Building and Environment 125 (2017) 439e450

Contents lists available at ScienceDirect

Building and Environment

journal homepage: www.elsevier.com/locate/buildenv

How smart do smart meters need to be?

lez c, d, JeeHang Lee a,
Nataliya Mogles a, Ian Walker b, Alfonso P. Ramallo-Gonza
Sukumar Natarajan , Julian Padget , Elizabeth Gabe-Thomas b, Tom Lovett a, Gang Ren a,
d a

Sylwia Hyniewska a, Eamonn O'Neill a, Rachid Hourizi a, David Coley d, *

a
Department of Computer Science, University of Bath, UK
b
Department of Psychology, University of Bath, UK
c
Faculty of Computer Science, Universidad de Murcia, Spain
d
Department of Architecture and Civil Engineering, University of Bath, UK

a r t i c l e i n f o a b s t r a c t

Article history: Governments across the world are investing in smart metering devices that report energy use to the user
Received 8 July 2017 with the aim of reducing consumption. However, the effectiveness of such In-Home Displays (IHDs) has
Received in revised form been questioned, since savings are small. This is possibly because informing the consumer of their
30 August 2017
consumption in kWh, or monetary units, fails to give context, or inform of possible actions to reduce
Accepted 6 September 2017
Available online 7 September 2017
consumption. We investigate, for the first time, the effect of replacing the simple statement of energy use
an IHD gives, with a detailed array of information specifically designed to improve consumer energy
literacy and suggest behaviour change through personalised actionable messages set against a series of
Keywords:
Energy feedback
psychological value systems for context, and which can identify potential profligacy. The results from a
In-Home-Displays (IHDs) carefully controlled field experiment show: 1) value framing and action prompts have a significant effect
Internal values on occupants' behaviour, with the mean temperature of homes being reduced from 22.4
C to 21.7
C,
Action prompts and a marked reduction in gas consumptiond22.0% overall and 27.2% in high consumers; 2) energy
Energy literacy literacy increasing from 0.52 to 1.28 (on a 0e4 scale); 3) it is possible to target potentially profligate
Smart meter households, without inappropriately messaging others; 4) engagement is high, with 82% of the partic-
ipants finding the system useful. These results emphasize the necessity of improving energy literacy
when encouraging energy efficient behaviours and point to a new generation of smart meters with the
potential to increase energy literacy, make much greater savings and impact climate change policy.
© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/).

1. Introduction Residential energy consumption is however a multidimensional

phenomenon embedded within a socio-cultural and infrastructure
The residential energy sector accounts for 23% of total energy context, and for this reason changing occupant behaviour might be
consumption worldwide, placing it third after industry at 37% and expected to be complex. The focus of the research reported here lies
transportation at 28% [1]. In developed countries the sector is even in inducing behavioural change with the help of energy demand
more important; in the US, for example, residential consumption feedback via smart meters or ambient displays. Opinions regarding
represents 25% of total energy use [2] and in the UK 29% [3]. This the effectiveness of such solutions to date are unfortunately not
translates to roughly 12% of UK greenhouse gas emissions [4]. In unanimous [6e12].
addition, it is projected that residential consumption worldwide Approaches to energy feedback have so far tended to be one-
will grow by an average of 1.4% a year from 2012 to 2040 [5]. Thus size-fits-all solutions; however, with new developments in en-
the sector is critical to both national energy policy and international ergy data management [13], more advanced feedback is now
climate change policy, and many attempts are being made to reduce possible. The main contribution of the current paper to digital en-
residential consumption by influencing the behaviour of occupants. ergy feedback research is twofold: i) in broad terms, it tests the
effectiveness of a novel smarter, building-aware, and more user-
personalised digital energy feedback in an experimental setting
for the first time; ii) in more specific terms, it evaluates the effect of
* Corresponding author. two new approaches to feedback e internal values and tailored
E-mail address: [email protected] (D. Coley).

http://dx.doi.org/10.1016/j.buildenv.2017.09.008
0360-1323/© 2017 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
440 N. Mogles et al. / Building and Environment 125 (2017) 439e450

action prompts, delivered via a computer tablet. It replaces the seems clear that users need something more to motivate and
normal user- and building-blind smart meter concept with one that engage them than plain energy feedback in kWh or cost if we are to
not only reports the energy used but recommends specific actions get energy reductions of 10% or more. Some research indicates that
and works with the personal values of the user. For example, it intelligent energy feedback that offers different feedback options
might recommend turning the heating thermostat down one might be effective [25].
degree-centigrade, and explain this not just in terms of kWh or Chiang et al. [26] have calculated that smart meters can pay back
financial savings, but with respect to environmental gain or other their installation costs in 4 years if the energy savings are 3% or
personal values. The system tested here is a simplified version of a more, although the observation that the utility company no longer
future building- and user-conscious approach that would report need to visit the home to read the meter might currently be the
complex and tailored information to the user in written or spoken main economic driving force for their installation. It is hence
sentences. For example, “I note that the heating turns off at 9am each interesting to ask whether we might move the focus from a device
weekday, yet your home appears to be unoccupied from 7am; would that is mainly of use to the energy company to one that has equal
you like me to change the heating timing? This might save you £47 per utility to the occupant.
year and the stop the emission of 218 kg of carbon dioxide”. Or, “I note The question of what best motivates householders has received
that your home might be over-ventilated when the heating is on, losing considerable attention from energy behaviour researchers, with
£98 of heat per year. You might like to keep most of your windows varying success being noted. For example, attempts to stimulate
closed when the heating is on”. Such an approach, which we term savings through social comparison and competition have not ach-
intelligent smart metering (or ISM), requires knowledge of which ieved notable success [25,27e29]. One possible reason for this is
values are most likely to prompt the user to act. These might be the so-called boomerang effect [30], whereby when households are
financial savings, the reduction of environmental damage, benefits told they are using less energy than average they start to use more
to future generations, wastefulness, or some other concern specific when they see what is ‘normal’ or ‘permissible’. This suggests that if
to the individual. messages are to be sent to households, only those where there is
ISM requires an energy (thermal) model of the building in order some evidence of potential energy profligacy should be targeted
to predict accurate savings based on the building and its use, rather with certain messages, and this is a key aspect of our study, and the
than on inaccurate typical values. To be cost effective, such a model one reason for its success.
would need to be automatically assembled from a minimum of The other key aspect is people's values. Different disciplines,
sensor information so as not to overly increase the cost of the smart including economics, psychology, philosophy, sociology and an-
meter, for example a mix of utility meter data, room temperature, thropology, have all tried to understand the role of personal values
sub-circuit or high-frequency electricity data to infer occupation in conservation behaviour. From an economic point of view, peo-
and maybe home CO2 concentration (as a proxy for ventilation ple's values correspond to long-term preferences, and can be
rate). The sensors needed and the accuracy of such an approach, explained by decision theory [31]. In anthropology, values are
which uses inverse modelling to obtain an accurate thermal model ‘cultural worldviews’ and their role in pro-environmental behav-
from a time series of data, have been reported by the authors in iour is studied within climate change risk perception [32]. In social
Refs. [14] and [15]. By creating such a model, the financial, and psychology, Schwartz [33] identified a number of personal values
other, impact of any suggestions, for example turning down the universal for all cultures and nations. From this perspective, per-
thermostat by one degree, can be calculated for the specific home sonal values are defined as superordinate goals that serve as
and reported to the user. In addition, inappropriate suggestions, for enduring guiding principles in peoples' lives [33]. Common to all
example suggesting a reduction in heated temperature in an these approaches is the idea that values are conceptually different
already under-heated home, can be avoided. from goals, opinions and attitudes: values reflect broad long-term
One can imagine other advantages of the ISM approach, for preferences and so provide unity across a broad range of behav-
example, the reporting of the presence of high U-Values in the iours (a person who above all values their own self-interest will be
fabric to utility companies for targeted intervention. self-interested in most settings); goals, opinions and attitudes, on
This paper attempts to discover if this new approach can be the other hand, are much more situation-specific and changeable
applied to a group of homes and if it generates changes in over time [34]. In the current study, we test personal values as
behaviour. motivators to energy-saving behaviour and focus on altruistic,
egoistic and biospheric values, representing, respectively, concern
2. Background for others, the self and the natural environment. These values were
chosen from the various Schwartz's personal value sets based on
Some research findings [16] suggest that continuous energy research by DeGroot and Steg [35,36], who identified altruistic,
feedback might be an effective driver of energy-related behaviour egoistic, biospheric and hedonic (pleasure-seeking) values as the
change. For example, Barbu et al. [17] suggest that energy feedback key value orientations highly correlated with pro-environmental
provided to users via smart meters could save 5e15% of total energy behaviour. In this study, hedonic values were dropped from this
costs. Similarly [18e20], suggest that energy feedback through set after pilot testing (with an opportunistic sample of 30 UK
advanced in-home displays (IHDs) could help to save up to 20% of adults) found much lower engagement with such messages
energy costs, either for electricity or total energy bills. However, compared to the other three value orientations (which pilot par-
reality seems to fall short of such predictions. For example, in ticipants found useful, and more engaging than energy messages
Ref. [21] a more modest average energy saving of 7% is reported expressed in standard kWh).
across multiple utility pilot programs aimed at electrical energy The final facet of this study concerned the nature of the action
conservation with the help of IHDs. prompts. The literature hints that energy feedback may be effective
Current technological solutions for real time energy feedback when it prompts concrete personalised actions (for example,
suffer from multiple issues [22,23], for example: unengaged users; “Please switch off unused appliances”, “Adjust your thermostat
failure to address users' personal motivations and needs embedded setting”), presumably because these do not require knowledge or
in daily routines and social practices; information comprehension problem-solving from the householder, as would a more general
issues caused by abstract numerical information in kWh or finan- action prompt like “Reduce your energy consumption”. However, at
cial costs; and inattention to users' personal characteristics [24]. It the moment we lack the empirical quantitative data needed to
 N. Mogles et al. / Building and Environment 125 (2017) 439e450 441

substantiate these claims [24]. Research on the effect of general 1. What is the overall effect of digital energy feedback
action prompts has shown that it has limited influence on behav- interventions?
iour, but with increased possibilities for specificity and context 2. What is the additional benefit of internal values embedded in
detection offered by modern technologies it might substantially energy feedback messages?
improve the effect of feedback [37]. Some findings also indicate that 3. What is the additional effect of personalised messages with
for the energy information to be useful, it should be tailored to action prompts?
personal contexts [38]. Thus, specific action prompts, tailored to 4. What effect has a digital energy feedback on cognitive variables,
specific households, might have a more profound effect on energy such as energy literacy and users' experience with a digital
related behaviour, and this idea is tested here. feedback?
Providing information on energy related actions and personal
context is directly related to the educational component of energy To answer these questions, we formulated the following
feedback and the concept of energy literacydwhich implies an in- hypotheses:
depth understanding of energy consumption [39] and is
Hypothesis 1. Digital feedback in general will have an effect on
frequently mentioned in pedagogical, educational and environ-
energy related behaviour compared to a baseline period before the
mental policy literature. Energy literacy has been called “a broad
interventions.
term encompassing content knowledge as well as a citizenship
understanding of energy that includes affective and behavioural Hypothesis 2. Energy feedback translated according to internal
aspects” [40], although arguably such a definition conflates values will have a significant effect on energy related behaviour.
knowledge about where energy is consumed with the motivation
Hypothesis 3. Personalised messages with action prompts will
to reduce this. Cognitive (knowledge, in-depth understanding),
have a significant effect on energy related behaviour.
affective (attitudes, values), and behavioural (social practices)
modules can be distinguished within a complex concept of energy Hypothesis 4. No matter which behavioural effect of a digital
literacy. It is known that promoting energy literacy can foster a shift energy feedback is observed, the digital interventions will influence
in knowledge and perception of energy and thus will facilitate energy related cognitive variables, such as energy literacy.
responsible energy related choices and behaviour [40e42].
To the authors' knowledge, there are no studies or commercially
available energy feedback systems that use action prompts tailored 4. Method
to users' behaviour in a systematic manner. The only work that
provides some tailored heating action prompts is our previous To answer the research questions defined in Section 3, we
study [43], however we did not aim to test different energy feed- designed an energy feedback system with different types of feed-
back approaches in an experimental setting. In the work of D'Oca back, we deployed the system in homes in the UK and performed a
et al. [44] tailored to users' contexts, newsletters with action sug- field experiment. The experiment had a 2 (Value framed vs. Non-
gestions were sent via email to participants, though the study only value framed) by 2 (No action prompt vs. Tailored message with
addressed electricity consumption, and action prompts were not action prompt) within-subject factorial design resulting in four
part of an energy feedback display. One work on energy con- energy feedback experimental conditions (see Table 1).
sumption feedback in a non-residential sector reports the effec-
tiveness of user-tailored context advice and action prompts [45], 4.1. Participants and procedure
though this study again was focused only on electricity consump-
tion and the tailored advice was sent by email not a smart meter. Our participants were residents of social housing recruited in
There are a lot of commercially available devices and applica- the first half of 2013 with the help of Exeter City Council. Initially,
tions that provide people with simply energy feedback and give an information package was sent to the households and later re-
general, not occupant- or building-specific tips and suggestions cruiters went door-to-door to identify interested households. After
regarding energy saving issues. In other words, the existing IHDs a consent form was signed by the occupants, the sensors were
are not smart enough to address building context or personal installed during JanuaryeMay 2014 in 73 homes. The first phase of
motivations. the project established the baseline: sensor data were collected but
no energy feedback was provided to the participants.
For the second phase of the project, an energy feedback exper-
iment was conducted over three months (January 2016eMarch
3. Pilot study and research hypotheses 2016) with the help of a new energy feedback application (iBert)
written by the authors and presented to participants via a tablet
Very few studies test the effect of internal personal values computer. The current paper describes the analysis of this second
embedded in a digital energy feedback within field experiments. phase of the project. Due to the length of the study, and the social
One exception being the previously mentioned work of Schultz group involved, some households dropped out of the study, moved
et al. [29] which compared feedback based on normative compar- houses or changed contact details, which resulted in 43 homes
ison and financial costs, against a standard feedback showing participating for the full study period.
consumption alone. We hypothesised that couching energy feed- In this sample, each household had on average 2.8 residents.
back messages in terms of several of the personal values cited in the Each home had a main contact, and the mean age of these was 50.6
environmental psychology literature would substantially improve years (with a minimum age of 27 and a maximum of 81), 14 were
users' engagement and increase their motivation to save energy male and 29 female. Seventy-five percent of households reported
since multiple values guide people's lives [34] and focussing only an income lower than the UK national median of £25,600 (identi-
on one value might be not effective to increase motivations. fied at a national UK level for 2015, with the London area excluded
In addition to this, we have included the facet of personalised [46]). Around 25% of households reported difficulties with paying
feedback i.e. messages that refer, and are valid only for, specific energy bills. Fourteen percent of households already had a basic
households. smart meter (could correspond to condition 1 in our study).
The main research questions that we wanted to answer were: The focus of the study was in part the response of occupants to
442 N. Mogles et al. / Building and Environment 125 (2017) 439e450

Table 1
Energy feedback study design (examples of information presented within each condition in italics e other non-monetary values, for example environmental impact, were also
used.

No Value Value:

No Action C1: Standard display C2: Standard display þ values

No (kWh and financial costs) Your home has used 95 kWh this week; this is equivalent to £12.
personalised Your home has used 95 kWh this week.
Action C3: Standard display þ C4: Standard display þ
Personalised þ tailored action prompts þ tailored action prompts þ values
If you reduced the thermostat temperature in your house one degree If you reduced the thermostat temperature in your house one degree you would save
you would save 11 kWh. 11 kWh; this is equivalent to £1.43.

feedback from personalised messages. For this reason, individual change in the future, and report back to the occupant via the same
basic thermal models of each of the homes were created. It had connection to the Wi-Fi enabled tablet in the home as a weekly
already been demonstrated by the authors that the sensors used digest. The frequency with which householders interacted with the
can auto-generate a suitable dynamic thermal model of each iBert app was recorded directly from each tablet and as well as a
building based only on the data received from the sensors (without record of the tailored textual messages sent to each house.
the need for information on the size of the homes or their con- The households were cycled through the four conditions shown
structions). This model is a lumped parameter model and the work in Table 1 with each condition lasting three weeks. Households
has been reported in another paper [15] and shown to produce were randomly assigned to one of the four conditions according to
accurate results. However, to make sure that no extra noise was the Latin-Square counterbalancing design of Table 2. For example,
added to the experiment and ensure the results reported were not during the first three weeks, one-quarter of the homes were under
affected by an auto-generated model which might well change condition C1 of Table 1 and in the following three weeks they had
during the experimental cycle, simple (steady-state) models were condition C2 applied to them.
created by using floor plans and information about the materials In conditions C3 and C4, tailored energy feedback messages are
and constructions from the city council and visual inspection. These generated once per week, based on energy-related information
models consisted of calculating the mean heat loss coefficient of aggregated over the previous week, and were displayed during the
each building. In future developments auto-generating models like subsequent week. If the algorithm for identifying energy-wasting
those of [15] could easily be used. events and variables did not detect any potentially profligate en-
The simple model obtained from the building geometry was ergy use in a particular household, that household did not receive
used by iBert to investigate the energy savings of various strategies any tailored messages that week.
(turning the thermostat down, closing windows, turning the This results is sixteen textual messages in total: four for personal
heating off when the building was not occupied, turning the lights value neutral condition C3 and twelve for value-framed condition
and some other electrical items off when the building was not C4 (egoistic, altruistic and biospheric values). A household could
occupied) that could be applied to each house if the household receive up to four different messages since a value message is
breached set action levels of: room temperature, ventilation rate, framed according to one of the three personal values and a value is
having the heating on when building was unoccupied; and high selected at random. Messages in conditions C3 and C4 are sent only
electricity use when the building was unoccupied. Depending on if there are energy-wasting events identified by the intelligent
the condition applied to the home at the time (see Tables 1 and 2), algorithms.
this gave iBert the ability to inform households of: their current Participants received an incentive of £150 in the form of su-
energy consumption; the total consumption over the last week; permarket vouchers, sent by post in three instalments of £50 (at the
their consumption expressed in accordance to other values (such as beginning of the project, at the end of year one and at the end of
an environmental cost); the savings that might occur if they took year two after the completion of the project). In addition, they were
action; and, most importantly, the specific actions they might take. allowed to keep the tablet once the study has ended. All partici-
Due to page limitations, full details of the electronics, sensors pants who completed a post-study phone survey received an
and coding that lie behind iBert will be described in a later paper. additional incentive with a chance to win a £20 supermarket
But in outline, the system consisted of sensors that measured real- voucher during a prize draw. After the end of the experiment the
time utility data, air temperature (in up to three rooms), radiator energy feedback app was deactivated and sensor equipment was
temperature (to sense if heating was on), humidity (in up to three collected from the participating homes four months after the end of
rooms), light, motion and CO2 levels, with this data collected the energy feedback interventions.
approximately every 5 min via Raspberry Pi and transmitted via the Prior to both phases of the study participants were given paper
home's broadband or 3G network to a secure cloud store. These information sheets, consent forms and surveys that contained
data were then used to see if action levels had been triggered, among others demographics and the first seven items from the
calculate any savings that might be made if behaviour were to energy literacy survey described in Ref. [40]. This is an updated

Table 2
Latin square counterbalancing design for the four experimental conditions: C1 to C4 across the four temporal phases of the experiment.
 N. Mogles et al. / Building and Environment 125 (2017) 439e450 443

version of a validated questionnaire and energy literacy assessment box with action prompts in text form about any potentially energy
framework developed in the by DeWaters and Powers [40]. To the profligate behaviour (room temperatures > 21
C, occupied CO2
authors' knowledge, it is the only validated questionnaire that levels of <800 ppm (possibly indicating excessive ventilation),
evaluates energy literacy according to a broader definition. The mean unoccupied electricity consumption in excess of 600 W, or
seven items used from this questionnaire addressed only the gen- heating on when the house was unoccupied) during the previous
eral knowledge of energy consumption. (The full energy literacy week. Action prompts described concrete actions with low per-
and user experience survey used for this work can be found in sonal costs, e.g. lowering a thermostat setting or leaving windows
Appendix A). open less frequently. A participant could receive from zero to a
As previously mentioned the interactions with iBert were also maximum of four messages. Messages consisted of four parts: (1)
monitoreddFig. 1. The reader can see that almost all houses had communication of factual information regarding energy related
substantial level of interactions with the tablet. Each interaction variables [for example, we have noticed that your thermostat was set
with the app consisted of a minimum of three recorded app related to 23
C]; (2) evaluation of the factual information [e.g. this value is
events, such as ‘app started’, ‘app resumed’, ‘app closed’ etc. unnecessarily high]; (3) a motivational component which expressed
wasted energy information aggregated over a longer time period
(over a whole winter, over a year) [e.g. in your house this may imply
4.2. The four experimental conditions
520 kWh use more per winter]; and (4) an advice part which con-
tained an action prompt [e.g. we recommend that you lower the
C1: Standard energy feedback. For the standard display con-
thermostat to 21 degrees centigrade and that you check if you are
dition (C1: No value e No action; see Table 1), we followed the
comfortable at that temperature].
recommendations of the minimal requirements for an In-Home
The trigger of 21
C was based on national recommendations
Display (IHD), i.e. we replicated a typical smart meter in the UK
[48]. The 800 ppm was chosen by considering EN 15251. This sug-
[47]. This displayed: current gas and electricity consumption; cu-
gests IDA 2 air quality (medium IAQ) is achieved below 1000 ppm.
mulative consumption for the past week; information in kWh; and
In an effort to be precautionary, a lower value of 800 ppm was used
a visual (i.e. non-numerical) presentation alongside a numeric one.
- this was the median value found during the baseline phase. The
C2: Standard feedback with personal values. In a personal
600 W was also derived from data gathered during the baseline
value only condition (C2: Value - No action) the display was the
phase. This showed that in all houses consumption was bi-modal:
same as in C1, but complemented the weekly summary information
low values during unoccupied periods or the night; and high
translated into one of the three internal values: egoistic e expressed
values during occupied periods. The mean figure that separated the
in money; biospheric e expressed in trees destroyed; or altruistic e
two being 600 W.
expressed as the cost to society, couched as the number of minutes
C4: Standard feedback with tailored actions and values. In
with a family doctor that could be provided at the same cost. So, for
experimental condition C4, households received the same infor-
example, a household might have the kWh display replaced with
mation as in C3, but the kWh part of the message was translated
information that their energy use in the last week was equivalent in
into one of the three internal personal values just like in C2. An
carbon terms as the destruction of 1.7 trees or 20 min with a doctor.
example of an energy feedback message in condition C4 where
Egoistic, biospheric or altruistic personal values for each household
energy information is translated into biospheric value concepts
were chosen randomly.
might be: “I think a lot of heat might be escaping from open windows.
C3: Standard feedback with tailored actions. This condition
The escaping heat results in wasted energy. This requires roughly 12
was the same as C1 except households also saw a tailored message

Fig. 1. Interactions heat mapdthe colour indicates the number of interactions. The House ID is the ID of the house within the database. (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this article.)
444 N. Mogles et al. / Building and Environment 125 (2017) 439e450

more trees to compensate for the extra pollution caused by your home did not decrease their internal temperature.
over a winter. This may be because there are too many windows open, The overall effect of iBert on electricity consumption and CO2
they are open too wide or for too long. Try changing how many win- concentration, with frequency of app usage and received textual
dows you open and for how long and you may save energy that over a messages as covariates, would appear not to be statistically
whole winter is equivalent to planting 12 trees. Try it, and check if this significant.
message disappears next week”.
Appendix B shows a screen shot of ibert as seen by a household. 5.2. Analysis 2: Post-intervention effect of messages

Temperature-related action prompts were triggered in homes

5. Results
over 21
C. To assess whether people responded to these, we con-
ducted four follow-up one-sample t-tests to compare internal
In this section, we show the results from the different analyses
temperatures to 21
C, before and after the intervention, for the
that were performed to evaluate the effectiveness of the iBert
group who received temperature-related action prompts and the
system.
group who did not. For the non-message group (the subgroup
whose home temperature was lower than 21
C and who therefore
5.1. Analysis 1: overall effect of iBert were never told to change the temperature even under experi-
mental conditions C3 and C4), internal temperature was signifi-
This analysis tests hypothesis 1, asking if digital feedback in cantly lower than 21
C both before (mean temperature ¼ 18.7
C,
general has an effect on energy-related behaviour compared to a SD ¼ 1.6, t (24) ¼ 6.99, p < 0.001) and after the experiment (mean
baseline period before the interventions. The purpose of this is to temperature ¼ 18.7, SD ¼ 1.9, t (16) ¼ 5.09, p < 0.001) e as would
examine if our findings are in line with those from other EHD trials. be expected. The group that received the temperature message
A repeated-measures ANOVA was used to evaluate the effect of because their baseline temperature was over 21
C (mean
iBert on internal temperature, electricity consumption and venti- temperature ¼ 22.3
C, SD ¼ 1.1, t (14) ¼ 4.38, p < 0.001), however,
lation rate (CO2 levels). had internal temperatures that essentially settled at 21
C after the
We found a significant effect of Time (i.e. study phase) on home messages (mean temperature ¼ 21.1
C, SD ¼ 2.00, t (11) ¼ 0.20,
internal temperature, F (2,58) ¼ 3.78, p ¼ 0.029, h2p ¼ 0.12. Post-hoc p ¼ 0.80), Fig. 3. In summary, people who received temperature-
pairwise comparisons using Holm-Bonferroni p value correction related action prompts saw a reduction in internal temperatures
reveal that home internal temperature before iBert was active is that put their home temperatures almost exactly at the level that
significantly higher than the internal temperature during the iBert was suggested; a similar change was not seen in those who
experimental phase (d ¼ 0.24, p < 0.001) and the difference be- received no prompts, suggesting the effect is likely not the result of
tween the internal temperature during the experiment and after weather-related changes. Note that both groups could still see the
the experiment is not significant either (p ¼ 0.97)dindicating the standard IHD module of the app after the experiment, although the
intervention had lasting impact (Table 3). We understand that tailored messages function was deactivated. No messages were
lowering the temperature can lead to larger issues such as comfort triggered based directly on gas consumption (but rather on internal
problems or even health risks in vulnerable households, it is for this temperature, heating times and excessive ventilation), however the
reason that this analysis is expanded later to check if this temper- use of iBert reduced the mean household gas consumption by
ature reduction was from all households, or only from ones with 14e29% with a mean of 22.0%.
initially high temperatures.
In addition, we conducted two analyses of covariance to see how 5.3. Analysis 3: Effect of values and action prompts
often occupants interacted with iBert and if occupants receiving
tailored textual feedback influenced these results. Frequency of To evaluate our second and third hypotheses, regarding the ef-
interaction with iBert proved not to be statistically significant. fect of internal values framing and tailored action prompts during
However, change in internal temperature was predicted by the energy feedback iBert phase, we conducted a two-way repeated
whether householders had received at least one tailored textual measures ANOVA on home internal temperature, electricity con-
feedback message, F (2,56) ¼ 3.45, p ¼ 0.04, h2p ¼ 0.11. This inter- sumption and CO2 level data with the factors Personal Values
action shows that the difference in internal temperature between (present or absent) and Action Prompts (present or absent). Base-
homes with and without a tailored textual feedback is not visible line measurements of these dependent variables, taken before iBert
during the digital interventions phase, but it is after the in- deployment, were used as covariates. After controlling for the
terventions. In the light of these results, we conducted a follow up baseline temperature, homes in the value framing conditions (C3
t-test on the differences in internal temperature between with- and C4) have a lower internal temperature (19.23
C) compared to
iBert and after-iBert phases for two groups of households: those homes in the conditions without value framing (C1 and C2)
who did not receive any temperature-related tailored textual (19.32
C).
message and those who did. An ANOVA on these internal temperature readings during the
This difference is significant, t (29) ¼ 2.29, p ¼ 0.03, which in- iBert phase, with baseline temperature as a covariate, revealed a
dicates that homes that received tailored textual messages with significant effect for Value Framing, F (1,33) ¼ 5.29, p ¼ 0.028,
action prompts had lowered their internal temperature after the h2p ¼ 0.14. The baseline temperature covariate was also significant, F
digital interventions; while homes which did not receive messages, (1,33) ¼ 4.94, p ¼ 0.033, h2p ¼ 0.13, suggesting the warmer homes at
the start of the study still tended be the warmer homes at the end.
Table 3
As shown later, iBert had a marked impact on energy literacy.
Mean internal temperature (in
C) for households before, during and after the digital
iBert energy feedback interventions.
Hence it is unlikely that action prompts were forgotten between
experimental phases. In light of this, the data were separated into
Temperature (
C) Standard deviation (
C)
two groups, before the issuing of action prompts and post issuing.
Baseline 20.4 2.07 Action prompts were found to have a statistically significant impact
IBert Phase 19.6 2.50 under a pairwise t-test (p ¼ 0.021) on the heating degree day
After 19.4 2.35
adjusted mean gas consumption (mean before action
 N. Mogles et al. / Building and Environment 125 (2017) 439e450 445

prompts ¼ 1239 W, SD ¼ 379; after 1120 W, SD ¼ 348), contributing considered as potentially over ventilated, and therefore it was
9.6% points to the overall reduction in gas consumption. assumed that some energy savings could be achieved by reducing
In contrast to internal temperature and gas consumption, there ventilation.
were no effects of Actions or Values on electricity consumption and To evaluate this, we again separated the homes into two groups:
CO2 concentration. the ones that showed high levels of ventilation in the baseline
(CO2 < 800 ppm) and those that did not (Fig. 3). In this case, all the
5.4. Analysis 4: Case studies of subgroups with high and low energy tests show that the CO2 concentrations were not changed in the
consumptions intervention; neither in the homes that showed high levels of CO2
nor in those that showed low levels. This is possibly because
To get more insight into the changes in internal temperature, gas changing the ventilation regime in a house is not a trivial task. Also,
and electricity consumption, and CO2 concentration, across all the it should be noted that CO2 is not only an indicator of ventilation,
stages of the study, we examined the box plots of the data for two but also an indicator of infiltration occurring through cracks and
subgroups: one being the households that had high energy use and other defects, and that is not under the control of the occupants.
the other those homes with low energy use. With respect to electricity, it also seems that the intervention
Fig. 2 shows that the median temperature of the houses that had failed to change the behavioural habits of the occupants, with no
a high thermostat temperature (upper graph) went down during statistically significant change seen in the behaviour of big con-
the iBert interventions and stayed low after the study. sumers shown in the top part of Fig. 4.
The temperatures before the intervention were shown to be The gas consumption however shows a clear indicator of the
statistically different to 21
C in the group with temperatures effect of iBert. Although iBert contained no direct messaging with
higher than 21
C in the baseline: with a p-value rejecting the Null respect to gas consumption, three statements within iBert were
hypothesis (they are equal) of p ¼ 0.0039 (<0.05); whereas the designed to reduce gas consumption: change in the thermostat
temperatures before and after in the group in which the tempera- temperature, change in the operating scheme (turning heating off
tures were already below 21
C cannot be considered different (p- when house unoccupied) and change in the ventilation regime. As
value of 0.46). Moreover, a t-test was used to prove if the hypothesis there was no direct messaging with respect to gas consumption, we
that the temperatures in the baseline in the houses in which the cannot separate high and low consumers based on whether they
starting temperature was larger than 21
C was statistically received a message. A reasonable alternative is to simply classify
different to 21
C after the study. The test proved that the tem- those in the top quartile of consumption as high consumers.
peratures of the group which behaviour needed to be modified Calculating the pre/post iBert household gas consumption for just
were not statistically different to 21
C after the study with a p- these high consumers shows a mean household saving of 27.2%,
value of 0.17 (the null hypothesis mean ¼ 21
C, can not be rejected) compared to 22.0% for all homes.
meaning that the behaviour of the occupants had been modified. It is unknown why there was no reduction in electricity use, but
A similar evaluation was done with the CO2 concentration. the following are possibilities: (1) for an occupant, reducing elec-
Houses in which the CO2 levels were less than 800 ppm were tricity consumption is more complex than simply turning a

Fig. 2. Home internal temperatures (in
C) before, across the iBert digital intervention phase and after the feedback. The graph has been separated as houses likely to receive a
personalised message (upper graph) and those that did not (lower). The blue box represents 50% of the data, the median is represented as a horizontal (red) line within the box and
the blue whiskers represent the 97.5% of the data. The thick green line is at 21
Cdthe trigger for receiving a message. The individual dots, and dashed lines, represent individual
homes. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
446 N. Mogles et al. / Building and Environment 125 (2017) 439e450

Fig. 3. Home CO2 concentrations before, across the iBert digital intervention phase and after the feedback. The graph has been separated into houses likely to receive a personalised
message (upper graph) and those that were not (lower). The thick green line is at 800ppmdthe trigger for receiving a message. (For interpretation of the references to colour in this
figure legend, the reader is referred to the web version of this article.)

Fig. 4. Box plots of daily electricity usage after separation of big consumers (top) and small consumers (bottom). The thick green line is at 600 Wdthe trigger for receiving a
message. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

thermostat down, or ensuring heating is off when leaving the devices in different homes, this could lead to high consumers not
building; (2) rational electricity consumption is more individual- necessarily being profligate; (3) it is known that people are unsure
istic than heating consumption, with a multitude of different of which electrical items use the most electricity, and hence which
 N. Mogles et al. / Building and Environment 125 (2017) 439e450 447

to manage more effectively. This hypothesis was confirmed. Homes with high internal
temperatures reduced their temperatures and gas consumption
5.5. Analysis 5: Energy literacy was reduced.
Hypothesis 2. Energy feedback translated according to internal
A paired sample t-test was conducted to evaluate the effect of
personal values will have a significant effect on energy related
digital energy feedback interventions on the cognitive component
behaviour.
of energy literacy. The results indicated that the mean energy lit-
This hypothesis was confirmed for home internal temperature
eracy score before the interventions (M ¼ 0.52, SD ¼ 0.71) was
measurements. Framing messages in value-based terms led to
significantly lower than the mean energy literacy score after the
greater changes in behaviour.
interventions (M ¼ 1.28, SD ¼ 1.06, t (24) ¼ 3.17, p ¼ 0.004). This
confirms our fourth hypothesis that the energy literacy of partici- Hypothesis 3. Personalised messages with action prompts will
pants will improve after digital feedback that is framed in an have a significant effect on energy related behaviour.
educational way. This hypothesis was confirmed, with an additional 9.6% point
reduction in gas consumption from the use of action prompts.
5.6. Analysis 6: User experience and engagement
Hypothesis 4. No matter which behavioural effect of a digital
energy feedback is observed, the digital interventions will influence
Three months after the end of the energy feedback study, par-
energy related cognitive variables, such as energy literacy.
ticipants were contacted by phone to arrange for the equipment to
This hypothesis was confirmed. We found a statistically signif-
be collected and were asked to complete the same items from the
icant positive effect of iBert interventions on energy literacy. One
energy literacy survey described in Refs. [49,50], again along with
may argue that the effect of energy literacy could be attributed to
nine questions related to system usability and user experience: four
other factors and not the tailored educative messages per se, e.g. by
items adopted from the System Usability Scale (SUS) [51]) which is
the Hawthorne effect [52]. This is a phenomenon where people
the most frequently cited scale for system usability evaluation, one
behave differently when they know they are being observed. Given
item addressing participants' preferences for different components
the fact that the sensors sets were installed in these homes two
of iBert and four items on the effect of iBert on participants'
years earlier and they had been monitored for a long period of time
behaviour. In total, 65% of the participants completed the post-
prior to the energy feedback intervention study, we believe that the
study survey. The answers were rated on a 3-point ordinal Likert
Hawthorne effect was minimal in comparison to other studies as
scale. The usability questions were:
many of them did not control for this effect.

“I think that I would like to use iBert frequently”

“I found the system very difficult to use”
“I needed to learn a lot of things before I could get going with the 7. Conclusions
system”
“How useful did you find iBert overall?” The main innovative aspect of this study is that it tests for the
first time a combination of two energy feedback strategies, per-
The results show that the majority of the participants liked to sonal values and tailored action prompts, applied to energy related
use iBert frequently, found the system easy to use and in general behaviour in an experimental setting. Value-framed motivational
found iBert useful. Eighty-two percent of the respondents found messages and action prompts tailored to user behaviour and
live energy consumption data useful, 77% liked the weekly energy building characteristics were embedded into an intelligent energy
summary and 78% the tailored messages; ANOVA analysis showed feedback system called iBert. The system gives advice according to
the three components of feedback were found roughly equally individual building context and problematic energy-related
useful, F (2, 32) ¼ 1.27, p ¼ 0.30. Some 50% of the respondents took behaviour detection, such as high internal temperature, excessive
energy saving actions suggested by iBert and roughly 36% of them ventilation, and heating and electricity usage while the home is not
took other energy saving actions not suggested by the system. User occupied.
experience with iBert is summarised in Table 4. Our findings suggest a positive effect of digital feedback on
home internal temperature, and a specific positive effect of internal
6. Discussion values and action prompts incorporated in energy feedback, from
digital energy feedback delivered by iBert. This is the first time a
In this section we discuss the findings in the light of the four smart metering approach has been shown to directly influence
hypotheses formulated prior to the study. room temperature. Unprecedented gas consumption saving were
achieved with much of this directly attributable to the use of action
Hypothesis 1. Digital feedback in general will have an effect on
prompts.
energy related behaviour compared to a baseline period before the
The 22.0% reduction in gas consumption (27.2% in high con-
interventions.
sumers) found is substantially greater than the reductions seen in
the literature; this provides strong evidence of the power of per-
Table 4
Self-reported user experience with iBert energy feedback. sonalised feedback, personal values, action prompts and an active
improvement of energy literacy, and could be industry-
Interaction Percentage of
transformative.
households
The digital feedback system we have demonstrated in this paper
Took energy saving actions suggested by iBert 50%
had a marked positive effect on the knowledge component of
Took additional energy saving actions NOT 36%
suggested by iBert participants' energy literacy. This result is very promising as it in-
Paid attention to iBert text messages 78% dicates that digital feedback (a standard IHD accompanied by
Other households members were aware of 86% tailored educational messages as in the current study) can improve
the iBert information general energy understanding. A better understanding of energy
Found iBert (partially) useful 82%
use would be a natural first step towards forming new attitudes and
448 N. Mogles et al. / Building and Environment 125 (2017) 439e450

foundations for energy-saving behaviour amongst occupants. 5. Which of the following actions, if everyone did this all the time,
In this paper we have evaluated a ground-breaking way of would save the most energy in the UK? (Please circle the most
informing occupants about their energy consumption with the important)
purpose of educating users and lowering their demand. By being Turn off lights when they are not in use
building- and occupant-aware, and only reporting messages when Turn down the heat in rooms
consumption is potentially high, many of the criticisms of smart Reduce water consumption
metering have been addressed. For the first time a smart metering Walk or cycle short distances instead of going by car
system has been shown to reduce room temperatures directly, with Turn appliances off at the plug
the temperature chosen by the occupants identical to that sug- 6. Which kind of lighting uses the least amount of energy? (Please
gested by the algorithm. The gas consumption savings found in this circle one answer)
two-year study have the potential to be transformative for both the Standard light bulbs
occupants and for climate change policy. Low energy light bulbs
Fluorescent lights
LED lights
Acknowledgements
Don't know

This work was done under the auspices of the ENLITEN project,
funded by EPSRC (grant reference EP/K002724/1). We thank Exeter
Part 2
City Council for their support provided during the project. The in-
ternal temperature and energy consumption data used for the
1. I think that I would like to use iBert frequently
analysis together with the energy literacy and usability surveys
Agree
data can be found here: http://doi.org/10.15125/BATH-00256.
Somewhat agree
Ramallo-Gonza lez would like to thank the program Saavedra
Disagree
Fajardo (20035/SF/16) funded by Consejería de Educacio n y Uni-
2. I found the system very difficult to use
versidades of CARM, via Fundacio n Se
neca-Agencia de Ciencia y
Agree
Tecnología de la Region de Murcia.
Somewhat agree
Disagree
Appendix A. Energy literacy and user experience survey 3. I needed to learn a lot of things before I could get going with the
system
Part 1 Agree
Somewhat agree
1. How much do you feel you know about energy? (Please circle Disagree
one) 4. In which extent did you find the following parts of the iBert
A lot (expert) system useful:
Quite a bit (informed) Live status panel
Not much (novice) Useless
Nothing, Nothing Moderately useful
2. Which of the following sources of information has contributed Very useful
most to your understanding of energy issues? (Please choose Weekly energy consumption summary
one option) Useless
Further or higher education Moderately useful
School Very useful
Books, newspapers or magazines Text messages with tips and advice/suggestions
Friends or family members (including parents) Useless
Internet Moderately useful Very useful
Television iBert system 5. How useful did you find iBert overall?
Other (please specify) Useless
3. The term renewable energy resources means (please circle one Moderately useful
answer) Very useful
Resources that are free and convenient to use 6. Did you pay attention to the text messages that were displayed
Resources that can be converted directly into heat and by iBert?
electricity, Resources that can be converted directly into heat Yes
and electricity No
Resources that do not produce air pollution Partially
Resources that are very efficient to use for producing energy 7. Did you take any energy saving actions suggested by iBert?
Resources that can be replenished by nature in a short period Yes
of time No
4. Most of the renewable energy in the UK comes from which of 8. Did you take any additional energy saving actions NOT sug-
the following sources? (Please circle one) gested by iBert during or after the study?
Solar Yes
Water (hydro/tidal/wave) power No
Wind Partially
Landfill gas 9. If there are other adults living in your home, did they attend/
Geothermal were aware of the information provided by iBert?
Don't know Not aware
 N. Mogles et al. / Building and Environment 125 (2017) 439e450 449

Aware Appendix B. iBert running on an android tablet

Partially aware
450 N. Mogles et al. / Building and Environment 125 (2017) 439e450

References [26] Teresa Chiang, Sukumar Natarajan, Ian Walker, A laboratory test of the effi-
cacy of energy display interface design, Energy Build. 55 (2012) 471e480.
[27] Hunt Allcott, Social norms and energy conservation, J. Public Econ. 95 (9)
[1] IEA, World Energy Statistics and Balances 2016, Technical report, International
(2011) 1082e1095.
Energy Agency, 2016.
[28] Sally Dibb, Marylyn Carrigan, Tim Harries, Ruth Rettie, Matthew Studley,
[2] EIA, Annual Energy Outlook 2016, Technical report, Energy Information
Kevin Burchell, Simon Chambers, Is social norms marketing effective? a case
Administration, US, 2014.
study in domestic electricity consumption, Eur. J. Mark. 47 (9) (2013)
[3] BEIS, Energy Consumption in the UK, Technical report, Department of Busi-
1458e1475.
ness, Energy and Industrial Strategy, UK, 2016.
[29] P. Wesley Schultz, Mica Estrada, Joseph Schmitt, Rebecca Sokoloski,
[4] DECC, 2014 UK Greenhouse Gas Emissions, Final Figures, Technical report,
Nilmini Silva-Send, Using in-home displays to provide smart meter feedback
Department of Energy and Climate Change, UK, 2014.
about household electricity consumption: a randomized control trial
[5] EIA, International Energy Outlook 2016, Technical report, US Energy Infor-
comparing kilowatts, cost, and social norms, Energy 90 (2015) 351e358.
mation Administration, 2016.
[30] P. Wesley Schultz, Jessica M. Nolan, Robert B. Cialdini, Noah J. Goldstein,
[6] Sarah Darby, Making it obvious: designing feedback into energy consumption,
Vladas Griskevicius, The constructive, destructive, and reconstructive power
in: Energy Efficiency in Household Appliances and Lighting, Springer, 2001,
of social norms, Psychol. Sci. 18 (5) (2007) 429e434.
pp. 685e696.
[31] Adam Corner, Ezra Markowitz, Nick Pidgeon, Public engagement with climate
[7] Stavroula Karatasou, Marina Laskari, Mat Santamouris, Models of behavior
change: the role of personal values, Wiley Interdiscip. Rev. Clim. Change 5 (3)
change and residential energy use: a review of research directions and find-
(2014) 411e422.
ings for behavior-based energy efficiency, Adv. Build. Energy Res. 8 (2) (2014)
[32] M. Douglas, Wildavski, Risk and Culture: an Essay on the Selection of Tech-
137e147.
nological and Environmental Dangers, Univ of California Press, 1983.
[8] Chris Foulds, Rosalyn A.V. Robison, Rachel Macrorie, Energy monitoring as a
[33] Shalom H. Schwartz, Universals in the content and structure of values:
practice: investigating use of the iMeasure online energy feedback tool, En-
theoretical advances and empirical tests in 20 countries, Adv. Exp. Soc. Psy-
ergy Policy 104 (2017) 194e202.
chol. 25 (1) (1992) 1e65.
[9] Thanasis G. Papaioannou, Dimos Kotsopoulos, Cleopatra Bardaki,
[34] Linda Steg, Jan Willem Bolderdijk, Kees Keizer, Goda Perlaviciute, An inte-
Stavros Lounis, Nikos Dimitriou, George Boultadakis, Anastasia Garbi,
grated framework for encouraging pro-environmental behaviour: the role of
Anthony Schoofs, IoT-enabled gamification for energy conservation in public
values, situational factors and goals, J. Environ. Psychol. 38 (2014) 104e115.
buildings, in: 2017 Global Internet of Things Summit (GIoTS), Geneva,
[35] Judith IM. De Groot, Linda Steg, Value orientations to explain beliefs related to
Switzerland, 2017, pp. 1e6.
environmental significant behavior how to measure egoistic, altruistic, and
[10] Heta Karoliina Kosonen, Amy Ahim Kim, Advancement of behavioral energy
biospheric value orientations, Environ. Behav. 40 (3) (2008) 330e354.
interventions in commercial buildings, Facilities 35 (5/6) (2017) 367e382.
 [36] Linda Steg, Goda Perlaviciute, Ellen Van der Werff, Judith Lurvink, The sig-
[11] Thorben Jensen, Georg Holtz, Carolin Baedeker, Emile J.L. Chappin, Energy-
nificance of hedonic values for environmentally relevant attitudes, prefer-
efficiency impacts of an air-quality feedback device in residential buildings:
ences, and actions, Environ. Behav. (2012), 0013916512454730.
an agent-based modeling assessment, Energy Build. 116 (2016) 151e163.
[37] Jon Froehlich, Promoting energy efficient behaviors in the home through
[12] Xingxing Zhang, Jingchun Shen, Tong Yang, Llewellyn Tang, Luying Wang,
feedback: the role of human-computer interaction, in: Proc. HCIC Workshop,
Yingqi Liu & Peng Xu, Smart meter and in-home display for energy savings in
vol. 9, 2009, pp. 1e11.
residential buildings: a pilot investigation in Shanghai, China, Intell. Build. Int.
[38] Neil Simcock, Sherilyn MacGregor, Philip Catney, Andrew Dobson,
(2016) 1e23. Published online: 29 Jul 2016, http://dx.doi.org/10.1080/
Mark Ormerod, Zoe Robinson, Simon Ross, Sarah Royston, Sarah Marie Hall,
17508975.2016.1213694.
Factors influencing perceptions of domestic energy information: content,
[13] Fernando Terroso-Saenz, Aurora Gonzalez-Vidal, Alfonso P. Ramallo-
lez, Antonio F. Skarmeta, An open IoT platform for the management and source and process, Energy Policy 65 (2014) 455e464.
Gonza
[39] Debby Cotton, Wendy Miller, Jennie Winter, Ian Bailey, Stephen Sterling,
analysis of energy data, Future Generat. Comput. Syst. Available online 6
Knowledge, agency and collective action as barriers to energysaving behav-
September 2017, ISSN 0167-739X. https://doi.org/10.1016/j.future.2017.08.
iour, Local Environ. (2015) 1e15.
046.
[40] Jan DeWaters, Susan Powers, Establishing measurement criteria for an energy
[14] T. Lovett, E. Gabe-Thomas, S. Natarajan, E. O'Neill, J. Padget, ‘Just Enough’
literacy questionnaire, J. Environ. Educ. 44 (1) (2013) 38e55.
sensing to ENLITEN: a preliminary demonstration of sensing strategy for the
[41] Lung-Sheng Lee, Yi-Fang Lee, James W. Altschuld, Ying-Ju Pan, Energy literacy:
‘energy literacy through an intelligent home energy advisor’ (ENLITEN)
evaluating knowledge, affect, and behavior of students in taiwan, Energy
project, in: Proceedings of the Fourth International Conference on Future
Policy 76 (2015) 98e106.
Energy Systems, ACM, Berkeley, CA, 2013, pp. 279e280.
lez, M. Brown, D.A. Coley, The reliability of inverse [42] Mageswary Karpudewan, Jamunah Ponniah, Ahmad Nurulazam, Md Zain.
[15] A.P. Ramallo-Gonza
Project-based learning: an approach to promote energy literacy among sec-
modelling for the wide scale characterization of the thermal properties of
ondary school students, Asia-Pacific Educ. Res. 25 (2) (2016) 229e237.
buildings, J. Build. Perform. Simulat. 1e19, Received 22 Sep 2014, Accepted 10
[43] Marika Vellei, Sukumar Natarajan, Benjamin Biri, Julian Padget, Ian Walker,
Dec 2016, Published online: 17 Jan 2017. http://dx.doi.org/10.1080/19401493.
The effect of real-time context-aware feedback on occupants heating behav-
2016.1273390.
iour and thermal adaptation, Energy Build. 123 (2016) 179e191.
[16] Wokje Abrahamse, Linda Steg, Charles Vlek, Talib Rothengatter, A review of
[44] Simona DOca, Stefano P. Corgnati, Tiziana Buso, Smart meters and energy
intervention studies aimed at household energy conservation, J. Environ.
savings in Italy: determining the effectiveness of persuasive communication
Psychol. 25 (3) (2005) 273e291.
in dwellings, Energy Res. Soc. Sci. 3 (2014) 131e142.
[17] A.-D. Barbu, N. Griffiths, G. Morton, Achieving Energy Efficiency through
[45] Andreas Kamilaris, Jodi Neovino, Sekhar Kondepudi, Balaji Kalluri, A case
Behaviour Change: what Does it Take?, Technical report, EEA, 2013.
study on the individual energy use of personal computers in an office setting
[18] Sarah Darby, Smart metering: what potential for householder engagement?
and assessment of various feedback types toward energy savings, Energy
Build. Res. Inf. 38 (5) (2010) 442e457.
Build. 104 (2015) 73e86.
[19] Tom Hargreaves, Michael Nye, Jacquelin Burgess, Making energy visible: a
[46] Office for National Statistics, Nowcasting Household Income in the UK:
qualitative field study of how householders interact with feedback from smart
Financial Year Ending 2015, Technical report, Office for National Statistics, UK,
energy monitors, Energy Policy 38 (10) (2010) 6111e6119.
2015.
[20] Vine Desley, Buys Laurie, Morris Peter, The effectiveness of energy feedback
[47] OFGEM, Smart Metering Implementation Programme: in-Home Display,
for conservation and peak demand: a literature review, Open Journal of En-
Technical report, Office of gas and electricity markets, UK, 2010.
ergy Efficiency 2013 (2013).
[48] CIBSE, Guide, A: Environmental Design, The Chartered Institution of Building
[21] Ahmad Faruqui, Sanem Sergici, Ahmed Sharif, The impact of informational
Services Engineers, London, 2006.
feedback on energy consumptiona survey of the experimental evidence, En-
[49] Debby RE. Cotton, Wendy Miller, Jennie Winter, Ian Bailey, Stephen Sterling,
ergy 35 (4) (2010) 1598e1608.
Developing students energy literacy in higher education, Int. J. Sustain. High.
[22] K. Buchanan, R. Russo, B. Anderson, The question of energy reduction: the
Educ. 16 (4) (2015) 456e473.
problem(s) with feedback, Energy Policy 77 (2015) 89e96.
[50] Jan E. DeWaters, Susan E. Powers, Energy literacy of secondary students in
[23] Geraldine Fitzpatrick, Greg Smith, Technology-enabled feedback on domestic
New York state (USA): a measure of knowledge, affect, and behavior, Energy
energy consumption: articulating a set of design concerns. Pervasive
Policy 39 (3) (2011) 1699e1710.
Computing, IEEE 8 (1) (2009) 37e44.
[51] John Brooke, et al., Sus-a quick and dirty usability scale, Usability Eval. in-
[24] Kathryn Buchanan, Riccardo Russo, Ben Anderson, Feeding back about eco-
dustry 189 (194) (1996) 4e7.
feedback: how do consumers use and respond to energy monitors? Energy
[52] John G. Adair, The Hawthorne effect: a reconsideration of the methodological
Policy 73 (2014) 138e146.
artifact, J. Appl. Psychol. 69 (2) (1984) 334.
[25] Corinna Fischer, Feedback on household electricity consumption: a tool for
saving energy? Energy Effic. 1 (1) (2008) 79e104.