Understanding Intention to Adopt Smart Thermostats: The Role of

Individual Predictors and Social Beliefs Across Five EU Countries

Mona Bielig

1,2 a

, Florian Kutzner

, Sonja Klingert

and Celina Kacperski

1,2 c

Seeburg Castle University, Seeburgstraße 8, 5201 Seekirchen am Wallersee, Austria

University of Konstanz, Universitätsstraße 10, 78464 Konstanz, Germany

University of Stuttgart, Keplerstraße 7, 70174 Stuttgart, Germany

Keywords: Technology Acceptance, Smart Thermostats, Social Beliefs.

Abstract: Heating of buildings represents a significant share of the energy consumption in Europe. Smart thermostats

that capitalize on the data-driven analysis of heating patterns in order to optimize heat supply are a very

promising part of building energy management technology. However, factors driving their acceptance by

buildings’ inhabitants are poorly understood although being a prerequisite for fully tapping on their potential.

In order to understand the driving forces of technology adoption in this use case, a large survey (N = 2250)

was conducted in five EU countries (Austria, Belgium, Estonia, Germany, Greece). For the data analysis

structural equation modelling based on the Unified Theory of Acceptance and Use of Technology (UTAUT)

was employed, which was extended by adding social beliefs, including descriptive social norms, collective

efficacy, social identity and trust. As a result, performance expectancy, price value, and effort expectancy

proved to be the most important predictors overall, with variations across countries. In sum, the adoption of

smart thermostats appears more strongly associated with individual beliefs about their functioning, potentially

reducing their adoption. At the end of the paper, implications for policy making and marketing of smart

heating technologies are discussed.

1 INTRODUCTION

Around 40% of energy in the EU is consumed in

buildings, and of building energy consumption about

80% is used for heating

. In order to meet the 2030

target of a 55% reduction in emissions compared to

1990, which heavily involves the building sector

(European Environment Agency (EEA), 2021), there

has been a push for the adoption of smart home

technologies, including smart energy management

technologies (European Commission, 2022). A smart

thermostat, a specific type of smart heating

technology (SHT), connects to the existing heating

system and detects behavioral patterns of residents, in

some cases allows for smart controls, and can save up

https://orcid.org/0000-0001-7535-8961

https://orcid.org/0000-0003-0653-003X

https://orcid.org/0000-0002-8844-5164

https://energy.ec.europa.eu/topics/energy-efficiency/

energy-efficient-buildings/energy-performance-

buildings-directive_en

to 30% energy, depending on the type (Lu et al., 2010;

Wang et al., 2020).

In principal, smart thermostats work in a loop of

detection of behavioral patterns through sensors, and

potentially with the prediction of external events or

temperature, in order to predict dynamic heating

needs under comfort constraints and then provide

optimal requirement to heating supply, as shown in

Figure 1 (Haji Hosseinloo et al., 2020).

Bielig, M., Kutzner, F., Klingert, S. and Kacperski, C.

Understanding Intention to Adopt Smart Thermostats: The Role of Individual Predictors and Social Beliefs Across Five EU Countries.

DOI: 10.5220/0013356200003953

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 14th International Conference on Smart Cities and Green ICT Systems (SMARTGREENS 2025), pages 36-47

ISBN: 978-989-758-751-1; ISSN: 2184-4968

Figure 1: Design of smart thermostats (Haji Hosseinloo et

al., 2020).

Smart home technology may not yet be widely

perceived as mainstream (Chang & Nam, 2021), and

its rate of adoption has been characterized as

relatively slow (Marikyan et al., 2019). However,

market trends suggest a gradual expansion in the

sector (Sovacool & Furszyfer Del Rio, 2020).

Eurostat figures from 2022 reveal that about 10% of

European households have incorporated smart home

technologies, including but not limited to energy

management systems.

Figure 2: Left: Smart heating technology used as the basis

for technical design. Right: General technology picture

shown to participants.

Notably, the widespread rollout of smart

thermostats not only requires the availability of

technical equipment, but also households’

willingness to adopt the smart thermostats, and a

consent to give up some control. In order to test

technology acceptance, a real life example was used

(see Figure 2) to describe the way that a smart

thermostats works to non-scientific participants of a

survey. In the following, we will first describe the

underlying theoretical models and related work on

predictors for smart thermostat adoption, resulting in

our proposed model and hypotheses. We will then

describe our study design to understand intentions to

adopt smart thermostats in five European Countries,

followed by an overview on our most important

We will further refer to the UTAUT-model comprising

research based on both UTAUT and UTAUT2.

results. We end with discussions and conclusions of

our work, including limitations and strengths as well

as policy implications.

2 RELATED WORK

Previous acceptance research has focused mainly on

smart home technologies in general (for a review, see

Li et al., 2021), with research on smart heating

technologies in particular under-represented. This is

problematic, as behavioral flexibility in heating is

lower than for other appliances (Spence et al., 2015).

The existing research on acceptance of smart heating

technology concentrates on individual and technical

factors (e.g., Girod et al., 2017); social drivers for

acceptance are either absent or results inconsistent

(Große-Kreul, 2022). As both the diffusion of

innovative technologies, and pro-environmental

decisions have been shown to be driven by social

aspects (Fritsche et al., 2018; Rogers, 2003), the

current research deepens the understanding of

acceptance of a smart heating technology, smart

thermostats, combining individual and social aspects

in one model.

2.1 UTAUT and Smart Technology

Acceptance

The Unified Theory of Acceptance and Use of

Technology (UTAUT) (Venkatesh et al., 2003) was

developed to summarize eight different technology

acceptance models and has been successfully

employed across multiple contexts such as mobility,

IoT in health care or mobile payment (Abrahão et al.,

2016; Arfi et al., 2021; Nordhoff et al., 2021). It was

extended to the UTAUT2 (Venkatesh et al., 2012) to

better align with a consumer context. The full model

includes seven predictors for behavioral intention to

use a technology, which only then translates into

actual user behavior. The seven predictors are

performance expectancy, effort expectancy, social

influence, facilitating conditions, hedonic motivation,

price value and habit.

While the UTAUT

, or sometimes a subset of its

predictors, have been studied in the context of

acceptance of different smart home technologies, the

predictive capacity of its components has been

inconsistent across studies, with the exception of

performance expectancy, which has repeatedly been

shown to best predict intention to adopt, both for

Understanding Intention to Adopt Smart Thermostats: The Role of Individual Predictors and Social Beliefs Across Five EU Countries

smart energy technologies in general (Gimpel et al.,

2020), and for smart thermostats in particular (Ahn et

al., 2016; Girod et al., 2017; Große-Kreul, 2022;

Mamonov & Koufaris, 2020).

Effort expectancy, that is the perceived ease of

use, was amongst the strongest predictors for

behavioral intention to adopt smart energy

technologies, i.e. energy-saving technologies

comprising sensors and automatic control in a Danish

sample (Billanes & Enevoldsen, 2022), and smart

homes in a sample from Jordan (Shuhaiber & Mashal,

2019). However, it showed no effects for the intention

to adopt smart thermostats in an US sample (Ahn et

al., 2016). Hedonic motivations were important for

the adoption of smart thermostats in a German sample

(Girod et al., 2017), while in a different German

study, they were irrelevant (Große-Kreul, 2022).

Similar inconsistencies can be found for price value,

which had no effect on intention to adopt a smart

thermostat (Girod et al., 2017), but was found

relevant in a discrete choice experiment (Tu et al.,

2021).

The UTAUT also includes the factor social

influence, i.e., the belief that important others think

an individual should use the technology, which will

be discussed in the next section (2.2). Beyond the

UTAUT, other technology acceptance factors such as

privacy concerns or compatibility have been

examined in studies on intention to use or adopt smart

thermostats, smart homes, or smart meters.

2.2 Social Beliefs as Predictors of

Smart Technology Acceptance

Within the UTAUT, the factor of “social influence”

depicts the belief that important others think an

individual should use the technology (Venkatesh et

al., 2003). Social influence is thus understood as a

social norm in the sense of the Theory of Planned

Behavior (TPB; Ajzen, 1985). This type of norm

describes an individual’s perception of what others

expect them do and reflects the normative belief or

social pressure of ‘ought’ of (important) others,

which is also labeled as an injunctive norm (Cialdini,

2007; Cialdini et al., 1990; Göckeritz et al., 2009;

Rivis & Sheeran, 2003). Studies for acceptance of

smart thermostats that included social influence

found varying effects (Ahn et al., 2016; Billanes &

Enevoldsen, 2022; Gimpel et al., 2020; Girod et al.,

2017): While it has predicted smart thermostat

Supplementary material can be found under

https://osf.io/ba2vf/?view_only=986065e170584cad90

98d0a2937e216b

adoption intention in Germany (Große-Kreul, 2022)

and smart meter adoption in Brazil (Gumz et al.,

2022), there were very small or no significant effects

in other studies (Ahn et al., 2016; Gimpel et al., 2020;

Girod et al., 2017). In the two studies in which social

influence had the strongest effect on behavioral

intention to adopt smart thermostats and smart meters

(Gumz et al., 2022), the operationalization included

additional aspects less reminiscent of an injunctive

norm, such as sheer perceptions of presence of smart

thermostats in media, or recommendations by the

government. An overview is given in the

supplementary material of our article (SM1)

The perception of what other do or believe

corresponds to the psychological social norm

approach (Berkowitz, 2004), which differentiates

between injunctive and descriptive norms. While

injunctive norms capture the priorly mentioned

normative belief of social pressure, descriptive norms

refer to an individual’s belief about the prevalence of

a behavior, i.e. of what is “normal”, therefore the

perception of others' own attitudes and behaviors in

the domain (Cialdini, 2007; Rivis & Sheeran, 2003).

Further, research shows that descriptive and

injunctive norms might differ in their effects in

changing behavior (e.g. Park & Smith, 2007; White

et al., 2009). As prior research failed to demonstrate

an effect of injunctive social norms for smart

thermostat uptake, e.g. in Ahn et al. (2016) or Girod

et al. (2017), we aim to investigate descriptive social

norm perceptions as a possible driver of technology

acceptance for smart heating.

Beyond descriptive social norms, there are several

other social beliefs that have been shown to influence

pro-environmental behavior, which might

complement or interact with social norms to promote

the adoption of smart thermostats. Prominently, the

social identity model of pro-environmental action

(SIMPEA) (Fritsche et al., 2018) depicts the influence

of social identity processes for pro-environmental

behavior, also related to the adoption of “green”

technologies. Relevant predictors are collective

efficacy beliefs, i.e. people’s beliefs in the

effectiveness of their combined ability to achieve

goals, and social identification, i.e. the degree to

which relevant group memberships are considered

important for the individual. Further, generalized

social trust, referring to general trust in others across

groups, and trust in the state, i.e. government and

institutions were found decisive for intentions of pro-

SMARTGREENS 2025 - 14th International Conference on Smart Cities and Green ICT Systems

environmental behavior (Cologna & Siegrist, 2020),

and individual energy-saving behavior (Caferra et al.,

2021). We therefore explore whether these social

beliefs might drive the adoption of smart heating

technology. The following section will explicate the

definition of our predictors, the prior findings we

build them on, then form our hypotheses.

3 MODEL PREDICTORS AND

HYPOTHESES

We examine predictors of the intention to adopt smart

heating technology, specifically combining

individual beliefs incorporated in the UTAUT, and

different social beliefs which have been shown to

affect pro-environmental behavior decisions, energy

saving intentions and uptake of smart technologies in

prior research. Our goal is to examine whether social

beliefs can explain the intention to adopt smart

heating technology in addition to beliefs about

technical aspects. We decided to exclude habit,

facilitating conditions and hedonic motivation from

our model, as no experience with the technology is

expected, no active usage is required, and no

additional infrastructure beyond the thermostat itself

is necessary.

Behavioral intention (BI) is our key dependent

variable, reflecting the willingness to adopt smart

heating technologies, and is the strongest predictor of

technology use, especially for technologies with

limited consumer experience (Venkatesh et al., 2012).

To explain behavioral intention, we consider the

following predictors, both based on literature of

technology adoption and wider pro-environmental

behaviors:

Performance expectancy (PE) refers to the

perceived usefulness of the technology in achieving

specific goals and was found to be a significant

predictor of intention to adopt smart energy

technologies in multiple studies (Venkatesh et al.,

2012; Gimpel et al., 2020; Ahn et al., 2016). Effort

expectancy (EE), another core UTAUT construct,

captures the perceived ease of use and technical

efficacy of the technology (Venkatesh et al., 2003),

and evidence suggests it plays a role in smart energy

technology adoption (Billanes & Enevoldsen, 2022;

Ahn et al., 2016). Price Value (PV) is the perceived

trade-off between the cost of technology and its

benefits, which has been shown to influence

technology adoption (Venkatesh et al., 2012; Tu et

al., 2021).

Social norms (SN) reflect the influence of others'

perceived behavior on individual intentions.

Descriptive neighborhood social norms have been

found to influence pro-environmental behaviors

(Farrow et al., 2017; Allcott, 2011). Social

identification (SI), can moderate this effect of social

norms on pro-environmental behavior, i.e. it interacts

with the originally expected influence of social norms

and might modify it (Cialdini & Jacobson, 2021;

Masson & Fritsche, 2014). Collective efficacy (CE),

or the belief in the collective ability to achieve

environmental outcomes, is another key predictor of

pro-environmental behavior (Bandura, 2000; Wang,

2018), and can compensate for low individual

efficacy (Jugert et al., 2016).

Lastly, trust is a significant predictor of pro-

environmental behavior, encompassing both general

trust in people (TP) and trust in state institutions (TS).

Higher trust in others and in institutions has been

shown to influence pro-environmental behaviors,

particularly energy-saving intentions (Cologna &

Siegrist, 2020; Caferra et al., 2021). Demographic

variables like age and gender are also considered as

control variables, given their influence on smart home

adoption (Shin et al., 2018; Sovacool et al., 2020).

This leads us to the following Hypotheses:

H1 Performance expectancy has a positive effect on

BI to adopt smart heating technology.

H2 Effort expectancy has a positive effect on BI to

adopt smart heating technology.

H3 Price value has a positive effect on BI to adopt

smart heating technology.

H4 Social norms within the neighborhood have a

positive effect on BI to adopt smart heating

technology.

H5 Social identification moderates the effect of

social norms, with stronger effects on BI in case of

higher neighborhood identification.

H6 Collective efficacy beliefs have a positive effect

on BI to adopt smart heating technology.

H7 General trust in people has a positive effect on

BI to adopt smart heating technology.

H8 General trust in state has a positive effect on BI

to adopt smart heating technology.

Understanding Intention to Adopt Smart Thermostats: The Role of Individual Predictors and Social Beliefs Across Five EU Countries

3.1 Model

Figure 3 depicts our research model, including both

the structural and measurement model. Predictors are

split into ‘individual (left) and ‘social’ (right) beliefs.

Figure 3: Research model, including both structural and

measurement model. All item abbreviations correspond to

the item codes in the supplementary material.

4 METHOD

Our survey consisted of items related to a smart

heating thermostat, its acceptance and the individual

and social beliefs that might influence intention to

adopt. The survey started with an introduction, a short

description of a smart heating device together with an

image (see Figure 1), and consent procedures.

Afterwards, participants were asked to answer all

items included in the questionnaire, ending with

demographic details

4.1 Sample Description

Our sample of N = 3227 was recruited by a

professional panel provider and stratified by age and

gender for five European countries (Austria,

Belgium, Estonia, Germany, Greece). Renumeration

was based on the provider’s usual rates. Data was

gathered through an online survey link between

21.07.2021 and 10.08.2021. Data collection was

anonymous and in line with the ethical guidelines of

the DGPs (DPGs, 2016). All items were translated

into each country’s native language by native

speakers and back-translated to check for accuracy;

Together with the device image, we randomly assigned

participants to a control group, and groups with financial

and environmental (CO2) savings information, and a

group that was presented an app that would facilitate

control of the smart device. We did not find any

significant differences between these groups on any of

our model or dependent variables and therefore will not

discuss this intervention further.

each translation was reviewed by a researcher with a

native speaker multiple times.

We excluded participants who did not finish the

survey, failed the attention check, had an average

relative speed index of >2, and used the careless

package (Yentes & Chevallier, 2021) to exclude

participants with a longstring > 25

. This led to a final

sample of N = 2250 (51.3% women) from five

different countries, with Austria N = 465 (49.7%

women), Belgium N = 414 (52.9% women), Estonia

N = 510 (52.5% women), Germany N = 425 (50.6%

women), and Greece N = 436 (50.7% women).

Overall sample size, distributed similarly between

countries, was chosen to enable both an overall

structural equation modeling (SEM, Hair et al., 2021),

as well as country-specific analyses

. Participants’

distribution across age brackets was 21.7% 18-30

years old, 34.6% between 30-50 years old, 42.1%

between 50-70 years old, and 0.4% over 70 years old.

Regarding their living situation, 1393 participants

(62%) owned their home, while 853 participants

(38%) indicated to rent their living space. Most

survey participants lived in households between two

to four people (87%), with 37% of the households

having children. 34% of participants reported heating

with a boiler, 22% reported district heating, 17%

reported electric heating, and 27% reported using

other ways of heating.

4.2 Measurements and Scales

We examined predictors of acceptance of smart

heating technology, specifically combining in one

model UTAUT predictors and social beliefs that have

been shown in the past to affect technology adoption

and/or pro-environmental decisions, in-line with the

model shown in Figure 1. Items were surveyed on a

seven-point Likert scale ranging from “strongly

disagree” to “strongly agree” unless indicated

otherwise. We measured Behavioral Intention with

four items, e.g., "If I had the opportunity, I would opt

for a smart heating appliance" (Abrahão et al., 2016;

Venkatesh et al., 2003). Performance Expectancy was

assessed with four items, e.g., "I believe by using such

a smart appliance, I would save a meaningful amount

of greenhouse gases" (Girod et al., 2017; Venkatesh

The cut-off criterion is number of items until the first item

was reverse recoded.

Based on an A-priori Sample Size Calculator for

Structural Equation Models, with the specifications of

our model, with a desired statistical power level of 0.8

and an estimated effect size of 0.2, we needed N = 425

observations to find an effect.

SMARTGREENS 2025 - 14th International Conference on Smart Cities and Green ICT Systems

et al., 2003). Effort Expectancy was measured using

four items, e.g., "I believe that using such smart

heating control would be: difficult – easy" on a 7-

point scale (Venkatesh et al., 2012, 2003). Price

Value included two items, e.g., "Such a smart heating

appliance is good value for the money" (Venkatesh et

al., 2012). Collective Efficacy was measured with one

item, "If a large portion of the population used the

smart device, we would have a positive effect on

society and the climate" (Chen, 2015; Wang, 2018).

Social Norms were assessed with three items, e.g., "I

believe most of my neighbours will adopt such a

technology" (Lazaro et al., 2020; White et al., 2009).

Social Identification was measured with a pictorial

representation to assess the relationship with the

community, based on the "Inclusion of the Other in

the Self" scale (Aron et al., 1992; Gächter et al.,

2015), ranging 5 points. Trust in People was

measured using two items, e.g., "Generally speaking,

most people can be trusted" (European Value Study

(EVS); Caferra et al., 2021), and Trust in State with

two items, e.g., "Please rate how much you trust in

your legal system" (EVS; Caferra et al., 2021).

4.3 Data & Statistical Procedure

Data was handled with R statistics (R Development

Core Team, 2008). Before conducting the analyses,

all variables were mean-centered. Descriptive

statistics are reported of raw scores. Building on a

recommended two-step approach by Anderson &

Gerbing (1988), we first conducted a Confirmatory

Factor Analysis (CFA) before structural modelling to

assess the fit of the measurement model. This step

aims to estimate the measurement relationships

between the observed variables and their underlying

latent variables that cannot be directly assessed. The

latent variables in our measurement model are

behavioral intention (BI) as dependent variable, and

performance expectancy, price value, effort

expectancy, social norms, trust in state and trust in

people, collective efficacy, and social identification.

To examine the psychometric properties of our

measurement model, we used indicator reliability,

internal consistency reliability, convergent validity,

and discriminant validity as quality criteria. To test

our hypotheses, a SEM was then calculated, which

through the weighting of factors makes it possible to

quantify the strength of the relationship between the

latent variables.

We calculated the model with and without trust in people,

which did not change the results.

5 RESULTS

The measurement model consists of the latent

variables and their underlying scale items for

observation. Psychometric properties of our latent

variables including item loadings can be found in the

supplementary material (SM2). All loadings

exceeded the recommended threshold of 0.7 on their

respective scales except the first ‘trust in people’

item. This is further reflected in the results, as both

McDonalds omega (ω, McDonald, 1999) and

Cronbach's alpha (α , Cronbach, 1951) exceeded the

recommended threshold of 0.7 for all cases, which

confirms internal consistency reliability, except for

trust in people. The same is true for the average

variance extracted (AVE), where all values exceeded

the recommended threshold of 0.5 (ranging from .69

to .90), while trust in people just met the minimum

criteria with an AVE of .50. We will therefore

interpret findings regarding trust in people with

caution. Nevertheless, as we assessed trust in people

through a well-established scale (based on the EVS),

we decided to keep it in our model

5.1 Psychometric Properties of the

Measurement Scales

Building on recommendations of Hair et al. (2018),

we assessed factor loadings for our constructs, as well

as reliability indicators and mean variance extracted

of items loading on the respective construct. We

summarized these results for all model components in

the supplementary material (SM2). Overall, the

reliability of the scales was strong, with Cronbach’s

α ranging from .87 to .94 for key constructs such as

Behavioral Intention, Price Value, Effort Expectancy,

and Performance Expectancy. The Average Variance

Extracted (AVE) values were generally high (.62-

.80), indicating good convergent validity for most

scales, though Trust in People had a lower reliability

(Cronbach's α = .62, AVE = .49). To evaluate overall

model fit, we used two absolute fit indices (RMSEA,

SRMR) and two incremental fit indices (CFI, TLI)

which are all recommended for models with large

sample sizes (Hair et al., 2018). The fit indices for

both our CFA model and SEM model showed a good

fit, as depicted in Table 1: All parameters of goodness

exceeded the pre-defined cut-off, based on Hair et al.

(2018).

RMSEA – Root Mean Square Error of Approximation;

SRMR – Standardized Root Mean Square Residual; CFI

– Comparative Fit Index; TLI – Tucker Lewis Index

Understanding Intention to Adopt Smart Thermostats: The Role of Individual Predictors and Social Beliefs Across Five EU Countries

Table 1: CFA and SEM fit results, including fit cut-off.

Measure CFA SEM Cutoff

CFI 0.928 0.922 >0.9

TLI 0.907 0.906 > 0.9

RMSEA 0.077 0.071 < 0.08

SRMR 0.036 0.043 < 0.08

For discriminant validity, we found that in most

cases, both the Fornell-Lacker Criterion (Fornell &

Larcker, 1981) and in all cases, the conservative

Hetereo-trait-mono-method (HTMT) criterion

(Henseler et al., 2015) were met: The AVE values of

each construct (in cursive) were higher than their

squared correlations and the inter-construct

correlations are below .85. Only for performance

expectancy, we found an AVE which is below the

squared correlations with behavioral intention, price

value and collective efficacy. Still in this case, the

conservative HTMT criterion was met (Henseler et

al., 2015), which is why we accept discriminant

validity to be given. Results are detailed in SM4.

5.2 Descriptive Results

Descriptive statistics of raw scores for the key

constructs are reported in Table 2. BI to accept the

smart heating technology was not normally

distributed (Shapiro-Wilk-test: p < .001), with a mean

value of 5.44 (SD = 1.51) in the overall sample

. On

average, participants’ ratings for both social and

smart constructs of our model were quite high. We

found the lowest average ratings for social norms.

Generalized trust in people, and effort expectations

were also above the scale mid-point. Social

identification scores were not as high, with on

average 2.4 on scale of 1 to 5.

Table 2: Descriptive statistics of raw scores.

Scale α M SD

Behavioral Intention .94 5.44 1.51

Price Value .88 5.08 1.49

Effort expectancy .90 5.33 1.43

Performance expectanc

.87 4.95 1.37

Social norms .87 4.35 1.50

Trust in

le .62 5.18 1.98

Trust in state .82 5.01 2.33

Social identification

2.39 1.09

Collective efficacy - 5.03 1.49

1 Note that ‘social identification’ was assessed on a scale from 1-5

As data cannot be assumed to be drawn from a normally

distributed population, we calculated all models (CFA

& SEM) with a robust estimator, but find no differences

in results.

Country-specific means and standard deviations

for they key constructs can be found in SM5. We find

a significant difference between countries for BI [F(4,

2245) = 46.9, p <.001]. Post hoc comparisons using

the Tukey HSD test indicated that the BI mean score

for Estonia was significantly higher than for Austria

(p < .001, 95% C.I. = [.21, .72]), Belgium (p < .001,

95% C.I. = [.35, .87]) and Germany (p < .001, 95%

C.I. = [.25, .77]), and BI mean score for Greece was

significantly higher than for all other countries

(Greece – Austria: p < .001, 95% C.I. = [.73, 1.26];

Greece – Belgium: p < .001, 95% C.I. = [.87, 1.41];

Greece – Estonia: p < .001, 95% C.I. = [.27, .77];

Greece – Germany: p < .001, 95% C.I. = [.76, 1.31]).

Further, Greek participants showed the highest

average ratings particularly for individual beliefs,

including price value, performance expectancy and

effort expectancy. Higher age was a negative

predictor for BI (ß = -.10, p = .008). We found no

differences between genders for smart heating

technology acceptance (ß = -.04, p = .480).

5.3 Structural Model

Behavioral intention in our structural model had an R

value of .707, which exceeded the cutoff value of

0.10 for acceptable explanatory power for

endogenous variable (Falk & Miller, 1992). R

values

of all predictor items, reflecting the variance

explained by the corresponding latent variable, were

> 0.5, except for trust in people with .246 Fit indices

for our model confirmed a good fit and Figure 2

shows the model results.

As depicted in Figure 4, for the overall sample,

we found that price value had the strongest positive

influence on BI, followed by performance expectancy

and effort expectancy, in line with H1, H2 and H3. Of

all our included social predictors, only trust in state

had a small negative effect with higher trust levels

reducing the intention to adopt, contradicting our

hypotheses H4, H5 and H6. Additionally, also

contrary to our hypothesis H4, we found a small

negative interaction between social identification and

social norms: social norms had a stronger effect on BI

when people felt less close to their neighbors.

To examine differences between national

samples, we calculated results grouped by country (A

detailed overview of results is in SM7). Across all

five groups, the general pattern was consistent, with

strong significant effects of price value and effort

SMARTGREENS 2025 - 14th International Conference on Smart Cities and Green ICT Systems

expectancy. Performance expectancy was a

significant predictor of BI for samples from

Germany, Estonia and Belgium, but not for those

from Austria and Greece. Additionally, collective

efficacy had a significant positive effect on BI for

Austrian participants; trust in state was a negative

predictor for BI for participants from Greece. Finally,

we found that the negative interaction between social

norms and social identification was only found for

our sample from Estonia, and the interaction was not

significant for participants from the other countries.

Figure 4: Model results for SEM.** p < 0.001; * p < 0.05.

6 DISCUSSION

By means of structural equation modelling, we

studied predictors of the intention to adopt a smart

heating technology. The model included the

technologies’ perceived effectiveness, price value,

effort expectancy as well as social beliefs including

social norms, collective efficacy and different types

of trust. Across five countries, the results indicate that

the individual beliefs of the UTAUT model are

suitable to predict the acceptance: price value,

performance expectancy and effort expectancy were

the most relevant predictors. Of the included social

predictors, only trust in state had a small negative

effect, and we found a small negative interaction

between social identification and social norms.

Within the specific country samples, some social

predictors reached significance (e.g., collective

efficacy in Austria) but overall, estimates were very

small. For the aggregated model, we therefore can

only accept H1 – H3, while hypotheses H4 – H6 must

be rejected.

This indicates that individual beliefs currently

better predict the intention to adopt smart heating

thermostats: for the overall sample, and within the

country samples, particularly financial aspects and

technology related beliefs (usefulness, ease of use)

influenced the intention to adopt a smart heating

devices, in line with findings from prior research

(Ahn et al., 2016; Girod et al., 2017; Tu et al., 2021).

Some of our results do not replicate evidence

from previous similar studies, though. For example,

social influence was the strongest predictor for the

intention to adopt a smart thermostat in a

representative consumer study in Germany (Große-

Kreul, 2022), but we do not find any significant

results for social beliefs – neither in the overall

sample, nor in the representative German sample.

One possible explanation lies in how social influence

is operationalized: as we already discussed, the

inconsistent results from social influence or social

norms might be driven by whether the concept is

understood as an injunctive or descriptive social

norm, and who is considered within as norm-related

group. Compared to Große-Kreul (2022), who

assessed social influence as usage in other people and

media presence (see Table 2), we used perceived

descriptive social norms within the neighborhood.

We included collective efficacy, and the

moderating effect of social identification, to broaden

the interpretation of social influence as a singular

construct, based on models and research of social

influences on pro-environmental behavior. Although

we did not find effects for most proposed social

indicators in the overall sample, we found a small

effect of collective efficacy in Austria. This might be

a promising start for future research. In general,

future work should consider the identified differences

between countries, and gain a deeper understanding

of this variations. Interestingly, we further found a

negative interaction between social identification and

social norms, which contradicts most earlier research

(Cialdini & Jacobson, 2021). This might be driven by

our choice of instrument: We used ‘closeness of

relations’ within the neighborhood as indicator for

social identification, which correlates highly with

other relationship indicators, including knowledge of

others’ goals (Gächter et al., 2015). This better

knowledge in turn might have limited the variability

of perceived norms and therefore its potential to

predict intentions. The effect size of the moderation

effect is very small and country-level analysis shows

it to be based in the Estonian sample.

6.1 Limitations

Firstly, we investigate how social beliefs relate to

smart heating technology adoption, but causal

conclusions can only be drawn within the limits of the

SEM methodology we used; our assumptions about

the causal impact of social belief predictors are here

not supported by the empirical data, while our

Understanding Intention to Adopt Smart Thermostats: The Role of Individual Predictors and Social Beliefs Across Five EU Countries

assumptions about the causal impact of individual

predictors find support in the associations within the

data, in line with prior empirical evidence.

Secondly, we did not measure adoption behavior,

but rather the intention to adopt. The validity of the

findings might be affected by the intention-behavior

gap, which is found widely in pro-environmental

consumer behavior (Carrington et al., 2010; Sheeran

& Webb, 2016). This reflects in our data in the sense

that we find very high adoption intention for smart

technology, but the real adoption rate of smart

thermostats in included countries has not yet reached

full potential

Lastly, the construct ‘trust in people’ was not

found to exceed critical thresholds, e.g. the AVE and

reliability criteria in our model, and findings

regarding it should be interpreted with caution.

Despite this, our CFA and SEM demonstrated a good

model fit and almost all constructs met reliability and

validity criteria. Thus, as we used a well-established

scale from the EVS we decided for its inclusion in the

final model.

6.2 Conclusion and Policy Implications

Overall, we conclude that the UTAUT model is well

suited to explain behavioral intention to use smart

thermostats. Our data did not yield support for an

extension of the UTAUT model to include social

beliefs derived based on evidence from previous pro-

environmental behavior research; however, this does

not indicate that they don’t play a role. Possible

explanations are the lack of publicness of thermostat

adoption and current marketing practices focusing

mainly on individual benefits. Social effects on pro-

environmental choices are stronger on highly visible

behaviors: studies find that visibility increases the

perception of social status (Uren et al., 2021),

collective efficacy affects public, but not private pro-

environmental behaviors (Hamann & Reese, 2020),

and people imitate visible behavior more (Babutsidze

& Chai, 2018). Visibility has also been found to

strengthen the relationship between a pro-

environmental social identity and behavioral

engagement (Brick et al., 2017). The adoption of a

smart thermostat is invisible behavior, conducted in

the private domain, so this might be a reason why

social beliefs do not impact it greatly. The proposed

model would benefit from implementation of explicit

comparisons of private and public pro-environmental

target variables to differentiate the effects of social

https://interpret.la/smart-home-sees-significant-growth-

in-western-europe/

beliefs, specifically with a focus on adoption of novel

technologies.

The conceptualization of smart thermostat

adoption as private sphere behavior also delivers a

potential explanation of the negative significant effect

of trust in state. As part of a meta-analysis, trust in

state has been found to correlate with public pro-

environmental behaviors (Cologna & Siegrist, 2020),

and a study which examined the effect of generalized

trust and trust in governments found that private

behaviors are negatively correlated to trust in

governmental institutions (Taniguchi & Marshall,

2018). This is explained with a theory of overreliance

on the state, which decreases the perceived need or

responsibility for own environmental action. It also

seems worth investigating how the adoption and use

of thermostats is framed across marketing campaigns

and public service announcements. In a previous

extensive qualitative analysis of product reviews for

five commercial smart thermostats, technology or

comfort related content categories were dominant

(Malekpour Koupaei et al., 2020). The marketing

emphasizes individual benefits, i.e. costs savings,

energy efficiency and technology features; this might

be one reason why social beliefs are not prevalent in

people’s cognitions regarding these devices. We

second suggestions by for exampleLi’s review on

smart home adoption (Li et al., (2021)), that

advertisements for such energy-efficiency devices

should consider including broader social benefits,

especially in light of potentially existing rebound

effects (Dütschke et al., 2018; Seebauer, 2018).

Finally, studies on saving devices and efficiency

technology often examine single individual’s

intention to adopt them. However, decisions about

thermal comfort often rely on household decisions

(Sintov et al., 2019). Sovacool et al. (2020)

specifically identified decision-making structures

around smart heating in households, displaying

conflicts between different household members

including between partners, roommates or parents

and children. Care should be taken when interpreting

results from our and previous literature about heating

technology adoption based on individual’s reported

intentions, as in most cases (and in 94% of our

sample), ‘a household is not a person’ (Seebauer &

Wolf, 2017). Future studies should design

measurements of target behaviors that are sensitive to

both individual and household-level decision-

making. Taken these findings into account, it is

imperative to design interventions to better study how

SMARTGREENS 2025 - 14th International Conference on Smart Cities and Green ICT Systems

to accelerate the diffusion of smart heating

technologies across Europe to the extent envisioned

by policymakers.

The data, analysis scripts and questionnaire with

stimulus materials can be downloaded at

https://osf.io/ba2vf/?view_only=986065e170584cad

9098d0a2937e216b

ACKNOWLEDGEMENTS

This research was supported by a European Union

Horizon 2020 research and innovation programme

grant (RENergetic, grant N957845; DECIDE, grant

N894255) awarded to MB, FK, SK, CK. Conflicts of

interests: none. MB: Conceptualization,

Methodology, Data curation, Formal analysis,

Visualization, Writing – original draft. FK:

Conceptualization, Methodology, Investigation,

Funding acquisition, Writing - review & editing. SK:

Conceptualization, Methodology Writing - review &

editing. CK: Conceptualization, Data curation,

Formal analysis, Writing - review & editing,

Supervision, Funding acquisition.

REFERENCES

Abrahão, R. de S., Moriguchi, S. N., & Andrade, D. F.

(2016). Intention of adoption of mobile payment: An

analysis in the light of the Unified Theory of

Acceptance and Use of Technology (UTAUT). RAI

Revista de Administração e Inovação, 13(3), 221–230.

https://doi.org/10.1016/j.rai.2016.06.003

Ahn, M., Kang, J., & Hustvedt, G. (2016). A model of

sustainable household technology acceptance:

Sustainable household technology acceptance model.

International Journal of Consumer Studies, 40(1), 83–

91. https://doi.org/10.1111/ijcs.12217

Ajzen, I. (1985). From Intentions to Actions: A Theory of

Planned Behavior. In J. Kuhl & J. Beckmann (Hrsg.),

Action Control (S. 11–39). Springer Berlin Heidelberg.

https://doi.org/10.1007/978-3-642-69746-3_2

Arfi, W. B., Nasr, I. B., Kondrateva, G., & Hikkerova, L.

(2021). The role of trust in intention to use the IoT in

eHealth: Application of the modified UTAUT in a

consumer context. Technological Forecasting and

Social Change, 167, 120688. https://doi.org/10.1016/

j.techfore.2021.120688

Babutsidze, Z., & Chai, A. (2018). Look at me Saving the

Planet! The Imitation of Visible Green Behavior and its

Impact on the Climate Value-Action Gap. Ecological

Economics, 146, 290–303. https://doi.org/10.1016/

j.ecolecon.2017.10.017

Berkowitz, A. D. (2004). The social norms approach:

Theory, research, and annotated bibliography.

Billanes, J., & Enevoldsen, P. (2022). Influential factors to

residential building Occupants’ acceptance and

adoption of smart energy technologies in Denmark.

Energy and Buildings, 276, 112524. https://doi.org/

10.1016/j.enbuild.2022.112524

Brick, C., Sherman, D. K., & Kim, H. S. (2017). “Green to

be seen” and “brown to keep down”: Visibility

moderates the effect of identity on pro-environmental

behavior. Journal of Environmental Psychology, 51,

226–238. https://doi.org/10.1016/j.jenvp.2017.04.004

Caferra, R., Colasante, A., & Morone, A. (2021). The less

you burn, the more we earn: The role of social and

political trust on energy-saving behaviour in Europe.

Energy Research & Social Science, 71, 101812.

https://doi.org/10.1016/j.erss.2020.101812

Carrington, M. J., Neville, B. A., & Whitwell, G. J. (2010).

Why Ethical Consumers Don’t Walk Their Talk:

Towards a Framework for Understanding the Gap

Between the Ethical Purchase Intentions and Actual

Buying Behaviour of Ethically Minded Consumers.

Journal of Business Ethics, 97(1), 139–158.

https://doi.org/10.1007/s10551-010-0501-6

Chang, S., & Nam, K. (2021). Smart Home Adoption: The

Impact of User Characteristics and Differences in

Perception of Benefits. Buildings, 11(9), 393.

https://doi.org/10.3390/buildings11090393

Cialdini, R. B. (2007). Descriptive Social Norms as

Underappreciated Sources of Social Control.

Psychometrika, 72(2), 263–268. https://doi.org/

10.1007/s11336-006-1560-6

Cialdini, R. B., & Jacobson, R. P. (2021). Influences of

social norms on climate change-related behaviors.

Current Opinion in Behavioral Sciences, 42, 1–8.

https://doi.org/10.1016/j.cobeha.2021.01.005

Cialdini, R. B., Reno, R. R., & Kallgren, C. A. (1990). A

focus theory of normative conduct: Recycling the

concept of norms to reduce littering in public places.

Journal of personality and social psychology, 58(6),

1015.

Cologna, V., & Siegrist, M. (2020). The role of trust for

climate change mitigation and adaptation behaviour: A

meta-analysis. Journal of Environmental Psychology,

69, 101428. https://doi.org/10.1016/

j.jenvp.2020.101428

Cronbach, L. J. (1951). Coefficient alpha and the internal

structure of tests. Psychometrika, 16(3), 297–334.

https://doi.org/10.1007/BF02310555

Dütschke, E., Frondel, M., Schleich, J., & Vance, C. (2018).

Moral Licensing—Another Source of Rebound?

Frontiers in Energy Research, 6, 38. https://doi.org/

10.3389/fenrg.2018.00038

European Commission. (2022, Juni 27). Joint Statement by

President von der Leyen and President Biden on

European Energy Security. https://ec.europa.eu/

commission/presscorner/detail/es/statement_22_4149

European Environment Agency (EEA). (2021, Oktober 26).

Greenhouse gas emissions from energy use in buildings

in Europe. https://www.eea.europa.eu/data-and-

maps/indicators/greenhouse-gas-emissions-from-

energy/assessment

Understanding Intention to Adopt Smart Thermostats: The Role of Individual Predictors and Social Beliefs Across Five EU Countries

Falk, R. F., & Miller, N. B. (1992). A primer for soft

modeling. University of Akron Press.

Fornell, C., & Larcker, D. F. (1981). Evaluating Structural

Equation Models with unobservable variables and

measurement error. Journal of Marketing Research,

18(1), 39–50.

Fritsche, I., Barth, M., Jugert, P., Masson, T., & Reese, G.

(2018). A Social Identity Model of Pro-Environmental

Action (SIMPEA). Psychological Review, 125(2), 245–

269. https://doi.org/10.1037/rev0000090

Gächter, S., Starmer, C., & Tufano, F. (2015). Measuring

the Closeness of Relationships: A Comprehensive

Evaluation of the „Inclusion of the Other in the Self“

Scale. PLOS ONE, 10(6), e0129478. https://doi.org/

10.1371/journal.pone.0129478

Gimpel, H., Graf, V., & Graf-Drasch, V. (2020). A

comprehensive model for individuals’ acceptance of

smart energy technology – A meta-analysis. Energy

Policy, 138, 111196. https://doi.org/10.1016/

j.enpol.2019.111196

Girod, B., Mayer, S., & Nägele, F. (2017). Economic versus

belief-based models: Shedding light on the adoption of

novel green technologies. Energy Policy, 101, 415–

426. https://doi.org/10.1016/j.enpol.2016.09.065

Göckeritz, S., Schultz, P. W., Rendón, T., Cialdini, R. B.,

Goldstein, N. J., & Griskevicius, V. (2009). Descriptive

normative beliefs and conservation behavior: The

moderating roles of personal involvement and

injunctive normative beliefs. European Journal of

Social Psychology, n/a-n/a. https://doi.org/10.1002/

ejsp.643

Große-Kreul, F. (2022). What will drive household

adoption of smart energy? Insights from a consumer

acceptance study in Germany. Utilities Policy, 75,

101333. https://doi.org/10.1016/j.jup.2021.101333

Gumz, J., Fettermann, D. C., Sant’Anna, Â. M. O., &

Tortorella, G. L. (2022). Social Influence as a Major

Factor in Smart Meters’ Acceptance: Findings from

Brazil. Results in Engineering, 15, 100510.

https://doi.org/10.1016/j.rineng.2022.100510

Hair, J. F., Hult, G. T. M., Ringle, C. M., Sarstedt, M.,

Danks, N. P., & Ray, S. (2021). An Introduction to

Structural Equation Modeling. In J. F. Hair Jr., G. T. M.

Hult, C. M. Ringle, M. Sarstedt, N. P. Danks, & S. Ray

(Hrsg.), Partial Least Squares Structural Equation

Modeling (PLS-SEM) Using R: A Workbook (S. 1–29).

Springer International Publishing. https://doi.org/

10.1007/978-3-030-80519-7_1

Haji Hosseinloo, A., Ryzhov, A., Bischi, A., Ouerdane, H.,

Turitsyn, K., & Dahleh, M. A. (2020). Data-driven

control of micro-climate in buildings: An event-

triggered reinforcement learning approach. Applied

Energy, 277, 115451. https://doi.org/10.1016/

j.apenergy.2020.115451

Hamann, K. R. S., & Reese, G. (2020). My Influence on the

World (of Others): Goal Efficacy Beliefs and Efficacy

Affect Predict Private, Public, and Activist Pro‐

environmental Behavior. Journal of Social Issues,

76(1), 35–53. https://doi.org/10.1111/josi.12369

Henseler, J., Ringle, C. M., & Sarstedt, M. (2015). A new

criterion for assessing discriminant validity in variance-

based structural equation modeling. Journal of the

Academy of Marketing Science, 43(1), 115–135.

https://doi.org/10.1007/s11747-014-0403-8

Li, W., Yigitcanlar, T., Erol, I., & Liu, A. (2021).

Motivations, barriers and risks of smart home adoption:

From systematic literature review to conceptual

framework. Energy Research & Social Science, 80,

102211. https://doi.org/10.1016/j.erss.2021.102211

Lu, J., Sookoor, T., Srinivasan, V., Gao, G., Holben, B.,

Stankovic, J., Field, E., & Whitehouse, K. (2010). The

smart thermostat: Using occupancy sensors to save

energy in homes. Proceedings of the 8th ACM

Conference on Embedded Networked Sensor Systems,

211–224. https://doi.org/10.1145/1869983.1870005

Malekpour Koupaei, D., Song, T., Cetin, K. S., & Im, J.

(2020). An assessment of opinions and perceptions of

smart thermostats using aspect-based sentiment

analysis of online reviews. Building and Environment,

170, 106603. https://doi.org/10.1016/

j.buildenv.2019.106603

Mamonov, S., & Koufaris, M. (2020). Fulfillment of

higher-order psychological needs through technology:

The case of smart thermostats. International Journal of

Information Management, 52, 102091. https://doi.org/

10.1016/j.ijinfomgt.2020.102091

Marikyan, D., Papagiannidis, S., & Alamanos, E. (2019). A

systematic review of the smart home literature: A user

perspective. Technological Forecasting and Social

Change, 138, 139–154. https://doi.org/10.1016/

j.techfore.2018.08.015

McDonald, R. P. (1999). Test Theory: A Unified Treatment

(1st ed.). Psychology Press. https://doi.org/10.4324/

9781410601087

Nordhoff, S., Madigan, R., Van Arem, B., Merat, N., &

Happee, R. (2021). Interrelationships among predictors

of automated vehicle acceptance: A structural equation

modelling approach. Theoretical Issues in Ergonomics

Science, 22(4), 383–408. https://doi.org/10.1080/

1463922X.2020.1814446

Park, H. S., & Smith, S. W. (2007). Distinctiveness and

Influence of Subjective Norms, Personal Descriptive

and Injunctive Norms, and Societal Descriptive and

Injunctive Norms on Behavioral Intent: A Case of Two

Behaviors Critical to Organ Donation. Human

Communication Research, 33(2), 194–218.

https://doi.org/10.1111/j.1468-2958.2007.00296.x

Rivis, A., & Sheeran, P. (2003). Descriptive norms as an

additional predictor in the theory of planned behaviour:

A meta-analysis. Current Psychology, 22(3), 218–233.

https://doi.org/10.1007/s12144-003-1018-2

Rogers, E. M. (2003). Diffusion of innovations/everett m.

Rogers. NY: Simon and Schuster, 576.

Seebauer, S. (2018). The psychology of rebound effects:

Explaining energy efficiency rebound behaviours with

electric vehicles and building insulation in Austria.

Energy Research & Social Science, 46, 311–320.

https://doi.org/10.1016/j.erss.2018.08.006

SMARTGREENS 2025 - 14th International Conference on Smart Cities and Green ICT Systems

Seebauer, S., & Wolf, A. (2017). Disentangling household

and individual actors in explaining private electricity

consumption. Energy Efficiency, 10(1), 1–20.

Sheeran, P., & Webb, T. L. (2016). The Intention-Behavior

Gap: The Intention-Behavior Gap. Social and

Personality Psychology Compass, 10(9), 503–518.

https://doi.org/10.1111/spc3.12265

Shuhaiber, A., & Mashal, I. (2019). Understanding users’

acceptance of smart homes. Technology in Society, 58,

101110. https://doi.org/10.1016/j.techsoc.2019.01.003

Sintov, N. D., White, L. V., & Walpole, H. (2019).

Thermostat wars? The roles of gender and thermal

comfort negotiations in household energy use behavior.

PLOS ONE, 14(11), e0224198. https://doi.org/10.1371/

journal.pone.0224198

Sovacool, B. K., & Furszyfer Del Rio, D. D. (2020). Smart

home technologies in Europe: A critical review of

concepts, benefits, risks and policies. Renewable and

Sustainable Energy Reviews, 120, 109663.

https://doi.org/10.1016/j.rser.2019.109663

Sovacool, B. K., Martiskainen, M., Osborn, J., Anaam, A.,

& Lipson, M. (2020). From thermal comfort to conflict:

The contested control and usage of domestic smart

heating in the United Kingdom. Energy Research &

Social Science, 69, 101566. https://doi.org/10.1016/

j.erss.2020.101566

Spence, A., Demski, C., Butler, C., Parkhill, K., & Pidgeon,

N. (2015). Public perceptions of demand-side

management and a smarter energy future. Nature

Climate Change, 5(6), 550–554. https://doi.org/

10.1038/nclimate2610

Taniguchi, H., & Marshall, G. A. (2018). Trust, political

orientation, and environmental behavior.

Environmental Politics, 27(3), 385–410.

https://doi.org/10.1080/09644016.2018.1425275

Tu, G., Faure, C., Schleich, J., & Guetlein, M.-C. (2021).

The heat is off! The role of technology attributes and

individual attitudes in the diffusion of Smart

thermostats – findings from a multi-country survey.

Technological Forecasting and Social Change, 163,

120508.

https://doi.org/10.1016/j.techfore.2020.120508

Uren, H. V., Roberts, L. D., Dzidic, P. L., & Leviston, Z.

(2021). High-Status Pro-Environmental Behaviors:

Costly, Effortful, and Visible. Environment and

Behavior, 53(5), 455–484. https://doi.org/10.1177/

0013916519882773

Venkatesh, Thong, & Xu. (2012). Consumer Acceptance

and Use of Information Technology: Extending the

Unified Theory of Acceptance and Use of Technology.

MIS Quarterly, 36(1), 157. https://doi.org/10.2307/

41410412

Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D.

(2003). User acceptance of information technology:

Toward a unified view.

MIS quarterly, 425–478.

Wang, C., Pattawi, K., & Lee, H. (2020). Energy saving

impact of occupancy-driven thermostat for residential

buildings. Energy and Buildings, 211, 109791.

https://doi.org/10.1016/j.enbuild.2020.109791

White, K. M., Smith, J. R., Terry, D. J., Greenslade, J. H.,

& McKimmie, B. M. (2009). Social influence in the

theory of planned behaviour: The role of descriptive,

injunctive, and in-group norms. British Journal of

Social Psychology, 48(1), 135–158. https://doi.org/

10.1348/014466608X295207

Yentes, R., & Chevallier, F. (2021). Careless: Procedures

for computing indices of careless responding. R

package version 1.2.1.

Understanding Intention to Adopt Smart Thermostats: The Role of Individual Predictors and Social Beliefs Across Five EU Countries