wearable device adoption model with tam and ttf hyung sik...

518 Int. J. Mobile Communications, Vol. 14, No. 5, 2016

Copyright © 2016 Inderscience Enterprises Ltd.

Wearable device adoption model with TAM and TTF

Hyung Sik Chang Graduate Program in Technology Policy, Yonsei University, 262 Seongsanno, Seodaemun-gu, Seoul 120-749, Korea Email: [email protected]

Seul Chan Lee* and Yong Gu Ji Department of Information and Industrial Engineering, Yonsei University, 262 Seongsanno, Seodaemun-gu, Seoul 120-749, Korea Email: [email protected] Email: [email protected] *Corresponding author

Abstract: Wearable devices have received great highlights as next core products for global information technology companies. Since the wearable-device market is in its early phase, major factors influencing the adoption behaviour have not been completely identified. The goal of this study, therefore, is to investigate the important factors affecting usage behaviour of wearable devices. To achieve this, we applied an integrated model based on the technology acceptance model and the task-technology fit model. Along with task and technology characteristics, factors of wearable device, social influence and user characteristic were also considered. The survey was conducted with 342 participants, and the results were analysed by the partial least squares method using a two-phase procedure. The research model explained 50.3% of the behavioural intention variance and the 13 out of 15 hypotheses were statistically supported. We identified the explanatory power of the proposed model. Interestingly, users did not expect wearable devices to provide communication functions or become fashion items.

Keywords: adoption behaviour; task-technology fit (TTF) model; technology acceptance model (TAM); wearable device.

Reference to this paper should be made as follows: Chang, H.S., Lee, S.C. and Ji, Y.G. (2016) ‘Wearable device adoption model with TAM and TTF’, Int. J. Mobile Communications, Vol. 14, No. 5, pp.518–537.

Biographical notes: Hyung Sik Chang is a PhD candidate in the Graduate Program in Technology Policy at Yonsei University, Korea. His research interests is technology policy in IT industry. He has over 20 years R&D management experience in global IT company.

Seul Chan Lee is a PhD candidate in the Department of Information and Industrial Engineering at Yonsei University, Korea. His research interests include human-computer interaction, human factors, and user experience, and interface design.

Wearable device adoption model 519

Yong Gu Ji is a Professor in the Department of Information and Industrial Engineering at Yonsei University, where he directs the Interaction Design Laboratory. He received his PhD in Industrial Engineering from Purdue University. His research interests include UX design in smart device, emotional design, accessibility, and the elderly in human-computer interaction.

1 Introduction

Intense competition for high-tech product market share encourages companies to release novel products. Among the different types of information technology (IT) devices, wearable devices have attracted attention as the next generation smart devices. Although markets are still growing, there are uncertainties in high-tech product market. Therefore, in order to decrease the uncertainty and achieve success in a high-technology market, it is critical to analyse the characteristics of these devices and explore the antecedents of adoption. It is important to analyse the adoption of wearable devices because they are expected to have an important role in the future of the IT device market. It is seemingly obvious that hardware technology market is stable and the speed of development is slow. Therefore, those in both academic and industrial fields are focusing on wearable devices that can lead to an extended and improved usage of previous smart devices such as smartphones and tablet PCs.

A wearable device is a smart device that has a new form factor. Several studies have defined wearable devices with a different terminology, such as wearable computers or wearable technology. Silina and Haddadi (2015) defined wearable devices as “a general term that currently refers to devices, worn on or around body, including, but not limited to garments, shoes, accessories, and jewelry, that have input, output, or both.” Gemperle et al. (1998) defined the term wearable as “implying the use of the human body as support for some product.” Seymour (2008) stated that the term wearable technology includes “the electrical engineering, physical computing, and wireless communications networks that make a fashionable wearable functional.” Bieber, Kirste and Urban (2012) defined smart watches as “wrist worn devices that have computational power, integrated sensors, connectivity to other devices or the Internet, and an integrated clock.” These explanations share several characteristics including ‘be worn’, ‘computer’, ‘sensor’, and ‘internet’. In summary, a wearable device is a smart device that can be worn.

Wearable devices have special characteristics unlike any other smart devices. First, wearable devices essentially have different forms as compared to prior existing devices. Previously, devices allowed users to perform an overall task on a large screen with internet access and sensor technologies. However, wearable devices typically have no display or only a small-sized display. Hence, users perceive wearable devices differently from prior smart devices. Moreover, if a user wants to utilise a wearable device, they must actually put on the device. People have the tendency to feel uncomfortable when faced with new products and could feel strange wearing a technological device, as was experienced with the first cellphones.

520 H.S. Chang et al.

These traits lead to different forms of discussion regarding the success of devices. Although the literature has focused on versatility or general usability for predicting the success of prior smart devices, specialised functions or characteristics are expected to play important roles for wearable devices (Chai et al., 2014). Recent investigations of wearable devices have focused on the expansion of previous devices or special objectives rather than versatility. Therefore, analysing the factors for the success of wearable devices provides meaningful milestone to predict the future direction of IT devices.

The main goal of this study is to predict the aspects of user adoption of wearable devices. Specifically, we proposed a theoretical model based on the technology acceptance model (TAM), task-technology fit (TTF) model, and external factors associated with the usage intention of wearable devices. The remainder of this paper is organised as follows. Section 2 provides a literature review for the research model. Section 3 presents the research model and hypotheses. Section 4 explains the methodology and Section 5 discusses the results. Finally, Section 6 presents discussion and the conclusion.

2 Literature review

2.1 Technology acceptance model

A number of theories have been developed to explain the IT usage behaviour. However, the representative theory to explain the determinants that influence the IT usage is the TAM, suggested by Davis (1986).

The TAM was theorised based on ‘behavioural intention (BI)’, ‘perceived usefulness (PU)’, and ‘perceived ease of use (PEOU)’. BI is defined as the key factor influencing actual use. An actual behaviour is determined by intention to use. The two antecedents, PU and PEOU, have a direct positive influence on BI. Accordingly, people will want to use the IT device if it will help them perform a task or the benefits from the IT system exceed the effort of using it. Furthermore, people consider that the more effort an IT system requires, the less its usefulness. The relationships between these constructs are found in studies related to IT (Lin, 2014; Tarhini, Hone and Liu, 2014).

Although the usefulness of the TAM has been demonstrated in many studies, the model has some weaknesses in understanding IT usage. The TAM focuses on the individual user’s attitude or belief towards IT, and does not incorporate the actual task aspects (Dishaw and Strong, 1999).

2.2 TTF model

According to Goodhue and Thompson (1995), the adoption of a new technology is dependent on an individual’s task. Hence, they suggested the TTF model to explain an individual’s adoption usage. The TTF model posits that the appropriateness between task requirements and technology functions leads to technology utilisation and higher performance results. The TTF is defined as “the degree to which a technology assists an individual in performing his or her portfolio of tasks.” Thus, the TTF theory can compensate for the weaknesses of the TAM. Up to now, several integrated models of the


TAM and the TTF model have been developed to explain user behaviour (Chang, 2008; Dishaw and Strong, 1999; Polančič, Heričko and Pavlič, 2011; Yen et al., 2010), and the usefulness of such integrated models has been demonstrated.

The TTF model includes two main variables: task and technology characteristics. Tasks refer to “the actions carried out by individuals in turning inputs into outputs,” and technologies are the “tools used by individuals in carrying out their tasks” (Goodhue and Thompson, 1995). Previous studies addressed these constructs in two manners. First, the two variables have been used directly as latent variables (Chang, 2008; Yen et al., 2010). Second, each variable has been divided into several sub-variables according to the context (Chung, Lee and Choi, 2015; Polančič, Heričko and Pavlič, 2011). For example, Chung, Lee and Choi (2015) proposed two task-related and three technology-related characteristics. They constructed relationships between these variables and the TTF variables. Polančič, Heričko and Pavlič (2011) suggested four technological characteristic factors in the development context: confidence, efficiency, adaptability, and understandability. We, however, developed another method to address these constructs because the task and technology characteristics of wearable devices are not familiar and must be defined for people to use the devices.

Three characteristics were derived based on the previous research. Pitzer et al. (2013) suggested six categories of wearable device tasks: fitness, medical, lifestyle, infotainment, gaming, and other. A Flextronics technical report (2014) categorised seven criteria as follows: security/safety, medical, wellness, sport/fitness, lifestyle, computing, and communication. We excluded security/safety because it is not pertinent to the personal purpose of using the device. Healthcare embraces similar categories such as fitness, medical, and wellness. We included lifestyle as a user characteristic factor; hence, we excluded the lifestyle task. Gaming was integrated into infotainment. In summary, we suggested three task characteristics, i.e., communication, infotainment, and healthcare.

Connectivity has been suggested as a technology characteristic of wearable devices in many studies. Connectivity describes the interaction between devices using Bluetooth or wireless network technology. Wireless network and sensor technology allow wearable devices to enhance human-computer interaction and synchronise personal data between devices (Bieber, Kirste and Urban, 2012). A significant amount of the literature has emphasised the importance of connectivity of wearable devices (Chai et al., 2014; Pitzer et al., 2013). Chai et al. (2014) pointed out poor connectivity issues must be solved for any meaningful research on wearable devices. According to Pitzer et al. (2013), wearable technologies are not new, but connectivity is one of the reasons that they are attracting much attention recently.

2.3 External factors

Aldhaban (2012) reviewed the research on the adoption of smartphones and categorised the external factors of smartphone adoption as device and services characteristics, facilitating conditions, social factors, and user characteristics. We selected external factors in accordance to this categorisation; however, we excluded facilitating conditions because they are already well constructed due to the popularisation of smartphones.


2.3.1 Wearable device factors As discussed previously, users must wear a wearable device to use it. Thus, the intention to adopt wearable devices must consider clothing characteristics. We chose two important factors, wearability and fashionability, after reviewing the following studies.

Many researchers considered wearability as a major evaluation factor of a wearable device. Motti and Caine (2014) identified 20 human factor considerations in the design of wearable devices; two clothing-related characteristics, wearability and fashion, were included. Gemperle et al. (1998) emphasised the importance of the wearability of wearable devices and defined the term wearability as the interaction between the human body and the wearable object. Marculescu et al. (2003) suggested that the concept of wearability includes being lightweight, breathable, comfortable, easy to wear and to remove, and easy to access wounds. They asserted that wearable products must exhibit these characteristics for adoption.

Additional research has suggested fashionability, the role as a fashion item, as important for users to accept a wearable device. Seymour (2008) defined fashionable technology as designed garments, accessories, or jewellery that combine aesthetics and style with functional technology. Wrist watches are examples of functional devices that are also fashion items (Narayanaswami et al., 2001). Pascoe and Thomson (2007) stated that the smartwatch may be a socially and fashionably acceptable computing device. Silina and Haddadi (2015) analysed wearable devices as a fashion item with technology in the fashion market. They analysed 187 wearable devices and suggested that more than half of the devices could be considered as fashion items.

2.3.2 Social influence factor Social influence refers to the perceived pressure to perform a certain behaviour (Fishbein and Ajzen, 1977). An individual’s behaviour is often not solely the consequence of their inner aspect. Individuals assess themselves depending on their social relationships and decide to act based on these assessments. Accordingly, different studies have suggested social influence factors as antecedents of the usage behaviour (Tan et al., 2015; Wang and Chou, 2016). Therefore, we considered the social influence factor category in our study.

2.3.3 User characteristic factors The first user characteristic factor we have considered is perceived privacy. Privacy is considered as important factor when users want to use new IT and privacy variables have received attention (Lai and Shi, 2015).

Moreover, there is a growing interest in privacy issues because many information leakage accidents have occurred recently. Consequently, perceived privacy is an increasingly important factor for adopting personal devices. Further, many applications in wearable devices require personal information such as GPS data and health records.

Privacy factors have been discussed in different forms, such as security, trust, and perceived privacy. In this study, we focus on perceived privacy. Perceived privacy can be defined as an individual’s ability to control the terms by which their personal information is acquired and used (Shin, 2010). Individual privacy anxieties result from the concern that personal information is not secure (Dinev and Hart, 2004), even if information systems provide a high level of security.


The second user characteristic factor we consider is innovativeness. Innovativeness is classified as a personal characteristic construct used to predict the personal trait to accept technology. Innovativeness explains the degree of acceptance of new ideas or technology in comparison to others and can be defined as an individual’s willingness to test new information technologies (Agarwal and Prasad, 1998). Individuals who have high innovativeness are more inclined to use new devices.

The third user characteristic factor we consider is lifestyle. With the rapid advances in IT, many new devices and services have been developed. However, the adoption and continuous usage phases are different depending on the culture, age, and gender of the prospect users. Accordingly, whether users accept new IT is highly influenced by personal lifestyle. Lifestyles can be defined as “the manner in which people conduct their lives, including their activities, interests and opinions” (Peter and Olson, 1994). Recent studies have used lifestyles to verify user adoption behaviour (Chan and Leung, 2005; Li, 2013).

3 Research model and hypotheses

3.1 Research model

Figure 1 illustrates the proposed research model in this study. As discussed, the model is constructed from a combination of the TAM and the TTF models. Moreover, we incorporate additional external factors to extend the original model. A total of 15 hypotheses are proposed based on the literature review.

Figure 1 Proposed research model


3.2 Hypotheses

3.2.1 TAM constructs

As discussed, many studies have demonstrated that PU and PEOU positively influence BI, even though the significance levels differ. Moreover, studies have demonstrated that the effect of PEOU on BI is mediated by PU. Therefore, these relationships may also be found in the wearable device adoption context.

H1: Perceived usefulness has a positive effect on behavioural intention.

H2: Perceived ease of use has a positive effect on behavioural intention.

H3: Perceived ease of use has a positive effect on perceived usefulness.

3.2.2 TTF constructs Two relationships between constructs were identified by reviewing the previous studies that utilised an integration model of the TAM and TTF model (Chang, 2008; Dishaw and Strong, 1999; Yen et al., 2010): the task-technology fitness positively influences the PU and PEOU. Therefore, we formulated the following hypotheses.

H4: Task-technology fitness has a positive effect on PU.

H5: Task-technology fitness has a positive effect on PEOU.

Past research has documented three task characteristics and two technology characteristics. The majority of the studies have concluded that task characteristics and technology characteristics have a positive relationship (Chang, 2008; Yen et al., 2010). Therefore, we hypothesise that individuals with a higher expectation level of task and technology characteristics are more likely to have more positive opinions of the TTF. Based on these perspectives, we formulated the following hypotheses.

H6: Communication task characteristics have a positive effect on TTF.

H7: Healthcare task characteristics have a positive effect on TTF.

H8: Infotainment characteristics have a positive effect on TTF.

H9: Connectivity task characteristics have a positive effect on TTF.

3.2.3 Wearable device characteristic constructs

Previous studies have confirmed that fashionability (Marculescu et al., 2003; Profita, 2011) and wearability (Narayanaswami et al., 2001; Pascoe and Thomson, 2007) are major constructs for predicting wearable types of IT devices. Therefore, we constructed the following hypotheses:


H10: Fashionability has a positive effect on behavioural intention.

H11: Wearability has a positive effect on behavioural intention.

3.2.4 Social influence constructs One of the major variables of social influence is subjective norms (SN). SN is defined as “the degree with which individuals perceived that people who are important to them think they should or should not use a certain system or perform a certain action” (Fishbein and Ajzen, 1977; Venkatesh and Davis, 2000). Empirical studies confirmed that SN positively influences BI (Lin, 2014; Tarhini, Hone and Liu, 2014). Therefore, we constructed the following hypothesis:

H12: Subjective norm has a positive effect on behavioural intention.

3.2.5 User characteristic constructs Many studies have found that a user’s belief regarding the protection level of their information positively influences their behavioural intention (Shin, 2010; Zhou, 2011). Based on previous research, we formulated the following hypothesis:

H13: Perceived privacy has a positive effect on behavioural intention.

Lifestyle is a significant predictor of PU. In accordance with the literature (Chan and Leung, 2005; Li, 2013), we considered that people who think wearable devices suit their lifestyle may want to use them. Thus, we constructed the following hypothesis:

H14: Lifestyle has a positive effect on BI.

Previous IT studies have confirmed the relationship between innovativeness and PEOU (Agarwal and Prasad, 1998). Hence, we proposed the following hypothesis.

H15: Innovativeness has a positive effect on PEOU.

4 Methodology

4.1 Questionnaire

We gathered and revised the measurement variables to validate the proposed model (Table 1). A total of 49 items of 14 latent variables were developed to identify users’ opinions based on a seven-point Likert scale (1: strongly disagree ~ 7: strongly agree). The majority of the items and variables were from the previous literature related to adoption usage behaviour; we developed additional items because task and technology characteristics had not been studied previously. However, we developed these items based on the composition of similar items that were used previously. We then verified the reliability and validity of all the measurement items.


In this study, we wanted to gather data from both experienced and unexperienced users. Accordingly, we added explanations to provide information on wearable devices before the questionnaire part for the unexperienced participants’ group. Therefore, the unexperienced participants conducted the questionnaire based on their prior experience on smart devices and the information we provided. Table 1 Construction of latent and measurement variables

Latent variables Measurement variables Behavioural intention (BI)

BI_1) Given the opportunity, I will use wearable devices BI_2) I am likely to use wearable devices in the near future BI_3) I am willing to use wearable devices in the near future BI_4) I intend to use wearable devices when the opportunity arises

Perceived usefulness (PU)

PU_1) Using wearable devices improves daily task performance PU_2) It is useful for me to use wearable devices every day PU_3) Wearable devices are beneficial to me

Perceived ease of use (PEOU)

PEOU_1) It is easy to handle the wearable device that I own PEOU_2) It is easy to use the wearable device that I own at any time PEOU_3) It is easy to learn how to use the wearable device that I own

TTF TTF_1) The functionalities of wearable devices were adequate TTF_2) In helping me to perform the assigned task, the functionalities of the wearable devices were adequate TTF_3) The functionalities of wearable devices were appropriate TTF_4) In general, the functionalities of wearable devices best fit the task

Connectivity Connectivity_1) It is useful to connect wearable devices with other smart devices such as a smartphone or tablet PC Connectivity_2) It is useful if the applications of my smart devices are synchronised Connectivity_3) It is useful to use wearable devices with other smart devices

Communication Communication_1) I need to check messages, e-mails, or phone calls using wearable devices Communication_2) I intend to use the wearable device to interact with my friends, family, and colleagues Communication_3) Messaging through wearable devices enables me to respond to my friends, family, and colleagues

Healthcare Healthcare_1) It is useful for me to use wearable devices for health management Healthcare_2) I am interested in managing my health Healthcare_3) It is useful for me to use healthcare applications through wearable devices

Infotainment Infotainment_1) It is useful for me to use wearable devices for checking SNS and surf the web Infotainment_2) It is useful for me to use entertainment applications through wearable devices Infotainment_3) I intend to use wearable devices to search information and play game applications


Table 1 Construction of latent and measurement variables (continued)

Latent variables Measurement variables Fashionability Fashionability_1) I think wearable devices are fashion items

Fashionability_2) The outward appearance of wearable devices is important Fashionability_3) The design and aesthetics of wearable devices are important to me

Wearability Wearability_1) The fit and overall comfort of wearable devices are important for me Wearability_2) I am reluctant to use wearable devices that feel uncomfortable Wearability_3) I am interested in using wearable devices regardless of how they feel

Subjective norm SN_1) People I am influenced by think I should use wearable devices SN_2) People who are important to me think that I should use wearable devices SN_3) My friends think I should use wearable devices

Perceived privacy

PP_1) I am confident that I know all the parties who collect the information I provide during the use of wearable devices PP_2) I am aware of the exact nature of information that will be collected during the use of SNS PP_3) I am not concerned that the information I submit on the wearable devices could be misused

Innovativeness Innovativeness_1) If I heard about a new information technology, I would look for ways to experiment with it Innovativeness_2) Among my peers, I am usually the first to try new information technologies Innovativeness_3) In general, I am hesitant to try new information technologies Innovativeness_4) I like to experiment with new information technologies

Lifestyle Lifestyle_1) Using wearable devices fits my lifestyle well Lifestyle_2) I think that wearable devices fit my lifestyle well Lifestyle_3) Using wearable devices is completely compatible with my current situation Lifestyle_4) Using wearable device is compatible with all aspects of my lifestyle

The questionnaire was written in Korean. We minimised possible misinterpretation or errors by forward translation and backward translation processes between Korean and English version. First, two researchers translated an English questionnaire into Korean. Then another researcher retranslated the Korean questionnaire into English. We confirmed that the meaning of all items was identical, with minor differences in wording.


4.2 Participants

We recruited 386 participants for a month in January, 2015. We performed two processes for controlling data quality. First, data were collected both online and offline. Data from the online group were collected using a web-version questionnaire; data from the offline group were collected using a printed-version. The two versions of the questionnaire were identical except for the page format. Statistical tests revealed no significant difference on the results between the online and offline groups. Second, we included four questions that required participants to answer with a designated number to distinguish their sincerity, e.g., “Please check Number 4 in this question”. If a participant checked the wrong number, his or her data were removed. The data from 37 participants were excluded according to this sincerity test. Similarly, seven participants were excluded because they missed several measurement items.

Table 2 is the overall information for the participants. We divided the participants into two groups: those who had experienced wearable devices and those who had not. The experienced group was composed of people who had experienced wearable devices including owning a device or handling a device directly. The unexperienced group participants were people who did not have previous knowledge of wearable devices or who only had an indirectly experienced.

Table 2 Descriptive statistics of participants

Construct Total With experience Without experience Number of participants 342 127 215 Gender Male 194 (56.7%) 91 (71.7%) 103 (47.9%) Female 148 (43.3%) 36 (28.3%) 112 (52.1%) Age 20 193 55 138 30 89 46 43 40 50 22 28 50 10 4 6 Smartphone experience (year) 4.70 (1.82) 5.03 (1.68) 4.48 (1.87)

4.3 Data preparation and analyses

We implemented partial least squares analysis with SmartPLS 3.0 to perform path analyses and test the hypotheses. A two-phase method was used to analyse the results. First, we assessed the reliability and validity of the measurement model. Then, we tested the structural model based on the explained variance (R²) of the dependent variables and path coefficients (β) by bootstrapping with a 500 re-sampling method. Finally, we conducted a multi-group analysis.


5 Results

5.1 Measurement model

A measurement model must be assessed before a structural model is examined. A measurement model can be assessed based on internal consistency, convergent and discriminant validities. Cronbach’s alpha was used to verify the internal consistency; a value of 0.7 or higher is recommended. Furthermore, it is recommended that item loadings are recommended to exceed 0.6 and the composite reliability (CR) values are recommended to exceed 0.7. The average variance extracted (AVE) value for each latent variable should exceed 0.5, and the square root of the AVE should be greater than the inter-construct correlations.

Three questionnaire items were excluded owing to low reliability: Fashionability_3, SN_3, and PP_3. Table 3 displays all item loadings, Cronbach’s alphas, CR, and AVE after removing these three items. All values are greater than the recommended threshold levels (item loading > 0.6, Cronbach’s alpha > 0.7, CR > 0.7, AVE > 0.5), except for the Cronbach’s alpha value for fashionability. Although the value, 0.670, does not exceed the threshold level, it is somewhat acceptable.

Table 4 depicts the correlation matrix for the discriminant assessment. Every square root of the AVE values is higher than the correlation values between the latent variables. Furthermore, the correlation values between the latent variables are less than 0.7, indicating that multicollinearity issues were avoided. Thus, the measurement model was proven reliable and valid for the study.

To test the common method variance bias, Harman’s one-factor statistical test was applied. An exploratory factor analysis was conducted on all measurement items. The results indicated that 16 factors with eigenvalues greater than one were extracted, and a single factor did not emerge. Therefore, our research data did not have problems related to the common method variance bias. Table 3 Scales for reliability and convergent validity

Construct Item Mean SD Loading α CR AVE

Behavioural intention

BI1 5.45 1.36 0.902 0.944 0.960 0.856 BI2 5.46 1.39 0.937 BI3 4.96 1.60 0.913 BI4 5.26 1.45 0.949

Perceived usefulness

PU1 5.04 1.08 0.892 0.865 0.917 0.787 PU2 4.70 1.22 0.876 PU3 4.76 1.21 0.894

Perceived ease of use (PEOU)

PEOU1 4.65 1.36 0.900 0.893 0.933 0.823

PEOU2 4.75 1.30 0.920

PEOU3 5.04 1.30 0.902


Table 3 Scales for reliability and convergent validity (continued)

Construct Item Mean SD Loading α CR AVE

TTF TTF1 4.49 1.33 0.850 0.849 0.898 0.688 TTF2 4.62 1.11 0.829 TTF3 5.02 1.49 0.821

TTF4 4.42 1.15 0.818 Connectivity Connectivity1 5.90 0.99 0.905 0.895 0.934 0.826

Connectivity2 5.89 1.02 0.898 Connectivity3 5.93 0.98 0.923

Communication Communication1 5.23 1.31 0.823 0.843 0.905 0.762 Communication2 4.92 1.48 0.906 Communication3 4.72 1.32 0.887

Healthcare Healthcare1 5.71 1.21 0.862 0.868 0.919 0.790 Healthcare2 5.62 1.31 0.904 Healthcare3 5.51 1.09 0.900

Infotainment Infotainment1 5.25 1.32 0.777 0.826 0.896 0.744 Infotainment2 4.67 1.59 0.902 Infotainment3 4.57 1.62 0.901

Fashionability Fashionability1 5.76 1.07 0.913 0.670 0.854 0.746 Fashionability2 5.71 1.21 0.812

Wearability Wearability1 6.35 0.86 0.809 0.737 0.850 0.654 Wearability2 6.58 0.70 0.812 Wearability3 6.43 0.86 0.809

Subjective norm SN1 2.86 1.48 0.950 0.905 0.954 0.913 SN2 2.91 1.59 0.961

Perceived privacy PP1 3.73 1.56 0.964 0.926 0.964 0.931 PP2 3.82 1.60 0.965

Innovativeness Innovativeness1 4.61 1.58 0.871 0.896 0.928 0.765 Innovativeness2 5.19 1.50 0.927 Innovativeness3 5.29 1.49 0.787 Innovativeness4 5.41 1.35 0.907

Lifestyle Lifestyle1 4.33 1.24 0.895 0.939 0.957 0.846 Lifestyle2 4.35 1.44 0.921 Lifestyle3 4.27 1.51 0.924 Lifestyle4 4.39 1.39 0.939

Note: α (Cronbach’s alpha); AVE, average variance extracted; BI, behavioural intention; CR, composite reliability; PEOU, perceived ease of use; PU, perceived usefulness; SD standard deviation


Table 4 Correlation matrix and discriminant validity


5.2 The structural model

Table 5 presents the results of verifying the hypotheses. A t-test was conducted to test the significance of the path coefficients based on a significance level 0.05. Figure 2 illustrates the structural path coefficient estimates (β) with explained variances (R²).

Table 5 Structural model results

Hypothesis Path β t-value Significance Support?

H1 PU → BI 0.509 12.528 * Supported H2 PEOU → BI 0.096 2.12 ** Supported H3 PEOU → PU 0.054 1.116 0.265 Not Supported H4 TTF → PU 0.282 5.319 * Supported H5 TTF → PEOU 0.257 4.639 * Supported H6 Connectivity → TTF 0.142 2.520 ** Supported H7 Communication → TTF 0.085 1.250 0.212 Not Supported H8 Healthcare → TTF 0.253 4.819 * Supported H9 Infotainment → TTF 0.240 3.753 * Supported H10 Fashionability → BI 0.044 1.084 0.279 Not Supported H11 Wearability → BI 0.123 3.192 * Supported H12 SN → BI 0.155 3.280 * Supported H13 PP → BI 0.168 3.745 * Supported H14 Innovativeness → PEOU 0.339 6.678 * Supported H15 Lifestyle → PU 0.434 9.473 * Supported

*p < 0.05, **p < 0.01. Note: β (path coefficient)

Figure 2 Results of path analysis


The research model explains 50.3% of the BI variance, 38.4% of the PU, and 27.9% of the PEOU. The task and technology characteristics explain 27.9% of the TTF construct. All hypotheses were verified except H3, H7, and H10.

Figure 3 presents the results of the group comparison. First, a t-test was conducted to examine the significance of the path coefficients for each group. Then, a multi-group analysis was applied if the path coefficients of both groups were significant. Six hypothesis paths were significantly common (H1, H4, H5, H8, H14, H15). Among these paths, H5 was significantly different between the two groups when the multi-group analysis was applied (t = 2.27, p < 0.05).

Figure 3 Group comparison results

6 Discussion and conclusion

The objective of this research was to understand the aspects of wearable device adoption during the early stages. We developed integrated model of the TAM and TTF, incorporating several other latent variables, which were divided into four groups: task and technology characteristic, wearable device characteristic, social influence, and user characteristic factors. These factors were suggested to extend the understanding of the user behaviour. A survey was conducted to collect users’ opinions, and the research model was statistically tested using this data.

We found several important results from the analysed data. First, PU had more influence on BI than did PEOU in the TAM construct. Further, PU did not meditate the relationship between BI and PEOU. Through these results, we recognised that customer expectations in using wearable devices depend on how useful the device is rather than how easy it is to use. This is because users are familiar with smart devices such as smartphones and tablet PCs and find it easy to use. However, it is important whether a wearable device can benefit the user. For that reason, it is hard to expect consumers to use a wearable device unless it provides greater benefits than those experienced before.

Task technology fit is similarly associated with PU and PEOU. These results are in line with the existing adoption studies of IT (Chang, 2008; Dishaw and Strong, 1999). In


the TTF model construct, 27.9% of the variance was explained by task and technology characteristic factors. Among these factors, communication was not a significant predictor of TTF. This can be scrutinised more precisely by considering the differences in the wearable device experience. Interestingly, users without experience expected wearable devices to have communication and healthcare functions based on connectivity; users with experience did not expect these communication functions and connectivity. That is, inexperienced users expected a wearable device to have overall functions similar to other smart devices. On the other hand, experienced users expected a specialised device to provide healthcare or infotainment. These results can be interpreted in two different manners. First, wearable devices that were already available in the current market did not satisfy users’ expectations. Hence, they only used their device for a particular purpose. Second, whereas users at first expected wearable devices to be similar to smart devices such as smartphones, they actually discovered that wearable devices were useful when they used them with the purpose of performing specific tasks.

Wearability and fashionability were selected as wearable device characteristic factors. Wearability significantly influenced BI; fashionability did not. This indicates that a wearable IT device must be designed for comfort. Unexpectedly, fashionability was not a significant antecedent of BI; people did not expect wearable devices to be fashion items. However, further research is needed to scrutinise the fashionability factors as many previous studies have documented that wrist watches are fashion items. As we are in the early phase of wearable device adoption, the primary concern for users is whether these devices provide useful functions. For this reason, we were able to explain the non-significance of fashionability as a predictor of BI. However, future research is necessary to determine whether fashionability will remain unimportant following wearable device market growth.

The SN is a significant factor to predict BI as presented in other studies (Lin, 2014; Tan et al., 2015; Tarhini, Hone and Liu, 2014). That is, people believe that they are expected to use wearable devices based on the behaviour of the people around them.

The results for the user characteristic factors indicated that all hypotheses were supported. Perceived privacy positively influenced BI, innovativeness positively affected PEOU, and lifestyle positively influenced PU. These results were in line with previous studies. Remarkably, lifestyle was strongly associated with PU, and 38.4% of the variance of PU was explained by TTF and lifestyle. Considering the effect of PU on BI, TTF and lifestyle are key factors for predicting BI for wearable devices.

In conclusion, the proposed model and research results provide implications to the academic and industrial fields. First, this is an early phase of the study on investigation of users’ intention towards wearable devices. Although there is significant literature on smartphones and tablet PCs, there are a minimal number of studies on wearable devices. Therefore, this provides an opportunity to extend previous findings of users’ adoption. Second, our results have an important meaning that extends the original TAM and TTF model. We verified the validity and reliability of an integrated model that other studies have attempted. Furthermore, we provided evidence that new factors that are elicited from distinctive aspects of wearable devices are significant. Finally, we implemented a systematic approach and identified several factors that promote the understanding of wearable device adoption by classifying factors into task, technology, user characteristics, social influence, and device characteristics. We expect that these findings will provide significant evidence that researchers and engineers should consider when they address wearable devices.


Although the results of this study provide meaningful findings on the adoption of wearable devices, several limitations must be considered in future studies. First, a generalisation issue could exist because the participants were not perfectly controlled. Although the gender ratio was well balanced, the age distribution was not even. The number of participants aged 20-39 was relatively higher than the participants aged 40–59, even though we did not limit participants’ age. Accordingly, the findings could be relatively skewed towards young participants who tend to prefer and do not have issues using new technology and devices. For example, PEOU did not have a significant influence on PU and had a relatively low path coefficient to BI. Second, the factors we identified are focused on the design aspects of devices including task, technology, and device characteristics. Therefore, future studies should consider other factors than the technological aspect such as cost and benefit. Third, we did not gather any qualitative data to strengthen the validity and reliability of our findings, although we identified the factors and verified the proposed model by statistical tests. The results would be more meaningful if findings could be supported with evidence from a longitudinal study.

References Agarwal, R. and Prasad, J. (1998) ‘A conceptual and operational definition of personal

innovativeness in the domain of information technology’, Information Systems Research, Vol. 9, No. 2, pp.204–215.

Aldhaban, F. (2012) ‘Exploring the adoption of smartphone technology: literature review’, 2012 IEEE Proceedings of PICMET'12 in Technology Management for Emerging Technologies (PICMET), pp.2758–2770, Vancouver, Canada.

Bieber, G., Kirste, T. and Urban, B. (2012) Ambient interaction by smart watches. ACM Proceedings of the 5th International Conference on Pervasive Technologies Related to Assistive Environments, p.39.

Chai, P.R., Wu, R.Y., Ranney, M.L., Porter, P.S., Babu, K.M. and Boyer, E.W. (2014) ‘The virtual toxicology service: wearable head-mounted devices for medical toxicology’, Journal of Medical Toxicology, Vol. 10, No. 4, pp.382–387.

Chan, J.K.C. and Leung, L. (2005) ‘Lifestyles, reliance on traditional news media and online news adoption’, New Media & Society, Vol. 7, No. 3, pp.357–382.

Chang, H.H. (2008) ‘Intelligent agent's technology characteristics applied to online auctions’ task: a combined model of TTF and TAM’, Technovation, Vol. 28, No. 9, pp.564–577.

Chung, S., Lee, K.Y. and Choi, J. (2015) ‘Exploring digital creativity in the workspace: the role of enterprise mobile applications on perceived job performance and creativity’, Computers in Human Behavior, Vol. 42, pp.93–109.

Davis, F.D., Jr. (1986) A Technology Acceptance Model for Empirically Testing New End-User Information Systems: Theory and Results, Doctoral dissertation, Massachusetts Institute of Technology.

Dinev, T. and Hart, P. (2004) ‘Internet privacy concerns and their antecedents-measurement validity and a regression model’, Behaviour & Information Technology, Vol. 23, No. 6, pp.413–422.

Dishaw, M.T. and Strong, D.M. (1999) ‘Extending the technology acceptance model with task–technology fit constructs’, Information & Management, Vol. 36, No. 1, pp.9–21.

Fishbein, M. and Ajzen, I. (1977) Belief, Attitude, Intention, and Behavior: An Introduction to Theory and Research, Addison-Wesley, Reading, MA.

Flextronics’s Technical Report. (2014) ‘Wearable technology, fashioning the future’, http://www.flextronics.com/Documents/White%20papers/WT%20Fashioning%20the%20Future%20-%20FINAL_AE.PDF


Gemperle, F., Kasabach, C., Stivoric, J., Bauer, M. and Martin, R. (1998). Design for wearability. Digest of Papers. IEEE Second International Symposium on Wearable Computers, pp.116–122.

Goodhue, D.L. and Thompson, R.L. (1995) ‘Task-technology fit and individual performance’, MIS Quarterly, pp.213–236.

Lai, I.K.W. and Shi, G. (2015) ‘The impact of privacy concerns on the intention for continued use of an integrated mobile instant messaging and social network platform’, International Journal of Mobile Communications, Vol. 13, No. 6, pp.641–669.

Li, S.C.S. (2013) ‘Lifestyle orientations and the adoption of Internet-related technologies in Taiwan’, Telecommunications Policy, Vol. 37, No. 8, pp.639–650.

Lin, H.C. (2014) ‘An investigation of the effects of cultural differences on physicians’ perceptions of information technology acceptance as they relate to knowledge management systems’, Computers in Human Behavior, Vol. 38, pp.368–380.

Marculescu, D., Marculescu, R., Zamora, N.H., Marbell, P.S., Khosla, P.K., Park, S., Jayaraman, S., Jung, S., Lauterbach, C., Weber, W., Kirstein, T., Cottet, D., Grzyb, J., Tröster, G., Jones, M., Martin, T. and Nakad, Z. (2003) ‘Electronic textiles: a platform for pervasive computing’, Proceedings of the IEEE, Vol. 91, No. 12, pp.1995–2018.

Motti, V.G. and Caine, K. (2014). ‘Human factors considerations in the design of wearable devices’, Proceedings of the Human Factors and Ergonomics Society Annual Meeting, Vol. 58, No. 1, pp.1820–1824.

Narayanaswami, C., Raghunath, M., Kamijoh, N. and Inoue, T. (2001) What Would You Do with a Hundred Mips on Your Wrist?’, IBM Research Division, Thomas J. Watson Research Center.

Pascoe, J. and Thomson, K. (2007). ‘On the use of mobile tools in everyday life’, ACM Proceedings of the 19th Australasian conference on Computer-Human Interaction: Entertaining User Interfaces, pp.39–47, Adelaide, Australia.

Peter, J.P. and Olson, J.C. (1994) Understanding Consumer Behavior, Burr Ridge, IL: Irwin. Pitzer, J.W., Garcha, K., Buss, C., Ju, S. and Carver, R. (2013) The Next Big Thing-Wearables Are

in Fashion. Americas. Polančič, G., Heričko, M. and Pavlič, L. (2011) ‘Developers’ perceptions of object-oriented

frameworks: an investigation into the impact of technological and individual characteristics’, Computers in Human Behavior, Vol. 27, No. 2, pp.730–740.

Profita, H.P. (2011) ‘Social acceptability of wearable technology use in public: an exploration of the societal perceptions of a gesture-based mobile textile interface’, Georgia Institute of Technology, http://hdl.handle.net/1853/44785

Seymour, S. (2008) Fashionable Technology: The Intersection of Design, Fashion, Science, and Technology, Springer Publishing Company, Incorporated, New York.

Shin, D.H. (2010) ‘The effects of trust, security and privacy in social networking: A security-based approach to understand the pattern of adoption’, Interacting with Computers, Vol. 22, No. 5, pp.428–438.

Silina, Y. and Haddadi, H. (2015) ‘New directions in jewelry: a close look at emerging trends & developments in jewelry-like wearable devices’, Proceedings of the 2015 ACM International Symposium on Wearable Computers, pp.49–56.

Tan, G.W.H., Siah, M.W., Ooi, K.B., Hew, T.S. and Chong, A.Y.L. (2015) ‘The adoption of PDA for future healthcare system: an emerging market perspective’, International Journal of Mobile Communications, Vol. 13, No. 1, pp.1–28.

Tarhini, A., Hone, K. and Liu, X. (2014) ‘The effects of individual differences on e-learning users’ behaviour in developing countries: a structural equation model’, Computers in Human Behavior, Vol. 41, pp.153–163.

Venkatesh, V. and Davis, F.D. (2000) ‘A theoretical extension of the technology acceptance model: four longitudinal field studies’, Management Science, Vol. 46, No. 2, pp.186–204.


Wang, E.S.T. and Chou, N.P.Y. (2016) ‘Examining social influence factors affecting consumer continuous usage intention for mobile social networking applications’, International Journal of Mobile Communications, Vol. 14, No. 1, pp.43–55.

Yen, D.C., Wu, C.S., Cheng, F.F. and Huang, Y.W. (2010) ‘Determinants of users’ intention to adopt wireless technology: an empirical study by integrating TTF with TAM’, Computers in Human Behavior, Vol. 26, No. 5, pp.906–915.

Zhou, T. (2011) ‘The impact of privacy concern on user adoption of location-based services’, Industrial Management & Data Systems, Vol. 111, No. 2, pp.212–226.