Acceptation of a Demand-response Enabling Technology for using

Electricity at Home upon a Simulated Marketing Campaign

Role of Sociodemographic Variables and Prior Energy Behaviors, in Tandem with

Expectations and Attitudes Formed to the Message Target

Maria São João Breda

and Marta Lopes

2,3

Institute of Cognitive Psychology, Faculty of Psychology and Educational Sciences, University of Coimbra,

R. Colégio Novo, 3001-802 Coimbra, Portugal

Dept. of Environment, ESAC - Polytechnic Institute of Coimbra, 3045-601 Coimbra, Portugal

INESC Coimbra, Rua Antero de Quental 199, 3000-033 Coimbra, Portugal

Keywords: Technology-acceptation, Attitudes to Technology, Socio-demographic Factors, Energy Behaviors.

Abstract: The present work addresses the usefulness of IT acceptance frameworks for studying consumers’

adoption intentions upon learning of a new technology described as affording demand-response and

energy conservation at home. A survey study was conceived which relied on the exposure of

respondents to a marketing campaign for this technology. In preceding steps Theory of Planned

Behavior and Technology Acceptance Unified Theory were applied to adequately model and test

predictive relations upon intention to adopt that technology and upon expectation of remaining with

known methods that compete with the technology, with a Partial Least Squares structural modeling

approach. Results suggest that the frameworks are useful for predicting intention to adopt this type of

technology. Especially important predictors were Effort Expectancy, Social Factors and Positive Attitudes.

Given validation of the nomological network, the goal is to comprehensively integrate differences in

adoption across socio-demographic strata and tied to consumers’ energy behaviors with the structural

network linking IT predictors to dependents.

1 INTRODUCTION

The present work complements a study on the

applicability of Information Technology (IT)

acceptance models Theory of Planned Behavior

(TPB) (Fishbein and Ajzen, 2005) and Unified

Theory of Acceptance and Use of Technology

(UTAUT) (Venkatesh et al., 2003) to the

understanding of residential consumers’ response, in

terms of intention to adopt a technological proposal

for energy efficiency and for demand-response to

dynamic tariffs, in the context of a simulated

marketing campaign. Empirical support for the

applicability and usefulness of these frameworks for

predictive purposes was obtained from questionnaire

data, i.e., from self-reported perceptions of the

message conveyed in the campaign, through

building and testing measurement and structural

models which model adoption intention, and

expectation of continuing to use known methods

alternative to the technology, as a function of a set of

expectations, attitudes, affect and social factors.

Two alternative models were developed and

tested, differing in the operationalization of attitude.

The dependent variables were intention to adopt

and expectation of remaining with known methods

that compete with the technology. Predictors

contemplated positive and negative attitudes

(bidimensionality), performance and effort

expectancies, social factors, and perceived behavior

control.

The present study draws on positive evidence on

the usefulness of the nomological network with its

associated blocks of indicator variables, and

inspects the role of consumer variables, e.g. socio-

demographic and prior energy behaviors, posited to

work in tandem with the cognitive-affective ones.

Gender and age have been found to draw associated

differences in adoption intention for technology in

workplace in the United States. Are they to be found

in regard to this particular technology, in a

296

Breda M. and Lopes M..

Acceptation of a Demand-response Enabling Technology for using Electricity at Home upon a Simulated Marketing Campaign - Role of Sociodemograp-

hic Variables and Prior Energy Behaviors, in Tandem with Expectations and Attitudes Formed to the Message Target .

DOI: 10.5220/0004941902960304

In Proceedings of the 3rd International Conference on Smart Grids and Green IT Systems (SMARTGREENS-2014), pages 296-304

ISBN: 978-989-758-025-3

 2014 SCITEPRESS (Science and Technology Publications, Lda.)

Portuguese sample? If so, is the structural model

proposed also useful to understand or represent

processes whereby such adoption differences build

across groups of respondents, or are these

independent effects (via other social, cognitive

and/or behavioral processes)? If they are not

independent, do the predictors and structural links in

the model convey groups’ differences (mediating

effect), or is it necessary to differentiate the

importance of specific predictors according to

groups of respondents (moderation effects)?

Energy behaviors are an important subject matter

in sustainability studies (Lopes et al., 2012). In light

of the concept of “Intelligent efficiency” (Elliott et

al., 2012) energy behaviors become challenged by

becoming more involved with technology systems.

This is increased with the ‘Smart grid’, and Demand

Response initiatives, (Santacana et al., 2010). These

behaviors belong to multiple types (Black et al.,

1985); (Stern, 2000) e.g., investment in efficient

equipment and whether and how they may impact

the willingness to acquire or use automating

technology is unknown. Prior energy behaviors of

respondents are likely to affect their acceptation

intention but they may do so through different

processes represented in alternative attitude-

behavior or behavior-attitude theories. Moreover

given the multiplicity of energy behaviors and their

potential similitude to one or the other conative

variables (adoption intention; expectation to remain

with known methods), various hypotheses are

studied.

Energy behaviors involving technology, similar

in that regard to adoption intention, may enhance

positive attitude towards adoption, according to

behavior-attitude theory, and thus promote intention

to adopt. Alternatively, behaviors similar to

adoption may strengthen certain heuristic processing

such that considerations on the cost or easiness of

behavior enactment rather than the considerations on

outcomes probabilities become more salient

(Kidwell & Jewell, 2008).

2 REVISION OF IT

ACCEPTANCE LITERATURE

The Attitude-behavior concepts pertaining to TPB

(Fishbein and Ajzen, 2005) have been successfully

applied in multiple behavioral domains (Armitage

and Christian, 2003), including the energy realm

when expanded with personal norm (Abrahamse and

Steg, 2009); (Gadenne et al., 2011) pro-

environmental behavior (Oreg and Katz-Gerro,

2011), as well as IT acceptation in the workplace

(Taylor and Todd, 1995). IT acceptance studies have

highlighted the role of users’ attitudes and beliefs

about the technology as antecedents of acceptance

versus rejection, and the importance of

understanding the factors that motivate or hinder

those attitudes. Acceptance refers to the willingness

of a potential user, with some degree of choice, to

intentionally employ a technology for the purposes

or tasks it was conceived for (Dillon, 2001). A

change in measurement paradigm towards

expectations/ appraisals referring to spheres of

experience (utilitarian, experiential, and enjoyment)

underlies Technology Acceptance Model and

Decomposed TPB (Davies et al., 1989); (Bagozzi et

al., 2002). Framework for measurement of users’

perceptions and attitude by operationalization of

core constructs after revision of the literature has

been proposed by Venkatesh et al., (2003). The

UTAUT was based upon seven different models of

technology acceptance.

As a novel application of IT acceptance

literature, studying the intention or willingness to

adopt automation technology for energy efficiency,

in the context of transitioning to smart grids, and

ensuing dynamic tariff policies for demand-response

(Livengood and Larson, 2009), implied conducting a

survey, and, subsequently, modeling the adoption

response, by broadening the scope of the variables,

both the dependent variables and attitudinal ones.

Table 1 summarizes the constructs of TPB and

UTAUT in IT studies for the prediction of intention

under circumstances of little experience with the

technology, and illustrates how constructs were

operationalized in our study.

2.1 Respondents, Preceding Steps and

Results

Respondents were 504 members from a marketing

panel from the center region of Portugal. They were

eligible to the survey if their educational level was

the 12

grade and if they paid their own electricity

bill. Sixty two percent (311) were women, age

ranged from 25 to 60 years old (mean of 33.5 and

standard deviation of 7.2); educational level was

balanced across the two levels (54% had higher

education). The sample was then randomly divided

into two halves, for exploratory factor and

confirmatory Partial Least Squares analyses (PLS).

The (future) smart grid context was presented to

respondents highlighting its purpose of efficiency,

peak reduction, and environmental protection. The

technology concept was introduced in terms of tasks

AcceptationofaDemand-responseEnablingTechnologyforusingElectricityatHomeuponaSimulatedMarketing

Campaign-RoleofSociodemographicVariablesandPriorEnergyBehaviors,inTandemwithExpectationsandAttitudes

FormedtotheMessageTarget

297

Table 1: Constructs and findings regarding TPB and UTAUT and the adapted operationalization in our survey. Adapted

from Venkatesh et al., (2003a).

to meet in demand-side initiatives and price policies.

The message and the survey followed the principles

of the Utilitarian Information and Persuasive

Communication. The survey was constituted by sets

of questions dealing with socio-demographic

information, residence and electricity consumption

and behaviors; adoption of the technology under the

current and future smart grid, anticipated feelings to

monitoring and to using automated decision aids.

Measurement and structural models’ tests were

performed with PLS path modeling, employing

SmartPLS version 2.0.3. (Ringle et al., 2006).

Adjusted models aimed at explaining dependent

variables of adoption intention (BehvIntenExpc), and

also resistance to adoption, e.g., expectations of

keeping using the methods already known to manage

electricity consumption at home

(Pessim_AltMethods). The models recognized the

bidimensionality of the attitudinal concept, making

place for positive and negative attitudes towards

adoption. Two models were developed and adjusted,

and simplified versions with only higher magnitude

path coefficients were retained. The first comprises

as predictors variables of Effort Expectancy, Social

Theory

Concepts

Definition Operationalization

(examples)

Operationalization adaptation for our study

(examples)

TPB

Attitude

towards

behavior

“an individual’s positive or negative

feelings (evaluative affect) about

performing the target behavior”

(Fishbein & Ajzen, 1975)

Global evaluative

ratings

(cf. infra) plus Beliefs: “By adopting the

EnergyBox at home, when the Smart Grid

exists, I will contribute to the reduction of the

greenhouse effect emissions to the

environment”; “By adopting this device at

home, I will contribute to reduce the grid

consumption peaks”; “Using the EnergyBox at

home will allow me to increase the habit of

saving energy and money”

Subjective

Norm

(S.N.)

“The person’s perception that most

people who are important to him

think he should or should not

perform the behavior” (Ibid)

“People who influence

my behavior think I

should use the system”

“People who are influent to me will think that I

ought to adopt this type of device”

Perceived

Behavior

Control

(P.B.C.)

“The perceived ease or difficulty of

performing the behavior (Ajzen,

1991); “Perception of internal or

external constraints on behavior”

(Taylor & Todd, 1995); Relates to

self-efficacy, and facilitating

conditions regarding resources, as

well as regarding the technology

“Given the resources,

the opportunities and

knowledge necessary, it

will be easy for me to

use the system”;

“Using this device will take too much time,

considering all I have to do”

UTAUT

Performance

Expectancy

(PE)

“the degree to which an individual

believes that using the system will

help him or her to attain gains in

job performance”. It integrates

constructs of: perceived

usefulness (TAM), extrinsic

motivation (MM), job-fit (MPCU),

relative advantage (IDT), and

outcome expectations (SCT) .

“I would find the system

useful in my job” “Using

the system enables me

to accomplish my tasks

more quickly”

“At home, using a monitor to keep track of the

domestic electric consumption will allow me to

save energy and money”; “At home, if I use

the Autonomous Decision aid I will probably

be able to displace the consumption to

whenever the energy is cheaper”: “Even when

using the Smart Grid, I trust my own ability to

place the consumption times to whenever

electricity is cheaper or more available at

home, by myself, without the help of the

Autonomous Decision aid”

Effort

Expectancy

(EE)

“The degree of ease associated

with the use of the system”

encompassing: perceived ease of

use (TAM), complexity (MPCU),

and ease of use (IDT)

“I would find the system

ease to use” “Learning

to operate the system is

easy for me”

“It will be easy to learn how to use the

EnergyBox monitor to check consumptions”;

“When adapting to the Smart Grid, it will be

easy to learn how to use the EnergyBox”

Social

Influence

“the degree to which an ind.

perceives that important others

believe he or she should use the

new system” and relates to

constructs of: S.N. (TPB), social

factors (MPCU), and image (IDT)

“People who influence

my behavior think I

should use the system”

“People from my group of colleagues and

friends who use the EnergyBox will have a

higher social status than the ones who don’t”

Attitude

towards using

technology

Positive feelings about performing

the behavior; Reflects constructs

of: Attitude towards behavior

(TPB), Intrinsic Motivation; Affect

towards Use; and Affect.

“Using the system is a

bad/ good idea” “The

system makes work

more interesting”

“Adopting the EnergyBox is a good idea”;

“Adopting the E. in a context of Smart Grid is a

wise idea”; “I will grow bored of using this type

of device”; “It will be hard for me to accept that

the Autonomous Decision aid is in charge of the

use of electricity at home”; “With the E.,

managing the expenses at home will be more

interesting”

Behavioral

Intention

(BehvIntent

Expc)

A self-generated instruction to

perform an action, “behavioral plan

making possible the achieving of a

behavioral goal” (Ajzen, 1986)

I intend to use the

system in the next ()

months

“I intend to be one of the first people to install

and use the device”; “I can foresee that I will

use this device as soon as it is available”

“I think I will use more or less the methods I

already use, without any specific devices, in

electrical consumption”

SMARTGREENS2014-3rdInternationalConferenceonSmartGridsandGreenITSystems

298

Image and Performance Expectancy, as well as two

attitudinal variables with different valences that

partially overlap conceptually with Performance

Expectancy, and are understood as a compound of

respectively, positive, and negative, affect and

expectancy, and, lastly, a variable of valuations of

outcomes of conserving electricity. The second

model (Fig 1) aimed at achieving a higher

discrimination between expectancy and attitudinal

blocks of items. It operationalized attitude as

anticipated emotions to performing behaviors

implied in the technology’s functionalities, namely,

checking on overall and detailed electricity

consumptions with a monitor; and relying on an

automated decision system to schedule and set in

functioning the equipment at home. Otherwise, it

was similar to the first model.

Estimation and evaluation of the measurement

component of the models afforded psychometrically

good measures. The two structural models displayed

adequate predictive capacity globally and upon

selected endogenous, by criteria of coefficient of

determination R

, R

change and cv-redundancy.

They converge in the pattern of results regarding

prediction of both variables. In the prediction of

Adoption Intention, three expectations factors about

spheres of experience with technology are

important: EE, S.I.; and also Positive affect

anticipated to the behaviors. In predicting

expectations of using already known methods that

are alternative to the technology, the major

explanatory factor is Negative Attitude/Negative

Affect; and Perceived Behavior Control.

3 SOCIODEMOGRAPHIC

DIFFERENCES

3.1 Research Goals and Hypotheses

The effects of gender and age (Venkatesh et al.,

2003a; b); (Venkatesh and Morris, 2000);

(Venkatesh et al., 2000) and of prior experience with

technology, are addressed within IT acceptance

studies. They have been addressed in terms of

moderator effects, i.e., group differences in

variations in importance of the different factors or

predictors that contribute to adoption. Within the

framework of TAM (Venkatesh et al., 2003b);

(Venkatesh and Morris, 2000), it has been found that

men had their intention to adopt more closely

associated to Perceived Usefulness (a Performance

Expectancy factor); and women to Perceived Ease of

Use (Effort Expectancy variable). Under the

contingency that experience was low, Subjective

Norm was more salient for women (Venkatesh et al.,

2003b). In our study experience can only exist with

other technology, e.g. with employing other

programming devices, experience is assumed to be

low by default. Within TPB model, Venkatesh et al.,

(2000) suggested that Attitude was more important

to men, and SN and PBC more for women when

experience is low. Regarding age, it was found that

younger workers had Attitude contribute more to

adoption intention, and older workers, PBC. SN was

more important for older women (Venkatesh et al.,

2003b).

3.2 Analysis and Results

Moderator analysis for gender and for age relies on

multi-group analysis (PLS_MGA), wherein our

choice was to employ the parametric approach (Keil

et al., 2000, cit in Hair et al., 2014). In the case of

age, the variable was discretized into 3 categories:

from 25 to 30 years old (212 respondents); higher

than 30 but below 41 years of age (215respondents),

and over 40 years up to 60 years of age (77

respondents).

Gender (being of female as opposed to male

gender) did not have a significant impact in adoption

intention (structural path coefficient beta = -0,038;

bootstrapping asymptotic t=0,71), nor in expectation

to keep with old methods (beta= 0,067; t=1,097).

For moderation analysis, the procedure was to

perform separate estimations for each group,

followed by a modified version of a 2-independent

samples t-test, to compare path coefficients across

the 2 groups of data. T test provides the statistical

significance of the difference between the two

groups’ path estimates.

With a simplified model, retaining only one

dependent variable (Adoption Intention) we find

male respondents compared to female respondents

give higher importance to Performance Expectancy

of the technology in forming Intention to adopt

(betas for the two groups respectively of 0,205

[t=2.38] and 0.116 [t=2.01]) and in PBC (betas

respectively of 0.314 [t=3.35] and 0.134 [t=1.94]),

and female respondents’ group had higher salience

of Effort Expectancy both in forming anticipated

positive affective feelings (betas respectively of

0.47 [t=4.29] and 0.135 [t=1.47])and in Intention to

adopt (betas respectively of 0.23 [t=4.14] and 0.14

[t=1.87]). One-tail t tests of the differences were

significant at minimum at 0.05 alpha level,

exception made of EE to BehIntentAdopt at 0.1

alpha level.

AcceptationofaDemand-responseEnablingTechnologyforusingElectricityatHomeuponaSimulatedMarketing

Campaign-RoleofSociodemographicVariablesandPriorEnergyBehaviors,inTandemwithExpectationsandAttitudes

FormedtotheMessageTarget

299

Figure 1: The second structural equation model tested in explanation of the dependent variables Adoption Intention

(BehvIntenExpc), and Pessimistic expectations and expectation to employ other methods (PessimPE&AltMeth).

Age categories were compared in MG approach.

Younger than 30 compared to respondents in their

30s: a) Had higher association between PE_SG

(performance expectancy items referring to smart

grid) and Anticipated Positive Affect to behaviors

(beta_category1: 0.25 (sig); beta_ category2: 0,09

(n.s.); p value associated to t test for the difference

in path coefficients (2 tail): 0.05; but they had not

more important paths from PE_SG to Adoption

Intention; b) Maintained less association between

PBC and PE-SG (beta_category1: 0.02 (n.s.);

beta_category2: 0.27 (sig); p value of t test: 0.005; c)

Had less salient path from Pessism_AltMethod to

Intention to Adopt: (beta_category1: 0.11 (n.s.);

beta_category2: 0.25 (sig); p value of t test: 0.06.

Other differences were not quantitatively relevant. In

synthesis, significant and relevant differences

between younger respondents as compared to ones

in their 30s, were that the younger group responded

to Intention to adopting in a way that was less

consistent with their expressed intention to employ

alternative methods, and they gave less importance

to valuations of outcomes of conserving in their

expressed intention to adopt the technology.

Respondents over 40 years of age compared to

the ones in the 30s: a) Had a higher association of

PBC to Intention to Adopt (beta_category3: 0.27

(sig) against beta_category2: -0.12 (n.s.); p value of

test for the difference 2-tail: 0.003); b) Lower

importance of Anticipated Positive Affect upon

Intention to Adopt (beta_category3: 0.08 (n.s.)

against beta_category2: 0.26 (sig); p value of t test

(2 tail): 0.03); c) lower importance of Performance

Expectation (items referring to current grid) to

Anticipated Positive Affect (beta_category3: -0.01

(n.s.) against beta_category2: 0.23 (sig); p value

associated to t test: 0,002).Thus, respondents older

than 40, compared to those in their 30s, had higher

salience of PBC in Adopting Intentions, but lower

salience of Anticipated Positive Affect to the

behaviors.

Group differences due to educational level and

occupational field were analyzed. Educational level

assumed 2 levels: 12

grade (educ_level 1) and

higher education (educ_level2). Occupational field

was classified according to the degree of importance

of ICT technology. Importance of computers and

technology for the occupation was coded after

Portuguese occupational taxonomies (e.g., taxon of

Technology field Portuguese Inventory of

Occupational Preferences). This lead to organizing 2

categories: lower use of technological knowledge

and skill about computers, software and

programming, or higher. There was some overlap

between technology load and educational level

categories (Pearson's Phi of 0.303).

These variables were first checked regarding

their potential role as predictors of the dependent

variables, and either non-significant or extremely

low, effects were obtained. Relevant differences

SMARTGREENS2014-3rdInternationalConferenceonSmartGridsandGreenITSystems

300

among occupational level and field/technology

categories were found in path coefficients (thus

manifesting moderating roles of these variables)

statistically significant at alpha level of 0.05, in a

two tail adapted t-test. Respondents with higher

education, comparatively with those with 12

grade,

gave lower importance to Performance Expectation

in forming Adopting intentions (beta_educlevel_1:

0.21 (sig); beta_educlevel_2: 0.07 (n.s.); t test (16.66

d.f.) =2.48, p(2 tail)=0.025). Thus for educational

level, a significant and relevant effect was that the

more highly educated respondents seemed to be less

driven in their adopting considerations by

performance expectations. Respondents whose

occupation were classified as less infused with ICT

tasks and skills (category 1) displayed higher paths

between PE_SG and two other constructs. One of

these constructs was PBC, so that performance

perception depended upon the PCB significantly or

the reciprocal, whereas this is not the case with more

ICT loaded occupations (Path_Categ1: 0.22 (sig);

Path_Categ2: 0.05 (n.s.), t(21.28 d.f.)=3.86,

p(2t)=0.001). The other construct more highly linked

to PE_SG by Category-1 group was Anticipated

Positive Affect: They accorded higher importance to

PE_SG in their ratings of positive anticipated affect

to adoption. (Path_categ1: 0.18 (sig);

Path_Category2: 0.05 (n.s.), t(25.19)= 2.64;

p=0.014.

The same former group also had more strong

negative association between Social Image and

Negative Affect (so that Social Image factors implied

the presence of more anticipated negative affect), a

link that was non-significant for the group with more

technologically loaded occupations. (Path_Categ1: -

0.30 (sig); Path_Categ2: -0.07 (n.s.); t(16.66)=3.44;

p(2t)=0.003).

4 ENERGY BEHAVIORS

The concept of Intelligent Efficiency positions

technology and people’s energy behaviors as

integrated parts of efficient solutions, to which

accrues demand-response management as policy on

tariffs and prices, in the transition toward more

intensely digitalized grids. Energy behaviors (Black

et al., 1985) are a set of different kinds of

behaviors theorized to impact the way energy is

consumed at home and the energy consumption

outcome e.g., switching off lights in unoccupied

rooms; checking the invoice, buying efficient

equipment; performing small thermal improvements

in the house, but to those we added the use of

automation or programming for setting goals (e.g.

temperature). For 431 participants for whom a

subset of energy behaviors were jointly applicable,

self-reported frequency of these behaviors was not

highly correlated overall. Instead, a subset of

behaviors could be used to compose a consistent

formative construct, namely: buy efficient

equipment, checking invoice; switch off equipment

at button and perform small improvements in

thermal insulation of the house. Two other behaviors

(regulating ambient temperature according to

season; and employing programmer) were studied,

treated with a single indicator variable approach.

Frequency in the case of highly repetitive behaviors

controlled by specific cues may constitute habit.

Hypothesis on relations between energy

behaviors and adoption intention can be derived

from TPB in Conner and Armitage’s (Conner and

Armitage, 1998); (Armitage and Christian, 2003)

extended version with past behavior as a factor,

possibly mediated by self-identity as discussed by

Smith, Terry and Manstead (2008). If this kind of

influence predominates, different sets of energy

behaviors may be direct predictors of each of the

dependent variables. Hypotheses relating those

behaviors - when accessible to memory - to attitudes

originate from frameworks in Attitude literature

favoring the behavior–attitude link, namely self-

perception theory, under conditions where pre-

existing attitudes might not be clear nor supported

by knowledge structures (Eagly and Chaiken, 1993);

(Olson and Stone, 2005). Thus self-report of specific

sets of energy behaviors is expected to covary with

Valuation of Outcomes of Conserving, and also with

Attitudes, either in favor or counter the adoption of

the technology depending upon the behaviors’

overlap with technology use. This association may

have different causes, such as salient motives for the

behaviors, consistency effects of self-perception and

self-inference of attitude, and even possible social

desirability bias.

Distinct hypotheses on potential moderating role

of prior behaviors of a similar kind as the

dependent(s), within TPB framework, are

formulated by Kidwell and Jewell (Kidwell and

Jewell, 2008) by theorizing their role as sources of

activation of specific heuristics in the information

processing for decision. As prior experience with the

behaviors calls attention to information obtained

from experience, it is then expected to change the

importance of TPB variables upon behavior

Intention/ behavior considerations, by decreasing the

importance of Attitude (outcome probability), and

by increasing the importance of behavioral control

AcceptationofaDemand-responseEnablingTechnologyforusingElectricityatHomeuponaSimulatedMarketing

Campaign-RoleofSociodemographicVariablesandPriorEnergyBehaviors,inTandemwithExpectationsandAttitudes

FormedtotheMessageTarget

301

considerations. But, conversely, if the accessibility

to memory of sets prior behaviors is the starting

point for consumers’ inference of own attitudes to

the technology adoption, the salience of attitude

(e.g., positive attitude) in adoption considerations is

not expected to decrease for the respondents with a

high frequency of similar behaviors.

Self-ratings of frequency of a sample of energy

behaviors were part of the survey. In order for

energy behaviors to be analyzed together, they

needed to be applicable in the respondents’ home

with an identical rate (because energy behaviors

derive from activation of energy services, they are

applicable if the corresponding services are chosen

and activated by consumer at home). Groups of

respondents who indicated the behaviors as applying

were the basis for partitioning the sample of

behaviors and the sample of respondents into

subsets. For two behaviors - employing

programming devices in setting parameters for

ambient temperature for a longer time ahead and

adjust ambient temperature according to

different benchmarks in winter and in

summertime – applicability largely overlapped. A

set of other behaviors afforded a consistent

formative construct: buy efficient, checking

invoice, switch off at button and thermal

insulation. Employing a programmer is

considered the most similar behavior to Adoption,

whereas AltMethods may be more similar behaviors

such as keeping doors closed of the rooms being

heated, or switching off equipment using the

button, and checking the invoice.

4.1 Analysis and Results

In a first step, a model where each behavior is

entered as a predictor of the main dependent variable

is estimated and analyzed. In a second step, two sets

of moderating effects are checked for: the potential

moderation of Attitude upon Intention, in the present

case, Positive Affect upon Adoption Intention, and

also Negative Affect upon Pessismim_AltMethod;

and potential moderation of PBC upon Adoption

Intention, and by extension, the moderation of EE

upon Adoption Intention and Pessism_Alt Method.

Employing a programmer had a significant

relation to Effort Expectancy (0.21; t=2.26); but not

to Positive Affect, and, differently from other energy

behaviors, no association to valuation of outcomes of

conserving electricity.

Adjusting ambient temperature according to

season had a significant association with Valuations

of outcomes of conserving (0.21; t=2.72),

Performance Expectancy_CG (0.17; t=2.95).

A formative construct composed of 4 energy

behaviors - buy energetically efficient equipment;

checking invoice; switch off equipment at button;

thermal insulation of the home, displays

significant paths with Valuations of conserving and

with Performance Expectancy_CG (0.24; t=4.88);

with Effort Expectancy (0.15; t=3.20) ; with Social

Image (0.10; t=2.12), with Negative Affect to

behaviors (0.11; t=2.68);

Results of moderator analysis: Employing a

programmer: does not significantly (at alpha 0.05)

interact with Pos Affect upon Intention to Adopt

(0.10; t= 1.14); nor with Neg Affect in predicting

Pesssimis-AltMethod (-0,08; t= 1,22) although there

is a trend in that direction. The second interaction is

significant at 0.11, in a one-tail test. The latter

interaction is negative, as predicted in the

hypothesis that prior behaviors lower the importance

of attitudes to adoption considerations. But,

regarding its potential role in interaction with PBC,

there is no significant interaction, nor with Effort

Expectancy.

Adjusting ambient temperature according to

season does not display significant interaction with

Positive affect

in prediction of Adoption Intention;

nor with Negative Affect in prediction of Alt Method.

No significant interaction exists with PBC. With EE,

however, the interaction reaches a statistically

significant coefficient of 0.196 in prediction of Alt

Method (t=2.89) which is high. Thus, with higher

behaviors of adjusting ambient temperature

seasonally there is a trend towards an increase in

importance of EE in choosing alternative methods.

5 CONCLUSIONS

Gender, age and occupational groups may differ

among each other in several aspects that potentially

impact adoption of technological solutions for

managing the use of electricity at home, and for

conserving in the current conditions, and in the

future scenario of the smart grid. Being of female, in

contrast to male, gender did not have a significant

impact in adoption intention, but it did change the

importance of considerations regarding outcomes

(performance) and behavior barriers (PBC),

lowering them, and the importance of perceptions of

ease of learning and of using the technology (EE),

increasing it.

The emphasis upon behavior control (behavioral

barriers) considerations of older respondents in

thinking about technology adoption appears as a

theme in several IT acceptance studies and also in

SMARTGREENS2014-3rdInternationalConferenceonSmartGridsandGreenITSystems

302

our study. Additionally, our results suggest that this

age group (above 40years of age) may also be less

concerned with outcome probabilities (performance)

and report less importance of hot, emotional,

thoughts around the issue of adopting.

Much younger groups (below 30s in comparison

with those in their 30s) may be less concerned with

consistency between performance expectancy i.e.,

outcome expectations and behavior control, and

more concerned with coherence between outcome

(performance) and affect to behaviors. Younger

respondents may be also less concerned with the

consistency between alternative (or competitive)

methods of managing the use of electricity.

Regarding education level, the more highly

educated respondents seemed to be less driven in

their adopting considerations by performance

expectations, compared to those with high school

education. A possible explanation is that more

highly instructed respondents may have lower

pressure to perform in reducing and controlling

energy use in their homes, and this is not the primary

motivational source of adoption; but a different

explanation is that they may require more precise

information to elaborate a perception that they can

feel as a bases for decision.

Educational and occupational groups differing in

IT knowledge do not differ in regard to the most

structural paths with higher coefficients in the

overall model: From Social Image to Anticipated

Positive Affect, and to Adoption Intention; from

Effort Expectancy to Positive Affect and to Adoption

Intentions; and from PBC to Negative Affect and to

Pessimism_Alternative Methods. The structural

model seems to hold for both groups analyzed.

However, less technologically informed respondents

had a more consistent or uniform pattern of

responses in relating Performance expectations

(outcome probabilities they express), PBC, and

positive affect to adoption behaviors, while more

instructed and technology informed respondents

have more differentiated patterns.

Energy technology acceptation takes part with

energy behaviors, e.g., conservation, monitoring,

investments, as well as the use of automation or

programming for setting goals (e.g. temperature) or

automating decision. An integrated view of energy

efficiency, under the conceptual umbrella of

Intelligent Efficiency, positions technology and

people’s energy behaviors as integrated parts of

efficient solutions, and furthermore, electrical grid

changes towards the digital grid are integral with

new market and economic policies on tariffs,

including demand-response. Hypotheses on the role

of prior energy behaviors of a similar kind as the

dependent(s), stem from their postulated role as

sources of inferences on own attitudes towards

adoption, or alternatively, as sources of activation of

specific heuristics in the information processing for

decision, by changing the importance of TPB

variables Attitude and PBC upon Intention/ behavior

considerations. But in the larger context, acceptation

of new technology is also designed to help

consumers change their energy behaviors.

Energy behaviors appear to constitute a

multifaceted construct. Prior energy behaviors are

correlated to some predictors of adoption intention,

or to alternative methods. A behavior that

presupposes automation technology (employing a

programmer) enhances EE, leading to higher

perception of ease of learning to use the new

technology, but they were not found to have a role in

attitudes, i.e. affect, nor to alter the importance of

(positive or negative) attitudes/affect and behavior

control, upon considerations of adoption. But

another behavior - Adjusting ambient temperature

according to season incremented Performance

Expectation, and led to a higher dependence of

Expectation to employ alternative methods from

Effort expectancy. Other behaviors studied enhance

almost all predictors of adoption, including absence

of negative anticipated affect. In this regard, the

results in our study are mixed, not leading to closure

in the competing hypotheses of processes whereby

energy behaviors and adoption considerations meet.

Goals and hypotheses are rooted in well-

established attitude, persuasion, and IT acceptance

literatures, and jointly contribute to advancing the

understanding of issues and processes in the attitude

formation for developing technologies for smart grid

adaptation, and identifying consumer segments in

regard to the adoption of this type of technologies.

ACKNOWLEDGEMENT

This work has been framed under the Energy for

Sustainability Initiative of the University of Coimbra

and partially supported by the Energy and Mobility

for Sustainable Regions Project (CENTRO-07-0224-

FEDER-002004) and by Fundação para a Ciência e

a Tecnologia (FCT) under grant

SFRH/BD/51104/2010 and project grants

MIT/SET/0018/2009 and PEst-C/EEI/UI0308/2011.

AcceptationofaDemand-responseEnablingTechnologyforusingElectricityatHomeuponaSimulatedMarketing

Campaign-RoleofSociodemographicVariablesandPriorEnergyBehaviors,inTandemwithExpectationsandAttitudes

FormedtotheMessageTarget

303

REFERENCES

Abrahamse, W., Steg, L., 2009, How do socio-

demographic and psychological factors relate to

households’ direct and indirect energy use and

savings?, Journal of Econ. Psychol., 30, 711-720.

Armitage, C. J. and Christian J., 2003, From attitudes to

behavior: Basic and applied research on the theory of

planned behavior. Current Psychol, 22, 187-95.

Bagozzi, R. Z., Gürnao C. and Priester, J., 2002, The

Social Psychology of Consumer Behav, O.U. Press.

Black, J. S., Stern, P. C. and Elworth, J. T., 1985, Personal

and Contextual Influences on Household Energy

Adaptations. Applied Psychology, 70(1): p. 3-21.

Conner, M. and Armitage, C., 1998, Extending the theory

of planned behavior: review and avenues for further

research, J. Appl. Social Psych, 28, 1429.

Davis, F. D., Bagozzi, R. P. and Warshaw, P. R.., 1989,

User Acceptance of Computer Technology,"

Management Science, 35, 982-1003.

Dillon, A., 2001, User Acceptance of Information

Technology, In Karwowski W (ed.), Encyclopedia of

Human Factors & Ergon., Taylor and Francis.

Eagly, A. H. and Chaiken, S., 1993, The impact of

behavior in attitude formation and change. In The

Psychology of Attitudes, Wadsworth.

Elliott N, Molina M, Trombley D, “A Defining

Framework for Intelligent Efficiency”, ACEEE

Research Report E125, 2012

Fishbein, M.; Ajzen, I., 2005, Predicting and changing

behavior: The Reasoned Action Approach, Taylor and

Francis.

Gadenne, D. Sharma,B., Kerr & Smith, 2011, The

influence of consumers’ environm. beliefs in attitudes

and energy saving behaviors, Ener.Pol..

Kidwell, B. & Jewell, R., 2008, The Influence of past

behavior on behavioral intent: an informat.-processing

explanation. Psychology and Marketing, 25,2,1151-66.

Livengood D. and Larson, R., 2009, The Energy Box:

Locally automated optimal control of residential

electricity usage”, Service Science, 1(1), 1-16.

Lopes, M. A. R., C. H. Antunes, and N. Martins, Energy

behaviors as promoters of energy efficiency: A 21st

century review. Renewable and Sustainable Energy

Reviews, 2012. 16(6): p. 4095-4104.

Olson, J. M.; Stone, J., 2005, The influence of behavior on

Attitudes, In D.Albarracín, B. T. Johnson & M. P.

Zanna (eds.), The Handbook of Attitudes, Taylor and

Francis.

Oreg S. and Katz-Gerro, T., 2011, Predicting pro-

environmental behavior cross-nationally, Environment

and Behavior, 38,4, 462-483.

Santacana E., Rackliffe G., Tang L., Feng X. (2010).

Getting Smart: IEEE, 8:2, 41-48.

Smith, J. R. Terry, D. J. Manstead, A. S. et al., 2008, The

attitude-behavior relationship in consumer conduct, J

Soc Psychol; 148(3), 311-33.

Stern, P. C., 2000 “Toward a coherent theory of

environmentally significant behavior”, Jour.Social

Issues, 56(3), 407–24.

Taylor, S. and Todd, P. A., 1995, Understanding

information technology usage, Information Systems

Research, 6, 2, 144-176.

Taylor, S.; Todd, P. A., Assessing IT usage role prior

experience. MIS Quarterly,19,4,561-70.

Venkatesh, V. Morris, M. G. and Ackerman, P. L., 2000,

A longitudinal field investigation of gender differences

in ind. technology adoption decision making

processes. Organiz. Behavior and Human Decision

Processes, 83,1, 33-60.

Venkatesh, V. & Morris, M. G., 2000, Why don’t men

ever stop to ask for directions? Gender, social infl. in

technology acceptance, MIS Quart, 24,1,115-39.

Venkatesh, V., Morris, M. G., Davies, G. B. And Davies,

F. D., 2003, User Acceptance of Information

Technology: Toward A Unified View. MIS Quarterly,

27,3, 425-478.

SMARTGREENS2014-3rdInternationalConferenceonSmartGridsandGreenITSystems

304