IEEE Technology and Society Magazine - Fall 2013 - 57

world of ubiquitous computing, full of networked
mobile and embedded technologies, is
approaching. The
benefits of this technology are
numerous, and act as the major
driving force behind its development. These benefits are brought
about, in part, by ubiquitous
monitoring (UM): the continuous and wide spread collection of
significant amounts of data about
users [1].
While UM is a source of many
of the benefits of the technology, it may also play a role in its
failure. To appreciate the potential negative aspects of this new
type of monitoring, we must look
at existing monitoring technologies and methods such as closed
circuit television cameras, human
observation, health-based systems in hospitals, and automated
city travel passes. These systems
share a common social implication
where they can often be found to
undesirably influence the behavior of those using them (i.e., being
observed) [2]-[5]. In addition these
systems cause feelings of stress and
distrust [2].
In light of this, we anticipate
that behavioral changes and other
undesirable effects are likely to
continue to occur in UM systems.
furthermore, given the reduction
in physical constraints (e.g., walls,
floors, and distance) that hinder
existing technologies, the effects of
the monitoring are likely to become
amplified. The problem this creates for the field of ubiquitous computing is that if the system used
to collect naturalistic data about
users is influencing their behaviors
and cognitive state, then the data
collected is unlikely to be accurate.
As a result, the system may provide
a sub-optimal response or, at worst,
suffer a failure.
In response to this issue, a number of predictive models have been
developed with the aim of minimizing or preventing undesirable

behaviors prior to a UM systems
development and implementation.
Such models would allow designers
to explore the potential behavioral
impacts of their system designs
without costly developments and
negative social reactions, which
could resolve many social issues
before they occur. The Perceptions
of System Attributes-Behavioral
Intention (PSA-BI) model [1],
is one such model, specifically
designed to predict user behaviors
in ubiquitously monitored environments. Although parts of the model
have been empirically tested [1],

the problem space, they fail to consider many of the different aspects
of large-scale monitoring systems. The Perceptions of System
Attributes - Behavioral Intention
(PSA-BI) (see fig. 1) model was
developed to address these issues
[1], through clear identification of
specific behavioral influencing
system characteristics.
Like the above models, the PSABI model is grounded in aspects of
sociological theories such as the
Theory of Planned Behavior (TPB)
[9] and Technology Acceptance
Model (TAM) [10]. The model was

Ubiquitous monitoring systems can
cause feelings of stress and distrust.
one key aspect that remains unexplored is the impact of culture on
user salient perceptions, attitudes,
and intentions. Thus far, studies
involving the model have been
limited to U.K. universities and
workplaces.
Hence, the focus of this article
is on exploring specific components of the PSA-BI model through
a quantitative study of office workers in Kyoto University, Japan. An
established set of questionnaire
measures was used to capture the
workers salient perceptions, attitudes, and intentions toward wearing a monitoring device. Using
structural equation modeling [6],
the relationships in the model were
tested, and a series of regression
coefficients observed that reaffirm
the relationships in the model.

Existing Frameworks
There are number of different
existing frameworks and predictive models in the literature such
as the ubiquitous computing acceptance model [7], and the awareness monitoring model [8]. These
highly relevant frameworks propose to help guide future research
or model the undesirable effects of
pervasive computer systems. While
both are useful for conceptualizing

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

fALL 2013

developed with the intention of predicting users responses to monitoring systems based on the systems
characteristics.
Exploring the model from left to
right, the technology and application spaces contain the design characteristics of a monitoring system
which influence users. These are
external variables in the sense that
they are directly controllable by the
designers and engineers who build
the monitoring systems.
The technology space includes
elements such as device obtrusion,
device control, device coverage,
and number of devices [11], focusing on the physical behavior influencing characteristics of a system.
In contrast, the application space
focuses on the use of the technology and includes characteristics
such as: frequency of data collection, data integration, user knowledge, application control, data
access, and data sharing [11].
Object based beliefs are the
beliefs/perceptions a user has about
a system, and its specific characteristics. Both application and technology perceptions directly correspond
to the equivalent characteristics.
for example, a designed open level
of access to the data collected will
influence user perceptions of the
|

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - Fall 2013