Soft Biometrics in Retail Service: Understanding Privacy Paradox and Cross-Cultural Differences Regarding 3D Body Scanning Technology

Abstract

This study aims to explore the privacy paradox and cultural difference associated with integrating soft biometric information through 3D body scanning technologies into personalized shopping services. By incorporating the theory of Concerns for Information Privacy and the Unified Theory of Acceptance and Use of Technology into the Privacy Calculus Theory, the purpose of this study is to understand fashion consumers’ privacy paradox and investigate cultural differences (U.S. vs. China). A partial least squares structural equation modeling and multigroup analysis was conducted from two data sets collected online (N _U.S.= 525, N _China = 484). Results show that consumer concerns toward data management (i.e., collection, error, secondary use) and service excellence expectations (i.e., perceived expectance, facilitation condition, hedonic motivation) explain the privacy calculus (i.e., risk and benefit) in consumer decisions. The results indicate that error management is critical to understanding Chinese consumers’ privacy calculus, while secondary use is important for U.S. consumers.

Keywords

soft biometrics privacy calculus theory service expectation crosscultural cultural difference personalized retail service 3D body scanning

Introduction

Is the loss of privacy the price consumers pay in a store where technology knows no boundaries? This question is particularly pertinent in the retail environment where extensive personal data can be collected and shared. Specifically, as 3D body scanning technology advances, it allows for the capture of the full shape of consumers’ bodies using various methods such as lasers, structured light, and other motion-capturing equipment (D’Apuzzo, 2007; Hassan et al., 2021). The importance of the scanning technology is underscored for fashion retailers as it is able to collect soft biometric data (D'Apuzzo, 2007; Hassan et al., 2021). Soft biometrics, as opposed to hard biometrics, encompass traits that can partially identify people based on physical and behavioral characteristics, including attributes such as height, weight, skin color, eye color, age, ethnicity, gender, and other behavioral features (Hassan et al., 2021). This form of biometrics holds significance for the fashion retail industry, given its inherent focus on personalization. For fashion retailers, the precise capture of soft-biometric data is not merely about enhancing product customization according to various body shapes; it also enables them to understand consumers’ immediate needs by maintaining their personal information, thus laying the foundation for a customer-centric business model (Jaha & Nixon, 2014; Wang et al., 2019).

In this study, our primary focus is on the critical role of 3D body scanning technology in integrating soft biometric data within the innovative realm of fashion retail and investigating how this integration shapes privacy concerns among consumers from different cultural backgrounds. It has been prominently observed among leading fashion brands in different cultural contexts, such as the United States (U.S.) and China (Owen, 2021). In the U.S., for instance, Nike introduced “Nike Fit” in 2019, which is a 3D scanning solution. This technology combines artificial intelligence (AI), computer vision, and machine learning to precisely capture the intricate dimensions of a consumer's body. This data-driven approach not only ensures that consumers receive the most suitable product size for their unique measurements but also enhances their overall shopping experience and illustrates the tangible benefits of retail technologies in contemporary fashion (Owen, 2021). Similarly, the Chinese market has embraced this trend, with notable Danish fashion brands Jack and Jones, and Vero Moda, launching smart stores in 2020. These stores are powered by AI technology and leverage 3D body scanning technologies to collect soft biometric information (Chowdhary, 2020). The comprehensive use of soft biometrics has expanded the range of services offered, leading to advances in virtual try-ons, style recommendation systems, and streamlined payment processes (Chowdhary, 2020).

While the global retail sector is currently undergoing a service revolution, this transformation has also highlighted privacy concerns across different cultures. It is important to recognize that the significance of these privacy concerns varies across cultural contexts (Salfino, 2021; Zhong et al., 2022). For instance, U.S. organizations such as the Consumer Federation of America have been actively addressing privacy issues, particularly concerning intrusive technologies adopted by Macy's, Nordstrom, and J.C. Penney (Salfino, 2021). Similarly, in China, there is a growing demand for stricter Data Security and Protection Laws, especially concerning scanning technologies (i.e., collecting facial information), reflecting the unique societal and regulatory perspectives in the country (i.e., Zhong et al., 2022). This underscores the importance of understanding potential cultural differences in privacy concerns and adoption of technology, emphasizing the need for a cross-cultural approach.

This research acknowledges that the concerns of consumers in the U.S. and China may be shaped differently, particularly in terms of perceptions related to how companies manage collected soft biometric data and the anticipated future usage of this data for personalized services. These differences are likely influenced by each country's unique cultural norms, regulatory frameworks, and societal values (Zhong et al., 2022). While existing research has explored the implications of retail services and privacy concerns in various contexts, such as self-checkout systems (Lee & Leonas, 2021) and voice recognition services (Park et al., 2021), there has been a lack of scholarly efforts that investigated the cultural differences regarding the privacy paradox in tech-driven fashion retail settings, particularly concerning the differences between individualistic and collectivistic cultures (U.S. vs. China). This highlights the necessity of exploring cultural nuances when deploying advanced fashion retail technologies.

This study aims to enhance our current understanding of cross-cultural dynamics in fashion retail environments by proposing an extension of the Privacy Calculus Theory (PCT). The PCT outlines the rational process through which individuals weigh potential risks against benefits when deciding to disclose private information (Smith et al., 1996). To elucidate the privacy calculus mechanism in a context where perceived risks and benefits often coexist (Youn et al., 2023), this study draws on two theoretical perspectives: (1) apprehensions around a company's data management, and (2) service expectations from the consumer's viewpoint. The first perspective stems from the Concerns for Information Privacy (CFIP) theory on a company's data management (Smith et al., 1996). CFIP identifies four dimensions of concerns related to data collection, error rectification, safeguarding from secondary use, and the prevention of improper access. The second perspective leverages The Unified Theory of Acceptance and Use of Technology (UTAUT), which encompasses consumer expectations of innovative services (Venkatesh et al., 2016). UTAUT proposes three critical determinants—performance expectancy, facilitating conditions, and hedonic motivations—to assess anticipated service excellence. Both theoretical frameworks help us to understand the privacy paradox in fashion retail, where consumers are anxious about how their personal information is handled by retailers (CFIP) while simultaneously anticipating benefits from innovative services (UTAUT). Given privacy concerns shaped within cultural backgrounds, this study's theoretical foundation contributes to the growing body of literature on technology that uses soft biometrics in the global retail sector (Vance et al., 2020; Zhong et al., 2022). Therefore, it introduces a cross-cultural framework that clarifies consumer perceptions in an era of increasingly personalized fashion retail services. Consequently, within our extended PCT framework, we propose that concerns about the management of soft biometric information and service expectations are distinct dimensions that explain privacy calculus perceptions regarding the disclosure of soft biometric information. To this end, this study examines cross-cultural differences between the U.S. and China within the proposed framework.

Literature Background

Soft Biometrics in Fashion Retail

Hard biometrics, a subfield of biometrics, involves the application of statistical analysis to distinctive and quantifiable biological characteristics, such as fingerprints, iris patterns, facial features, palm prints, and vein patterns (Hassan et al., 2021). Its primary utility lies in the verification of individual identities. In contrast, soft biometrics encompass a range of non-identifying biological traits, including attributes like age, gender, ethnicity, height, hair color, and eye color (Hassan et al., 2021). These soft biometric traits serve to differentiate individuals but lack the capability for precise biological identification (Hassan et al., 2021). Initially, research in the field of biometrics classified soft biometrics into various categories, including demographic attributes (e.g., age, gender, ethnicity, hair color, and eye color), anthropometric and geometric attributes (e.g., body dimensions), and health-related characteristics (e.g., weight, height, and wrinkles) (Hassan et al., 2021).

While hard biometric information has been extensively applied in sectors such as healthcare and travel for identity verification purposes, soft biometric data, which encompasses physical characteristics, is increasingly applied in the fashion retail industry (Carolis et al., 2018; Wang et al., 2019). Advanced retail technologies such as interactive displays, touch screens, and virtual mirrors require soft biometric information to tailor retail services to the individual fashion consumer's product preferences (Li et al., 2016; Youn et al., 2023). Previous scholars suggested that soft biometrics enhance user experiences and identify consumer segments with shared characteristics, which help retailers better understand consumer needs during the point of sale (Carolis et al., 2018; Wang et al., 2019).

Particularly in the realm of fashion retail, scholars suggested that the acquisition of soft biometric information can be accomplished through various methods (D'Apuzzo, 2007; Kim & Forsythe, 2008, Zhang et al., 2022). For instance, in digital settings, manual data input enables consumers to contribute information like clothing size and preferred styles (Kim & Forsythe, 2008). In the retail store setting, retailers employ facial recognition technology for collecting demographic data, thereby enhancing personalization upon arrival to the store (Zhang et al., 2022). Moreover, body scanning technologies yield soft biometric data to retailers, including precise body measurements of their customers (D'Apuzzo, 2007). The body scanning technologies not only offer insights into consumers’ emotional states through real-time biometric data, such as heart rate (Carolis et al., 2018), but also facilitate the provision of customized products and virtual try-on services for personalized shopping experiences in the fashion retail sector (Carolis et al., 2018; Pizzi & Scarpi, 2020; Youn et al., 2023).

PCT: Privacy Paradox of 3D Body Scanning Technology

In the context of PCT, the concept of privacy calculus encompasses an individual's assessment of potential risks (i.e., loss of privacy), in comparison to the benefits gained from providing private information (Smith et al., 1996). This evaluation hinges on two primary constructs: perceived privacy risk and perceived benefit. Perceived privacy risk reflects the extent to which individuals perceive potential losses associated with disclosing personal information, while perceived benefit represents the cognitive evaluation of potential gains resulting from the disclosure of such information (Zhong et al., 2022).

While the retail landscape has undergone significant data transformations, the emergence of data-driven retail technologies capable of gathering detailed personal information raises privacy concerns. Pizzi and Scarpi (2020) explored factors in shaping privacy perceptions in the retail setting, which include consumer perceptions toward retailers (i.e., distributive fairness), retail technologies (i.e., smart mirrors, facial recognition), and their personal traits (i.e., anxiety). Specifically, the application of PCT holds particular relevance within the fashion retail sector, where considerations related to privacy risks and benefits significantly influence consumer behavior in the adoption of advanced retail services (Youn et al., 2023).

In particular, the deployment of innovative retail services that capture soft biometric information is increasingly prominent. The core of this technological evolution lies in 3D body scanning technology, a foundational tool for collecting precise soft biometric data (Hassan et al., 2021). 3D body scanning technology involves a method for capturing the three-dimensional shape, dimensions, and in some cases, even the texture of the human body or its components, employing techniques such as lasers, structured light, or other depth-sensing systems (Youn et al., 2023). These detailed data points contribute to retail services, including virtual try-on experiences (e.g., personalized avatars, smart mirrors) and AI-driven recommendation services (e.g., fit recommendations), thereby enhancing the perceived benefit toward personalized experiences (Lee & Leonas, 2021; Youn et al., 2023). While these technologies promise to revolutionize the shopping experience, the intimate and detailed nature of the data they capture, particularly the precise measurements and body profiles obtained through 3D body scans, can increase privacy concerns (Youn et al., 2023). This gives rise to a multifaceted privacy calculus challenge that fashion retailers must adeptly address (Youn et al., 2023). In expanding the application of PCT within our research model, we have incorporated elements of CFIP and the UTAUT. These theories encompass a wide array of factors that influence consumer decisions regarding the disclosure of personal information. The CFIP and UTAUT are explained below.

CFIP Toward Data Management

CFIP can be defined as the perception of how a company or organization collects, manages, and controls private information (Smith et al., 1996). The CFIP framework elucidates four distinct dimensions of privacy concerns, which encompass collection, errors, improper access, and secondary use (Smith et al., 1996). Specifically, collection pertains to concerns related to data collection practices by entities that amass large quantities of identifiable information in databases (Park et al., 2021). Error refers to apprehensions regarding the lack of adequate safeguards for rectifying data inaccuracies (Smith et al., 1996). Improper access relates to the extent to which individuals are worried about the security of stored information and the potential for unauthorized access (Smith et al., 1996). Lastly, secondary use covers concerns that collected personal information may be shared or used by third parties without the consent or authorization of consumers (Smith et al., 1996).

Within the context of fashion retail, the emergence of personalized services, such as virtual try-ons and AI-driven recommendation services, has accelerated the use of the 3D scanning technology that entails the collection of soft biometric information (Jaha & Nixon, 2014; Lee & Leonas, 2021; Wang et al., 2019; Youn et al., 2023). While the personalized services collecting soft biometrics via the body scanning technology hold the promise of enhancing user experiences, they also bring forth challenges associated with the collection and management of personal and private data (Zhong et al., 2022). Consumers frequently express anxieties about their limited control over the acquisition of their data, leading to heightened privacy concerns (Park et al., 2021). Each of these dimensions serves to intensify perceived risks while simultaneously diminishing the perceived benefits associated with these technologies (Lin et al., 2021). From a privacy calculus perspective, heightened privacy concerns correspondingly amplify the perceived risk of data disclosure and diminish the perceived benefits, such as the potential for improved personalized shopping experiences (Youn et al., 2023). Thus, we proposed:

H1–H4: For the fashion retail service, the consumer's privacy concerns toward the retailer's soft biometric data management—(H1) collection, (H2) error, (H3) secondary use, and (H4) improper access—will increase (a) the perceived risk and decrease (b) the perceived benefit of information disclosure of using the service.

Service Excellence in UTAUT

Consumer perceptions toward new technology services play a crucial role in their technology adoption decisions and are often assessed using frameworks such as the Technology Acceptance Model (TAM) (Venkatesh et al., 2016). The UTAUT was developed as an extension of TAM and incorporates various factors that assess service excellence, influencing the intention to use technology (Venkatesh et al., 2016). Because the perception of service excellence is a multifaceted and holistic experience, scholars have identified various dimensions within UTAUT (Venkatesh et al., 2016; Tomić et al., 2022). For instance, Vimalkumar et al. (2021) identified six dimensions for invasive services (i.e., Google Voice). Specifically, performance expectancy is the degree of expectation a consumer has of a specific technology that improves performance to achieve a goal (Venkatesh et al., 2016). Effort expectancy is the extent to which others perceive a particular service's ease of use (Venkatesh et al., 2016). Social influence refers to how much others affect a consumer's decision to adopt a specific technology (Venkatesh et al., 2016). Facilitating conditions are defined as the degree to which consumers perceive the resources needed to support the use of a technology (Tomić et al., 2022). Hedonic motivation refers to the degree of pleasure the consumer has in using a specific technology (Venkatesh et al., 2016). The price value is the consumer's cognitive trade-off between perceived benefits (the reward) and perceived costs (the price) of a specific technology (Venkatesh et al., 2016).

UTAUT in the Context of Soft Biometrics-Based Retail Services

In the context of 3D body scanning technology in retail services, effort expectancy is not considered for this study as the personalized service does not involve complex operations within the in-store service context. As there are no extra charges for consumers for additional services in this setting, pricing considerations are not included. Furthermore, due to the nature of the service being highly personalized and individual-specific, this study focuses on the consumer experience that is typically private and individualized (e.g., Vimalkumar et al., 2021). Thus, the impact of social factors like peer opinion or societal norms is not a focus of this study. Therefore, for our study, we adopted performance expectancy, facilitating conditions, and hedonic motivation to understand the consumers’ service expectations. By incorporating UTAUT into privacy calculous approaches, we posit that consumers evaluate the perceived privacy risk and benefit of information disclosure based on their assessment of service excellence. The perception of a personalized service influences the privacy calculus, intensifying the perceived cost (i.e., risk) of information disclosure, while concurrently enhancing the anticipated benefits (Vimalkumar et al., 2021). Despite the requirement for personal information sharing within the retail context to access innovative services, consumers remain aware of potential associated risks (Youn et al., 2023). Heightened service expectations mitigate the perceived risk of private information sharing while concurrently boosting the perceived benefit from a privacy calculus perspective (Li et al., 2016). This rationale serves as the foundation for the following hypotheses:

H5–H7: For the fashion retail service, the consumer's perception of service excellence—(H5) performance expectancy, (H6) facilitating conditions, and (H7) hedonic motivation—will decrease (a) perceived risk and increase (b) perceived benefit of soft biometric information disclosure for using the service.

Effect of Perceived Benefit and Risk on Behavior Intention

In the context of fashion retail setting, consumers’ perception of privacy risk associated with personalized retail technologies (i.e., scanning technologies, smart mirrors) diminishes the perceived value of disclosing personal information, which in turn erodes trust and leads to decreased repurchase intentions (Pizzi & Scarpi, 2020; Lee & Leonas, 2021). Conversely, when consumers anticipate benefits, such as personalized products tailored to their unique needs (e.g., body shapes and sizes), the perceived advantages are likely to enhance their intentions to adopt these services (Pizzi & Scarpi, 2020; Youn et al., 2023). Thus, we proposed:

H8: (a) Consumers’ perceived risk will decrease adoption intentions to use the retail service, whereas (b) perceived benefit will increase the adoption intentions.

Cultural Difference in Privacy Calculus: the U.S. vs. China

According to theories of cultural values (Hofstede, 2001), cultural differences between individualism and collectivism are evident in various dimensions, including self-concept and social values. This cultural differentiation also extends to perceptions of privacy across two countries: the U.S. and China. Research has indicated that Chinese consumers, influenced by their collectivist cultural values, tend to view privacy concerns as the responsibility of industries rather than individual consumers. Zhong et al. (2022) explained that Chinese consumers exhibit greater restraint in actions that might contravene societal expectations or norms, in contrast to their U.S. counterparts. Conversely, U.S. consumers, stemming from a culture of individualism, place greater emphasis on personal control over information, with clear expectations for individual permission for the usage of personal data by digital services such as Google and Facebook (Vimalkumar et al., 2021). Thus, for U.S. consumers, a company's data management practices play a crucial role in assessing perceived risk and benefit, as they tend to be more concerned about personal freedom to control their information. In contrast, Chinese consumers view data management issues not as a personal concern but as an industrial responsibility, reflecting the salience of collectivistic values (Zhong et al., 2022).

In addition to privacy concerns, it is anticipated that expectations of service excellence, as viewed from the UTAUT perspective, will influence perceptions of risk and benefit. However, in contrast to privacy concerns, these influences are predicted not to exhibit significant variations between U.S. and Chinese consumers. According to Zhong et al. (2022), expectations of service excellence stem from universal consumer desires for efficiency, effectiveness, and satisfaction in service delivery, regardless of cultural background. This perspective is supported by studies that have found consistent effects of service excellence on consumer perceptions across diverse cultural contexts. For instance, Nam et al. (2021) observed similar impacts of service quality on customer satisfaction (e.g., perceived trust) toward online apparel sites among consumers in South Korea and the U.S. Other scholars have also suggested that aspects of service quality (e.g., responsiveness, reliability, tangibles, and assurance) appear to impact positive attitudes toward services across various cultures (e.g., Youn et al., 2023). This implies that while privacy concerns can vary due to cultural factors, expectations of service excellence and their effects on risk and benefit perceptions could be relatively consistent across two cultures. This led us:

H9: There will be a cultural difference (U.S. vs. China) in the effect of consumer concerns toward soft biometric data management on the perception of risk and benefit of the information disclosure, while there will be no cultural difference in the influence of consumer expectation of service excellence on the risk and benefit perception.

Method

Survey Development and Data Collection

Survey Development. For this study, we employed online survey questionnaires as the primary data collection method. The questionnaires were designed to begin with the introduction of a realistic shopping scenario. To create this scenario, we curated a pool of retail cases from various fashion retail services that collect diverse forms of soft biometric data. Our primary reference for this scenario was the in-store service approach called “Body Scan Jeans,” where consumers of H&M and Levi's were invited to undergo a 3D body scanning process to access advanced personalized retail services for shopping (Roberts-Islam, 2020). Jeans were selected as the focal product due to their relevance to personalized retail services based on soft biometric information (e.g., Youn et al., 2023).

In the scenario presented in the study, participants were asked to imagine themselves shopping for jeans at a retail store. They were introduced to a personalized retail service without disclosing the retailer's name to prevent any confounding effects. The service aimed to provide accurate fit recommendations but required participants to share their soft biometric data through a 3D body scanning process. We provided a clear definition of soft biometrics and included pictures of the body scanning process (Roberts-Islam, 2020) to create a realistic situation. After the scenario, participants were informed that the collected data would be combined with their shopping history to offer more advanced personalized retail services, such as recommending new products that fit their body shape and size.

We then confirmed participants’ perceived realism regarding the scenario and their understanding of the situation where shared soft biometric data could be used for future retail services through screening questions. Subsequently, survey questions were administered to assess participants’ privacy concerns about data management (i.e., collection, errors, secondary use, improper access; α ranged from 0.82 to 0.93; Smith et al., 1996), as well as expected service excellence (i.e., performance expectancy, facilitating conditions, hedonic motivation; α ranged from 0.88 to 0.94; Venkatesh et al., 2016). Participants were asked about their perceived risk (α = 0.91) and benefit (α = 0.92) of disclosing personal information (Youn et al., 2023) and their adoption intentions toward the personalized retail service (α = 0.92; Kim & Forsythe, 2008). All questions were presented on a 5-point Likert Scale (e.g., 5 = strongly agree; see Table 1). Participants’ demographic information was also collected.

Table 1.

Assessment of Measurement Model on Factor Loading, CR and AVE.

	Complete (N = 1,009)			U.S. (N = 525)			China (N = 484)
Construct	Loading	CR	AVE	Loading	CR	AVE	Loading	CR	AVE
Collection (CL)—four items (α = 0.93)		0.935	0.782		0.949	0.822		0.914	0.726
CL01: It would bother me when the retailer asks me for the information.	0.910			0.916			0.875
CL02: When the retailer asks me for the information, I would think twice before providing it.	0.804			0.891			0.754
CL03: It would bother me to give the information to the retailer.	0.909			0.911			0.882
CL04: I'm concerned that the service is collecting too much the information about me	0.909			0.909			0.890
Errors (ER)—four items (α = 0.84)		0.895	0.680		0.904	0.702		0.846	0.579
ER01: I concern that all the information in their databases would not be double-checked.	0.838			0.857			0.821
ER02: I concern that the information in their files is not accurate.	0.812			0.827			0.722
ER03: I concern that the retailer's procedure would not correct errors in the information.	0.809			0.806			0.764
ER04: I concerned that the retailer would not devote enough time and effort to verify the accuracy of the information.	0.839			0.861			0.732
Secondary use (SU)—four items (α = 0.82)		0.891	0.672		0.902	0.697		0.820	0.535
SU01: I concern that the retailer would use the information for purposes not authorized by consumers who provided the information.	0.834			0.852			0.738
SU02: I concern that the retailer would use the information for any other reason.	0.845			0.847			0.724
SU03: I concern that the retailer would sell the information in their computer databases to other retailers.	0.764			0.791			0.760
SU04: I concern that the retailer would share the information with other retailers unless it has been authorized by consumers who provide information.	0.834			0.847			0.800
Improper access (IA)—three items (α = 0.91)		0.867	0.685		0.880	0.709		0.809	0.589
IA01: I concern that the retailer would not devote enough time and effort to preventing unauthorized access to the information.	0.857			0.868			0.805
IA02: I concern that computer databases that contain the information would not be protected from unauthorized access	0.821			0.809			0.636
IA03: I concern that the retailer would not take enough steps to make sure that unauthorized people cannot access the information	0.806			0.848			0.847
Performance expectancy (PE)—four items (α = 0.89)		0.902	0.696		0.919	0.738		0.819	0.531
PE01: I find the service is useful for the shopping.	0.852			0.878			0.717
PE02: Using the service increases my chances to shop effectively.	0.820			0.859			0.750
PE03: Using the service helps the shopping more quickly.	0.824			0.844			0.691
PE04: Using the service increases my shopping efficiency.	0.842			0.856			0.754
Facilitating conditions (FC)—three items (α = 0.88)		0.842	0.640		0.831	0.622		0.853	0.658
FC01: I can have the resources necessary to use the service.	0.779			0.740			0.834
FC02: I can have the knowledge necessary to use the service.	0.822			0.815			0.813
FC03: I can ask help from others to use the service.	0.799			0.808			0.787
Hedonic motivation (HM)—three items (α = 0.94)		0.913	0.778		0.922	0.798		0.843	0.641
HM01: Using the service is fun	0.886			0.891			0.807
HM02: Using the service is enjoyable	0.880			0.897			0.803
HM03: Using the service is very entertaining	0.880			0.893			0.793
Perceived risks (PR)—three items (α = 0.92)		0.937	0.832		0.931	0.817		0.927	0.809
PR01: Providing the retailer with my information would involve many unexpected problems.	0.908			0.904			0.886
PR02: It would be risky to disclose my information to the retailer.	0.914			0.902			0.912
PR03: There would be high potential for loss in disclosing my information to the retailer.	0.915			0.906			0.901
Perceived benefits (PB)—three items (α = 0.91)		0.876	0.701		0.891	0.731		0.839	0.634
PB01: The service can provide me with personalized products tailored to my preferences	0.852			0.873			0.799
PB02: The service can provide me with more relevant information tailored to my personal interest for shopping.	0.821			0.836			0.789
PB03: By providing my information, the service can provide me with the kind of products that I like to shop.	0.839			0.856			0.800
Adoption intention (AI)—four items (α = 0.92)		0.926	0.758		0.942	0.803		0.841	0.570
AI01: I would consider using smart retail services for my shopping.	0.886			0.917			0.735
AI02: The likelihood of using smart retail services for my shopping.	0.854			0.869			0.771
AI03: My willingness to use smart retail services for my shopping would be high.	0.881			0.906			0.780
AI04: The probability I would consider using smart retail services for my shopping would be high.	0.861			0.892			0.732

Note. Regarding convergent validity, all item loadings exceed the recommended threshold of 0.7. Furthermore, the construct CR values are greater than 0.7, and the AVE values surpass the established threshold of 0.5, following Hair et al. (2017).

CR=composite reliability; AVE=average variance extracted.

Given that the participants included both U.S. and Chinese consumers, and the original measurement scales were in English, a back-translation procedure was employed to collect data from Chinese consumers. Two experts proficient in both Chinese and English compared the translated questionnaires and made necessary modifications and adjustments. This rigorous process ensured that the meaning of the questions remained consistent for both groups.

Data Collection . After the Institutional Review Board (IRB) approval, we collected the data from U.S. consumers via Amazon Mechanical Turk (MTurk; January 22–25, 2022), and from Chinese consumers through Credamo (January 24–28, 2022) which is an online platform in China similar to MTurk. After excluding incomplete responses (i.e., less than 90% of completion rate; N _U.S. = 8; N _China= 12) and those that took an unusual amount of time (i.e., less than 5 min or more than 1 h; N _U.S. = 11; N _China= 16), we had 525 U.S. and 484 Chinese responses. Major assumptions (i.e., multivariate normality, multicollinearity) associated with structural equation modeling were successful. In addition, potential common method errors were checked using statistical remedies such as Harman's single factor test (Harman, 1967), partial correlation procedures, and controls for unmeasured latent methods factors. The results, with a variance ratio under 50%, confirmed that our data are not influenced by common method bias.

The U.S. and Chinese consumers’ samples were not significantly different in terms of age and gender (U.S. consumers: M_age = 28.7 years old, 51% women; Chinese consumers: M _age = 29.9 years old, 50% women). Most U.S. participants were Caucasian (81%) with a college education (72%), while 77.1% of the Chinese sample were college-educated (see Table 2).

Table 2.

Demographic Characteristics: U.S. vs. China.

	U.S. (N = 525)	China (N = 484)
Gender
Male	248	240
Female	277	244
Age
18–29	126	123
30–39	217	232
40–49	95	65
50–59	49	47
60–69	32	14
Over 70	6	3
Race
White or European	388	0
Asian	13	484
Asian American	5	0
Native American	36	0
Black or African American	48	0
Hispanic or Latino	28	0
Native Hawaiian or Pacific Islander	1	0
Two or more races	5	0
Do not wish to disclose	1	0
Education
High School or Equivalent	26	9
Some College	89	33
College Graduate (4year)	327	373
Advanced Degree (post-graduate)	79	69
Others	4	0
Yearly income level (USD)
Less than 24,999	51
25,000–49,999	178
50,000–74,999	154
75,000–99,999	77
100,000–149,999	43
150,000–199,999	11
200,000 or more	11
Monthly income level (RMB)
Less than 2,000		21
2,001–4,000		26
4,001–6,000		55
6,001–8,000		105
8,001–10,000		122
10,001–12,000		80
12,001–14,000		60
14,001 or more		15

Note. 1 RMB = 0.14 USD.

Results

Research Design and Data Analysis

In this research, we used SmartPLS 3.0 software to conduct partial least squares structural equation modeling (PLS-SEM). The objective of our study was to gain a comprehensive understanding of and predict consumer attitudes towards information disclosure. We integrated perspectives from both the CFIP and the UTAUT frameworks within the unique cultural contexts of the U.S. and China. PLS-SEM was chosen as the statistical method for our study due to its suitability for research designs that prioritize the maximization of explained variance, which was crucial in our complex model. Additionally, PLS-SEM is well-suited for multigroup analysis (MGA), allowing for nuanced comparisons across culturally distinct groups, even when dealing with variations in distribution characteristics or subgroup sizes (Hair et al., 2017). The research design and associated hypotheses are presented in Figure 1.

Figure 1.

Research model.

Assessment of Validity and Reliability

To assess convergent validity, we examined factor loadings, composite reliability (CR), and average variance extracted (AVE) for the total and sub-groups samples, with all values exceeding the recommended threshold (Table 1). For discriminant validity, the Fornell-Larcker criterion indicated satisfactory levels for the entire and sub-groups (Table 3).

Table 3.

Assessment of Discriminant Validity (Fornell-Larcker).

	Complete (n = 1,009)
	CL	ER	SU	IA	PE	FC	HM	PR	PB	AI
CL	0.884^a
ER	0.075	0.825 ^a
SU	−0.046	0.289	0.820 ^a
IA	0.057	0.432	0.610	0.828 ^a
PE	−0.278	0.283	0.138	0.158	0.835 ^a
FC	−0.201	0.267	0.240	0.257	0.511	0.800 ^a
HM	−0.246	0.278	0.134	0.180	0.692	0.428	0.882 ^a
PR	0.808	0.026	0.102	0.015	−0.273	−0.199	−0.241	0.912 ^a
PB	−0.258	0.313	−0.229	0.246	0.669	0.536	0.582	−0.247	0.837 ^a
AI	−0.379	0.272	0.058	0.104	0.791	0.475	0.707	−0.351	0.615	0.871 ^a

	USA (N = 525)										China (N = 484)
	CL	ER	SU	IA	PE	FC	HM	PR	PB	AI	CL	ER	SU	IA	PE	FC	HM	PR	PB	AI
CL	0.907^a										0.852^a
ER	0.274	0.838^a									−0.067	0.761^a
SU	−0.037	0.232	0.835^a								0.241	0.223	0.732^a
IA	0.147	0.409	0.621	0.842^a							0.133	0.347	0.464	0.768^a
PE	−0.185	0.249	0.104	0.118	0.859^a						−0.292	0.165	−0.067	0.025	0.729^a
FC	−0.133	0.263	0.337	0.338	0.545	0.788^a					−0.258	0.255	0.012	0.091	0.481	0.811^a
HM	−0.165	0.239	0.032	0.064	0.724	0.484	0.894^a				−0.156	0.114	0.092	0.228	0.438	0.325	0.801^a
PR	0.764	0.257	0.081	0.131	−0.152	−0.138	−0.134	0.904^a			0.722	0.130	0.191	0.094	−0.311	−0.233	−0.155	0.900^a
PB	−0.184	−0.284	−0.277	0.277	0.700	0.572	0.628	−0.155	0.855^a		−0.289	−0.282	−0.031	0.063	0.565	0.486	0.413	−0.289	0.796^a
AI	−0.313	0.235	−0.004	0.039	0.713	0.520	0.755	−0.255	0.645	0.896^a	−0.401	0.179	−0.044	0.054	0.624	0.412	0.411	−0.407	0.502	0.755^a

Note. CL=collection; ER=errors; SU=secondary use; IA=improper access; PE=performance expectancy; FC=facilitating conditions; HM=hedonic motivation; PR=perceived risk of information disclosure; PB=perceived benefit of information disclosure; AI=adoption intention.

Square root of AVEs shown diagonally, which are greater than correlation coefficients presented off-diagonally (Hair et al., 2017). This indicated discriminant validity. The discriminant validity of the model's constructs was also confirmed using the Heterotrait-Monotrait Ratio (HTMT) approach for three groups: Complete (n = 1,009), U.S. (n = 525), and China (n = 484). All HTMT values were below the 0.90 threshold, confirming that the constructs were distinct and not overly correlated (Hair et al., 2017).

Regarding model fit, it is important to note that while covariance based structural equation modeling emphasizes stringent model fit criteria, PLS-SEM focuses primarily on explaining the variance of dependent constructs and predicting path coefficients (Hair et al., 2017). Nevertheless, to provide a comprehensive assessment of the model fit, PLS-SEM experts suggest considering additional metrics, specifically the standardized root mean square residual (SRMR) and the normed fit index (NFI) (Hair et al., 2017). An SRMR value below 0.08 and an NFI value above 0.85 are considered indicative of good model fit (Hair et al., 2017). The obtained values for SRMR (0.07) and NFI (0.87) fall within the acceptable ranges, signifying the robustness and adequacy of the proposed model for each group.

Measurement Invariance

To ensure the validity of multigroup comparisons between the U.S. and China, a measurement invariance of composite models (MICOM) procedure was implemented. Having effectively satisfied all three stages of the MICOM procedure, namely, configural invariance, compositional invariance, and equal means and variances, our analysis has convincingly established that the measurement constructs exhibit invariance across the two groups. This foundation supports conducting group comparisons (Hair et al., 2017; Table 4).

Table 4.

Measurement Invariance: U.S. vs. China.

Composite	Configural invariance	Compositional invariance			Composite equality
Composite	Configural invariance	C value (=1)	95% CI	Composition invariance	Difference of mean value (=0)	95% CI	Equality of means	Difference of variances ratio (=0)	95% CI	Equality of variances
CL	Yes	1.000	[1.000; 1.000]	Yes	0.006	[–0.130; 0.129]	Yes	–0.004	[–0.116; 0.118]	Yes
ER	Yes	0.998	[0.998; 0.995]	Yes	0.077	[–0.126; 0.127]	Yes	–0.261	[–0.218; 0.279]	Yes
SU	Yes	0.998	[0.997; 0.993]	Yes	–0.069	[–0.125; 0.126]	Yes	0.252	[–0.304; 0.305]	Yes
IA	Yes	0.999	[0.996; 0.896]	Yes	–0.097	[–0.128; 0.127]	Yes	0.099	[–0.325; 0.334]	Yes
PE	Yes	1.000	[1.000; 0.999]	Yes	–0.126	[–0.129; 0.128]	Yes	0.211	[–0.347; 0.351]	Yes
FC	Yes	1.000	[0.999; 0.997]	Yes	–0.115	[–0.125; 0.128]	Yes	0.258	[–0.300; 0.300]	Yes
HM	Yes	1.000	[1.000; 0.999]	Yes	–0.117	[–0.132; 0.125]	Yes	0.235	[–0.300; 0.314]	Yes
PR	Yes	1.000	[1.000; 1.000]	Yes	0.041	[–0.130; 0.126]	Yes	0.020	[–0.109; 0.110]	Yes
PB	Yes	1.000	[1.000; 0.999]	Yes	–0.089	[–0.125; 0.124]	Yes	0.175	[–0.321; 0.326]	Yes
AI	Yes	1.000	[1.000; 1.000]	Yes	–0.037	[–0.129; 0.125]	Yes	0.030	[–0.324; 0.331]	Yes

Note. Before conducting MGA, measurement invariance must be assessed using the measurement invariance of composites (MICOM) approach (Hair et al., 2017). The MICOM process consists of three steps: (a) evaluation of configuration invariance; (b) evaluation of compositional invariance; and (c) evaluation of equal means and variances. Following the process, the present study fully established configure and compositional invariance. This shows that the measurement invariance is supported (Henseler et al., 2016). Since at least two steps were satisfied, we proceeded to assess whether there were significant differences among two groups (i.e., U.S. vs. China).

CL=collection; ER=errors; SU=secondary use; IA=improper access; PE=performance expectancy; FC=facilitating conditions; HM=hedonic motivation; PR=perceived risk of information disclosure; PB=perceived benefit of information disclosure; AI=adoption intention.

Structural Model Analysis and Hypotheses Testing (H1–H8)

A single-group SEM (N = 1,009) was conducted to test H1–H8 (Table 5). Consumers’ concern about the collection of privacy information increased perceived risk (H1a: β = .790, p < .001) and decreased perceived benefit of information disclosure (H1b: β = −.071, p < .05), accepting H1a and H1b. Error management concern did not influence risk perception, but it reduced benefit perception (H2b: β = −.083, p < .01), rejecting H2a while accepting H2b. Concern about the secondary use of privacy information significantly enhanced risk perception (H3a: β = .082, p < .01, accepted) while it did not explain the benefit, rejecting H3b. Improper access concern had no impact on perceived risk or benefit, rejecting H4. Regarding expected service excellence, performance expectancy (H5b: β = .392, p < .001), facilitating condition (H6b: β = .205, p < .001), and hedonic motivation (H7b: β = .170, p < .001) all positively explained the perceived benefit but not risk, partially supporting H5–H7. Perceived risk reduced (H8a: β = –.212, p < .001), while benefit enhanced intention to adopt the service (H8b: β = .563, p < .001), accepting H8.

Table 5.

Path Results of the Complete Group (H1–H8) and Multigroup Analysis Results (H9).

Path results (H1–H8)										U.S. vs. China (H9)
Complete (N = 1,009)										U.S. (N = 525)	China (N = 484)
Relationship		ß	t	Confidence interval (95%)		VIF	f ²	R ²	Q ²	ß	ß	\|diff\|	p-value
H1a	CL → PR	0.790***	18.986	[0.648	0.727]	1.148	1.591			0.723***	0.798***	0.075	.090
H1b	CL → PB	−0.071*	2.998	[–0.119	–0.026]	1.148	0.021			–0.073*	–0.088*	0.015	.772
H2a	ER → PR	0.065	0.404	[–0.035	0.054]	1.353	0.001			0.009	0.076*	0.141	.002**
H2b	ER → PB	–0.083**	2.181	[–0.135	–0.008]	1.353	0.031			–0.049	–0.178**	0.129	.028*
H3a	SU → PR	0.082**	3.004	[0.024	0.121]	1.626	0.029			0.103**	0.006	0.108	.024*
H3b	SU → PB	–0.054	1.677	[–0.119	0.007]	1.626	0.004			–0.108*	–0.025	0.133	.037*
H4a	IA → PR	0.027	1.012	[–0.023	0.080]	1.833	0.001			–0.079	–0.010	0.068	.217
H4b	IA → PB	0.041	1.086	[–0.030	0.118]	1.833	0.002			–0.086	0.040	0.126	.102
H5a	PE → PR	–0.032	1.119	[–0.092	0.026]	2.219	0.001			–0.001	–0.049	0.078	.180
H5b	PE → PB	0.392***	8.889	[0.306	0.475]	2.219	0.169			0.404***	0.335***	0.070	.095
H6a	FC → PR	–0.005	0.211	[–0.046	0.037]	1.460	0.000			–0.045	0.028	0.074	.095
H6b	FC → PB	0.205***	5.688	[0.137	0.278]	1.460	0.142			0.153**	0.205***	0.051	.514
H7a	HM → PR	–0.016	0.594	[–0.068	0.035]	1.990	0.000	0.659		–0.009	0.001	0.010	.843
H7b	HM → PB	0.170***	3.988	[0.083	0.259]	1.990	0.061	0.535		0.229***	0.178**	0.048	.580
H8a	PR → AI	–0.212***	7.556	[–0.268	–0.159]	1.068	0.083			–0.159***	–0.286***	0.128	.026*
H8b	PB →AI	0.563***	15.674	[0.490	0.632]	1.068	0.513	0.421	0.342	0.620***	0.420***	0.201	.005**

Note. Potential common method bias in the PLS-SEM approach was addressed by employing the full collinearity variance inflation factors (VIF) method (Hair et al., 2017). All VIF scores were found to be under the threshold of 3, thus affirming no common method bias across datasets. In the subsequent analysis, the significant relationship of each path was examined, and their effect sizes (f²) were assessed, which measure the relative strength and practical relevance of an effect. With a threshold of 0.02 for minimal effect and 0.35 for a large effect (Hair et al., 2017), the analysis delved deeper. Additionally, R² values, indicative of the variance explained for the sample, were evaluated. With values greater than 0.2 for dependent variables (i.e., PR, PB, and BI), they align well with behavioral research standards (Hair et al., 2017). The blindfolding procedure was implemented to determine the predictive relevance (Q²) of the path model. A predictive relevance greater than 0 is indicative of acceptable predictive quality for each model (Hair et al., 2017).

n = 3,000 subsample; ***p < .001; **p < .01; *p < .05.

Multigroup Analysis: U.S. vs. China (H9)

Analyzing MGA—U.S. vs. China (Table 5)—consumer concern about collecting private information enhanced perceived risk (β _U.S.= .723, p < .001; β _China= .798, p < .001) but lowered perceived benefit (β _U.S.= −.073, p < .05; β _China= −.088, p < .05). However, there were no significant differences across two groups. Cultural differences were found in the impacts of perceived data management errors and secondary use. Interestingly, for the China group, effects of the perceived errors on perceived risk (β _U.S.= .009, p > .05; β _China= .076, p < .05; |diff|= .141, p < .01) and benefit (β _U.S.= –.049, p > .05; β _China= –.178, p < .01; |diff|= .129, p < .05) were significantly greater when compared to the U.S. group. Conversely, for the U.S. group, effects of the perceived secondary use on perceived risk (β _U.S.= .103, p < .01; β _China= .006, p > .05; |diff|= .108, p < .05) and benefit (β _U.S.= −.108, p < .05; β _China= –.025, p > .05; |diff|= .133, p < .05) were significantly greater when compared to the China group.

Regarding the influence of service-related factors, significant differences were not detected across the two groups (U.S. vs. China). The performance expectancy had no influence on privacy calculus (i.e., risk and benefit) for both groups. For expected service excellence, performance expectancy (β _U.S.= .404, p < .001; β _China= .335, p < .001), facilitation condition (β _U.S.= .153, p < .01; β _China= .205, p < .001), and hedonic motivation (β _U.S.= .229, p < .001; β _China= .178, p < .01) positively enhanced the perceived benefit of information disclosure without reducing risk perception for both groups. Thus, H9 was supported.

In addition to testing hypotheses, we found that perceived risk had a stronger negative effect on future service use intention in China (β _U.S.= –.159, p < .001; β _China= −.286, p < .001; |diff|= .128, p < .05), while the U.S. showed a stronger positive effect of perceived benefit (β _U.S.= .620, p < .001; β _China= .420 p < .001; |diff|= .201, p < .01).

Discussion and Conclusions

From the perspective of CFIP, our findings shed light on the significance of company practices in collecting extensive private information and the potential use of collected data (secondary use). Consumers perceive risk when disclosing and sharing their information, and this perception is negatively linked to the expected benefit of such disclosures (e.g., Lin et al., 2021). This study provides empirical evidence that highlights the importance of transparency in management practices, especially concerning data collection and future data usage, in addressing consumers’ privacy concerns. Moreover, our findings suggest that consumers’ concerns toward proper error management by a company do not induce risk but are related to a reduction in perceived benefit. This implies that consumers tend to believe that errors in shared datasets should be managed to improve services, which, in turn, encourages them to share their personal information to adopt the service (e.g., Zhong et al., 2022). Additionally, our results indicate that, unlike other data management concerns, concerns related to improper access do not significantly explain the privacy calculus. This may be attributed to the rapid development of secure systems that can detect unauthorized access, such as data encryption (e.g., Vance et al., 2020). Consumers are less concerned about improper access because companies have implemented security systems to protect against unauthorized data breaches, including hacking incidents.

From the perspectives of the UTAUT, our findings reveal that while expectations of service excellence, which encompass performance expectancy, facilitating conditions, and hedonic motivation, influence the perception of benefit regarding information disclosure, they do not significantly impact risk perception. This suggests that within the context of personalized retail services, which prioritize enhancing the shopping experiences of fashion consumers, the privacy calculus mechanism operates differently regarding service expectations. Unlike previous research that emphasized utilitarian motivations for smart services in fields like healthcare (Li et al., 2016), our study underscores the prominence of experiential motivations in fashion retail. Therefore, retailers should focus on enriching consumers’ in-store experiences, not necessarily to diminish their risk perceptions, but to accentuate the benefits of disclosing personal information.

Significantly, our findings provide empirical evidence supporting cultural differences in consumers’ concerns about the management of soft biometric information, while cultural differences were not detected in the influence of expected service factors on the privacy calculus. Specifically, our results indicate that fashion retail practices related to the collection of personal information (i.e., collection) play a critical role in understanding the privacy calculus within both U.S. and Chinese cultures (e.g., Lin et al., 2021). Given the availability of retail services across different cultures, international retail practitioners must increase transparency in data collection procedures to increase benefits and decrease consumers’ privacy concerns.

Importantly, we found that consumer concerns about improper error management are important for understanding Chinese consumers’ privacy calculus (i.e., risk/benefit), while consumer concerns about secondary use are crucial for understanding the privacy calculus of U.S. consumers. To our acknowledge, this is the first attempt in identifying cultural differences in the perception of risk associated with information disclosure, particularly in the context of retail services utilizing soft biometrics. This cultural difference in data management practices adds novel evidence to the literature. China's collectivist culture, which values social rules and selflessness for the good of the community, leads Chinese consumers to entrust companies with the principled use of their shared private information. Chinese consumers’ privacy concerns (i.e., error management) may be rooted in their prevention-focused mindset, which is often associated with collectivistic or interdependent individuals (Zhong et al., 2022). In contrast, U.S. consumers are more concerned with preventing secondary data usage by companies. This indicates that the belief of U.S. consumers is distinctive from Chinese consumers as U.S consumers emphasize the maintenance of the original purpose of data use (i.e., Zhong et al., 2022). These findings emphasize the need for retailers to adapt their data handling and marketing approaches to the distinct preferences of each country. In the U.S., a focus on protective measures against unauthorized secondary data usage and open dialogue on data practices is essential. Meanwhile, retailers in China should pay closer attention to reducing information errors to amplify perceived advantages and curtail potential concerns.

Finally, our research findings highlight discernible factors influencing adoption intentions in the context of invasive retail services among Chinese and U.S. consumers. Notably, Chinese consumers, possibly influenced by a cultural inclination toward placing trust in corporations for responsible personal data management, exhibit a greater degree of risk aversion, carefully considering the consequences of disclosing their data (Zhong et al., 2022). Conversely, U.S. consumers appear to be primarily motivated by the immediate advantages they gain, aligning with prior research that underscores benefit-driven behaviors (Zhong et al., 2022).

Theoretical Contributions

This study adopts two theoretical approaches, CFIP and UTAUT, to provide a theoretical lens for understanding consumers’ perceptions of the risks and benefits of disclosing soft biometric information within the extended PCT framework. By contextualizing the adoption of biometrics-based services within the wider realm of information service literature, this research opens new theoretical avenues. It lays the groundwork for exploring the broader integration of soft biometric information into various services and sectors. Future research could extend this model to other areas where personalization and data privacy intersect. Additionally, the extended PCT framework could be expanded to explore variables such as trust in technology and the legal and ethical implications of data use.

Practical Implications

This research provides valuable insights for the retail service industry. First, it explores the use of soft biometric information in fashion retail, an area that has not been fully investigated in previous research. This offers actionable insights to fashion retailers, highlighting the potential for leveraging soft biometric data to significantly enhance shopping personalization while improving transparency in the management of personal body information. Retailers should consider how they can enhance the shopping experience and address privacy concerns related to the use of soft biometric data. Second, the study underscores the importance of recognizing cultural differences in consumers’ privacy concerns and perceived benefits when deploying personalized and potentially invasive retail services (e.g., mobile body scanning app). International retailers and companies must tailor their data handling and marketing strategies to align with the distinct preferences and cultural values of each country. For example, focusing on protective measures against unauthorized secondary data usage in the U.S. and ensuring transparency in data collection procedures in China are crucial considerations.

Limitations and Future Studies

While this study provides novel insights, it has certain limitations. It focuses on body scanning technology, and the risks and benefits may be specific to this technology and its objectives. The findings may not be generalizable beyond this context. Future research should consider situational and demographic factors related to the technology and explore a more diverse range of cultures to assess the generalizability of the results. Additionally, addressing issues with convenience sampling and incorporating a broader range of cultural contexts would strengthen the research.

Footnotes

Declaration of Conflicting Interests

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

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The investigation involving Chinese participants received support from the National Yang Ming Chiao Tung University’s Higher Education Sprout Project, as well as the Ministry of Education, Taiwan.

ORCID iD

Song-yi Youn

Author Biographies

Song-yi Youn is an Assistant Professor in the Department of Textile and Apparel Management, University of Missouri. She holds her PhD in Consumer Behavior from Purdue University, and another PhD from Hanyang University. The research areas include consumer experiences with retail technologies and concerns potentially associated with social problems. Her work provides knowledge in consumer behaviors in the technology-driven market.

Chi-Cheng Luan is an Assistant Professor in the Department of Management Science, National Yang Ming Chiao Tung University (NYCU), Taiwan. Currently the aim of his study is how situational/environmental factors and personality traits may influence socially responsible behavior. The topics include virtual reality application on socially responsible consumption and the effect of corporate social responsibility on consumer evaluation.

References

Carolis

B. D.

Macchiarulo

Palestra

(2018, October). A comparative study on soft biometric approaches to be used in retail stores. In International Symposium on Methodologies for Intelligent Systems, pp.120–129.

Chowdhary

(2020). Is facial recognition the future of retail? Retrieved from http://apparelresources.com/technology-news/retail-tech/facial-recognition-future-retail/

D'Apuzzo

(2007, January). 3D Body scanning technology for fashion and apparel industry. In Videometrics IX (Vol. 6491, pp. 203–214). SPIE.

Hair

Hollingsworth

C. L.

Randolph

A. B.

Chong

A. Y. L.

(2017). An updated and expanded assessment of PLS-SEM in information systems research. Industrial Management & Data Systems, 117(3), 442–458. https://doi.org/10.1108/IMDS-04-2016-0130

Harman

H. H.

(1967). Modern factor analysis. University of Chicago Press.

Hassan

Izquierdo

Piatrik

(2021). Soft biometrics: A survey. Multimedia Tools and Applications, https://doi.org/10.1007/s11042-021-10622-8

Hofstede

(2001). Culture's Consequences: Comparing Values, Behaviors, Institutions, and Organizations Across Nations (2nd ed.). Sage Publications. https://doi.org/10.1016/S0005-7967(02)00184-5

Jaha

E. S.

Nixon

M. S.

(2014). Soft biometrics for subject identification using clothing attributes. In Proceedings of the IEEE International Conference on Biometrics (pp. 1–6). IEEE. https://doi.org/10.1109/BTAS.2014.6996278

Kim

Forsythe

(2008). Adoption of virtual try-on technology for online apparel shopping. Journal of Interactive Marketing, 22(2), 45–59. https://doi.org/10.1002/dir.20113

10.

Lee

Leonas

K. K.

(2021). Millennials’ intention to use self-checkout technology in different fashion retail formats: Perceived benefits and risks. Clothing and Textiles Research Journal, 39(4), 264–280. https://doi.org/10.1177/0887302X20926577

11.

Gao

Shi

(2016). Examining individuals’ adoption of healthcare wearable devices: An empirical study from the privacy calculus perspective. International Journal of Medical Informatics, 88, 8–17. https://doi.org/10.1016/j.ijmedinf.2015.12.010

12.

Lin

M. Y. C.

B. R.

Nguyen

T. T.

Cheng

J. M. S.

(2021). Effects of personal innovativeness and perceived value of disclosure on privacy concerns in proximity marketing: Self-control as a moderator. Journal of Research in Interactive Marketing, 16(2), 310–327. https://doi.org/10.1108/JRIM-04-2021-0112

13.

Nam

Cho

Kim

Y. D.

(2021). Cross-cultural examination of apparel online purchase intention: SOR paradigm. Journal of Global Fashion Marketing, 12(1), 62–76. https://doi.org/10.1080/20932685.2020.1845766

14.

Owen

. (2021). Artificial Intelligence at Nike. Retrieved from https://emerj.com/ai-sector-overviews/artificial-intelligence-at-nike/

15.

Park

Choi

Jung

(2021). Users’ cognitive and affective response to the risk to privacy from a smart speaker. International Journal of Human–Computer Interaction, 37(8), 759–771. https://doi.org/10.1080/10447318.2020.1841422

16.

Pizzi

Scarpi

(2020). Privacy threats with retail technologies: A consumer perspective. Journal of Retailing and Consumer Services, 56, 102160. https://doi.org/10.1016/j.jretconser.2020.102160

17.

Roberts-Islam

(2020). H&M Group Experiments with Made-to-Order ‘Fast Fashion’ to Tackle Overstock. Forbes. Retrieved from https://www.forbes.com/sites/brookerobertsislam/2020/11/26/hm-group-experiments-with-made-to-order-fast-fashion-to-tackle-overstock/?sh=232bbaba23fa

18.

Salfino

(2021). Why fashion retail better prepare now for data privacy laws. Retrieved from https://sourcingjournal.com/topics/lifestyle-monitor/data-privacy-laws-ecommerce-fashion-apparel-amazon-mckinsey-gdpr-264544/

19.

Smith

H. J.

Milberg

S. J.

Burke

S. J.

(1996). Information privacy: Measuring individuals’ concerns about organizational practices. MIS Quarterly, 20(2), 167–196. https://doi.org/10.2307/249477

20.

Tomić

Kalinić

Todorović

(2022). Using the UTAUT model to analyze user intention to accept electronic payment systems in Serbia. Portuguese Economic Journal, 1–20.

21.

Vance

Siponen

M. T.

Straub

D. W.

(2020). Effects of sanctions, moral beliefs, and neutralization on information security policy violations across cultures. Information & Management, 57(4), 103212. https://doi.org/10.1016/j.im.2019.103212

22.

Venkatesh

Thong

J. Y. L.

(2016). Unified theory of acceptance and use of technology: A synthesis and the road ahead. Journal of the Association for Information Systems, 17(5), 328–376. https://doi.org/10.17705/1jais.00428

23.

Vimalkumar

Sharma

S. K.

Singh

J. B.

Dwivedi

Y. K.

(2021). Okay google, what about my privacy?’: User’s privacy perceptions and acceptance of voice-based digital assistants. Computers in Human Behavior, 120, 106763. https://doi.org/10.1016/j.chb.2021.106763

24.

Wang

Zhao

Ubul

(2019). Hunting for fashion via large-scale soft biometrics analysis. In Proceedings of International Conference on Biometrics (pp. 1–6). IEEE.

25.

Youn

S. Y.

Hwang

Zhao

Kim

J. B.

(2023). Privacy paradox in 3D body scanning technology: The effect of 3D virtual try-on experience in the relationship between privacy concerns and mobile app adoption intention. Humanities and Social Sciences Communications, 10(1), 1–13. https://doi.org/10.1057/s41599-022-01483-z

26.

Zhong

Sun

Zhou

Xie

(2022). Privacy matters: Reexamining internet privacy concern among social media users in a cross-cultural setting. Atlantic Journal of Communication, 1–18. https://doi.org/10.1080/15456870.2022.2099548