Marketing low-altitude tourism: How value-based drivers and safety risk shape consumer adoption

Abstract

Low-altitude tourism is emerging as an innovative form of travel enabled by new aviation technologies and policy support, yet its consumer adoption remains underexplored. This study integrates the Extended Norm Activation Model and the Emotion-Driven Behavior Model to examine how value-based drivers shape motivation and behavioral intention. A two-stage SEM-ANN analysis was conducted with 355 respondents in China. Results show that outcome expectancy, environmental moral awareness, and perceived self-efficacy strengthen consumer attitudes, while only outcome expectancy and moral awareness directly enhance desire. Both attitude and desire significantly predict intention. Unexpectedly, perceived safety risk reinforces rather than weakens these effects, suggesting that risk can serve as a motivator in hedonic or novel tourism contexts. ANN analysis further identifies desire and self-efficacy as the most influential predictors. The findings contribute to theory on risk-sensitive adoption while offering marketing implications for positioning low-altitude tourism as an exciting, sustainable, and emotionally engaging experience.

Keywords

Low-altitude tourism consumer adoption value-based drivers perceived safety risk SEM-ANN

Introduction

Low-altitude aviation, once confined to military and governmental use, is rapidly expanding into civilian applications through advances in unmanned aerial vehicles (UAVs), electric vertical takeoff and landing (eVTOL) aircraft, and supportive airspace reforms (CAAC, 2023; Liu et al., 2025). These developments have enabled new forms of mobility and experience, such as aerial photography, short-distance travel, and immersive sightseeing. Within this context, low-altitude tourism has emerged as an innovative travel mode that combines technological novelty with sustainable potential (Tang, 2026). According to the China Low-Altitude Economy Development Report, the sector is projected to exceed RMB 1 trillion by 2030, with tourism as a major growth driver (CAAC, 2023; Luo, 2025).

Despite its promise, the sector faces critical challenges in safety regulation, airspace coordination, and consumer acceptance (Ravich, 2020). Digital tools—such as route previews, booking platforms, and real-time tracking—may help reduce perceived risks (Alwateer et al., 2019), but consumer confidence remains uncertain. This underscores the need to better understand the psychological mechanisms that drive adoption.

In tourism innovation research, behavioral intention (BI) is widely used as a proxy for actual adoption. Prior studies highlight the roles of perceived usefulness, trust, convenience, novelty, and normative pressure in shaping technology acceptance (Yoo et al., 2018; Kim and Hwang, 2020; Wut et al., 2023). Symbolic meaning and identity values have also been shown to influence desire and intention (Zhang et al., 2022). However, most research has focused on contexts such as drone delivery or virtual reality travel, leaving low-altitude tourism adoption largely underexplored. Moreover, the role of perceived safety risk (PSR)—a recurring barrier in aviation-related services—remains ambiguous (Yoo et al., 2018; Zhu et al., 2020).

Mainland China provides a particularly fertile context for investigation. Strong policy support, rapid technology deployment, and high digital engagement create favorable conditions for adoption, while evolving regulatory frameworks and trial operations heighten perceptions of risk (Jin, 2024). Exploring how motivational factors interact with PSR is therefore both timely and significant for theory and practice.

To address these gaps, this study integrates the Extended Norm Activation Model (ENAM) (Schwartz, 1977; Stern, 2000) and the Emotion-Driven Behavior Model (EDBM) (Bagozzi, 1992; Perugini and Bagozzi, 2001). Specifically, it examines three value-based drivers—outcome expectancy (OE), environmental moral awareness (EMA), and perceived self-efficacy (PSE)—and their effects on consumer attitude (CA) and desire, with PSR as a moderator of the CA/Desire-BI relationship. A dual-stage SEM-ANN approach is employed to test hypotheses and evaluate predictor importance.

This study contributes by identifying the psychological drivers of adoption in a novel tourism domain and clarifying the paradoxical role of safety risk. Beyond theoretical integration, it offers marketing insights into how risk can be reframed as part of the appeal of innovative, technology-enabled, and experience-driven tourism products.

Theoretical background and hypotheses

Extended Norm Activation Model (ENAM)

The Extended Norm Activation Model (ENAM) builds on Schwartz's (1977) Norm Activation Theory (NAT), which explains altruistic behavior through personal norms activated by awareness of consequences and responsibility (Stern, 2000). ENAM expands this framework by incorporating cognitive elements such as OE and self-efficacy, recognizing that moral obligation alone seldom leads to action unless individuals also perceive the behavior as effective and feasible (De Groot and Steg, 2009; Jansson et al., 2010).

OE reflects the belief that an action yields desirable results, while self-efficacy denotes confidence in one's ability to execute it successfully (Bandura, 1986; Stern, 2000). The latter parallels perceived behavioral control in the Theory of Planned Behavior (TPB) (Ajzen, 1991). ENAM has been widely applied in sustainability studies—eco-friendly consumption, low-carbon mobility, and pro-environmental travel (Han, 2015; Lindenberg and Steg, 2007). Given its technological and environmental relevance, ENAM offers a suitable lens to examine OE, EMA, and PSE as drivers of low-altitude tourism adoption.

Emotion-Driven Behavior Model (EDBM)

The EDBM emphasizes desire as a key antecedent of BI, especially in hedonic contexts (Bagozzi, 2007). Desire is a motivational state oriented toward performing a behavior or reaching a goal (Perugini and Bagozzi, 2004), representing an affective force beyond rational planning.

In this framework, attitude is the evaluative appraisal of behavior, desire captures the motivational pull, and intention reflects planned action (Ajzen, 1991; Warshaw and Davis, 1985). EDBM posits attitude as a distal cause and desire as a proximate cause of BI, with attitude's effect mediated by desire (Leone et al., 1999).

Empirical evidence supports this mechanism: desire predicts adoption of drone food delivery (Hwang and Kim, 2021), interaction with social robots (Piçarra and Giger, 2018), and bicycle tourism (Han et al., 2017). Following EDBM, this study incorporates both attitude and desire to capture cognitive and affective pathways in low-altitude tourism adoption.

Value-based cognitive drivers

Drawing on ENAM, three value-based cognitive drivers are proposed: OE, EMA, and PSE.

OE: anticipated benefits such as novelty, convenience, or experiential satisfaction (Han, 2015; Jansson et al., 2010). In tourism, it reflects the belief that low-altitude travel yields rewarding experiences, which specifically involve enjoying a panoramic, three-dimensional aerial perspective of the destination, increased travel efficiency by circumventing ground congestion, and the novelty of utilizing cutting-edge eVTOL technology.

PSE: confidence in one's ability to manage novel services and technologies (Bandura, 1986). For low-altitude tourism, which involves digital booking, safety protocols, and navigation, higher PSE facilitates adoption (Kim et al., 2025; Pai et al., 2021).

EMA: concern for sustainability and the moral impact of travel (Lindenberg and Steg, 2007; Stern, 2000). Consumers may perceive low-altitude tourism, particularly with electric aircraft, as a greener alternative to conventional travel (Liang et al., 2025).

Effect of value-based cognitive drivers on consumer attitude

OE is a key determinant of attitudes toward new technologies and service innovations, as anticipated benefits such as novelty, convenience, and enhanced experiences typically foster positive evaluations (Huang et al., 2016; Jansson et al., 2010). In the low-altitude context, these enhanced experiences include unique, non-traditional visual immersion and efficient, direct point-to-point mobility enabled by advanced low-altitude vehicles, offering distinct value propositions that extend beyond traditional travel modes. PSE also shapes attitudes: consumers confident in their ability to complete tasks are more likely to view a behavior positively, especially in unfamiliar or technology-driven settings (Baker-Eveleth and Stone, 2008; Stone and Baker-Eveleth, 2013). In low-altitude tourism, PSE enhances assurance in handling booking platforms, safety measures, and digital interactions. Finally, EMA fosters favorable attitudes in sustainability research, as consumers concerned with ecological impacts tend to view environmentally aligned options more positively (Han, 2015; Stern, 2000).

Based on this reasoning, the following hypotheses are proposed:

H1a. OE has a positive influence on CA.

H1b. EMA has a positive influence on CA.

H1c. PSE has a positive influence on CA.

Effect of value-based cognitive drivers on desire

In the EDBM, desire represents an affective motivational state linking cognitive evaluations with behavioral goals (Perugini and Bagozzi, 2001). It is more likely when individuals anticipate meaningful benefits and feel capable of acting (Sekerka and Bagozzi, 2007). Evidence shows that PSE, conceptually aligned with perceived behavioral control, enhances desire by reinforcing confidence in performing the behavior (Yi et al., 2020). In low-altitude tourism, perceiving participation as manageable strengthens emotional motivation. Similarly, EMA can increase desire by aligning tourism choices with ecological values, thereby activating intrinsic motivation (Han, 2015; Lindenberg and Steg, 2007). OE, through expected novelty and satisfaction, also enhances consumers’ affective drive to adopt new travel modes.

Based on this reasoning, the following hypotheses are proposed:

H2a. OE has a positive influence on desire.

H2b. EMA has a positive influence on desire.

H2c. PSE has a positive influence on desire.

Effect of consumer attitude on desire

Within EDBM, attitude functions as a distal predictor of desire: favorable evaluations of a behavior create the psychological readiness that transforms cognitive appraisal into affective motivation (Perugini and Bagozzi, 2001). While attitude reflects reasoned evaluation, desire captures the emotional urge to act, and the former often precedes and amplifies the latter (Bagozzi, 2007). Empirical evidence in technology and tourism supports this pathway: positive attitudes toward social robots (Piçarra and Giger, 2018) and drone delivery services (Hwang et al., 2019) have been shown to stimulate a stronger desire for engagement.

In the context of low-altitude tourism, consumers who view the service as safe, novel, or environmentally responsible are more likely to experience an affective pull to participate. Thus, a favorable attitude is expected to enhance desire, strengthening the motivational chain from cognition to action.

Based on these theoretical and empirical insights, the following hypothesis is proposed:

H3. CA has a positive influence on desire.

Effect of consumer attitude and desire on behavioral intention

The link between CA and BI is central in theories of planned and goal-directed behavior. Attitude reflects overall evaluation of a behavior, and more favorable attitudes consistently predict stronger intentions across tourism, technology, and sustainability domains (Ajzen, 1991; Brand et al., 2019; Chen and Hung, 2016; Tsai and Tiwasing, 2021). In the context of low-altitude tourism, a positive evaluation of novelty, safety, or sustainability is therefore expected to reinforce adoption intentions.

H4a. CA has a positive influence on BI.

Beyond attitude, desire serves as a proximate antecedent of BI, capturing the motivational force that transforms appraisal into commitment (Perugini and Bagozzi, 2001). Research shows that desire significantly predicts intentions in emotionally salient or hedonic settings, such as drone delivery, service robots, and sharing economy platforms (Hwang et al., 2019; Piçarra and Giger, 2018; Yi et al., 2020). Desire is thus expected to be a direct and powerful driver of adoption in low-altitude tourism.

H4b. Consumer desire has a positive influence on BI.

The moderating influence of perceived safety risk

Despite its potential, low-altitude travel raises concerns due to its reliance on emerging aviation technologies. In China, the low flying altitude of most small aircraft heightens public visibility and safety awareness (Ren and Cheng, 2020). Prior studies highlight that limited information and uncertain outcomes increase perceived risk and reduce trust in technology-based services (Herzenstein et al., 2007; Lee, 2009).

PSR is defined here as the subjective expectation that low-altitude tourism may involve failure, harm, or loss of control (Cox and Rich, 1964; Klauser and Pedrozo, 2017). Research suggests that most consumers tend to avoid uncertainty rather than pursue potential gains (Bauer, 1969; Im et al., 2011). Thus, even if consumers hold favorable attitudes, high PSR may discourage adoption, consistent with evidence that operational risks weaken technology acceptance (Sah et al., 2021; Sun, 2014). Similarly, concerns about accidents, privacy, or service reliability can overshadow positive evaluations (Parker et al., 2016; Zhu, 2019). Given these insights:

H5a. PSR negatively moderates the positive influence of CA on BI.

Although PSR is typically expected to suppress behavioral intention, recent perspectives suggest that risk can also serve as a source of arousal and engagement under certain hedonic or innovative contexts (Atkinson, 1957; Belk et al., 2003). When consumers already hold favorable attitudes and a strong desire, moderate levels of perceived risk may heighten involvement and reinforce their commitment—a mechanism consistent with cognitive consistency theory (Kruglanski et al., 2018). Therefore, PSR may not only deter but also, under specific conditions, enhance the effects of CA and desire on BI. By analogy, high PSR is expected to reduce the strength of desire in shaping intention. Therefore, the following is proposed:

H5b. PSR negatively moderates the positive influence of desire on BI.

All proposed hypotheses are summarized in the research model shown in Figure 1.

Methodology

Questionnaire design

A structured questionnaire was developed to test the proposed research model and hypotheses. The instrument comprised three parts: demographic information, two preparatory tasks, and scale-based measures of the constructs.

Respondents first reported demographic details including age, gender, education, income, and tourism experience. A brief screening question assessed their familiarity with aviation-related tourism to ensure relevance.

To standardize understanding, two preparatory tasks were employed to simulate exposure to low-altitude tourism. In Task 1, participants read an informational text designed to familiarize them with low-altitude tourism. The text defined the concept, outlined its current development prospects—that is, existing applications in aerial sightseeing, short-distance mobility, and other early-stage tourism services to illustrate market viability—and addressed key safety considerations. In Task 2, they viewed three short videos illustrating aerial views, sightseeing experiences, and user interactions. Playback was fully controllable, reflecting the voluntary nature of media consumption. Participants could proceed only after self-assessing that they felt adequately informed. To objectively verify comprehension and attention, an instructional attention check (Meade and Craig, 2012) was administered immediately afterward, requiring correct responses to a simple factual question based on the materials. Only data from participants who passed this validity check were retained for analysis.

In this study, the pre-exposure materials comprised a concise textual introduction and three authentic broadcast news videos about low-altitude tourism. The text provided a shared baseline on the concept, development prospects, sustainability potential, and safety assurances. The videos—sourced from Chongqing Satellite TV, Changsha Cultural & Tourism Channel, and Hubei Comprehensive TV—each lasted approximately two to three minutes and depicted real cases without experimental manipulation. Safety information appeared naturally in the text and broadcasts to foster a realistic and balanced understanding of aviation-related services. To minimize confounds (e.g., from pricing or itinerary cues), commercial details were intentionally omitted. All texts, screenshots, and source links are provided in Appendix A (A1, A2, A3).

Following these tasks, respondents completed validated multi-item scales measuring OE, EMA, PSE, CA, desire, BI, and PSR. This task-driven procedure served as a scenario-based simulation, a well-established approach in consumer research on novel service innovations (e.g., Li et al., 2023), to enhance perceived task realism and comprehension by mirroring how consumers typically learn about new services. It also functioned as a procedural remedy to reduce potential common method bias (Podsakoff et al., 2003).

Measurement items were adapted from established studies and refined to the low-altitude tourism context. All items were translated into Chinese using Brislin's (1980) back-translation method to ensure semantic equivalence. A pilot test with 57 participants confirmed clarity and flow; minor adjustments were made before deploying the final questionnaire.

Data collection

Data were collected in February 2025 via major Chinese social media platforms—WeChat, QQ, Baidu Post Bar, Weibo, and Zhihu—to achieve broad demographic coverage across everyday communication, social media, and knowledge-sharing communities. This diverse platform portfolio was a deliberate strategy to access both general and technology-interested users who are most relevant to evaluating an emerging service such as low-altitude tourism.

A total of 417 responses were obtained. To ensure data integrity, several objective quality-control checks were implemented prior to analysis. Duplicate entries identified by participant IDs or IP addresses were removed to prevent multiple submissions. An instructional attention check was embedded immediately after the stimulus materials; respondents who failed this comprehension question were excluded. In addition, surveys completed in under 8 min were classified as likely cases of speeding or inattentive responding. This threshold was derived from pilot testing (median = 10.5 min), representing approximately 75% of the median completion time—a conservative standard widely recommended for maintaining online data quality (Brühlmann et al., 2020; Meade and Craig, 2012; Wang et al., 2023). The order of multiple-choice items was randomized to minimize response bias. After excluding 62 invalid cases, 355 valid questionnaires remained, yielding an effective response rate of 85.1%. The demographic characteristics of respondents are summarized in Table 1.

Table 1.

Demographic characteristics.

Measure	Category	Frequency	Percentage
Gender	Male	192	54.08%
Gender	Female	163	45.92%
Age	<18	29	8.17%
	18–25	112	31.55%
	26–35	101	28.45%
	36–45	70	19.72%
	>45	43	12.11%
Education	High school or below	52	14.65%
	Bachelor's or specialist degree	213	60.00%
	Master's degree or above	90	25.35%
Monthly income (RMB/month)	< 2000	32	9.01%
	2000–5000	151	42.54%
	5001–10 000	124	34.93%
	>10 000	48	13.52%
Travel Frequency (trips per year)	<3	64	18.03%
	3–5	141	39.72%
	6–8	93	26.20%
	>8	57	16.05%

Note: Income values are expressed in Chinese Yuan (¥) per month. Travel frequency denotes the number of trips taken per year.

Measurement

All constructs were measured using multi-item scales validated in prior research. Specifically, OE (5 items) and PSE (4 items) were adapted from Ratten and Ratten (2007); EMA (4 items) from Doran et al. (2015); CA (3 items) from Yoo et al. (2018), and desire (3 items) from Hwang et al. (2019). BI (4 items) was adapted from Hwang et al. (2019) and Yoo et al. (2018), while PSR (5 items) was derived from Neuburger and Egger (2021).

All items used a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Control variables included gender, age, education, monthly income, and travel frequency, each measured with a single item (Ayeh et al., 2013; Kim et al., 2020). Reliability and validity were assessed in subsequent analyses, with full measurement items provided in Appendix B.

Data analysis and results

Data analysis was conducted using SPSS 25.0 and AMOS 23.0. The analytical procedure consisted of three steps: (1) testing for common method bias, (2) validating the measurement model, and (3) assessing the structural model and hypotheses through SEM.

Common method biases test

Because the survey data were self-reported, common method bias was assessed using the common latent factor approach (Podsakoff et al., 2003). We compared model fit indices with and without the latent factor. The differences were small (ΔTLI = 0.018, TLI: Tucker Lewis Index; ΔIFI = 0.016, IFI: Incremental Fit Index; ΔCFI = 0.011, CFI: Comparative Fit Index; ΔNFI = 0.018, NFI: Normed Fit Index; ΔRMSEA = 0.022, RMSEA: Root Mean Square Error of Approximation and ΔSRMR = 0.016, SRMR: Standardized Root Mean Square Residual), all below the 0.03 threshold (Williams et al., 2010). These results indicate that common method bias is not a significant concern in this study.

Validation of measurement model

Reliability was evaluated through Cronbach's α and composite reliability (CR). As shown in Table 2, Cronbach's α values ranged from 0.837 to 0.885 and CR values from 0.839 to 0.886, both above the recommended 0.800 threshold (Taber, 2018). Convergent validity was also confirmed in Table 2, with all factor loadings above 0.700 and average variance extracted (AVE) values above 0.500 (Asmelash and Kumar, 2019). Model fit indices in Table 3 further indicate that the measurement model has a good fit.

Table 2.

Reliability and validity.

Variables	Items	Factor loading (FL > 0.6)	Cronbach's α > 0.8	CR > 0.8	AVE > 0.5
Outcome expectancy	OE1	0.811	0.885	0.886	0.608
	OE2	0.778
	OE3	0.748
	OE4	0.790
	OE5	0.771
Environmental moral awareness	EMA1	0.799	0.841	0.841	0.570
	EMA2	0.722
	EMA3	0.749
	EMA4	0.749
Perceived self-efficacy	PSE1	0.813	0.849	0.851	0.588
	PSE2	0.770
	PSE3	0.736
	PSE4	0.746
Consumer attitude	CA1	0.829	0.851	0.853	0.659
	CA2	0.794
	CA3	0.811
Desire	DE1	0.835	0.837	0.839	0.635
	DE2	0.731
	DE3	0.821
Behavioral intention	BI1	0.841	0.861	0.863	0.612
	BI2	0.766
	BI3	0.761
	BI4	0.758

Table 3.

Values of measurement model fit indices.

	$χ^{2} / d f$	RMSEA	TLI	IFI	CFI	NFI
This study	1.180	0.023	0.989	0.991	0.991	0.942
Reference	<3.000	<0.080	>0.900	>0.900	>0.900	>0.900

A confirmatory factor analysis (CFA) was conducted in AMOS 23.0 using the maximum likelihood (ML) estimation method to validate the measurement model. The results showed that all constructs achieved satisfactory model fit, with key indices (e.g., CFI, TLI, IFI, RMSEA, and NFI) within the conventional cut-off values (CFI, TLI, IFI > 0.90; RMSEA < 0.06) (Hu and Bentler, 1999). Although the fit indices appear relatively high, they remain methodologically reasonable because (i) all constructs were adapted from well-validated scales, resulting in high factor loadings; (ii) the model is moderately complex, containing seven first-order constructs with limited cross-loadings; and (iii) the sample size (N = 355) was adequate for stable parameter estimation. These factors jointly contribute to a robust and parsimonious measurement model.

Discriminant validity was tested using the Fornell–Larcker criterion. As shown in Table 4, the square root of each construct's AVE was greater than its correlations with other constructs, supporting discriminant validity. Collectively, these findings confirm that the measurement model is reliable, valid, and appropriate for SEM analysis.

Table 4.

Correlations between constructs (note: ***P < .001).

	OE	EMA	PSE	CA	Desire	BI
OE	0.608
EMA	0.383***	0.570
PSE	0.341***	0.349***	0.588
CA	0.434***	0.424***	0.441***	0.659
Desire	0.440***	0.353***	0.291***	0.451***	0.635
BI	0.263***	0.276***	0.277***	0.412***	0.348***	0.612
$\sqrt{A V E}$	0.780	0.755	0.767	0.812	0.797	0.782

Validation of structural model

Model path test

SEM was employed to test the hypothesized relationships, with results shown in Tables 5 and 6 and Figures 2 and 3. Table 5 indicates that the six fit indices of the structural model meet acceptable thresholds, confirming good model fit.

Table 5.

The fit indices of structural model.

	$χ^{2} / d f$	RMSEA	TLI	IFI	CFI	NFI
This study	1.192	0.023	0.988	0.990	0.990	0.940
Reference	<3.000	<0.080	>0.900	>0.900	>0.900	>0.900

Table 6.

Path coefficients and significant results.

Validation of main effects
Path	Standardization coefficient	SE	CR	P
OE→CA	0.207	0.068	3.219	**
OE→Desire	0.298	0.078	4.336	***
EMA→CA	0.252	0.081	3.629	***
EMA→Desire	0.162	0.093	2.192	*
PSE→CA	0.338	0.081	5.011	***
PSE→Desire	0.027	0.094	0.366	.714
CA→Desire	0.248	0.083	3.211	**
CA→BI	0.394	0.070	5.821	***
Desire→BI	0.259	0.066	3.829	***
Validation of control variables
Gender→BI	−0.053	0.118	−0.943	.346
Age→BI	0.100	0.051	1.760	.078
Education→BI	−0.050	0.071	−0.843	.399
Monthly Income→BI	0.026	0.074	0.440	.660
Travel Frequency→BI	0.037	0.050	0.662	.508

Note: ***P < .001, **P < .01, *P < .05

As reported in Table 6 and Figure 2, value-based cognitive drivers significantly influence motivational appraisals. OE (β=0.207, P < .01), EMA (β=0.252, P < .001), and PSE (β=0.338, P < .001) positively affect CA, supporting H1a-H1c. In addition, OE (β=0.298, P < .001) and EMA (β=0.162, P < .05) positively affect desire, while PSE (β=0.027，P = .714) is nonsignificant. Thus, H2a-H2b are supported and H2c is rejected. This non-significant effect may reflect that desire, as an affective construct, is less responsive to cognitive evaluations of capability, even though PSE strengthens confidence and attitudes.

Finally, CA (β=0.394, P < .001) and desire (β=0.259, P < .001) significantly predict BI, supporting H4a and H4b. CA also directly predicts BI (β=0.248, P < .01), supporting H3. Control variables have no significant effect (P > .001). The revised structural model is shown in Figure 3.

Mediating effect test

The mediating roles of CA and desire were examined using the bootstrap method in AMOS (5000 samples, 95% confidence interval (CI)) (Mackinnon et al., 2004; Preacher and Hayes, 2008). Results are reported in Table 7.

Table 7.

Standardized bootstrap mediating effect test.

Path	Effect	SE	Bias-corrected 95% CI		Percentile 95% CI
Path	Effect	SE	Lower, Upper	P	Lower, Upper	P
OE→CA→BI	0.082	0.032	[0.023, 0.151]	.004	[0.020, 0.147]	.005
OE→Desire→BI	0.078	0.028	[0.030, 0.146]	.000	[0.027, 0.140]	.000
OE→CA→Desire→BI	0.014	0.008	[0.004, 0.038]	.002	[0.002, 0.032]	.007
EMA→CA→BI	0.099	0.034	[0.042, 0.181]	.000	[0.038, 0.172]	.000
EMA→Desire→BI	0.044	0.026	[0.005, 0.106]	.025	[0.003, 0.102]	.033
EMA→CA→Desire→BI	0.017	0.009	[0.006, 0.043]	.001	[0.004, 0.037]	.002
PSE→CA→BI	0.134	0.040	[0.065, 0.221]	.000	[0.064, 0.217]	.000
PSE→CA→Desire→BI	0.023	0.011	[0.008, 0.053]	.001	[0.006, 0.047]	.002

For OE, significant indirect effects on BI were observed through CA (Effect = 0.082, SE = 0.032) and desire (Effect = 0.078, SE = 0.028, SE: standard error of estimate). A sequential path via CA and desire was also significant (Effect = 0.014, SE = 0.008). For EMA, indirect effects were significant through CA (Effect = 0.099, SE = 0.034) and desire (Effect = 0.044, SE = 0.026), as well as the serial path through both mediators (Effect = 0.017, SE = 0.009). For PSE, mediation occurred only via CA (Effect = 0.134, SE = 0.040), with a smaller sequential effect through CA and desire (Effect = 0.023, SE = 0.011).

Together, these findings validate hypotheses H1a–H1c, H2a–H2b, H3, and H4a–H4b, confirming that CA and desire act as key mediators linking cognitive evaluations to behavioral intention. Importantly, the results highlight that while PSE enhances attitudes, its influence on BI is realized primarily through cognitive rather than affective pathways.

Moderating effect test

The moderating role of PSR was assessed using Hayes’ PROCESS macro (Model 1) with 5000 bootstrap samples and 95% confidence intervals (CI). This procedure was adopted after estimating the main structural model in AMOS. The use of PROCESS was appropriate because PSR was treated as a composite observed variable rather than a latent construct. PROCESS allows straightforward estimation of interaction terms and provides simple slope analyses for more intuitive interpretation of moderation effects. In addition, this approach is consistent with the overall two-stage SEM-ANN analytical framework, serving as an auxiliary tool to examine moderation effects at the linear SEM stage before proceeding to the nonlinear predictive analysis. A supplementary latent-interaction check in AMOS produced consistent results, confirming the robustness of the findings.

As shown in Tables 8 and 9, both interaction terms were significant: CA * PSR (β = 0.142, P < .01) and desire * PSR (β = 0.109, P < .05). Figures 4 and 5 illustrate steeper slopes under high PSR, confirming a positive moderating effect.

Table 8.

Moderating effect test of PSR (CA).

	BI		BI (with interaction item)
	β	t	β	t
CA	0.323***	6.118	0.363***	6.72
PSR	0.258***	4.878	0.274***	5.22
CA * PSR			0.142**	2.899
R²	0.252		0.269
Adjusted R²	0.247		0.263
F	59.161***		43.072***

Table 9.

Moderating effect test of PSR (desire).

	BI		BI (with interaction item)
	β	t	β	t
Desire	0.254***	4.946	0.262***	5.116
PSR	0.312***	6.074	0.335***	6.437
Desire * PSR			0.109*	2.270
R²	0.226		0.237
Adjusted R²	0.221		0.230
F	51.325***		36.338***

Contrary to H5a and H5b, PSR was found to amplify rather than weaken the effects of CA and desire on BI. This suggests that when consumers already hold favorable attitudes or a strong desire, higher perceived risk may heighten engagement and readiness, reinforcing rather than diminishing behavioral intention. This unexpected result aligns with research on sensation-seeking and risk-related tourism, where risk is not merely a deterrent but can serve as an additional source of excitement and challenge (Atkinson, 1957).

Thus, PSR in this context functions not as a barrier but as a catalyst for adoption, particularly among consumers who are motivated by novelty, adventure, or technological innovation. This finding represents a significant theoretical contribution, as it challenges conventional risk-perception theory and underscores the complexity of consumer decision-making in emerging tourism markets.

Artificial neural network analysis results

Artificial neural network approach

ANN is a computational model consisting of interconnected neurons arranged in input, hidden, and output layers. Owing to its ability to learn nonlinear relationships and adaptively adjust weights, ANN has been widely used in classification, regression, and pattern recognition. However, its applications in the field of low-altitude tourism remain scarce.

This study employs a hybrid SEM-ANN approach to overcome the limitations of using either method in isolation. SEM is effective for testing linear causal relationships and validating hypotheses, but it assumes compensatory effects and may oversimplify complex decision-making (Xiong et al., 2022). In contrast, ANN excels at capturing nonlinear and non-compensatory patterns, making it better suited for prediction and importance ranking. Its limitation, however, lies in its “black box” nature, which restricts its explanatory capacity for theory testing.

To integrate the strengths of both methods, a two-stage SEM-ANN procedure was applied. In the first stage, SEM was used to test hypotheses and identify significant predictors. In the second stage, these predictors were entered as input neurons in the ANN, where sensitivity analysis was conducted to determine their relative importance (Leong et al., 2024). This hybrid design improves both the explanatory power and predictive accuracy of the research model, providing a more comprehensive understanding of consumer adoption mechanisms.

Validations of neural networks

Compared with the confirmatory purpose of SEM, the ANN models were employed to capture nonlinear and non-compensatory patterns and to maximize variance explained, which accounts for the higher R² values reported in this section. The observed R² gap between the SEM and ANN results is therefore methodological rather than empirical. In SEM, R² values are derived from hypothesis-driven linear models that test specific causal relationships, while ANN optimizes predictive accuracy through iterative learning and a broader set of predictors. Consequently, higher R² values in ANN are methodologically expected and reflect its predictive optimization capacity, not inconsistency across methods. This complementarity underscores the rationale for adopting the two-stage SEM-ANN analytical framework, where SEM validates theory and ANN enhances predictive robustness.

Based on the SEM results, three ANN models were constructed using SPSS 22, as illustrated in Figure 6: Model 1: Desire and CA predicting BI; Model 2: OE, EMA, and PSE predicting desire; Model 3: OE, EMA, and PSE predicting CA.

All models adopted a multi-layer perceptron architecture trained with the feedforward backpropagation algorithm. The sigmoid activation function was used, and the number of hidden nodes was automatically generated by the system. To mitigate overfitting, 10-fold cross-validation was applied, with 90% of the data used for training and 10% reserved for testing.

Model accuracy was evaluated using root mean square error (RMSE), with smaller values indicating stronger predictive performance. As reported in Table 10, all three ANN models achieved low RMSE values: Model 1 (training = 0.1471, test = 0.1386), Model 2 (training = 0.1541, test = 0.1603), and Model 3 (training = 0.1527, test = 0.1615). In addition, the R² values indicated that the predictors explained 80.58% of the variance in BI, 83.75% in desire, and 83.42% in CA, confirming the high predictive and explanatory capability of the ANN models.

Table 10.

RMSE values for ANN models.

	Model 1 (R² = 80.58%)						Model 2 (R² = 83.75%)						Model 3 (R² = 83.42%)
	Train			Test			Train			Test			Train			Test
Network	N	SSE	RMSE	N	SSE	RMSE	N	SSE	RMSE	N	SSE	RMSE	N	SSE	RMSE	N	SSE	RMSE
1	312	6.696	0.1465	43	0.804	0.1367	310	7.518	0.1557	45	1.016	0.1503	318	7.483	0.1534	37	0.748	0.1422
2	318	6.987	0.1482	37	0.719	0.1394	315	7.080	0.1499	40	1.352	0.1838	319	7.369	0.1520	36	1.002	0.1668
3	319	6.987	0.1480	36	0.719	0.1413	321	7.882	0.1567	34	0.592	0.1320	315	7.438	0.1537	40	0.934	0.1528
4	315	6.520	0.1439	40	0.897	0.1497	315	7.547	0.1548	40	0.825	0.1436	316	7.746	0.1566	39	0.939	0.1552
5	316	7.011	0.1490	39	0.516	0.1150	316	7.388	0.1529	39	1.022	0.1619	309	6.789	0.1482	46	1.479	0.1793
6	309	7.483	0.1556	46	0.863	0.1370	311	7.073	0.1508	44	1.429	0.1802	325	7.686	0.1538	30	0.651	0.1473
7	325	6.949	0.1462	30	0.448	0.1222	322	7.477	0.1524	33	1.093	0.1820	318	7.331	0.1518	37	0.960	0.1611
8	318	6.436	0.1423	37	0.986	0.1632	315	7.481	0.1541	40	0.903	0.1502	311	7.252	0.1527	44	0.988	0.1498
9	311	6.782	0.1477	44	0.747	0.1303	319	8.227	0.1606	36	0.731	0.1425	325	7.522	0.1521	30	1.207	0.2006
10	325	6.735	0.1440	30	0.688	0.1514	312	7.313	0.1531	43	1.340	0.1765	310	7.192	0.1523	45	1.151	0.1599
Mean	316.8	6.859	0.1471	38.2	0.739	0.1386	315.6	7.499	0.1541	39.4	1.030	0.1603	316.6	7.381	0.1527	38.4	1.006	0.1615
SD	5.4	0.297	0.0037	5.4	0.164	0.0142	4.1	0.347	0.0031	4.1	0.279	0.0191	5.6	0.272	0.0021	5.6	0.233	0.0173

Notes:

1. N = number of samples, SSE = sum square of error, RMSE = root mean square of error, SD = standard deviation.

2. In Model 1, desire and consumer attitude served as the input neurons, while behavioral intention served as the output neuron.

3. In Model 2, outcome expectancy, environmental moral awareness, and perceived self-efficacy served as the input neurons, while desire served as the output neuron.

4. In Model 3, outcome expectancy, environmental moral awareness, and perceived self-efficacy served as the input neurons, while consumer attitude served as the output neuron.

5. R² = 1 – RMSE/S². S² is the variance of the desired output for the test data.

Sensitivity analysis

Sensitivity analysis was conducted to assess the relative importance of predictors in the ANN models by quantifying how changes in input variables influence the output. The analysis used normalized importance scores, which allow direct comparison of predictor contributions.

Table 11 summarizes the results. In Model 1, CA emerged as the most influential factor for BI, while desire also contributed substantially, though at a lower importance level (78.1%). In Model 2, OE was identified as the strongest predictor of desire, followed by EMA (83.7%) and PSE (70.0%). In Model 3, PSE played the dominant role in predicting CA, with EMA (97.1%) and OE (81.3%) also exerting strong effects.

Table 11.

Sensitivity analysis.

	Model 1		Model 2			Model 3
Network	Desire	CA	OE	EMA	PSE	OE	EMA	PSE
1	0.402	0.598	0.394	0.359	0.248	0.229	0.390	0.380
2	0.478	0.522	0.377	0.297	0.326	0.326	0.336	0.338
3	0.438	0.562	0.374	0.356	0.269	0.298	0.301	0.401
4	0.418	0.582	0.388	0.338	0.275	0.283	0.394	0.323
5	0.466	0.534	0.462	0.305	0.233	0.323	0.358	0.319
6	0.387	0.613	0.375	0.355	0.271	0.302	0.345	0.353
7	0.435	0.565	0.398	0.365	0.237	0.296	0.353	0.351
8	0.465	0.535	0.420	0.267	0.313	0.295	0.352	0.353
9	0.469	0.531	0.381	0.348	0.271	0.285	0.377	0.337
10	0.428	0.572	0.373	0.309	0.318	0.283	0.281	0.436
Mean	0.439	0.561	0.394	0.330	0.276	0.292	0.349	0.359
Normalized Importance	78.1%	100%	100%	83.7%	70.0%	81.3%	97.1%	100%

These findings are noteworthy for two reasons. First, they are consistent with SEM results, particularly in showing that PSE had no significant direct effect on desire; ANN further corroborated this by ranking PSE as the weakest predictor in that pathway. Second, they highlight the added value of ANN: while SEM reveals linear causal relationships, ANN provides a more nuanced picture of predictor strength under non-compensatory conditions (Xiong et al., 2022). The convergence of these two methods underscores the robustness of the research model and demonstrates how ANN enriches theory testing by identifying the hierarchical importance of cognitive and motivational drivers in low-altitude tourism adoption.

Discussion and conclusion

Main findings

Low-altitude tourism is emerging as a novel form of travel driven by advancements in aviation technologies and supportive governmental policies. As the success of this market ultimately depends on consumers’ willingness to adopt and continue using the service, understanding their behavioral mechanisms is crucial. Drawing on the ENAM and the EDBM—and incorporating insights from the perceived value and service quality literature—this study explored the cognitive and motivational antecedents that shape consumers’ adoption intentions. Using a two-stage SEM-ANN approach, the analysis yielded several theoretically and practically meaningful findings, discussed in the following subsections.

Factors influencing CA toward low-altitude tourism

The results revealed that all three cognitive evaluations—OE, EMA, and PSE—positively influence CA toward low-altitude tourism. This confirms the fundamental role of cognitive appraisals in shaping overall evaluations of novel services.

Among these, OE emerged as a particularly strong determinant, aligning with technology adoption research where anticipated positive outcomes represent a key driver of favorable attitudes (Chang et al., 2025). Consumers are more likely to hold a positive attitude toward low-altitude tourism when they expect both hedonic benefits (e.g., excitement, unique scenery) and utilitarian benefits (e.g., convenience, time-saving).

Similarly, EMA exerted a significant influence on CA, consistent with research emphasizing moral and environmental considerations in behavioral models (Yoo et al., 2018). Consumers tend to strengthen their attitudes when they perceive low-altitude tourism as aligning with their environmental values—such as supporting sustainable aviation technologies or offering a cleaner travel alternative.

Finally, PSE was also found to positively affect CA. Self-efficacy provides consumers with the confidence that they can successfully navigate the process of adopting and using an unfamiliar technology (Ren et al., 2025). When individuals feel capable of managing the information, safety procedures, and operational aspects associated with low-altitude tourism, they are more likely to overcome initial cognitive barriers and develop favorable attitudes toward the service.

Factors influencing desire for low-altitude tourism

The analysis confirmed that value-based cognitive drivers play a central role in shaping desire for low-altitude tourism. Specifically, OE, EMA, and CA each exhibited significant positive effects on desire.

The dominant role of OE underscores that the anticipation of rewarding and often hedonic outcomes is crucial for generating affective motivation (Pool et al., 2016). The promise of novel experiences and unique aerial perspectives acts as a strong psychological incentive, reinforcing that perceived rewards serve as the primary motivational source in experiential tourism.

EMA was also validated as a key factor enhancing desire, integrating the intrinsic moral dimension emphasized by ENAM. This indicates that consumers’ motivation to adopt low-altitude tourism is not purely utilitarian but also value-driven. When individuals believe that low-altitude tourism supports environmental sustainability—through cleaner technologies or lower emissions—the sense of moral alignment provides a powerful, internalized motivation to act (Yi et al., 2020; Yoo et al., 2018).

Furthermore, CA was shown to positively influence desire. A positive attitude functions as the cognitive foundation that facilitates the emotional formation of desire (Bernecker, 2023). Conversely, while PSE strongly shapes CA, its lack of a direct effect on desire suggests a divergence between perceived competence (a cognitive readiness) and affective motivation.

Factors influencing bi toward low-altitude tourism

Both CA and desire exerted significant positive effects on BI to adopt low-altitude tourism. This dual influence highlights that both stable cognitive evaluations and dynamic emotional motivations jointly drive adoption of novel, experience-oriented services.

First, CA positively affects BI, consistent with the TPB and related models that emphasize attitude as a stable and robust predictor of intention (Hwang, Kim et al., 2021). When consumers perceive low-altitude tourism as beneficial, feasible, and aligned with their values, they are more likely to develop the intention to try or recommend the service.

Second, desire also strongly and directly influences BI, underscoring the power of affective motivation in converting intention into action (Yi et al., 2020). In experiential tourism, this emotional pull—rooted in hedonic anticipation and moral congruence—translates into concrete willingness to participate.

A particularly noteworthy and theoretically intriguing finding concerns PSR. Contrary to conventional technology adoption studies where perceived risk typically suppresses behavioral intention, PSR in this study exhibited no significant direct effect on BI. This outcome suggests a context-dependent dynamic shaped by the characteristics of the aviation sector and consumer psychology.

Given the highly regulated nature of aviation and the strong institutional credibility associated with flight safety, consumers may perceive sufficient safeguards—both technological and governmental—to neutralize risk aversion (Yi et al., 2020). Moreover, for consumers already exhibiting positive attitudes and strong desire, moderate levels of perceived risk may actually heighten arousal and reinforce engagement, consistent with cognitive consistency and sensation-seeking theories (Belk et al., 2003; Kruglanski et al., 2018).

This nuanced interpretation indicates that perceived risk in low-altitude tourism may not act as a straightforward deterrent but can instead function as a motivational amplifier—stimulating curiosity and emotional involvement under conditions of perceived safety and institutional trust. This finding contributes a more critical and contextually grounded understanding of risk perception in the adoption of emerging air tourism technologies.

Theoretical implications

This study advances theory in several important ways. First, it reinforces the applicability of the ENAM in technology-mediated tourism. OE and PSE are confirmed as key predictors of CA, although PSE does not directly influence desire or BI. While this diverges from earlier findings (Kim et al., 2020; Nel and Heyns, 2017), it supports studies showing that PSE primarily strengthens attitudes rather than affective states (Hsu et al., 2004; Stone and Baker-Eveleth, 2013). Attitude thus emerges as a pivotal mediator linking cognitive evaluations to adoption intention, particularly for digitally literate, risk-tolerant consumers.

Second, the findings extend the role of EMA. EMA not only enhances CA but also strengthens desire, confirming that moral and ecological values are central to pro-environmental behavior (De Groot and Steg, 2009; Stern, 2000). This study contributes by demonstrating that environmental concern can generate both cognitive approval and affective commitment in low-altitude tourism, a sector increasingly promoted as a sustainable alternative.

Third, consistent with the EDBM, desire is shown to be a critical motivational force in technology adoption. Shaped by attitudes, desire strongly predicts BI, particularly in hedonic settings such as low-altitude tourism. This supports the argument that consumer adoption is often fueled by a persistent desire for novelty and unique experiences (Belk et al., 2003).

Fourth, the findings challenge conventional perspectives on risk perception. Contrary to studies suggesting that risk reduces adoption (Chiu et al., 2014; Sah et al., 2021), PSR in this study amplifies intention when consumers already hold favorable attitudes or strong desire. This aligns with cognitive consistency theory (Kruglanski et al., 2018), suggesting that consumers resolve dissonance by discounting risks and reinforcing commitment. The result also resonates with sensation-seeking and adventure tourism literature, where risk enhances rather than deters motivation.

Finally, the integration of ENAM and EDBM explains over 71% of BI variance, outperforming earlier models in sustainable tourism (Han, 2015). The hybrid SEM-ANN approach further demonstrates the value of combining linear hypothesis testing with nonlinear predictor ranking (Hew et al., 2016; Paul et al., 2016). This methodological contribution underscores the robustness of the findings and provides a framework for examining technology-mediated, risk-sensitive consumer behavior in tourism markets, particularly in digitally advanced and sustainability-oriented regions such as East Asia.

Practical implications

The findings provide actionable insights for tourism firms, platform providers, and marketers seeking to expand low-altitude tourism, particularly among younger, digitally engaged consumers.

First, service design should be aligned with consumers’ environmental values and lifestyle preferences, which often vary across regions. Providers should assess expectations regarding sustainability, convenience, and novelty, and emphasize features such as carbon reduction, aerial perspectives, and immersive experiences when relevant. Short-form video platforms and youth-oriented social media are effective channels for communicating these values, while lifestyle-based segmentation can help firms identify and target consumer groups more precisely (Hwang and Kim, 2021).

Second, operators should enhance OE by highlighting experiential uniqueness, accessibility, and value for money—attributes that resonate strongly with younger, price-sensitive consumers (Kim et al., 2025). Although eVTOL-based services may reduce costs through efficiency gains, context-specific benefits are likely to vary. For example, in remote destinations, consumers may particularly value time-saving and scenic advantages, whereas in urban markets, it may be beneficial to explore the integration of aerial experiences with cultural attractions or influencer-led routes (Gao et al., 2025). These possibilities should be regarded as theoretical extrapolations rather than empirical conclusions, offering directions for future market testing.

Third, the results demonstrate that PSR does not weaken adoption intentions when attitudes and desire are favorable. This challenges the traditional view that safety concerns inevitably deter adoption (Clothier et al., 2015; Zhu, 2019). Instead, firms should focus on other salient risks—such as data privacy, noise, or over-commercialization—which may remain influential (Lowry et al., 2017; Tussyadiah et al., 2016). By reframing manageable safety concerns as part of the novelty and excitement of new experiences, providers can communicate transparency and control, thereby transforming potential anxiety into confidence and curiosity. In practice, this reframing can be achieved through transparent communication of safety measures, gradual experience design (e.g., short scenic or introductory flights), and interactive features that enhance travelers’ perceived control. These actions help translate perceived risk into a sense of managed adventure rather than danger.

Fourth, the ANN results (Table 11) provide additional managerial insights by identifying CA as the strongest predictor of BI and PSE as the most critical antecedent of CA. Accordingly, the most efficient way to strengthen BI is to enhance consumers’ PSE, which in turn reinforces positive attitudes and subsequent adoption intentions. To achieve this, practitioners should focus on strategies that strengthen consumers’ confidence and perceived competence in engaging with low-altitude tourism. These may include offering educational and experiential activities such as simulator training, guided trial flights, or virtual demonstrations to improve perceived mastery; providing clear and reassuring communication about safety standards, procedures, and pilot expertise; and designing progressive participation options that allow consumers to tailor their experience level and develop a sense of control. Strengthening PSE in these ways can elevate CA and, consequently, BI, representing the most direct and effective lever for increasing consumer adoption.

Finally, marketers could consider positioning low-altitude tourism within an aspirational lifestyle that emphasizes innovation, exclusivity, and immersive adventure. Such positioning is conceptually consistent with the hedonic and value-based motivations identified in this study, but should be validated by further empirical research. Story-driven campaigns, influencer collaborations, and augmented previews of flight routes may help frame low-altitude tourism as a trendsetting and emotionally engaging activity (Belk et al., 2003; Hwang and Kim, 2021). These strategies should be seen as future-oriented implications, intended to guide practice and subsequent research rather than as conclusions derived directly from the data.

Limitations

This study has several limitations. First, it examines BI rather than actual usage. Although the SEM-ANN approach improves predictive robustness, it cannot replace longitudinal analysis. Future research could adopt time-series or panel data to verify behavioral consistency. Second, the data were collected only in mainland China, which limits generalizability to other regions. Cross-cultural and cross-regulatory comparisons are needed. Third, the model excludes factors such as privacy concerns, trust in automation, and infrastructural support, which may also influence adoption. Finally, the study focuses on the demand side and does not account for supply or policy constraints such as cost, governance, and operational feasibility. Future research should integrate these elements to link consumer interest with practical implementation.

Footnotes

ORCID iDs

Boyi Zhu

He Zhu

Ziming Wang

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Appendix

References

Ajzen

(1991) The theory of planned behavior. Organizational Behavior and Human Decision Processes 50(2): 179–211.

Alwateer

Loke

Zuchowicz

(2019) Drone services: issues in drones for location-based services from human-drone interaction to information processing. Journal of Location Based Services 13(2): 94–127.

Asmelash

Kumar

(2019) Assessing progress of tourism sustainability: developing and validating sustainability indicators. Tourism Management 71: 67–83.

Atkinson

(1957) Motivational determinants of risk-taking behavior. Psychological Review 64(6, Pt.1): 359–372.

Ayeh

Law

(2013) Predicting the intention to use consumer-generated media for travel planning. Tourism Management 35: 132–143.

Bagozzi

(2007) The legacy of the technology acceptance model and a proposal for a paradigm shift. Journal of the Association for Information Systems 8(4): 243–254.

Bagozzi

(1992) The self-regulation of attitudes, intentions, and behavior. Social Psychology Quarterly 55(2): 178–204.

Baker-Eveleth

Stone

(2008) Expectancy theory and behavioral intentions to use computer applications. Interdisciplinary Journal of Information, Knowledge, and Management 3: 135–146.

Bandura

(1986) Social foundations of thought and action: A social cognitive theory. Englewood Cliffs, NJ, US: Prentice-Hall, Inc. pp. xiii, 617.

10.

Bauer

(1969) Consumer behavior as risk taking. the 43rd. Conference of the American Marketing Association, Chicago. https://www.scirp.org/reference/ReferencesPapers?ReferenceID=1409819

11.

Belk

Ger

Askegaard

and article.], [David Glen Mick served as editor and Eric J. Arnould served as associate editor for this. (2003). The fire of desire: a multisited inquiry into consumer passion. Journal of Consumer Research, 30(3), 326–351.

12.

Bernecker

(2023) Balancing desires and control. Nature Reviews Psychology 2(12): 722–722.

13.

Brand

Anable

Morton

(2019) Lifestyle, efficiency and limits: modelling transport energy and emissions using a socio-technical approach. Energy Efficiency 12(1): 187–207.

14.

Brislin

(1980) Translation and content analysis of oral and written materials. In: Handbook of Cross-Cultural Psychology: Methodology. Allyn & Bacon, 2, 389–444.

15.

Brühlmann

Petralito

Aeschbach

, et al. (2020) The quality of data collected online: an investigation of careless responding in a crowdsourced sample. Methods in Psychology 2: 100022.

16.

Chang

Huang

, et al. (2025) A study on public acceptance of urban air traffic based on extended TAM theory. Journal of Transport Information and Safety 43(3): 171–180.

17.

Chen

S-C

Hung

C-W

(2016) Elucidating the factors influencing the acceptance of green products: an extension of theory of planned behavior. Technological Forecasting and Social Change 112: 155–163.

18.

Chiu

C-M

Wang

ETG

Fang

Y-H

, et al. (2014) Understanding customers’ repeat purchase intentions in B2C e-commerce: the roles of utilitarian value, hedonic value and perceived risk: understanding customers’ repeat purchase intentions. Information Systems Journal 24(1): 85–114.

19.

Civil Aviation Administration of China (CAAC). (2023). White Paper on Low-Altitude Airspace Reform and Civil UAV Integration. https://www.caac.gov.cn/English/

20.

Clothier

Greer

, et al. (2015) Risk perception and the public acceptance of drones. Risk Analysis: An Official Publication of the Society for Risk Analysis 35(6): 1167–1183.

21.

Cox

Rich

(1964) Perceived risk and consumer decision-making: the case of telephone shopping. Journal of Marketing Research 1(4): 32–39.

22.

De Groot

JIM

Steg

(2009) Morality and prosocial behavior: the role of awareness, responsibility, and norms in the norm activation model. The Journal of Social Psychology 149(4): 425–449.

23.

Doran

Hanss

Larsen

(2015) Attitudes, efficacy beliefs, and willingness to pay for environmental protection when travelling. Tourism and Hospitality Research 15(4): 281–292.

24.

Gao

Wang

Cai

, et al. (2025) Unveiling the nexus: how perceived intelligence shapes the willingness to use robot. International Journal of Human–Computer Interaction 41: 1–10.

25.

Han

(2015) Travelers’ pro-environmental behavior in a green lodging context: converging value-belief-norm theory and the theory of planned behavior. Tourism Management 47: 164–177.

26.

Han

Meng

Kim

(2017) Emerging bicycle tourism and the theory of planned behavior. Journal of Sustainable Tourism 25(2): 292–309.

27.

Herzenstein

Posavac

Brakus

(2007) Adoption of new and really new products: the effects of self-regulation systems and risk salience. Journal of Marketing Research 44(2): 251–260.

28.

Hew

T-S

Leong

L-Y

Ooi

K-B

, et al. (2016) Predicting drivers of mobile entertainment adoption: a two-stage sem-artificial-neural-network analysis. Journal of Computer Information Systems 56(4): 352–370.

29.

Hsu

Chiu

(2004) Determinants of continued use of the WWW: an integration of two theoretical models. Industrial Management & Data Systems 104(9): 766–775.

30.

Bentler

(1999) Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal 6(1): 1–55.

31.

Huang

Backman

, et al. (2016) Exploring the implications of virtual reality technology in tourism marketing: an integrated research framework. International Journal of Tourism Research 18(2): 116–128.

32.

Hwang

Kim

(Markham), Kim

, et al. (2021) Investigation of perceived risks and their outcome variables in the context of robotic restaurants. Journal of Travel & Tourism Marketing 38(3): 263–281.

33.

Hwang

Kim

(2019) Investigating motivated consumer innovativeness in the context of drone food delivery services. Journal of Hospitality and Tourism Management 38: 102–110.

34.

Hwang

Kim

(2021) Expected benefits with using drone food delivery services: its impacts on attitude and behavioral intentions. Journal of Hospitality and Tourism Technology 12(3): 593–606.

35.

Hong

Kang

(2011) An international comparison of technology adoption: Testing the UTAUT model. Information & Management 48(1): –8.

36.

Jansson

Marell

Nordlund

(2010) Green consumer behavior: determinants of curtailment and eco-innovation adoption. Journal of Consumer Marketing 27(4): 358–370.

37.

Jin

(2024) The evolution and challenges of low-altitude economy: Insights from experience in China. In Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management (MLSCM 2024), p. 32–36. 10.5220/0013228700004558

38.

Kim

Hwang

(2020) Merging the norm activation model and the theory of planned behavior in the context of drone food delivery services: does the level of product knowledge really matter? Journal of Hospitality and Tourism Management 42: 1–11.

39.

Kim

Radic

Ariza-Montes

, et al. (2025) Cars are ready to fly: urban air mobility and psychological process of sustainable travel mode choices. The International Journal of Aerospace Psychology 35(2): 64–84.

40.

Kim

Lee

C-K

Jung

(2020) Exploring consumer behavior in virtual reality tourism using an extended stimulus-organism-response model. Journal of Travel Research 59(1): 69–89.

41.

Klauser

Pedrozo

(2017) Big data from the sky: popular perceptions of private drones in Switzerland. Geographica Helvetica 72(2): 231–239.

42.

Kruglanski

Jasko

Milyavsky

, et al. (2018) Cognitive consistency theory in social psychology: a paradigm reconsidered. Psychological Inquiry 29(2): 45–59.

43.

Lee

M-C

(2009) Factors influencing the adoption of internet banking: an integration of TAM and TPB with perceived risk and perceived benefit. Electronic Commerce Research and Applications 8(3): 130–141.

44.

Leone

Perugini

Ercolani

(1999) A comparison of three models of attitude–behavior relationships in the studying behavior domain. European Journal of Social Psychology 29(2–3): 161–189.

45.

Leong

L-Y

Hew

T-S

Ooi

K-B

, et al. (2024) An SEM-ANN approach—guidelines in information systems research. Journal of Computer Information Systems 65(6): 1–32.

46.

Zhu

(2023) The impact of cue-interaction stimulation on impulse buying intention on virtual reality tourism E-commerce platforms. Journal of Travel Research 63(5): 1256–1279.

47.

Liang

Zhang

, et al. (2025) Assessing the deployment of electric aircraft from energy, environmental, and economic perspectives. Sustainability 17(12): 5453.

48.

Lindenberg

Steg

(2007) Normative, gain and hedonic goal frames guiding environmental behavior. Journal of Social Issues 63(1): 117–137.

49.

Liu

Hao

Geng

, et al. (2025) Adopting electric vertical takeoff and landing aircraft for green transport transformation. Fundamental Research. 10.1016/j.fmre.2025.04.002.

50.

Lowry

Dinev

Willison

(2017) Why security and privacy research lies at the centre of the information systems (IS) artefact: proposing a bold research agenda. European Journal of Information Systems 26(6): 546–563.

51.

Luo

(2025) Service model of smart tourism cities integrated with low-altitude technology. Engineering Management in Production and Services 17(1): 92–110.

52.

Mackinnon

Lockwood

Williams

(2004) Confidence limits for the indirect effect: distribution of the product and resampling methods. Multivariate Behavioral Research 39(1): 99.

53.

Meade

Craig

(2012) Identifying careless responses in survey data. Psychological Methods 17(3): 437–455.

54.

Nel

Heyns

(2017) The in-store factors influencing the useintention of proximity mobile-payment applications. Management Dynamics : Journal of the Southern African Institute for Management Scientists 26(2): 2–20.

55.

Neuburger

Egger

(2021) Travel risk perception and travel behaviour during the COVID-19 pandemic 2020: a case study of the DACH region. Current Issues in Tourism 24(7): 1003–1016.

56.

Pai

Kang

Liu

, et al. (2021) An examination of revisit intention based on perceived smart tourism technology experience. Sustainability 13(2): Article 2.

57.

Parker

Rains

Ivanov

(2016) Examining the “blanket of protection” conferred by inoculation: the effects of inoculation messages on the cross-protection of related attitudes. Communication Monographs 83(1): 49–68.

58.

Paul

Modi

Patel

(2016) Predicting green product consumption using theory of planned behavior and reasoned action. Journal of Retailing and Consumer Services 29: 123–134.

59.

Perugini

Bagozzi

(2001) The role of desires and anticipated emotions in goal-directed behaviours: broadening and deepening the theory of planned behaviour. The British Journal of Social Psychology 40(Pt 1): 79–98.

60.

Perugini

Bagozzi

(2004) The distinction between desires and intentions. European Journal of Social Psychology 34(1): 69–84.

61.

Piçarra

Giger

J-C

(2018) Predicting intention to work with social robots at anticipation stage: assessing the role of behavioral desire and anticipated emotions. Computers in Human Behavior 86: 129–146.

62.

Podsakoff

MacKenzie

Lee

J-Y

, et al. (2003) Common method biases in behavioral research: a critical review of the literature and recommended remedies. Journal of Applied Psychology 88(5): 879–903.

63.

Pool

Sennwald

Delplanque

, et al. (2016) Measuring wanting and liking from animals to humans: a systematic review. Neuroscience & Biobehavioral Reviews 63: 124–142.

64.

Preacher

Hayes

(2008) Asymptotic and resampling strategies for assessing and comparing indirect effects in multiple mediator models. Behavior Research Methods 40(3): 879–891.

65.

Ratten

(2007) Social cognitive theory in technological innovations. European Journal of Innovation Management 10(1): 90–108.

66.

Ravich

T. M.

(2020). On-Demand aviation: governance challenges of urban air mobility (“UAM”). European Journal of Information Systems (EJIS) 124(3): 657–689.

67.

Ren

Yin

Zhang

, et al. (2025) Exploring the nonlinear relationship between techno-overload and digital creativity in Industry 4.0 companies: the effects of digital leadership and digital self-efficacy. Behaviour & Information Technology 44(19): 4853–4867.

68.

Ren

Cheng

(2020) Model of third-party risk index for unmanned aerial vehicle delivery in urban environment. Sustainability 12(20): 8318.

69.

Sah

Gupta

Bani-Hani

(2021) Analysis of barriers to implement drone logistics. International Journal of Logistics Research and Applications 24(6): 531–550.

70.

Schwartz

(1977) Normative influences on altruism. In: Berkowitz

(eds) Advances in Experimental Social Psychology. Academic Press, 10, 221–279. 10.1016/S0065-2601(08)60358-5

71.

Sekerka

Bagozzi

(2007) Moral courage in the workplace: moving to and from the desire and decision to act. Business Ethics: A European Review 16(2): 132–149.

72.

Stern

(2000) Toward a coherent theory of environmentally significant behavior. Journal of Social Issues 56(3): 407–424.

73.

Stone

Baker-Eveleth

(2013) Students’ expectation, confirmation, and continuance intention to use electronic textbooks. Computers in Human Behavior 29(3): 984–990.

74.

Sun

(2014) How risky are services? An empirical investigation on the antecedents and consequences of perceived risk for hotel service. International Journal of Hospitality Management 37: 171–179.

75.

Taber

(2018) The use of cronbach’s alpha when developing and reporting research instruments in science education. Research in Science Education 48(6): 1273–1296.

76.

Tang

(2026) Can low altitude economy development bring economic and environmental dividends? –evidence from Chinese cities. Journal of Air Transport Management 131: 102919.

77.

Tsai

Y-T

Tiwasing

(2021) Customers’ intention to adopt smart lockers in last-mile delivery service: a multi-theory perspective. Journal of Retailing and Consumer Services 61: 102514.

78.

Tussyadiah

Wang

Jia

(2016) Exploring the persuasive power of virtual reality imagery for destination marketing. In: Travel And Tourism Research Association (TTRA). Conference.

79.

Wang

Chen

(2023) Association between resilience and subjective well-being among community old adults. Chin J Public Health 39(2): 201–205.

80.

Warshaw

Davis

(1985) Disentangling behavioral intention and behavioral expectation. Journal of Experimental Social Psychology 21(3): 213–228.

81.

Williams

Hartman

Cavazotte

(2010) Method variance and marker variables: a review and comprehensive CFA marker technique. Organizational Research Methods 13(3): 477–514.

82.

Wut

Lee

(2023) Does attitude or intention affect behavior in sustainable tourism? A review and research agenda. Sustainability 15(19): Article 19.

83.

Xiong

Wang

(2022) Predicting mobile government service continuance: a two-stage structural equation modeling-artificial neural network approach. Government Information Quarterly 39(1): 101654.

84.

Yuan

Yoo

(2020) The effect of the perceived risk on the adoption of the sharing economy in the tourism industry: the case of airbnb. Information Processing & Management 57(1): 102108.

85.

Yoo

Jung

(2018) Drone delivery: factors affecting the public’s attitude and intention to adopt. Telematics and Informatics 35(6): 1687–1700.

86.

Zhang

Guo

, et al. (2022) Consumer purchase intention of intangible cultural heritage products (ICHP): effects of cultural identity, consumer knowledge and manufacture type. Asia Pacific Journal of Marketing and Logistics 35(3): 726–744.

87.

Zhu

(2019) Segmenting the public’s risk beliefs about drone delivery: a belief system approach. Telematics and Informatics 40: 27–40.

88.

Zhu

Pasch

Bergstrom

(2020) Understanding the structure of risk belief systems concerning drone delivery: a network analysis. Technology in Society 62: 101262.