You are on page 1of 12

Journal of Cleaner Production 311 (2021) 127560

Contents lists available at ScienceDirect

Journal of Cleaner Production


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

The influence of group-level factors on individual energy-saving behaviors


in a shared space: The case of shared residences
Jiaolan Zhu a, b, Md Morshed Alam a, Zhikun Ding b, Palaneeswaran Ekambaram a, Jie Li b,
Jiayuan Wang b, *
a
Engineering and Center for Sustainable Infrastructure, Swinburne University of Technology, Melbourne, Australia
b
Sino-Australia Joint Research Center in BIM and Smart Construction, Shenzhen University, Nanshan, Shenzhen, 518000, China

A R T I C L E I N F O A B S T R A C T

Handling editor: M.T. Moreira In business and service buildings, shared spaces constitute large areas where occupants share energy resources
and manage them. Changing people’s behaviors toward conscious and efficient energy use within shared spaces
Keywords: has great potential to conserve energy and reduce greenhouse gas emissions. The occupants in a shared area
Energy-saving behaviors closely interact with each other and their energy use behaviors have a higher probability of being influenced by
Group dynamic theory
the group environment. Although the significant role of groups in occupant behavior has been highlighted in
Group-level factors
some studies, little is known about how group environments affect individual energy-saving behaviors in a
Shared space
Structural equation modeling shared space. This study explored the influencing mechanism of group-level factors on individual energy-saving
behavior in a shared space and developed a theoretical model based on dynamic group theory. A literature
review was carried out to identify group-level factors that may influence individual energy-saving behavior. An
empirical study was conducted in the shared space contextualized in college student residences in Zhuhai, China,
to understand the mechanism of how the group-level factors influence individual energy-saving behaviors. Data
collected from the questionnaire survey was analyzed through structural equation modeling. The results showed
that descriptive norms i.e., actions of other group members, have the highest impact on individuals’ energy
saving behavior in a shared space. Interestingly, any direct feedback or suggestions provided to an individual to
save energy has a very minimum influence on individuals’ energy-saving behavior. In shared spaces, what others
actually do is more influential in motivating individuals to engage in energy-saving behaviors than any energy-
saving instructions given to an individual. Group interaction is conducive to increasing the visibility of
descriptive norms and thus strengthening individuals’ behavioral intention for energy saving. These findings will
contribute to developing energy-saving strategies in a shared space focusing on ‘descriptive norms’, replacing the
current instruction-based approach. Besides, the empirical research findings could provide a guideline for policy
makers to develop behavioral change strategies regarding energy conservation.

1. Introduction space, work lounges, classrooms and so on, in which multi occupants
(non-relatives) live or perform their daily tasks. In such shared spaces,
The buildings sector consumed 36% of total energy and resulted in energy consumption rates were found to be higher compared to indi­
39% of total energy-related greenhouse gas emissions worldwide in vidually billed apartments or private dwellings (Driza, 2014). A report
2018 (IEA, 2019). The ever-increasing greenhouse gas emission is from China Association Building Energy Efficiency (2019) also indicated
responsible for global warming and environmental deterioration (Tam that energy use intensity of public buildings was 2–3 times higher than
et al., 2018). Commercial and service buildings predominantly consist of that of residential buildings in 2017 (CABEE, 2019). Previous studies
shared spaces and account for 38% of the total building energy con­ focused on achieving energy conservation in shared spaces of buildings
sumption (Chathura et al., 2017) (CABEE, 2019). These shared spaces through the improvement of human-technology interaction (Wayes T.
include dorms, open-plan offices, meeting or conference rooms, team et al., 2017), an approach via the installation of energy-efficient

* Corresponding author.
E-mail addresses: zhujiaolan@126.com (J. Zhu), mmalam@swin.edu.au (M.M. Alam), ddzk@szu.edu.cn (Z. Ding), pekambaram@swin.edu.au (P. Ekambaram),
jennilion@163.com (J. Li), wangjy@szu.edu.cn (J. Wang).

https://doi.org/10.1016/j.jclepro.2021.127560
Received 8 June 2020; Received in revised form 20 April 2021; Accepted 17 May 2021
Available online 25 May 2021
0959-6526/© 2021 Elsevier Ltd. All rights reserved.
J. Zhu et al. Journal of Cleaner Production 311 (2021) 127560

equipment. However, the rebound effect has weakened this effort and found that people in shared offices behaved less sustainably in terms of
has even led to higher energy usage in the building sector (Zhang and energy consumption than in private offices. The lights in unoccupied
Peng, 2017). This effect, in turn, highlights the significance of another rooms are switched off less frequently in shared offices than that in
approach through occupant-centric interventions, which can help oc­ private offices, suggesting that shared responsibility for an action may
cupants take more efficient energy-use actions. However, existing effort be affected by the presence of other occupants. For occupants in a shared
focused on energy-saving behaviors in private households, which could space, a sense of personal responsibility for energy conservation may be
not be directly applied in shared spaces due to the difference in social inhibited due to the shared utility and uncertainty about the individual
environment and functions of these spaces (Zhang et al., 2019). For contributions of energy savings (Bulunga and Thondhlana, 2018). Social
example, energy use information feedback which was found to be loafing may occur in a shared space due to the presence of others and
effective in private homes (Delmas et al., 2013), is not applicable in a lack of motivations to the group goal. The group that an individual be­
shared space since the appliances and equipment are shared by all oc­ longs to has been found to guide one’s pro-environmental behaviors, like
cupants (Chen et al., 2016). Individual occupants in shared spaces are recycling behaviors, buying electric vehicles, re-use towels (Baddeley,
typically not responsible for their own energy consumption and there­ 2016). Studies have found that individuals are willing to improve their
fore adopt conventional energy-use characteristics, causing more energy pro-environmental behaviors by opinions and actions of their peers
waste (Chen et al., 2016). Kavulya and Becerik-Gerber (2012) reported within a group (Mulville et al., 2016), and the behaviors of people are
significant energy waste in office due to appliances (computers, printers, favorably guided by good examples set by their group members (Testa
laptop etc.) being put on standby mode. Occupants’ energy-saving et al., 2016). Thereby, individuals’ energy use behavior in shared space
behavior plays a key role to improve building energy efficiency (Alam should be studied in group contexts rather than in isolation (Forsyth,
et al., 2017). Previous research showed that adoption of efficient 2017). The group-level factors, such as group value, group norms, group
behavioral practices could result in up to 20% reduction in energy interaction, etc. will hence unavoidably affect occupants’ energy atti­
consumption (Azar and Al Ansari, 2017). Simulation studies of occupant tudes and behaviors consciously and unconsciously (Jans et al., 2018).
behavior in open-plan offices concluded that “wasteful” work-style oc­ Understanding the group dynamics occurring within share space con­
cupants used twice more energy of the standard (non-wasteful, texts is significant because they can affect occupants’ behaviors to a
non-austere) consumers, while consumers with a “austere” work style large extent (Midden et al., 2011). Although some scholars have high­
consumed 50% less energy of the standard (Hong et al., 2017). Thus, lighted the importance of groups and their process on energy conser­
changing people’s behavior toward conscious and efficient energy use vation behaviors (Hong et al., 2017; Sovacool et al., 2015), little is
within shared spaces is greatly significant to reduce energy waste. known about how groups in a shared space influences an individual’s
Behavioral-based interventions will be more successful to encourage energy-saving behaviors. Therefore, this study aims to develop a
sustainable energy use if significant drivers of behavior are understood research model to further understand the intrinsic relationship and
and targeted (Lopes et al., 2019). Hence, the key issue for scholars and interaction among those factors. Followings are the two objectives to
practitioners is to understand which factors drive occupants to imple­ achieve the research aim:
ment energy-saving behaviors in shared spaces. Only a handful of
studies have paid attention to the individual motivations affecting (1) To identify group-level factors affecting individual energy-saving
shared space users’ behaviors, such as personal values, norms, capabil­ behaviors in shared spaces and establish a hypothetical model
ities, attitudes, and habits. For example, Scherbaum and Popovich indicating their interactions.
(2008) explored the impact of individual-level factors on employees’ (2) To explore key group-level factors and the mechanism of their
energy-conservation behaviors at workplaces based on influence on individual energy-saving behaviors in shared spaces.
value-belief-norm theory (VBN). Their results showed that personal
environmental norms strongly influenced energy-saving behaviors and To achieve the above objectives, this study applied group dynamic
intentions. Chathura et al. (2017) found that personal capabilities theory to investigate the effect of group-level factors on individual en­
played a significant role in shaping shared space users’ energy-saving ergy conservation behaviors in a shared space. An empirical study was
behaviors. Lo et al. (2014) used the extended TPB model to examine conducted in shared student dormitories in Zhuhai, China. Data was
the influence of social-cognitive factors and habits on collected by questionnaire survey, and then factorial interactions were
energy-conservation behaviors in single and multi-occupant offices. examined through structural equation modeling. As a unified collective
Habit was found to be the greatest influential factor of the behaviors accommodation for students is implemented in Chinese universities, a
turning off lights and monitors. Although some studies have investigated student usually shares a dorm with others. Understanding university
the determinants behind individuals’ decisions to adopt energy-saving students’ energy use behavior in shared residences is conductive to
behaviors in shared spaces, they mostly focused on the perspective of improving building energy efficiency of the university campus and thus
VBN or TPB. These theories are fundamentally based on the rational promoting the sustainable development of Chinese society. Besides, the
proposition. Therefore, the explanatory value of above-mentioned shared residences are a typical kind of shared scenario, which can pro­
studies is limited because they just consider the rational factors. As we vide reference for further study of energy saving in other kinds of shared
know, the adoption of energy-saving behaviors cannot be just inter­ spaces.
preted as an individual decision, and it may be motivated by the social Compared with previous studies, this study contributes to the
environment. For instance, a study shows that employees usually leave following aspects. First, our research contributes to clarifying the
the lights on in unoccupied rooms at workplaces but not at home (Lo mechanism of how groups influence individuals’ energy-saving behav­
et al., 2014). This is because of the environment at work where occu­ iors in a shared space. Research on promoting occupants’ energy-saving
pants always see lights burning everywhere and no one is switching behaviors is extended from in private homes to in shared spaces, which
them off. However, the extant literature focused on analysis of share provides a new perspective for motivating energy savings on the con­
space users’ behavior seen as being apart from the group, while their sumption side. Second, while the group dynamic theory has been widely
behavior in relation to the social interaction with the group has been applied to investigate other phenomena, this study contributes to
ignored or underestimated. extending its applicability by adapting it to explain energy use behav­
Energy consumption in a shared space is largely a group phenome­ iors. To substantiate this theoretical contribution, we frame the group
non as it is the result of actions by joint group members. For people in dynamics theory by integrating a set of antecedents and intervening
shared space, their behaviors can easily be observed by each other, and variables that are unique to energy behaviors of occupants in shared
occupants in shared spaces interact mutally and have higher chances to settings. Third, the uniqueness of this research is that it extended the
be affected by their group members (Elie A. et al., 2014). Lo et al. (2014) focus from an analysis of individuals’ behavior seen as being apart from

2
J. Zhu et al. Journal of Cleaner Production 311 (2021) 127560

the group to considering the individuals’ behavior in relation to social regarding energy-saving behaviors, Jans et al. (2018) highlighted the
interaction with the group. Research on the influencing factors of significant role of group values and group norms on energy-related
occupant behaviors is extended from individual level to group level. sustainable behavior theoretically. In addition, one of the common
Finally, the empirical findings could practically help policy makers characteristics of a group is that group members interact with each
design behavioral change strategies for energy conservation in shared other. Positive group interaction was suggested to improve
space contexts. energy-saving behaviors (Hori et al., 2013). Therefore, the group-level
factors like group value, group norm, and group interaction were
2. Theoretical principle and research hypothesis selected and focused on in this study to explore their effects and routines
of shaping individual energy-saving behaviors in a shared space.
2.1. Theoretical principle
2.2. Research hypothesis
The theory of group dynamic proposed by Kurt Lewin in 1951 for
understanding human behaviors in groups has been applied in various
2.2.1. Group values
disciplines such as anthropology, business and industry, communica­
Group values are defined as group members’ moral principles and
tion, psychology, and sociology (Forsyth, 2017). Based on the principle
beliefs that they think are important, including self-enhancement group
of interactionism, the field theory assumes that each person’s behavior
values and self-transcendence group values. Self-transcendence group
in groups is determined by the interaction of the person and the envi­
values related to group members’ environmental beliefs highlight the
ronment (Lewin, 1951), which is summarized by the interactionism
welfare of others and nature (e.g., helpfulness, respecting nature, and
formula B = f(P, E). In a group context, this formula implies that in­
environmental protection) (Jans et al., 2018). Generally, consumers
dividual behavior (B) is a function (f) of his or her personal character­
with self-transcendence values will actively perform moral obligations
istics (P), and the social environmental factors (E), which include
and personal norms (Song et al., 2019b). Some studies have stressed the
features of the group, the group members, and the situation. In the field
role of interconnected group in individual perception and choice to take
of psychology and sociology, the group dynamic theory has been applied
pro-environmental actions (Steg et al., 2015; Testa et al., 2016; Zhang
to understand individual behavior, perceptions of other people, influ­
et al., 2018). They found that one’s perception regarding the values of
ence of norms on behavior. Previous studies have confirmed that groups
others in general affected his/her own values and contributed to shaping
have a profound impact on individuals’ actions, thoughts, and feelings
the individual tendency to implement a specific behavior. In other
(Abramst et al., 1990; Princes and Manurung, 2020). Group-level factors
words, an individual within a group is likely to be influenced by other
such as group values, group norms, group interaction, cohesion and
group members’ values, viz., group values. The stronger an individual of
leadership were considered as influential factors of individual behaviors
a shared space perceives that his or her group members endorse
(Forsyth, 2017).
self-transcendence values, the more likely the individual is willing to
The group dynamic theory is appropriate for this study for the
take energy-saving actions.
following reasons. First, energy consumption of shared spaces is essen­
Then, the following hypothesis is hence formulated:
tially a group phenomenon since shared space users’ actions jointly
determine the group energy performance. Shared space users could H1. The individual’s intention to engage in energy-saving behaviors is
affect each other mutually through information exchange, forming positively affected by group members’ self-transcendent values in a
norms, making joint decisions of utilization of systems and equipment. shared space.
The group dynamics occurring within shared spaces are important to
understand because the powerful interpersonal forces can significantly 2.2.2. Group norms
shape members’ actions. Second, previous studies primarily focused on Group norms activated by group values, indicating what behaviors
the perspective of TPB or VBN in explaining individuals’ energy-saving are socially acceptable within a group, are able to serve as a normative
behaviors. However, the explanatory value of these literatures is limited regulatory guide for individual behavior (Baddeley, 2016). Descriptive
because those studies only focused on individual-level factors of be­ norms included in group norms focus on what most people actually do in
haviors (Steg and Vlek, 2009). Environmental factors, particularly the a given settings. People’s behaviors are consistently guided by their
group contextual factors, are ignored. Unlike other theories (i.e., TPB, perceived values (Song et al., 2019a). As self-transcendent group values
VBN), the group dynamic theory emphasizes the influence of environ­ refer to the aggregate of group members’ personal non-egotism values,
mental factors on people’s behaviors, particularly the group contextual they could direct the behaviors of group members, viz., descriptive
factors. Thus, this theory is particularly well suited for understanding norms. Zhang et al. (2018) found that households with stronger envi­
the influence of group-level factors on individuals’ energy-saving be­ ronmental values were more likely to implement energy-saving behav­
haviors in shared spaces. iors. Environmental and ethical consumer values, especially those such
To encourage energy-saving behavior, the first point is to understand as self-transcendence and altruism, positively promoted
what drives or prevents such behavior, viz., behavioral intention. Ac­ pro-environmental behaviors (Steg et al., 2015; Steg and Vlek, 2009;
cording to the theory of group dynamics, individual behavior (B) does Zhao et al., 2019). Based on the same logic, it can be speculated that the
not depend on his or her personal characteristics alone. The social stronger group members endorse self-transcendence values, the more
context-and particularly the group contextual factors may facilitate or likely they will take energy-saving actions in a shared space. Therefore,
constraint individual behaviors. In group contexts, group dynamics it is hypothesized that:
occur within and between groups over time. The dynamics influence
H2. Self-transcendence group values (viz., group members’ values) are
how individuals treat themselves, with whom they interact, and what
positively related to descriptive norms (viz., the behaviors of group
kinds of beliefs they think are appropriate in some situations. The self-
members) in a shared space.
categorization theory explains the cognitive mechanisms that align
members’ self-conceptions with their conception of the groups they In addition, descriptive norms refer to the behaviors of others, which
belong to (Haslam and Reicher, 2015). As individuals have a sense of may affect individual’s decisions to implement a certain behavior or not.
group belonging, they are generally unwilling to be isolated by the For example, a roommate observes that most roommates turn off their
group they belong to. When the situation of the group changes, the in­ computers when leaving the dorm. Then he would think that such a
dividual is likely to adjust himself/herself accordingly. In other words, behavior is appropriate and imitate his roommates as a norm. In pro-
individuals’ attitudes and behaviors are more likely to be influenced by environmental research area, the role of descriptive norms in behav­
what the majority of their group members say, think, do. Specifically, iors has been highlighted. Descriptive norms were revealed to exert a

3
J. Zhu et al. Journal of Cleaner Production 311 (2021) 127560

significant influence on adolescents’ recycling behavior (Grønhøj and Taken together, we assume that group interaction is influential to in­
Thøgersen, 2012). For energy-saving behaviors, Ding et al. (2018) dividual energy-saving intention and behaviors in shared spaces.
indicated that a household’s beliefs about whether his or her social ac­ Therefore, two hypotheses are following proposed:
tors (such as relatives, friends, celebrities) engaged in heating
H6. Group interaction positively affects the intention of individuals’
energy-saving behavior had little correlation with an individual’s
energy-saving behavior in shared spaces.
behavioral intention and actual behavior in private homes. However,
individuals in shared spaces experience a different environment from H7. Group interaction positively affects individuals’ actual energy-
private homes. Their actions may be influenced by the presence of other saving behavior in shared spaces.
shared space users. According to self-categorization theory, individuals
tend to categorize themselves into a specific group and imitate what 2.2.4. Behavioral intention
others do in groups to follow their shared group membership. As in­ Behavioral intention, which reflects the extent that an individual is
dividuals have a sense of group belonging, they are generally unwilling willing to implement a certain behavior, was regarded as the most direct
to be isolated by the group they belong to. When an individual clearly antecedent variable of behaviors in some studies (AJZEN, 2002; Zhang
knows about their shared space users’ behavioral norms in a particular et al., 2018; Zhang et al., 2019). From the cognitive perspective, the
situation, the influence of the behavioral norms of others on his or her behavioral intention within a group becomes understandable by inves­
behavioral intentions and behavior may be strengthened. Therefore, two tigating the cognitive processes that allow members to collect infor­
hypotheses are added as follows: mation and make sense of it. According to these mental appraisals, the
actual behavior of individuals occurs (Forsyth, 2017). Ajzen (2002)
H3. Descriptive norms of the group positively affect individuals’
pointed out that intention could directly influence behavior and play a
intention of energy-saving behavior in shared spaces.
mediating role between other factors and actual behaviors. In the areas
H4. Descriptive norms of the group positively affect individuals’ actual of psychology and sociology, the positive and intermediary effects of
energy-saving behavior in shared spaces. intention on behaviors have been validated by some researchers. Zhang
et al. (2019) have indicated that the willingness of green consumption
Moreover, one of the most important characteristics of norms is that
exerted a significant direct and indirect influence on residents’ green
they do not exist if they are not shared with others (Cialdini and Gold­
consumption behavior. Zhang (2018) found that energy-saving in­
stein, 2004). Descriptive norms refer to the actions that are taken by
tentions directly influenced residents’ energy-saving behaviors and
group members regularly in a situation. In order to reinforce such
played an intermediary role in individual subjective factors and external
norms, group members with arbitrary patterns of behaviors are more
influencing factors. Thus, the hypothesis is to be as follows:
likely to pass their behavioral norms to others. The exchange and
inculcation of normative behavior can be accomplished via communi­ H8. Individual’s actual energy saving behavior is positively affected
cation. For example, they may communicate with others about the by their intention to conserve energy.
behavior patterns that are effective, relevant, and informative. In­
According to the above-mentioned hypothesis, the model shown in
dividuals who accepted and internalized the norms will discourage any
Fig. 1 was then built to present interrelationships of factors affecting
deviant tendencies by voicing what other members “should” or “ought
energy-saving behaviors.
to” do. Thus, the behavioral norms can be transmitted to others through
communication and therefore reinforced. In a shared space, individuals
3. Research Methodology
with higher normative consciousness are more likely to activate group
interaction. They attempt to communicate with and persuade others as
3.1. Measurements
well as cooperate and act collectively to pursue their group’s interests.
To sum up, the more arbitrary shared space users’ energy-saving
A questionnaire was designed to study individual behaviors and
behavior pattern is, the more likely they will interact with others
group-level factors related to energy saving, which is contextualized in
about energy saving.
students’ shared dormitories in this research. Each room accommodates
Therefore, the following hypothesis is proposed:
four to six students. Pilot interviews and literature bibliometric were
H5. Descriptive norm of a group promotes group interaction in a conducted to learn about their energy-saving behaviors and potential
shared space. group-level influencing factors. Then, a questionnaire was prepared
based on the pilot study and the hypothesized model established, as
2.2.3. Group interaction shown in the Appendix A. It comprises two parts, viz., (1) Part1: re­
Group interaction as a common characteristic of groups, is patterned spondents’ demographic information, including gender, grade, native
by group norms. Bales (1970) categorized group interaction into two place, social identity and living expenses. (2) Part2: the scale of group-
basic types of interactions: relationship interaction and task interaction, level factors affecting individual energy-saving behavior and intention.
including different action types, such as asking for or giving information As shown in Table 1, there are 24 measuring items designed in the
and opinions, providing suggestions and expressing solidarity. Group questionnaire. Each factor in the prior hypothesized model was
members exchange information with each other, via both verbal and measured using the three to five items. All items were scored with five-
nonverbal communication. Previous literatures mainly stressed the role point Likert scale, and respondents were required to rate from 1 to 5 (i.
of information provided by close social actors (i.e., family, friends, e., 1 = strongly agree or always, 2 = agree or often, 3 = neutral or oc­
neighbors) in shaping individual environmental-friendly behaviors casionally, 4 = disagree or rarely, 5 = strongly disagree or never) on
(Baddeley, 2016; Testa et al., 2016). Testa et al. (2016) found that how well each measurement described their energy-saving practices and
households were likely to improve their energy-saving behaviors by those factors.
opinions of their family and friends. Midden et al. (2011) highlighted Particularly, due to the difficulty of measuring actual behavior, the
persuasive power in group energy saving. In another survey of Hori et al. self-reported behaviors were often used to reflect energy-saving
(2013), researchers also found that social interaction showed dissimilar behavior (Li et al., 2015). Hence, five items were utilized to measure
effects on energy-saving behaviors due to different contexts which respondents’ energy-saving behaviors through self-report (Table 1). This
people were in. As the basis of community, social interactions in energy is a high-efficiency way to learn about the extent to which the re­
conservation were emphasized by Chen and Knight (2014), who sug­ spondents act on simple energy-saving practices with respect to lights,
gested that norms may be reinforced through interpersonal interaction air conditioners, computers and washing machines that are often used in
in workplace context, and thus significantly affect behavioral intention. dormitories. Behavioral intention toward energy saving was measured

4
J. Zhu et al. Journal of Cleaner Production 311 (2021) 127560

Fig. 1. The initial hypothetical model based on group dynamic principle.

by five items. Two items were adopted from the findings of Testa et al. conditioning systems. The dormitories have same layout and are made
(2016), asking the respondents to indicate their levels of daily from identical building materials. To ensure a high rate of response, we
energy-saving behaviors. In addition, whether the respondents would used face-to-face survey. 500 questionnaires were distributed randomly,
like to participate in energy-saving organizations and activities could with 430 valid returns finally collected yielding a high survey response
reflect their willingness to implement energy-saving measures. rate up to 86.00%. The description of the survey sample is summarized
In shared spaces, group-level factors can be addressed as the aggre­ in Table 2. Respondents are all from multi-occupancy rooms, with
gate of shared space users’ thought and behaviors. Group self- 53.49% of males and 46.51% of females. Most of them come from the
transcendence values were measured by asking an individual’s percep­ south of China (89.50%), mainly from Guangdong Province. There are
tion about his or her roommates’ beliefs regarding helpfulness, 83.50% of respondents whose average monthly living expense is from
respecting nature and environmental protection. Besides, the measure­ RMB1000 to RMB2000. Only around one-fifth of respondents (23.95%)
ment on descriptive norms toward energy saving was referenced from have management positions in students’ organizations or departments.
the literatures of Gao et al. (2017) and Tang et al. (2019). In shared The sample has homogeneity in demographic characteristics, which
rooms, occupants are likely to imitate their roommates’ behavior since aligns with the actual situation of most shared spaces. The people
there are close interactions among them. Five items were then designed sharing a space generally have strong similarities in sociodemographic
from behavior-based perspectives of the respondents’ roommates. Their characteristics, such as age, gender, social class, living expense status,
roommates’ performance of habitual behaviors in energy use was rated and education level. Especially for students in shared dorms, previous
by the respondents. Concerning group interaction, it was measured by studies have demonstrated that there are no significant differences
four items with reference to the research of Bales (1970) and Forsyth regarding their energy-related characteristics in different sociodemo­
(2017), including reminding roommates, sharing energy-saving experi­ graphic variables but gender (Wang et al., 2019). Hence, the influence of
ence, giving or asking energy-saving suggestions within a group. these factors on individuals’ energy-saving intention and behavior could
be taken no account in this research. It, therefore, justifies that the
sample used in this study with consideration of gender differences is
3.2. Data collection relatively representative. Besides, the sample size met the stipulation of
a large sample (normally more than 10 times of measuring items) (Xiong
Questionnaire was used to gather sample data about energy-saving et al., 2015).
behaviors and potential influencing group-level factors. Survey was
conducted among students in shared dormitories of a university campus
in Zhuhai, China. In China, most university students live in shared res­ 3.3. Data analysis
idences, and often about four to six students share a room. Each room
was equipped with similar manually controlled lighting and air The gathered data was analyzed by AMOS 24.0 software using

5
J. Zhu et al. Journal of Cleaner Production 311 (2021) 127560

Table 1 Table 2
Measures of group-level factors and self-reported energy-saving behaviors. Descriptive characteristic of the survey sample.
Factor Constructs Item Measurement References Demographic variables Category Frequency Percentage
No. (%)

Self-transcendence SV1 My roommates endorse Hiratsuka et al. Gender Male 230 53.49
group values respecting nature and (2018) Female 200 46.51
protecting environment. Jans et al. (2018) Grade Sophomore 27 6.30
SV2 My roommates believe Junior 403 93.70
preventing pollution is Native place Southern area in 385 89.50
important to human beings. China
SV3 My roommates advocate Other areas 45 10.50
living in harmony with nature. Average monthly living Less than RMB1500 199 46.30
SV4 My roommates place great expense RMB1500 ~ 160 37.20
value on social justice. RMB2000
SV5 My roommates care about the More than 71 16.50
interest of other people. RMB2000
Social identity leader 103 23.95
Descriptive norms DN1 My roommates switch off Gao et al. (2017);
Non-leader 327 76.05
lights when leaving Tang et al.
unoccupied room. (2019)
DN2 My roommates shut doors and
windows before turning on air
which implied the data are normally distributed (Kline, 2011). Thus, ML
conditioner. could be chosen as an estimation method for analysis in SEM. As a
DN3 My roommates turn off air comprehensive structural equation model usually includes several
conditioner when leaving measurement models and a structural model, a two-phase approach was
unoccupied room.
applied to analyze the data. First, the measurement model was estab­
DN4 My roommates switch off their
computer when not in use. lished to test reliability and validity of the constructs. Second, the
DN5 My roommates select energy structural model was built to explore interrelationships among latent
efficient mode of washing variables.
machine when in use.

Group interaction GI1 My roommates remind me of Bales (1970); 4. Research results


saving energy. Forsyth (2017)
GI2 My roommates share energy-
saving experience and tips 4.1. Measurement model
with me.
GI3 My roommates ask me for Confirmatory factor analysis (CFA) was carried out for reliability and
some opinions about energy validity tests of constructs in the measurement model. Through the CFA,
conservation.
GI4 My roommates give me some
the observed variables (i.e., DN5, SV5 and BI1) whose corresponding
suggestions related to energy factor loadings (FL) were less than acceptable level (FL ≥ 0.5) were
saving. removed from their constructs (Burton and Mazerolle, 2011). Based on
Behavior intention BI1 I will tend to use energy- Testa et al. the FL shown in Table 3, reliability and validity of constructs can be
efficiency appliances. (2016); evaluated by composite reliability (CR), average variance extracted
BI2 I am ready to adjust my daily Gao et al. (2017); (AVE) and square root of the AVE. The CR value are between 0.831 and
lifestyle to conserve energy. Tang et al. 0.889, which are greater than the recommended standard minimum
BI3 I intend to attend activities (2019)
related energy saving.
threshold of 0.7 (Nunnally and Bernstein, 1978), indicating a good
BI4 I would like to join in relevant reliability for all the constructs. In addition, the AVE value for all the
energy-saving organizations constructs presented in Table 3 ranged between 0.551 and 0.672, which
or departments. are larger than the benchmark value of 0.5, indicating a satisfactory
BI5 I will spare no efforts to help
my roommates to conserve
energy. Table 3
Results of the reliability and convergent validities of constructs.
Self-reported AB1 I switch off the lights when Scherbaum and
behaviors leaving unoccupied room. Popovich (2008) Constructs Items FL CR AVE
AB2 I shut doors and windows
Self-transcendent group value (SV) SV1 0.882 0.889 0.672
before turning on air
SV2 0.893
conditioner.
SV3 0.871
AB3 I turn off air conditioners
SV4 0.595
before leaving unoccupied
Descriptive Norm (DN) DN1 0.747 0.846 0.580
room.
DN2 0.792
AB4 I take my computer power off
DN3 0.809
when not in use.
DN4 0.694
AB5 I select energy-efficient
Group interaction (GI) GI1 0.737 0.831 0.551
running mode of washing
GI2 0.702
machine.
GI3 0.756
GI4 0.773
Energy-saving behavior intention (BI) BI2 0.831 0.880 0.648
Structural Equation Modeling (SEM). Before a structural model is built, BI3 0.840
the characteristics of the data need to be tested. The normality of data is BI4 0.806
a significant prerequisite when employing the default estimation BI5 0.740
method of maximum likelihood (ML) in SEM (Curran et al., 1996). In Actual energy-saving behavior (AB) AB1 0.646 0.871 0.578
AB2 0.781
this study, with aid of SPSS 21.0, the results of normality assessment
AB3 0.879
indicated that the kurtosis (K) values and skewness (Sk) values were both AB4 0.705
fell well within the proposed guideline (K ≤ 10, Ks ≤ 3) by Kline (2011), AB5 0.769

6
J. Zhu et al. Journal of Cleaner Production 311 (2021) 127560

convergent validity. Furthermore, discriminant validity can be evalu­ Table 5


ated by comparing the square root of each construct’s AVE with its GOF indices of the model.
highest correlation with other constructs (Fornell and Larcker, 1981). Fit index Evaluation criteria a Initial model Final model
Table 4 shows that the square roots of each construct’s AVE (values in 2
Absolute fit X /df <3 good 3.308 2.958
parentheses) are greater than its highest correlation coefficients, indi­ RMSEA <0.08 acceptable 0.075 0.069
cating sufficient discriminant validity of the constructs. The above GFI >0.8 good 0.875 0.890
analysis and results indicate that the constructs in the measurement AGFI >0.8 good 0.841 0.860
model have adequate reliability and validity. As survey data used in this Incremental fit NFI >0.9 excellent 0.877 0.901
IFI >0.9 excellent 0.911 0.925
study came from a single respondent, the issue of common method bias CFI >0.9 excellent 0.910 0.924
needs to be checked before analysis of causal relations (Narayanan and Parsimonious fit PNFI >0.5 acceptable 0.760 0.772
Narasimhan, 2014). Factor analysis of our items with SPSS21.0 was PGFI >0.5 acceptable 0.689 0.701
conducted to examine common method bias by performing the Har­ a
Evaluation criteria is adapted from (Xiong et al., 2015).
man’s test (Podsakoff et al., 2003). The five multiitem constructs of this
study were analyzed via the unrotated factor solution, finding each
Fig. 2 shows the key paths of group-level factors affecting individual
construct with an Eigenvalue higher than 1. The factor accounting for
energy-saving behavioral intention/behavior. To further clarify the role
the maximum variance contributed only 34.51% to the total variance,
of group-level factors on individual energy-saving behavior in shared
which is less than the suggested average of 41.7% (COTE and BUCKLEY,
spaces, the standardized direct, indirect, and total effects of these factors
1987). This indicates that the common method bias could not be an
on BI and AB are summarized in Table 7. It could be concluded that
important concern in our analysis.
descriptive norm is the most critical group-level factor in determining
individual energy-saving behavior in shared spaces, followed by self-
4.2. Structural model transcendence group value, while the factor of group interaction
makes the least contribution. Among these three group-level factors,
Following the CFA, a preliminary structural model was built to only descriptive norm has positive direct effect (0.441) on energy-saving
explore the proposed causal relations in the theoretical model in Fig. 1. behaviors, while the other two factors indirectly affect it by mediating
The model included five established dimensions, and each latent vari­ effects from individual energy-saving behavioral intention in shared
able was measured by several manifest variables, with four for SV, four spaces. According to the output of squared multiple correlation co­
for DN, four for GI, four for BI, and five for AB. efficients, the parameters of BI and AB are shown 0.544 and 0.373,
The hypothesized initial model was analyzed and refined by AMOS respectively. These coefficients imply that the proposed model
24.0. The modeling results are presented in Table 5 and Table 6. It could explained approximately 54.40% of the variances of BI and 37.30% of
be seen from Table 5 that some goodness of fit (GOF) indexes of the the variances of AB.
initial model are not within a preferably acceptable range (Xiong et al.,
2015), indicating that the initial model did not fit the sample data well 5. Discussions
and need some modifications. Firstly, the insignificant path whose
P-value was higher than 0.05 needs to be deleted (Hou et al., 2004). As This study used the group dynamic theory to examine the influence
shown in Table 6, two paths (DN→BI and GI→AB) with P-value higher of group contextual factors on an individual’s energy-saving behaviors
than 0.05 indicate an insignificant relationship. The path from GI to AB in a shared space. Descriptive norm was found to be the most influential
with the highest P-value of 0.928 was removed first. direct group-level factor of predicting participants’ energy conservation
Similarly, DN→BI path was deleted since the P-value of it seemed to behavior. This is contrary to the results of Ding et al. (2018) and Arimura
be higher than others in the second-round model revision (P = 0.336). et al. (2016), who found that descriptive norm insignificantly affected
After deleting prior paths, all remaining paths in the model were found residents’ energy-saving intentions and behaviors. The underlying
to be significant. Therefore, there is no need to remove any paths for reason is that individuals experience different circumstances. For in­
model improvement. Meanwhile, Modification indices (MIs) were dividuals in private homes, the behavioral information of their social
applied to relax constraints to improve scalar invariance (Mafimisebi actors (i.e., friends, neighbors, celebrities) in relatively closed family
et al., 2018). The parameter with a high value of MI was then freed from settings cannot be known by individuals, thus the influence of behav­
constraints, which contributed to improvement in fit. Priorities were ioral norms from these actors on individuals’ energy-saving practices is
given to the largest value of MIs of “e8<− >e7” (99.021) and relatively limited. Besides, if an individual is not participating in
“e7<− >e6” (52.919), respectively, and the model was modified in turn. energy-saving activities at home, he or she may not feel pressure from
Finally, relaxing both constraints induced statistically significant friends and surrounding groups. However, for individuals in shared
improvement in fit. spaces, they have a direct observation of each other’s behaviors. When
Regarding the final model, all GOF indices showed good fit according individuals observe that the majority of their shared space users take
to the test criterion, as presented from Table 5. Table 6 summarizes energy-saving actions, they may feel pressured if they do not perform
whether all the prior hypotheses are tenable, indicating that the original the same. Particularly for individuals in shared residences, the behaviors
H3 and H7 were rejected. The standardized estimation coefficient of of their group members are closely related to the individuals’ benefit
other paths in the final structural model reached the significant level of since they jointly share the billings of group energy consumption. Only
0.001. The results reveal that group-level factors play a prominent role when they see others collectively doing the right thing, they may be
in individual intentions to conserve energy and energy-saving behavior. motivated to behave in the same way. The findings indicate that in­
dividuals are more likely to follow the behaviors of their group members
in a shared space. This aligns with Latane’s (1996) opinions, indicating
Table 4
Correlation coefficients and square roots of the AVEs. that our tendency to be influenced most heavily by those who are closest
to us in physical space, which therefore produces local agreement about
Construct SV DN GI BI AB
important values and attitudes.
SV (0.820) To accurately understand the influence of descriptive norms, a
DN 0.263 (0.762)
follow-up online survey was conducted to the students from shared
GI 0.261 0.392 (0.742)
BI 0.246 0.482 0.711 (0.805) residences in China during October 2020. A total of 300 questionnaires
AB 0.365 0.529 0.367 0.428 (0.760) were issued via questionnaire star and 235 valid returns were analyzed.

7
J. Zhu et al. Journal of Cleaner Production 311 (2021) 127560

Table 6
Results of hypothesis path.
Hypothesis Path Estimate S.E. C.R. P Decision

H1 BI ← SV 0.502 *** Supported


H2 DN ← SV 0.308 0.066 5.347 *** Supported
H3 BI ← DN 0.090 0.093 0.961 0.336 Not Supported
H4 AB ← DN 0.441 0.069 7.352 *** Supported
H5 GI ← DN 0.420 0.070 6.801 *** Supported
H6 BI ← GI 0.480 0.076 9.649 *** Supported
H7 AB ← GI 0.006 0.064 0.091 0.928 Not Supported
H8 AB ← BI 0.294 0.035 5.594 *** Supported

Note: p < 0.001 ***.

Fig. 2. Results of final energy-saving model.

As presented in Table 8, over 90% of students thought that their energy-


Table 7
related behaviors were influenced by their roommates (Q1). This sug­
The effects of predictors on energy-saving intention and actual behavior.
gests that others’ presence plays an important role on individuals’
Path Direct effect Indirect effect Total effect energy-use behaviors. The results also agree with the findings of Ding
AB←SV … 0.301 0.301 et al. (2019) who found that more than 90% of students considered their
AB←DN 0.441 0.059 0.501 roommates’ feelings when using air conditioning. Our survey data
AB←GI 0.141 0.141

indicated that approximately 45.10% of students usually or always
BI←SV 0.502 0.062 0.564
BI←GI 0.480 … 0.480
imitated their roommates’ energy use behaviors (Q2). More than 70% of
BI←DN … 0.202 0.202 respondents were inclined to comply with the opinions and behaviors of
their co-occupants (Q3). The results from further face-to-face interviews
with some respondents found that they tended to keep conformity and
compliance with the actions of their roommates. The reason behind is
that they wanted to be accepted and liked by others. Following the

8
J. Zhu et al. Journal of Cleaner Production 311 (2021) 127560

Table 8
The influence of roommates on individuals’ energy use behaviors.
Never (%) Rarely (%) Sometimes Usually Always (%)
(%) (%)

My energy use behaviors are affected by the behaviors of my roommates (Q1) 8.9 14.0 32.2 31.1 13.6
I imitate my roommates’ energy consumption behaviors (Q2) 14.5 13.2 27.2 30.6 14.5
Regarding energy use, I feel the pressures from my roommates (Q3) 25.1 26.4 22.1 19.1 7.2
Regarding energy consumption, I comply with my roommates’ opinions and 5.5 6.0 23.4 41.3 23.8
behaviors (Q4)
Strongly disagree Disagree Neutral (%) Agree (%) Strongly agree
(%) (%) (%)

I am affected by the energy-saving climate of my dorms (Q5) 3.4 9.8 29.8 33.6 23.4
My roommates’ good energy-saving behaviors would inspire me to learn from 6.0 7.2 33.6 35.7 17.4
them (Q6)

behaviors of the majority can help them build and maintain good re­ inconsistency may be due to the different environments that individuals
lationships with others and can avoid being isolated. Another reason was are experiencing. Individuals may care more about their self-image in
due to the group pressure. Only 25.10% indicated that they never felt shared spaces, and they are therefore more likely to match their beliefs
pressured with respect to energy use (Q4). In the face of group pressure, with their actions in such circumstance.
it is difficult for them to keep independence in the presence of others. Group interaction has the least positive effect on energy-saving
Only 13.20% students made it clear that they would not be affected by behavior, but it could not be ignored. It is the second most powerful
the energy-saving climate of dorms (Q5). More than 50% of the re­ direct factor to affect intention of energy conservation, with effect as
spondents indicated that they would be motivated to engage in large as 0.48. Group interaction is conducive to improving people’s
energy-saving behaviors by good practices of their roommates (Q6). willingness to conserve energy. The finding is consistent with those in
These results further confirmed the validity of above-mentioned findings Chen and Knight (2014)’s research that interpersonal interaction among
related to descriptive norms in this study. Thus, it is concluded that there coworkers was influential to behavioral intention concerning energy
is a phenomenon of social conformity and compliance on energy-saving conservation in workplaces. Through interaction among members, such
behaviors in shared contexts. as giving or asking opinions, suggestions and sharing experience related
In addition, descriptive norms not only exert the largest direct in­ to energy-saving behaviors, it contributes to internalizing their intention
fluence on energy-saving behavior but also produce an indirect influ­ to conserve energy. However, another unexpected result from our study
ence by mediating group interaction and energy-saving behavioral is that group interaction cannot directly promote individuals to engage
intentions. As descriptive norms significantly influence group interac­ in energy-saving behaviors, which is dissimilar with the findings of Lu
tion, it implies that individuals with higher normative consciousness are (2018) in private homes. This finding means that the expected effect of
more likely to activate group interaction. They will try to communicate some commonly used interventions (i.e., instructions, communication,
with and persuade individuals to pursue their joint group interests, prompts) may not be ideal in persuading individuals to participate in
which can strengthen individuals’ intention to conserve energy and thus energy-saving activities in shared spaces. As group interaction has less
induce them to conform to the communicated behavior. The findings influence than descriptive norms to promote individuals’ energy-saving
from Gronhoj and Thogersen (2012) support this opinion that whether behaviors in a shared space, this indicates that individuals’ behaviors
parents’ actions affecting adolescents’ pro-environmental behavioral are more likely to be motivated by what others do than what they say.
intention depends on the extent to which their parents’ behaviors are For occupants in shared spaces, they are more concerned about people’s
visible and unambiguous and the children’s perception of their parents’ actual actions related to energy saving. Therefore, future studies could
actions. Some scholars also deemed that when individuals explicitly investigate the effect of normative feedback by expressing the extent of
perceive the desired behaviors of group members via written informa­ how people are involved in energy-saving behaviors in shared spaces.
tion, they are more willing to keep conformity with what others do in
groups (Anderson and Lee, 2016). The results of our study show clear 6. Implications
evidence that descriptive norm will affect individual behavioral inten­
tion only when norms are internalized in people’s mind. As attention is 6.1. Theoretical implications
malleable and non-durable, group interaction helps increase visibility of
descriptive norms and promote people’s behavioral intention, and thus This study has several theoretical implications for scholars. First,
motivate them to implement energy-saving activities. existing research on energy-saving behaviors in shared spaces remains
Self-transcendence group values, as the second most important relatively less. This study explores the driving factors in individual
influencing factor of actual energy-saving behaviors, affect individual energy-saving behaviors in shared spaces, which offers a brand-new
energy-saving behavior indirectly via descriptive norm, group interac­ group-level perspective for energy saving in consumption side. Previ­
tion and behavioral intention, with the total effect of 0.301. This in­ ous studies on energy saving in shared spaces mainly focused on the role
dicates that group values have an imperceptible influence on of rational factors in individual energy-saving behaviors, while the sig­
individuals’ energy-saving behaviors in a shared space. The stronger an nificant role of irrational factors (i.e., social environment) have been
individual perceives self-transcendence group values, the more willing overlooked. Particularly, this research extends the focus from analyzing
they are to adopt energy-saving actions. It aligns with the view of Jans the individual’s behavior seen as being apart from the group to studying
et al. (2018), who underlined the importance of group values in the the individual’s behavior in relation to the social interaction with the
sustainable energy behaviors without empirically testing for it. Our group, with the influencing factors of occupant behaviors extended from
research fills this key gap in the extant studies. Further, a recent study individual level to group level. Our research develops a research model
has indicated that nonmaterialistic values insignificantly impact resi­ to explore the mechanisms of how groups act on individual energy-
dents’ energy-saving behaviors (Hong et al., 2019). However, dissimilar saving behaviors in shared spaces based on group dynamics theory. It
with previous literature, this study showed that group members’ values is suggested that this theoretical model has a higher explanatory power
(viz., self-transcendence group values) significantly affect group mem­ (37.30%) than the commonly used model TPB (34.9%) provided by Gao
bers’ behaviors (viz., descriptive norms) in shared settings. This et al. (2017). It contributes to extending the application of group

9
J. Zhu et al. Journal of Cleaner Production 311 (2021) 127560

dynamic theory in clarifying the influencing routines and effects of individual and subgroup connections. Given the mutual interaction
factors on individuals’ energy-saving behavior, which enriches the among people, policy makers can make full use of the experience of
theoretical development in sustainable energy field. In future studies, selected units (i.e., sustainable groups or organizations) to promote
integration of both perspectives can provide a holistic understanding of work in the entire area.
occupant behaviors. Second, the results of the study provide theoretical
support for further study of designing norm-based interventions to 7. Conclusion
promote energy saving in shared spaces. As group norms influence in­
dividual energy-saving behaviors, individuals are more likely to follow In summary, this paper contributes to clarifying the mechanism of
the behaviors of their group members. Research on the effect of how group-level factors affect individual energy-saving behaviors in
normative information feedback on energy saving are highly recom­ shared spaces based on Group Dynamic Theory. This Theory has a
mended in shared spaces. An interesting empirical finding is that higher explanatory power than the traditional Theory of Planned
descriptive norms have greater influence on energy-saving behaviors Behavior (TPB) to explain individual energy-saving behaviors in a
than group interaction. This implies that the actual energy-saving ac­ shared setting. Our results show that group-level factors play a signifi­
tions of group members are being more influential than persuasion to cant role in individuals’ energy-saving behaviors in a shared space.
motivate individuals’ energy-saving behaviors in shared spaces. How­ Among group-level factors, descriptive norms have the highest direct
ever, this influence process could be negative or positive, so it is rec­ impact on individual’s energy-saving behavior, which means that in­
ommended that future studies integrate both sides to develop strategies dividuals are more likely to imitate the behaviors of their shared space
to foster growth in more positive directions and move people away from users. This indicates that there exists a phenomenon of social conformity
negative habits. Finally, as human beings are group-oriented, they are and compliance on energy-saving behaviors in shared settings. Mean­
likely to be changed after interacting with a close group member. The while, descriptive norms also indirectly affect individuals’ energy-
groups to which people belong plays a significant role in how they think saving behaviors through the mediating effects of group interaction
and behave. Thus, unlike the focus of changing personal factors (i.e., and behavioral intentions towards energy conservation. This suggests
knowledge, capability) in previous studies, our research suggests that group interaction is conducive to increase visibility of descriptive
changing the group contexts, under which behavioral choices are made, norms and thus strengthens individuals’ behavioral intention for energy
contribute more to improving people’s behavior, particularly in shared saving. However, group interaction has no direct impact on energy-
spaces. saving behaviors, which means that individuals could not be directly
persuaded to implement energy-saving behaviors through instruction or
6.2. Managerial implications communication only. The individuals with higher behavioral norm
consciousness are more likely to induce others to engage in energy-
The findings of this study also provide valuable practical implica­ saving actions. Moreover, self-transcendence group values exert indi­
tions for policy makers to design strategies for behavioral improvement rect influence on energy-saving behaviors by mediating behavioral
in shared spaces. As individuals’ behaviors regarding energy saving are intention regarding energy saving. It implies that group members’ be­
influenced by a group of contextual factors, their behaviors largely liefs have a subtle influence on individuals’ energy-saving behaviors in a
depend on group values, group norms, and group interaction in shared shared space. Group members’ behaviors are directly influenced by
spaces. An energy-saving action guideline should be designed in shared group members’ beliefs. Overall, the results conclude that individuals’
space first. This guideline is suggested to focus on how to improve group energy-saving behaviors in a shared space are influenced by what group
values, group norms and group interaction in a shared space. First, given members think, do and speak. Particularly, our study suggests that what
the influence of descriptive norms, improving the behaviors of a group is others actually do in a shared space is more influential to motivate in­
a prerequisite to improving individuals’ energy consumption behaviors dividuals to engage in energy-saving behaviors than what they say.
in a shared space. One approach is to provide group-level normative Findings from this study can contribute to designing holistic and effec­
feedback information to motivate the group to engage in energy-saving tive behavior-based strategies and policies for energy conservation in a
behaviors. Another approach is to set an example of recommended be­ shared space.
haviors to motivate the group to conserve energy. Second, considering Despite some insights on energy-saving behaviors obtained from
the importance of self-transcendence group value and group interaction, group level perspective, there are certain limitations. First, the re­
group-based activities like lectures, courses, competition regarding en­ spondents in this study are only undergraduate students in shared resi­
ergy saving could cultivate or activate their self-transcendence values dences from Zhuhai, China due to some restricted research conditions.
and strengthen group interaction. On the one hand, these measures Broader sample size in different types of shared space (i.e., open plan
facilitate improving the group’s environmental awareness and re­ offices) or regions or countries is recommended in future research to
sponsibility of energy conservation. Besides, the perceived self- validate and generalize our findings. Second, this study only focused on
transcendence group values could be strengthened by the exposure to energy-saving behavior using cross-sectional data. However, people’s
group members who attend group-based activities. On the other hand, behavior may change with time and the environment they experience.
the group-based activities not only enrich individual energy-saving Thus, future studies could investigate behavioral change mechanisms
knowledge and skills, but also provide more chances for their commu­ through longitudinal time-series data.
nication, thereby encouraging them to behave in a sustainable way.
Finally, considering the effect of human-human interactions, a prom­ CRediT authorship contribution statement
ising approach is to encourage shared space users to join sustainable
groups or organizations. In these groups and organizations, individuals’ Jiaolan Zhu: Conceptualization, Methodology, Data curation,
group values and group norms regarding energy saving can be inter­ Writing – review & editing. Md Morshed Alam: Design, Writing – re­
nalized, and then greatly influence the thoughts and behaviors of their view & editing, Revision, Methodology, Software, Writing – review &
shared space users and foster communication and cooperation among editing. Zhikun Ding: Writing – review & editing. Jie Li: Revision,
them. Meanwhile, these shared space users can establish connections Writing – review & editing. Jiayuan Wang: Design, Investigation, Su­
with other individuals or subgroups and promote desirable energy- pervision, All authors contributed to writing the paper.
saving practices to diffuse among them. This concept is like the spread
of infectious diseases that a link between two individuals or social Declaration of competing interest
groups is enough to spread the disease. Absolutely, the diffusion of
sustainable behaviors is desired and could be achieved by increasing The authors declare that they have no known competing financial

10
J. Zhu et al. Journal of Cleaner Production 311 (2021) 127560

interests or personal relationships that could have appeared to influence Gao, L., et al., 2017. Application of the extended theory of planned behavior to
understand individual’s energy saving behavior in workplaces. Resour. Conserv.
the work reported in this paper.
Recycl. 127, 107–113.
Grønhøj, A., Thøgersen, J., 2012. Action speaks louder than words: the effect of personal
Acknowledgments attitudes and family norms on adolescents’ pro-environmental behaviour. J. Econ.
Psychol. 33 (1), 292–302.
Haslam, S.A., Reicher, S.D., 2015. Self-Categorization Theory. International
This study is supported by the National Nature Science Foundation of Encyclopedia of the Social & Behavioral Sciences, pp. 455–459.
China (Grant No. 71772125), Foundation from Zhuhai Federation of Hiratsuka, J., et al., 2018. Testing VBN theory in Japan: relationships between values,
Social Sciences, China (Grant No. 2019YB031), National Nature Science beliefs, norms, and acceptability and expected effects of a car pricing policy.
Transport. Res. F Traffic Psychol. Behav. 53, 74–83.
Foundation of China (Grant No.71974132) and the Natural Science Hong, J., et al., 2019. Impact of psychological factors on energy-saving behavior:
Foundation of Guangdong Province, China (Grant No. moderating role of government subsidy policy. J. Clean. Prod. 232, 154–162.
2018A0303130037). The authors are grateful to Professor Patrick X.W. Hong, T., et al., 2017. Ten questions concerning occupant behavior in buildings: the big
picture. Build. Environ. 114, 518–530.
Zou for his comments and suggestions. Hori, S., et al., 2013. The determinants of household energy-saving behavior: survey and
comparison in five major Asian cities. Energy Pol. 52, 354–362.
Appendix A. Supplementary data Hou, J., et al., 2004. Structural Equation Model and its Application. Education and
Science Press, Beijing.
IEA, 2019. Global status report for buildings and construction 2019. https://www.iea.or
Supplementary data to this article can be found online at https://doi. g/reports/global-status-report-for-buildings-and-construction-2019.
org/10.1016/j.jclepro.2021.127560. Jans, L., et al., 2018. A part of the energy "in crowd": changing people’s energy behavior
via group-based approaches. IEEE Power Energy Mag. 16 (1), 35–41.
Kavulya, G., Becerik-Gerber, 2012. Understanding the Influence of Occupant Behavior on
References Energy Consumption Patterns in Commercial Buildings, Asce Workshop of
Computing in Civil Engineering.
Abramst, D., et al., 1990. Knowing what to think by knowing who you are: self- Kline, R.B., 2011. Principles and practice of structural equation modeling. J. Am. Stat.
categorization and the nature of norm formation, conformity and group polarization. Assoc. 101 (12).
Br. J. Soc. Psychol. (29), 97–119. Latané, B., 1996. Dynamic Social Impact:The creation of culture by communication.
Ajzen, I., 2002. Perceived behavioral control, self-efficacy, locus of control, and the J. Commun. 46, 13–25.
theory of planned behavior. J. Appl. Soc. Psychol. 32 (4), 665–683. Lewin, K., 1951. Field Theory in Social Science. Harper, New York.
Alam, M.M., et al., 2017. Behavior Change of Building Users and Energy Consumption. Li, J., et al., 2015. Designers’ attitude and behaviour towards construction waste
Encyclopedia of Sustainable Technologies, pp. 189–196. minimization by design: a study in Shenzhen, China. Resour. Conserv. Recycl. 105,
Anderson, K., Lee, S., 2016. An empirically grounded model for simulating normative 29–35.
energy use feedback interventions. Appl. Energy 173, 272–282. Lo, S.H., et al., 2014. Only reasoned action an interorganizational study of energy-saving
Arimura, T.H., et al., 2016. Do social norms matter to energy-saving behavior? behaviors in office buildings. Energy Effic (7), 761–775.
Endogenous social and correlated effects. J. Assoc. Environ. Resour. Econ. 3 (3), Lopes, J.R.N., et al., 2019. A new model for assessing industrial worker behavior
525–553. regarding energy saving considering the theory of planned behavior, norm activation
Azar, E., Al Ansari, H., 2017. Framework to investigate energy conservation motivation model and human reliability. Resour. Conserv. Recycl. 145, 268–278.
and actions of building occupants: the case of a green campus in Abu Dhabi, UAE. Lu, Y., 2018. Occupant Behavior in Building Energy Management: Behavioral
Appl. Energy 190, 563–573. Characterization, Intervention and Forecasting. University of Maryland.
Baddeley, M., 2016. Behavioral approaches to managing household energy consumption. Mafimisebi, I.B., et al., 2018. A validated low carbon office building intervention model
In: Beckenbach, F., Kahlenborn, W. (Eds.), New Perspectives for Environmental based on structural equation modelling. J. Clean. Prod. 200, 478–489.
Policies through Behavioral Economics. Springer International Publishing Midden, C., et al., 2011. Persuasive power in groups: the influence of group feedback and
Switzerland, pp. 213–235. individual comparison feedback on energy consumption behavior. In: Proceedings of
Bales, R.F., 1970. Personality and Interpersonal Behavior. Holt, Rinehart and Winston, the 6th International Conference on Persuasive Technology Persuasive Technology
New York. and Design: Enhancing Sustainability and Health ACM.
Bulunga, A.A.L., Thondhlana, G., 2018. Action for increasing energy-saving behaviour in Mulville, M., et al., 2016. Energy-saving occupant behaviours in offices: change
student residences at Rhodes University, South Africa. Int. J. Sustain. High Educ. 19 strategies. Build. Res. Inf. 45 (8), 861–874.
(4), 773–789. Narayanan, S., Narasimhan, R., 2014. Governance choice, sourcing relationship
Burton, L.J., Mazerolle, S.M., 2011. Survey instrument validity part I: principles of characteristics, and relationship performance. Decis. Sci. J. 45 (4).
survey instrument development and validation in athletic training education Nunnally, J.C., Bernstein, I.H., 1978. Psychometric theory. Am. Educ. Res. J. 5 (3), 83.
research. Athl. Train. Educ. J. 6 (1), 27–35. Podsakoff, P.M., et al., 2003. Common method biases in behavioral research: a critical
CABEE, 2019. Research Report on building energy consumption in China 2019. http: review of the literature and recommended remedies. J. Appl. Psychol. 88 (5),
//www.cabee.org/site/content/23565.html. 879–903.
Chathura, W., et al., 2017. In: Design Challenges in Energy Conservation Strategies for Princes, E., Manurung, A.H., 2020. Taking advantage of social conformity in
Shared Spaces, ICED17: 21st International Conference on Engineering Design. entrepreneurship. J. Manage. Mark. Rev. 5 (1), 64–73.
University of British Columbia, Vancouver, Canada. Scherbaum, C.A., Popovich, P.M., 2008. Exploring individual-level factors related to
Chen, C.-f., Knight, K., 2014. Energy at work: social psychological factors affecting employee energy-conservation behaviors at work. J. Appl. Soc. Psychol. 38 (3),
energy conservation intentions within Chinese electric power companies. Energy 818–835.
Res. Soc. Sci. 4, 23–31. Song, Y., et al., 2019a. Assessing customers’ perceived value of the anti-haze cosmetics
Chen, X., et al., 2016. In: Poster Abstract: Personal Energy Footprint in Shared Building under haze pollution. Sci. Total Environ. 685, 753–762.
Environment, IPSN ’16: Proceedings of the 15th International Conference on Song, Y., et al., 2019b. Does haze pollution promote the consumption of energy-saving
Information Processing in Sensor Networks, pp. 1–2. appliances in China? An empirical study based on norm activation model. Resour.
Cialdini, R.B., Goldstein, N.J., 2004. Social influence: compliance and conformity. Annu. Conserv. Recycl. 145, 220–229.
Rev. Psychol. 55 (1), 591–621. Sovacool, B.K., et al., 2015. Integrating social science in energy research. Energy Res.
Cote, J.A., Buckley, M.R., 1987. Estimating trait, method, and error variance: Soc. Sci. 6, 95–99.
generalizing across 70 construct validation studies. J. Mar. Res. 24 (3), 315–318. Steg, L., et al., 2015. Understanding the human dimensions of a sustainable energy
Curran, P.J., et al., 1996. The robustness of test statistics to nonnormality and transition. Front. Psychol. 6, 805.
specification error in confirmatory factor analysis. Psychol. Methods 1 (1), 16–29. Steg, L., Vlek, C., 2009. Encouraging pro-environmental behaviour: an integrative review
Delmas, M.A., et al., 2013. Information strategies and energy conservation behavior: a and research agenda. J. Environ. Psychol. 29 (3), 309–317.
meta-analysis of experimental studies from 1975 to 2012. Energy Pol. 61, 729–739. Tam, W.Y.V., et al., 2018. A review on contemporary computational programs for
Ding, Z., et al., 2019. Agent-based model for simulating building energy management in Building’s life-cycle energy consumption and greenhouse-gas emissions assessment:
student residences. Energy Build. 198, 11–27. an empirical study in Australia. J. Clean. Prod. 172, 4220–4230.
Ding, Z.H., et al., 2018. Factors affecting heating energy-saving behavior of residents in Tang, Z., et al., 2019. Understanding employees’ energy saving behavior from the
hot summer and cold winter region. Nat. Hazards 95 (2). perspective of stimulus-organism-responses. Resour. Conserv. Recycl. 140, 216–223.
Driza, P.-J.N., 2014. Optimal Building Performance: Exploring Human Behavior Impacts Testa, F., et al., 2016. Determining factors of curtailment and purchasing energy related
on Energy and Water Consumption in Campus Residence Halls. UNIVERSITY OF behaviours. J. Clean. Prod. 112, 3810–3819.
FLORIDA. Wang, J., et al., 2019. Typical energy-related behaviors and gender difference for cooling
Elie, A., et al., 2014. Framework to evaluate energy-saving potential from occupancy energy consumption. J. Clean. Prod. 238, 1–16.
interventions in typical commerical buildings in the United States. J. Comput. Civ. Wayes, T., et al., 2017. Policy design for controlling set-point temperature of ACs in
Eng. 28 (1), 63–78. shared spaces of buildings. Energy Build. 134 (JAN), 105–114.
Fornell, C., Larcker, D.F., 1981. Evaluating structural equation models with unobservable Xiong, B., et al., 2015. A critical review of structural equation modeling applications in
variables and measurement error. J. Market. Res. 18 (1), 39–50. construction research. Autom. ConStruct. 49, 59–70.
Forsyth, D.R., 2017. In: Group Dynamics, seventh ed. Cengage Learning, Boston. Zhang, C.Y., et al., 2018. Impact factors of household energy-saving behavior: an
empirical study of Shandong Province in China. J. Clean. Prod. 185, 285–298.

11
J. Zhu et al. Journal of Cleaner Production 311 (2021) 127560

Zhang, M., et al., 2019. Study on the impact of haze pollution on residents’ green Zhao, C., et al., 2019. Exploring the influence of severe haze pollution on residents’
consumption behavior: the case of Shandong Province. J. Clean. Prod. 219, 11–19. intention to purchase energy-saving appliances. J. Clean. Prod. 212, 1536–1543.
Zhang, Y.-J., Peng, H.-R., 2017. Exploring the direct rebound effect of residential
electricity consumption: an empirical study in China. Appl. Energy 196, 132–141.

12

You might also like