the following hypothesis:

Overall, the application of TPB in educational and learning behavior research has fully demonstrated its effectiveness in predicting individual behavioral intentions. These studies validate the importance of attitude, subjective norm, and perceived behavioral control as core predictive variables. They also showcase TPB’s potential to enhance explanatory power by integrating external variables. TPB offers a strong theoretical foundation for understanding how non-art majors form intentions to persistently selectl in esthetic education courses before or early in the course selection process. Such choices are based on evaluations of esthetic education courses, social influence, and perceived self-control.

However, TPB primarily focuses on the formation of behavioral intentions, with relatively limited explanatory power regarding the persistence of behavior after it occurs, as well as the influence of experiences and satisfaction on future behavior. This is precisely the aspect that ECM can address.

Based on the integration of ECM and TPB, the research model and hypotheses are presented in Figure 1.

The integration of ECM and TPB has been validated across multiple domains, particularly in technology adoption and educational behavior studies. For instance, Venkatesh et al. (2011) combined UTAUT predictors (such as effort expectancy, social influence, and facilitators) with ECM when examining sustained use of e-government technologies. They found that expectation-confirmation indirectly influenced intention for sustained use through perceived usefulness and satisfaction. On the other hand, contextual factors (e.g., trust) further enhanced the model’s explanatory power (Venkatesh et al., 2011). Similarly, Wandira et al. integrated ECM and TPB to examine usage intentions toward cloud-based academic information systems, revealing that satisfaction, perceived usefulness, and perceived ease of use collectively drive behavioral intentions (Wandira et al., 2024). These studies demonstrate that integrating TPB and ECM can provide more comprehensive behavioral prediction models by combining psychological factors influencing initial decisions with experiential factors following usage.

Within the context of esthetic education courses, integrating the Extrinsic Motivation Concept (ECM) and Theory of Planned Behavior (TPB) can establish a theoretical framework encompassing initial course selection decisions, course experiences, and sustained enrollment intentions. Specifically, the three core variables of TPB—attitude, subjective norm, and perceived behavioral control—can be used to explain students’ intentions to selectl in esthetic education courses for the first time. Perceptions of the courses’ value, such as enhancing esthetic literacy and enriching extracurricular life, may influence students’ attitudes toward esthetic education courses. Subjective norms may stem from support and expectations of peers, teachers, or parents. Perceived behavioral control may involve course accessibility, scheduling flexibility, and students’ learning capabilities.

We adopted all studies and models from prior research, making minor adjustments to adapt them to the context of elective esthetic education courses. The questionnaire design for this study comprised two main sections. The first part was a demographic survey that asked for basic information about the students, such as their school, gender, grade level, and major. The second part measured latent variables using a 5-point Likert scale. Specifically, the scales for identification, satisfaction, and continuance intention were adapted from Bhattacherjee (2001), while the perceived usefulness scale was modified from Sun (2025). Scales for perceived ease of use, perceived enjoyment, and attitude were adapted from Davis (1989). The social norm scale was adapted from Mathieson (1991). The scales for interpersonal influence and external source influence were derived from Brown and Venkatesh (2005). The scale for perceived behavioral control was adapted from Taylor and Todd (1995a, 1995b). All latent variable scales in this study refer to internationally established scales, with appropriate modifications made for the specific context and research practices.

The survey questionnaire for this study was collected online via the Questionstar platform. This platform provides services comparable to Amazon Mechanical Turk. We selectled participating students in our Esthetic Education course during the second semester of the 2024 academic year. Before the final session of the esthetic education course concluded, students were asked to scan a pre-generated QR code displayed on a large screen to access and complete the questionnaire. The survey remained open from May 5, 2024, until June 5, 2024, spanning 1 month. No specific incentives were offered; instead, the questionnaire stated that valid responses would assist faculty and institutions in enhancing esthetic education programs. Students completed a total of 536 questionnaires. Based on three criteria—completion time, completion rate, and consistency in responses to identical questions—we retained 519 valid questionnaires (Table 1).

In this study, SPSS version 24 and AMOS version 23 software were employed for further data analysis. Following the two-step approach recommended by Anderson and Gerbing (1988), a confirmatory factor analysis was conducted using maximum likelihood estimation to generate a measurement model for assessing measurement validity. Subsequently, a structural model was constructed to test the proposed theoretical framework.

This study first examined the reliability and validity of the scale, with validity subdivided into convergent validity and discriminant validity.

This study employed a questionnaire survey method for data collection. The homogeneity of data from this collection method can lead to common method bias. Therefore, the Harman single-factor test was used to examine whether common method bias existed in the measurement. The specific procedure was as follows: All measurement items were loaded onto a single common factor to construct a single-factor structural equation model. Confirmatory factor analysis was conducted using Amos 23.0. Results indicated poor overall model fit, with the following fit indices: χ²/df = 11.292 > 5; CFI = 0.554 < 0.850; NFI = 0.532 < 0.850; TLI = 0.525 < 0.850; IFI = 0.555 < 0.850; RMSEA = 0.141 > 0.08. These findings indicate no significant common method bias.

Table 4 presents the model fit results for this study. The findings indicate that χ²/df = 1.115 < 3; CFI = 0.996 > 0.9; NFI = 0.959 > 0.9; TLI = 0.995 > 0.9; IFI = 0.996 > 0.9; RMSEA = 0.015 < 0.05. These findings indicate that the measurement model exhibits a favorable fit.

To validate the hypotheses, this study first analyzed the fit indices of the structural equation model. As shown in Table 4, χ²/df = 2.134 < 3; CFI = 0.953 > 0.9; NFI = 0.915 > 0.9; TLI = 0.948 > 0.9; IFI = 0.953 > 0.9; RMSEA = 0.047 < 0.05. All six fit indices of the structural model meet the commonly accepted standards for structural equation modeling. This indicates that the constructed structural equation model exhibits a satisfactory fit. Subsequently, the structural equation model was tested using Amos 23.0. The standardized path coefficients and their significance levels are presented in Table 5.

Confirmation has a significant positive effect on satisfaction (β = 0.642, p < 0.001). Satisfaction has a significant positive effect on continued intention (β = 0.182, p < 0.001). Confirmation has a significant positive effect on perceived usefulness (β = 0.698, p < 0.001). Confirmation has a significant positive effect on perceived ease of use (β = 0.521, p < 0.001). Confirmation has a significant positive effect on perceived enjoyment (β = 0.568, p < 0.001). Perceived usefulness has a significant positive effect on attitude (β = 0.446, p < 0.001). Perceived ease of use has a significant positive effect on attitude (β = 0.139, p < 0.01). Perceived enjoyment significantly and positively influenced attitude (β = 0.203, p < 0.001). Attitude significantly and positively influenced continued intention (β = 0.238, p < 0.001). Interpersonal influence significantly and positively influenced subjective norm (β = 0.345, p < 0.001). External sources’ influence significantly and positively affects subjective norm (β = 0.360, p < 0.001). Subjective norm significantly and positively affects continued intention (β = 0.228, p < 0.001). Perceived behavior control significantly and positively affects continued intention (β = 0.240, p < 0.001).

This indicates that all hypotheses are supported. The R² for satisfaction is 41.3%, for attitude is 39.3%, for subjective norm is 36.5%, and for continued intention is 37.1%, demonstrating that the research model possesses a high degree of explanatory power. Figure 2 presents the measurement results of the overall structural model.

This study confirms the relevance and robust explanatory capacity of the ECM-TPB integrated model in elucidating the enduring learning behavior of non-art majors in esthetic education courses. It also reveals the psychological mechanisms and behavioral logic underlying non-art majors’ continued enrollment in esthetic education courses.

The study systematically examined the mechanisms underlying non-art majors’ intentions and behaviors in persistently selecting in esthetic education courses by integrating TPB with ECM. Based on the analysis and discussion, the following key findings were identified:

First, among the three components of TPB, the students’ positive attitudes toward serve as key drivers. Core variables influencing sustained selection intentions include students’ positive attitudes toward esthetic education courses (e.g., content appeal, personal interest) and their perceptions of course selection feasibility (e.g., content difficulty, time flexibility, resource accessibility). The role of subjective norms (e.g., peer recommendations, social pressure) is relatively weak. This study indicates that non-art majors primarily base their decision-making on personal interest when selecting esthetic education courses.

Second, the ECM pathway indirectly influences continued course selection intention through expectation confirmation and satisfaction. The degree to which the actual course experience matches expectations (expectation confirmation) indirectly strengthens continued course selection intentions by enhancing satisfaction. However, satisfaction does not have a significant direct effect on continued behavior.

Third, in the integrated model, expectation confirmation indirectly influences attitudes through perceived usefulness, perceived ease of use, and perceived enjoyment. This subsequently affects the intention to continue selecting, demonstrating greater explanatory power: the integrated model combining TPB and ECM sigsignificantly outperforms either model alone in explaining both the intention to continue selecting and actual behavior (R² = 37.1%). This validates the effectiveness of the integrated model.e complementary roles of psychological motivation and experiential feedback within the context of esthetic education courses.

However, the integrated ECM and TPB model also faces some challenges, and this study has the following limitations, which may affect the generalizability of its conclusions.

First, although the complexity of the model allows for detailed explanations, it also increases the difficulty of data collection and analysis. Future research needs to develop more comprehensive scales to measure the relationships among various variables. Of particular note is that the diversity of esthetic education courses may cause differences in expectation confirmation and satisfaction. Future studies should consider course types as important moderating or control variables.

Second, the cross-sectional design of this study is a key limitation. This design only collects data at a single point, which can reveal correlations between variables but makes it difficult to establish causal relationships and does not allow for dynamic tracking of the evolution from “satisfaction” to “continuance intention.” For example, the impact of initial course experience satisfaction on long-term intention may change over time or with subsequent experiences. Future longitudinal studies will be better able to test the dynamic predictive power of the model.

Third, the study sample was drawn from only three universities, and the characteristics of their regions, school types, and student populations may limit the generalizability of the results. Moreover, this study is rooted in the specific higher education and cultural context of China. Chinese society’s perception of esthetic education, university course offerings and evaluation systems, and the potentially utilitarian learning atmosphere faced by students all create a unique cultural context. These factors may make the mechanisms of variables such as “subjective norms” and “perceived usefulness” culturally specific. Therefore, we must cautiously validate the conclusions of this study when applying them to other cultural or educational contexts.

Finally, to deepen the understanding of these mechanisms, future research could introduce additional variables, such as students’ baseline interest in art, deep learning motivation, or flow experiences, to explore their mediating or moderating roles within the model.