1. Introduction
Smartphones have evolved to become minicomputers that support a variety of information services, often in the form of mobile apps that are available to use anytime and anywhere [
1,
2]. The functionality of such information services plays an important role in tourism consumers’ decision-making. Indeed, prior research indicates that the readily available information on smartphone mobile apps assists tourism consumers [
2,
3,
4,
5,
6]. In particular, Wang, Park and Fesenmaier [
2] found that mobile apps enable tourism consumers to cope with unexpected events and to engage in activities more effectively and efficiently. Thus, tourism consumers can enhance the quality of travel through smartphone mobile apps [
5,
7].
Although mobile apps have positive effects on tourism consumers’ decision-making, many tourism consumers do not use them. In order to determine whether tourists use mobile apps on their trips, we ran a simple survey. The survey respondents (
n = 100) were American tourists in Korea. The result showed that 55% had used a mobile app and 45% had not (
p > 0.1). Thus, approximately half of the respondents had not used a mobile app. This result is similar to that of the Expedia/Egencia survey in which 58% of respondents reported that they had used a mobile app [
8]. Furthermore, when we asked our survey participants about the type of searching tool that they mainly used during their trip to Korea, 62% said that they had used the mobile Web and 13% said that they had used a mobile app (
p < 0.001). In particular, this result showed that none of the respondents had used the tourism mobile app developed by the Korea Tourism Board. This suggests that although tourism mobile apps are developed in order to attract tourists, tourists do not use them. To the best of our knowledge, no research has explored this issue. Consequently, we investigate this important topic.
According to technology readiness of Parasuraman [
9], when people embrace and use new technology, they focus not only on positive aspects such as optimism and innovativeness but also on negative aspects such as discomfort and insecurity. Consumers’ technology readiness affects their perception of technology (perceived usefulness and perceived ease of use) and intention to use the technology. However, although technology readiness is an essential variable for consumers’ acceptance of technology, previous studies tended to focus on the consequences of technology readiness [
10]. In other words, there was limited research on antecedents of technology readiness. This situation raises two questions. First, what type of tourism consumers focus on positive rather than negative aspects? Second, when do tourism consumers focus on negative rather than positive aspects?
This research approaches these questions from the perspective of self-construal and perceived risk in the context of mobile apps. We propose that tourism consumers’ technology readiness for mobile apps differs according to self-construal. Specifically, tourism consumers with independent self-construal, which focuses on gains, place relatively more weight on the positive aspects of mobile apps. In contrast, tourism consumers with interdependent self-construal, which focuses on losses, place relatively more weight on the negative aspects of mobile apps. However, we also propose that the relationship between self-construal and technology readiness is qualified by perceived risk. If perceived risk is high, tourism consumers place relatively more weight on negative aspects and less weight on positive aspects regardless of self-construal because a high level of perceived risk prompts thoughts of loss.
4. Results
4.1. Manipulation Check
We analyzed the data using SPSS 21.0. A one-way analysis of variance (ANOVA) test on the perceived risk revealed a significant effect on risk level (F = 10.295, p < 0.01). Participants who read the description of the high-risk mobile app had more to say about risk (M = 3.75) compared to those who read the description of the low-risk mobile app (M = 3.30).
4.2. Technology Readiness
Table 3 presents the correlations for all variables. A technology readiness index was calculated by subtracting the inhibitors from the drivers. Since independent self-construal and interdependent self-construal were measured as continuous variables rather than categorical variables, a regression analysis was performed, instead of an analysis of variance. In addition, to eliminate the possibility of multicollinearity, the two self-construal values were mean-centered. To test our hypotheses, we regressed the technology readiness index on participants’ mean-centered independent and interdependent self-indices; the perceived risk level (coded as high risk = 1 and low risk = 0); and the two interaction terms of perceived risk level × independent self and perceived risk level × interdependent self (see
Table 4). The overall model was significant (R
2 = 0.278, F (5, 278) = 22.846,
p < 0.001). Furthermore, the main effects on risk level, independent self-index, and interdependent self-index were significant. The results showed the positive effect of the independent self-index (β = 0.404,
p < 0.001) together with the negative effects of the interdependent self-index (β = −0.366,
p < 0.001) and the perceived risk level (β = −0.406,
p < 0.001) on technology readiness. Thus, the first hypothesis is supported. More importantly, the interaction effects between perceived risk level and independent self-index (β = −0.303,
p < 0.001), and the interaction effect between perceived risk level and interdependent self-index (β = 0.289,
p < 0.001), were significant.
To achieve a better understanding of the interaction effects, we examined the effects of the two self-construal indices on participants’ technology readiness toward the mobile apps at each perceived risk level. First, for the high perceived risk level, the independent self-index (β = −0.046, p > 0.1) and interdependent self-index (β = 0.049, p > 0.1) were not significant. For the low perceived risk level, the independent self-index was positive and significant (β = 0.372, p < 0.001), whereas the interdependent self-index was negative and significant (β = −0.328, p < 0.001). These results provide evidence that independent self-construal had a positive influence on technology readiness toward the mobile app with the low perceived risk level but had no effect on technology readiness toward the mobile app with the high perceived risk level. However, interdependent self-construal had a negative influence on technology readiness toward the mobile app with the low perceived risk level but had no effect on technology readiness toward the mobile app with the high perceived risk level. Thus, the second hypothesis is also supported.
We also conducted a spotlight analysis at one standard deviation below and above the mean of the independent self-index (M = 4.73) while holding the interdependent self-index (M = 4.17) constant at its mean [
59]. The result showed that participants with a low level of independent self-construal did not differ in technology readiness toward the high-risk and low-risk mobile apps (β = 0.010,
p > 0.1), whereas participants with a high level of independent self-construal perceived technology readiness more positively toward the low-risk mobile app than the high-risk mobile app (β = −0.619,
p < 0.001) (see
Figure 1a).
We conducted a similar spotlight analysis at one standard deviation below and above the interdependent self-index (M = 4.17) while holding the independent self-index (M = 4.73) constant at its mean. The result showed that participants with a low level of interdependent self-construal perceived technology readiness more positively toward the low-risk mobile app than the high-risk mobile app (β = −0.715,
p < 0.001). However, participants with a high level of interdependent self-construal did not differ in technology readiness toward the high-risk and low-risk mobile apps (β = −0.271,
p > 0.1) (see
Figure 1b).
4.3. Use Intention
We regressed the use intention on participants’ mean-centered independent and interdependent self-indices; the perceived risk level; and the two interaction terms of perceived risk level × independent self and perceived risk level × interdependent self (see
Table 5). The overall model was significant (R
2 = 0.144, F (5, 278) = 10.540,
p < 0.001). The results showed the positive effect of the independent self-index (β = 0.417,
p < 0.001) together with the negative effects of the interdependent self-index (β = −0.239,
p < 0.01) and the perceived risk level (β = −0.216,
p < 0.001) on use intention toward the mobile apps. More importantly, the interaction effects between the perceived risk level and the independent self-index (β = −0.266,
p < 0.01), and between the perceived risk level and the interdependent self-index (β = 0.255,
p < 0.01), were significant.
To understand the interaction effects, we examined the effects of the two self-construal indices on participants’ use intention toward the mobile apps at each of the perceived risk levels. For the high perceived risk level, the independent self-index (β = −0.003, p > 0.1) was not significant whereas, for the low perceived risk level, the independent self-index was positive and significant (β = 0.390, p < 0.01). The interdependent self-index was not significant regardless of the perceived risk level (p > 0.05 for all).
We conducted a spotlight analysis at one standard deviation below and above the mean of the independent self-index (M = 4.73) while holding the interdependent self-index (M = 4.17) constant at its mean. The result showed that participants with a low level of independent self-construal did not distinguish between the high-risk and low risk mobile apps (β = 0.016,
p > 0.1), whereas participants with a high level of independent self-construal perceived use intention more positively toward the low-risk mobile app than the high-risk mobile app (β = −0.319,
p < 0.05) (see
Figure 2a).
We conducted a similar spotlight analysis at one standard deviation below and above the interdependent self-index (M = 4.17) while holding the independent self-index (M = 4.73) constant at its mean. The result showed that participants with a low level of interdependent self-construal had a more positive use intention toward the low-risk mobile app than the high-risk mobile app (β = −0.569,
p < 0.001), whereas participants with a high level of interdependent self-construal did not differ in their use intention toward the high-risk and low-risk mobile apps (β = −0.219,
p > 0.1) (see
Figure 2b).
4.4. Mediation Analysis
To test our assumption that differences in use intention toward mobile apps stem from self-construal, we examined mediator models [
60]. In this context, we expect that the effect of self-construal on use intention is mediated by technology readiness. Thus, we examined such mediation for the different conditions of perceived risk levels using bootstrapping methods (10,000 resamples, PROCESS macro model 4) [
61].
A bootstrap analysis for the high-level perceived risk condition showed that the effects of independent and interdependent self-construal on use intention are not mediated by technology readiness, using a 95% confidence interval (CI) that includes 0 (for independent self-construal, CI = −0.2531 to 0.1232; for interdependent self-construal, CI = −0.0654 to 0.1463). The findings for the low-level perceived risk condition, however, showed that the mean indirect effects via technology readiness are significant. Specifically, the mean indirect effect via technology readiness on the relationship between independent self-construal and use intention is positive and significant (a × b = 0.30, 95% CI = 0.1415 to 0.5286) excluding 0. In addition, the mean indirect effect via technology readiness on the relationship between interdependent self-construal and use intention is negative and significant (a × b = −0.19, CI = −0.3520 to −0.0859) excluding 0.
5. Discussion and Conclusions
Mobile apps are excellent information tools for tourism consumers. However, the greater the diversity of tourism information services, the greater the perceived risk. For example, tourism consumers could perceive the login and booking services of mobile apps as personal and financial risks. In this context, the technology readiness theory suggested by Parasuraman [
9] proposed that, when people encounter new technologies, they focus not only on the positive aspects but also on the negative. Further, Parasuraman [
9] argued that people who focus on the positive aspects of new technologies could use the new technologies, whereas those who focus on the negative aspects probably do not use them. Thus, tourism consumers with a high level of technology readiness could use tourism mobile apps. However, despite the burgeoning interest in technology readiness, research on its antecedents and their implications is sparse. In this study, we asserted that individual self-construal is the antecedent of technology readiness. We therefore examined the extent to which technology readiness differs across the concepts of self-construal (an individual variable) and perceived risk (a situational variable).
The findings showed that compared to consumers with interdependent self-construal, consumers with independent self-construal have a positive perception of technology readiness because they focus on what they gain from a mobile app. Consumers with interdependent self-construal focus on the losses they associate with a mobile app. More importantly, the results showed that the effect of self-construal on technology readiness is qualified by the perceived risk level. Specifically, the effect of self-construal on technology readiness is not significant in a high-level perceived risk condition because consumers’ thoughts of losses are activated by a high perceived risk level regardless of their self-construal. In addition, the findings showed that the relationship between self-construal and use intention toward mobile apps is mediated by technology readiness only when perceived risk is at a low level.
The current research has theoretical implications for self-construal and technology readiness research. First, this study expands the boundaries of technology readiness research by demonstrating that self-construal can lead to technology readiness. In this regard, the study adopts a different approach to technology readiness by examining its antecedents rather than its consequences. The academic interest in the antecedents of technology readiness was insufficient, and the previous studies simply presented demographic variables as the antecedents. For example, Elliott et al. [
62] suggested studying the antecedents of technology readiness by examining the effect of culture (e.g., American vs. Chinese) on technology readiness. Dutot [
63] investigated the effects of consumers’ age on technology readiness. By examining the role of an individual variable (self-construal), this research considers antecedents that have a more direct effect on technology readiness than those of prior studies.
This study also extends the prior research on self-construal and perceived risk by examining another research dimension. Aaker and Lee [
29] suggested that individuals with an independent self-construal place relatively more weight on gains rather than losses, whereas individuals with an interdependent self-construal place relatively more weight on losses rather than gains. Further, Lin, Chang and Lin [
52] examined the effect of message framing on product evaluation and showed that a high perceived risk level prompted a person to focus on losses no matter what type of individual self-construal applied. Our results show that these effects of self-construal and perceived risk apply similarly to research about the adoption of new technology.
In addition, this study has important practical implications for the tourism industry and mobile technologies. Tourism mobile apps can influence all stages of tourism consumers’ decision-making [
64], including the choice of destination, booking accommodation, and sharing travel experiences. This means that making mobile apps available for tourism in specific countries can increase the number of tourism consumers who visit such countries. Indeed, many countries have recently developed tourism mobile apps to attract tourism consumers’ interest. Examples are There’s Nothing Like Australia, Visit Norway, Your Singapore Guide, and Visit Korea 3.0. In this regard, the study’s findings show that tourism mobile apps should be developed, focusing on reducing tourism consumers’ perceived risk. When the level of perceived risk is high, regardless of self-construal, consumers focus on the negative side of the technology. Based on the study’s experimental stimulation, whether to approve the GPS function, the reliability of the payment system, and requests for personal data affect consumers’ perceived risk. Therefore, it is necessary to consider ways to lower the perceived level of risk in relation to these functions. Furthermore, advertisements viewed by consumers through tourism mobile apps can be a good profit model for companies developing mobile apps; however, advertisements for companies and products that are unfamiliar to consumers can increase the perceived risk. Companies should develop tourism mobile apps after considering the awareness and image of a company or brand presented in advertisements can also affect consumers’ perceived risk.
In addition, companies or countries that develop tourism mobile apps should implement differentiated strategies according to consumers’ self-construal. Consumers are interested in different types of information according to their self-construal. If consumers searching for tourism-related information have independent self-construal, the tourism information presented to them through the tourism mobile app should be related to the benefits. For independents, it would be more effective to communicate the convenience and efficiency of the mobile app. However, if consumers have an interdependent self-construal, the information presented to them must be related to the loss. In other words, for them, it is necessary to suppress the induction of negative perceptions or emotions by increasing the reliability of the information obtained through the mobile app. At this time, information, such as other customers’ reviews, can be an effective tool for lowering their perceived risk or uncertainty. To this end, companies should provide services that are discriminating in providing information after advance identification of consumers’ self-construal.
An important limitation of this research relates to the stimulus materials and the measurement of perceived risk. As we presented a fictitious tourism mobile app to participants, tourism consumers’ perceived risk could differ from actual situations, and although we measured perceived risk as overall risk, the type of risk that tourism consumers perceive could be different according to the functions of the mobile apps. For example, tourism consumers who use the booking services of mobile apps could be concerned about financial risk rather than other risk factors. Similarly, tourism consumers who use mobile apps to search for information about unfamiliar destinations could be concerned about performance and social risks rather than other risk factors. Future research could consider the weight of risk factors according to the different functions of real rather than fictitious tourism mobile apps.