Wearable device adoption model with TAM and TTF

Abstract

Wearable devices have received great highlights as next core products for global information technology companies. Since the wearabledevice 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.

Full text

518
I
nt. 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: hschang@samsung.com
Seul Chan Lee* and Yong Gu Ji
Department of Information and Industrial Engineering,
Yonsei University,
262 Seongsanno, Seodaemun-gu,
Seoul 120-749, Korea
Email: seulchan@yonsei.ac.kr
Email: yongguji@yonsei.ac.kr
*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

Wearable device adoption model 521
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.

522
H
.S. Chang et al.
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.

Wearable device adoption model 523
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

524
H
.S. Chang et al.
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:

Wearable device adoption model 525
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.

526
H
.S. Chang et al.
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

Wearable device adoption model 527
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.

528
H
.S. Chang et al.
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.

Wearable device adoption model 529
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

530
H
.S. Chang et al.
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

Wearable device adoption model 531
Table 4 Correlation matrix and discriminant validity

532
H
.S. Chang et al.
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

Wearable device adoption model 533
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

534
H
.S. Chang et al.
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.

Wearable device adoption model 535
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%20Futu
re%20-%20FINAL_AE.PDF

536
H
.S. Chang et al.
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.

Wearable device adoption model 537
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.