IEEE Technology and Society Magazine - Fall 2014 - 77

with anecdotal evidence provided by
the main resident in each dwelling.
The anecdotal evidence suggests a
variety of reasons for the increase
in consumption observed. Some of
the reasons stated included: change
in personal circumstances such as
new/additional people in a property
(a new partner, returning children),
child-minding activity, increase in
both the type of consumption and
time of use of consumption, such as
people working in shifts, children
using laptops/gaming consoles,
excessive use of electronic/household goods, ill-health resulting in
people staying more at home, and
using more energy than what they
would use otherwise.
We also found that displaying
community feedback did not result
in a decrease in overall consumption. This could be attributed to
two factors. New technology often
excites people. However, subsequent to adoption, their interest
dwindles over a period of time,
leading them back to their old habits. This is known as the "novelty
effect." Thus, any efforts made in
using less energy, after the introduction of the tablet intervention, might
not have been sustained in the long
term. Secondly, an increase in consumption in most dwellings subsequent to the community feedback
intervention could also point to the
existence of a seasonal component.
In this case, the community feedback intervention was conducted in
October, which marks the onset of
winter. Thus the average, before the
intervention, corresponds to the end
of summer months, and the average
after corresponds to the beginning
of winter months. This seasonal
increase could be attributed to residents resorting to the use of portable
heating devices.
In observing the average monthly
off-peak in the electricity consumption of apartments, i.e, energy used
to power storage heaters, most
dwellings exhibited a significant
increase in consumption during the
winter months. Thus, during the

winter months, it can be generalized that households are less likely
to reduce energy consumption. A
closer look at the monthly consumption of these dwellings reveals that,
in some dwellings, the consumption in summer months was significantly lower in comparison to that
in winter months.
Fig. 5 shows a plot of the percentage savings of the non-heating
electricity throughout the whole
project, including both the terraced
houses and the apartments. The
two periods of comparison are the
"measured baseline" and the "tablet
intervention," and thus aim to show
the influence of the tablet interface
of the users consumption. Each line
represents one household, and they
are sorted from negative reduction, the participants' consumption

the variation in the use of the tablet
interface over time, and second the
effect of community workshops on
the tablet interface use. To enable
monitoring of the above parameters, the application software on
the tablet was designed to record
the participant's activity usage of
the interface. One recorded when
buttons were pressed on the main
part of the interface concerned
with monitoring energy consumption and captured this information
in the database. The other recorded
which web application buttons
were pressed on the right hand side
of the interface concerned with the
later enhancements (e.g., new web
apps introduced) to the interface.
Fig. 6 shows how interface usage
for participants, in each of the
dwellings considered in this study,

New technology generally captures
the attention of users, but their
interest gradually fades over time.
increased across the intervention,
to positive reduction. The red lines
indicate that the comparison test
between the baseline and intervention samples was statistically significant at the five percent level. The
black lines indicate a lower level of
confidence, and the grey lines are
ones for which it was not possible to
calculate the confidence correctly,
due to insufficient samples. The
majority of households saved nonheating electricity, and the median
falls around the 6% saving level.
There is a good subset of households
(between the 60% and 80% levels in
Fig. 5) that show both statistically
significant and economically significant reduction at the 10-20% level.
Level of Participant Engagement
In the context of this study, it is
of interest to identify the level of
participant engagement that was
achieved. To address this aspect of
the study, we examined two main
perspectives. First, we examined

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

FALL 2014

varied over time. Interface activity
is measured as the number of times
any relevant web page relating to
energy feedback was clicked on per
day. This calculated a total number
of interface hits per day for each
house, which provided the total
number of hits for all households
per day. The total hits were calculated for a 3-month and 6-month
period after interface activation. A
clear decaying effect can be seen
in the graph indicating a reduction
in initial enthusiasm for the interface. Interestingly, after about 60
days, the interface activity seems
to stabilize around the 100 hits per
day. This finding is encouraging as
it suggests that there is a non-zero
level of activity that is present in
the trial. We also found that the
terraced houses had a flatter decay
rate while the apartments showed
a far steeper decay rate in terms of
the interface use. Determining causality is very difficult in this case as
it could be attributed to a variety of
|

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - Fall 2014