Social Sciences & Humanities Open 11 (2025) 101612

Contents lists available at ScienceDirect

Social Sciences & Humanities Open

journal homepage: www.sciencedirect.com/journal/social-sciences-and-humanities-open

Regular Article

Does education affect energy behavior? Investigating the influence of

educational attainment in Indonesia
Lilia Endriana a,b , Djoni Hartono a,c,* , Khoirunurrofik a,c ,
Irfani Fithria Ummul Muzayanah a,c
a
Department of Economics, Faculty of Economics and Business, Universitas Indonesia, Depok, West Java, Indonesia
b
BPS-Statistics Indonesia, Jakarta Pusat, Indonesia
c
Research Cluster on Energy Modeling and Regional Economic Analysis, Department of Economics, Faculty of Economics and Business, Universitas Indonesia, Depok,
West Java, Indonesia

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

Keywords: This study investigates the causal relationship between educational attainment and various energy behaviors,
Energy behavior including clean cooking fuel use, energy conservation practices, and energy-efficient technology adoption.
Education Additionally, this study explores potential underlying mechanisms and regional heterogeneity. Utilizing a natural
Indonesia
experiment from the 1978 school year extension policy in Indonesia, this study applies a regression discontinuity
Clean fuel
Energy conservation
design and a two-stage least squares approach. Based on the 2017 National Socio-Economic Survey of Social
Energy efficiency Resilience Module data, the study results show that education significantly impacts the adoption of energy-
efficient technologies. However, education has a limited impact on clean cooking fuel use, showing signifi­
cance only in urban areas, and it does not significantly affect energy conservation behavior. These differences
arise mainly due to variations in the costs involved, the technical understanding required, and structural factors
such as affordability and accessibility. Further analysis reveals that education enhances income and access to
information sources but does not substantially improve energy-saving knowledge. These varying effects suggest
that education alone is insufficient in driving sustainable energy behavior. This implies the need for comple­
mentary measures such as clean energy subsidies, targeted behavioral campaigns, and expanded access to clean
energy infrastructure (particularly in rural areas). Moreover, given education’s role in improving access to in­
formation, public communication through mass and digital media could play a vital role in encouraging sus­
tainable energy practices.

1. Introduction Indonesia—a developing nation with the world’s fourth-largest popu­

lation (United Nations, 2019), estimated at approximately 270 million
Human activities, including energy behavior, which refers to both people (BPS, 2021)—where various demographic and socioeconomic
energy consumption and the behavioral aspects of energy use (Lopes factors influence energy behavior in varying ways. These differences
et al., 2012), have markedly elevated greenhouse gas emissions, highlight the need for policy strategies that consider regional disparities
resulting in increased global temperatures and accelerating climate and behavioral drives to enhance energy management efficacy.
change (Allen et al., 2018; IPCC, 2023). These environmental changes As environmental challenges intensify, understanding energy
have caused serious problems, including ecosystem degradation, food behavior is essential to promoting sustainable practices (Qiao & Yin,
insecurity, and negative effects on public health and economic stability 2021). Energy behavior includes unsustainable practices, which
(United Nations Environment Programme, 2023). As these challenges contribute to environmental degradation and global warming (IPCC,
become more pressing, policymakers worldwide are placing greater 2023; Santangelo & Tondeli, 2021; Steg & Vlek, 2009), and sustainable
focus on reducing their impact by promoting sustainable energy prac­ practices, which support the transition to a cleaner and more sustainable
tices. However, addressing this issue is particularly challenging in future (Steg, 2008; Steg & Vlek, 2009). Considering this dual

* Corresponding author. Department of Economics, Faculty of Economics and Business, Universitas Indonesia, Depok, West Java, Indonesia.
E-mail addresses: [email protected] (L. Endriana), [email protected] (D. Hartono), [email protected] (Khoirunurrofik), [email protected]
(I.F.U. Muzayanah).

https://doi.org/10.1016/j.ssaho.2025.101612
Received 2 January 2025; Received in revised form 21 May 2025; Accepted 21 May 2025
Available online 26 May 2025
2590-2911/© 2025 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
L. Endriana et al. Social Sciences & Humanities Open 11 (2025) 101612

characteristic, analyzing energy behavior from several viewpoints and between education and energy behavior in Indonesia, a major energy
determining its principal influencing factor are essential for compre­ consumer with a crucial role in global emission reduction efforts. The
hending how individuals make energy-related decisions. findings provide insights applicable not only to Indonesia but also to
Energy behavior is complex and has been studied from various emerging nations confronting analogous issues in education and energy
disciplinary perspectives, including psychology (Gifford & Nilsson, practices. Second, this study seeks to establish a causal link by
2014), energy intervention effectiveness (Spandagos et al., 2021), cul­ addressing the potential endogeneity of education, an issue that has
ture (Stephenson et al., 2010), and economics (Narasimha Rao & Reddy, received limited attention in previous research. To this end, we exploit a
2007). The literature suggests that energy behavior is influenced by natural experiment: the 1978 education policy in Indonesia, which
diverse factors, including cognitive (attitudes, beliefs, values), norma­ serves as a source of exogenous variation in educational attainment. We
tive (personal, social, descriptive, and injunctive norms), instrumental implement a regression discontinuity design (RDD) to strengthen the
(environmental awareness/knowledge, opportunities), emotional (trust, reliability of our causal estimates (Cattaneo et al., 2019; DiNardo & Lee,
environmental concern), and sociodemographic elements (income, 2011; Lee & Lemieux, 2010). Third, this study extends prior research by
gender, education, geographical location) (Frederiks et al., 2015; empirically analyzing how education influences energy behavior (spe­
Mundaca et al., 2022; Schlindwein & Montalvo, 2023). Among these, cifically, clean cooking fuel use, energy conservation practice, and the
education stands out as a crucial factor, as it shapes an individual’s adoption of energy-efficient technology) and how these effects vary
ability to address challenges and make informed decisions, particularly across regions. By integrating these aspects, this study offers a more
in household energy use, by improving knowledge and cognitive pro­ comprehensive understanding of education’s role in shaping energy
cessing (Marzano & Kendall, 2008). However, research on energy behavior, particularly in the context of policy design for sustainable
behavior in Indonesia has focused primarily on other factors, such as development.
residential preferences (Pratiwi et al., 2023), the effect of COVID-19 This study is structured as follows. The second section presents an
(Hartono et al., 2021), accessibility and affordability (Hartono et al., overview of education and energy behavior, along with Indonesia’s
2020), population density (Muzayanah et al., 2022), prepaid electricity 1978 education reform. The third section outlines the data and empirical
systems (Nugroho et al., 2017), and new technology adoption (Chou & methodology employed. The fourth section presents the results, and the
Yutami, 2014; Nakano et al., 2018). The influence of education in fifth section contains a discussion. The sixth section provides a conclu­
shaping energy behavior remains largely unexplored. sion to the study.
Research worldwide has shown a positive association between edu­
cation and modern energy use, including energy transition (Farsi et al., 2. Context study
2007; Liao et al., 2021; Narasimha Rao & Reddy, 2007; Pachauri &
Jiang, 2008; Rahut et al., 2019), energy-saving behavior, and the 2.1. Education and energy behavior
adoption of energy-efficient technologies (Broadstock et al., 2016;
Flamm, 2009; Mills & Schleich, 2012). However, most studies focus on Energy behavior encompasses all human actions that influence fuel
correlation rather than causation. Standard regression methods such as use to meet daily needs, including both physical actions and cognitive
OLS, multinomial logit, ordered probit, and Tobit do not account for processes in acquiring, utilizing, and managing energy-related materials
biases arising from unobserved personal traits, such as social conscience, and technology (Santangelo & Tondeli, 2021), as defined by the Inter­
individual and social responsibility, and risk perception, which can national Energy Agency Demand Side Management Energy Efficiency
affect both education and energy behavior at the same time. Conse­ (IEA DSM Energy Efficiency, 2014). Therefore, energy behavior includes
quently, existing estimates may be biased, making it difficult to deter­ both energy consumption patterns and behavioral dimensions (Lopes
mine whether education truly influences energy behavior or if et al., 2012), which are observable through the amount and type of
unobserved factors are responsible for the correlation. energy consumed, along with efforts in energy conservation and the
While these limitations challenge the validity of existing estimates, adoption of energy-efficient devices. However, inconsistencies exist in
research on the causal link between education and energy behavior re­ some studies regarding the distinction between ‘energy conservation’
mains underexplored. Some studies employ an instrumental variable and ‘energy efficiency’ (Oikonomou et al., 2009). Energy conservation
approach to address this gap, relying on changes in the compulsory involves reducing energy consumption through more economical usage
education law (Meyer, 2015; Powdthavee, 2021) and teacher avail­ and managing resources wisely (Zainudin et al., 2021). Meanwhile,
ability (Chankrajang & Muttarak, 2017) as instruments to estimate the energy efficiency minimizes energy use through technological ad­
impact of education. However, findings are mixed; some suggest that vancements that lower consumption without requiring behavioral
education reduces energy consumption and promotes environmentally changes (Oikonomou et al., 2009; Zainudin et al., 2021) or enable the
conscious buying (Meyer, 2015), while others find no statistically sig­ same function with less energy (Radzi & Hassan, 2021). Although both
nificant impact on energy conservation (Chankrajang & Muttarak, 2017; concepts aim to save energy, conservation emphasizes behavioral ad­
Powdthavee, 2021). justments, whereas efficiency relies on technological improvements.
Moreover, much of the causal literature tends to place energy Considering these differences, this study investigates how education
behavior within the broader context of general pro-environmental be­ influences energy behavior, especially in clean cooking choices, and
haviors (such as waste management, recycling, and the use of personal distinguishes between energy-efficient technology adoption and energy
shopping bags) rather than treating it as a distinct behavior category conservation practices.
(Chankrajang & Muttarak, 2017; Meyer, 2015; Powdthavee, 2021). Regarding how education influences energy behavior, the literature
Although informative, this approach often overlooks the unique char­ indicates both direct and indirect mechanisms. Education directly en­
acteristics and policy relevance of energy behavior itself. As highlighted hances individuals’ knowledge, skills, and awareness, thereby influ­
by Esfandiar et al. (2022), there is a growing need for more targeted encing their attitudes, behaviors, culture, and preferences related to
studies that analyze energy behavior within specific contexts and among energy consumption. The notion of bounded rationality, which posits
particular populations. that decision-making is constrained by information availability (Simon,
Taken together, the limitations in empirical methods, mixed find­ 1955, 1957), is pertinent in this context. Education helps individuals
ings, and the narrow treatment of energy behavior underscore the need overcome limitations by expanding access to information and improving
for further research that can more precisely identify the causal impact of their capacity to engage with new knowledge (Grossman, 2006).
education on distinct forms of energy behavior. Increased knowledge and awareness gained through education
This study makes three contributions to the literature. First, to the strengthen awareness of the negative impacts of energy use and tradi­
best of our knowledge, this study is the first to examine the relationship tional fuels, influencing choices for cleaner fuels (Farsi et al., 2007; Liao

2
L. Endriana et al. Social Sciences & Humanities Open 11 (2025) 101612

et al., 2021; Rahut et al., 2019; S. Wang et al., 2020), promoting more critical energy behavior variables, the 2017 dataset meets these criteria,
economical energy usage, and encouraging the adoption of reinforcing its suitability for this study. It remains the most recent and
energy-efficient technologies (Flamm, 2009). Education generally im­ comprehensive dataset for analyzing household energy behavior. It
proves socioeconomic status or income, increasing individuals’ capacity covers 74,253 households across 34 provinces and 514 municipalities in
to access essential resources (Chankrajang & Muttarak, 2017; Rahut Indonesia and provides extensive demographic and geographic repre­
et al., 2019). Higher incomes facilitate greater access to clean energy sentation. The dataset uniquely captures energy sources, energy-saving
and energy-efficient equipment (Farsi et al., 2007). Furthermore, edu­ practices, the adoption of energy-efficient technology, environmental
cation improves access to various information sources, which in turn, knowledge, and sources of environmental information. Appendix A
with more comprehensive knowledge about energy use impacts, emis­ presents the summary statistics of the variables.
sion reduction methods, and appropriate adaptations, empower in­ To further evaluate the causal impact of education on energy
dividuals to improve their behavior more sustainably (Chankrajang & behavior, this study leverages a policy-driven variation in schooling
Muttarak, 2017). duration. Specifically, it uses a birth year to determine whether in­
dividuals experienced an extended schooling period during the 1978
2.2. The 1978 school year system change and education policy in academic calendar reform. In Indonesia, children typically begin
Indonesia elementary school at the age of seven. Consequently, those born in 1972
or later followed the standard schooling duration, whereas those born in
In mid-1978, the Indonesian government introduced Ministerial 1971 or earlier remained in school for an additional period—unless they
Decree No. 0211/U/1978 to improve the efficiency and effectiveness of exited the education system before 1978. Since the reform assigned
the education system, organize and stabilize teaching and learning ac­ schooling duration based on birth year rather than individual choice, it
tivities, and align the academic year with the fiscal calendar (Solihat & created a quasi-random variation in educational attainment. This study
Syukur, 2020). This policy mandated a nationwide shift in the academic exploits this natural experiment to estimate how education influences
calendar, changing the school year cycle from January–December to energy behavior.
July–June. As a result, all primary and secondary schools were required
to extend the academic year by six months, keeping the students at the 3.2. Empirical strategy
same grade level until June 1979.
Although the reform did not alter the curriculum, the additional six Unobserved individual characteristics such as personal and social
months were used to reinforce previously taught material, helping responsibility, social conscience, and risk perception can influence both
address low student absorption rates. This extension gave teachers more educational attainment and energy behavior. This poses a potential bias
time to review lessons, provide targeted assistance to students with in estimating the impact of education on energy behavior. Therefore, to
learning difficulties, and strengthen overall comprehension. As a result, minimize this bias and establish a causal relationship, this study utilizes
students showed improved academic performance, which in turn the exogenous variation that influences education but is uncorrelated
increased parental appreciation for education. Studies have shown that with energy behavior, specifically the 1978 academic year extension
the reform contributed to lower grade repetition rates, increased policy reform, as previously implemented in other studies (Masuda &
educational attainment, and improved graduation rates in middle and Yudhistira, 2020; Parinduri, 2014, 2017; Samarakoon & Parinduri,
high schools (Masuda & Yudhistira, 2020; Parinduri, 2014, 2017; 2015). The academic year extension was announced in mid-1978, so
Pischke, 2007; Samarakoon & Parinduri, 2015). parents could not intervene in whether their children fell into the
Beyond the 1978 reform, the government introduced additional affected group or not, because children born in 1971 had already started
policies to expand education access and quality, such as the SD Inpres school by January 1978, and those born in 1972 had not yet started their
program (1973), the six-year compulsory education policy (1984), and education (Parinduri, 2017). Therefore, the setup in this study meets the
the nine-year compulsory education policy (1994). However, these ini­ criterion of ‘as good as randomly assigned’.
tiatives did not affect this study’s identification strategy, as they did not The 1978 academic year extension policy created a discontinuity in
create different educational outcomes for students who commenced the likelihood of experiencing an extended period of schooling for
primary school in the 1978/79 academic year (Masuda & Yudhistira, certain cohorts (birth years 1971 and 1972), which fits well with the
2020; Parinduri, 2014, 2017; Samarakoon & Parinduri, 2015). RDD approach. The estimation process involves determining the poly­
nomial order, choosing the bandwidth, and calculating the RDD esti­
3. Methods mates (Cattaneo et al., 2019). This study follows the recommendation by
Gelman and Imbens (2019), choosing to use local linear and quadratic
3.1. Data polynomials instead of global high-order polynomials, due to their
greater stability, lower sensitivity to model specification choices, and
This study analyzes the impact of education on energy behavior more reliable confidence interval coverage. The sample is restricted to
using data from the 2017 National Socio-Economic Survey of Resilience individuals born between 1960 and 1984, with a bandwidth of 12 years
Module (Susenas Modul Hansos 2017), collected and managed by Sta­ on either side of the cutoff (born in 1972). This restriction ensures that
tistics Indonesia (BPS). This dataset provides comprehensive variables the oldest cohort in the sample (born in 1960) was in the educational
on education and energy behavior, making it well-suited for this anal­ system when the policy was implemented, whereas individuals born
ysis. The Susenas Modul Hansos is typically conducted every three years. before 1960 were likely not affected by the policy, having already
However, the 2020 survey was modified due to the COVID-19 pandemic, graduated from high school, considering that the school year extension
resulting in the removal of key energy behavior questions to streamline policy applied only from primary to high school levels.
data collection and reduce respondent burden. As a result, essential This study employs RDD with two-stage least squares (2SLS) esti­
variables for this study were unavailable. Furthermore, the 2023 survey mation to evaluate education’s impact on energy behavior, consistent
was not conducted due to budget constraints and policy refocusing by with prior studies (Masuda & Yudhistira, 2020; Meyer, 2015). In the first
BPS, preventing access to a more recent dataset with comparable energy stage, the study estimates the impact of the academic year change on
behavior information. educational achievement. In the second stage, it estimates the impact of
The use of older datasets in academic research is justified under educational attainment on energy behavior.
certain conditions, such as the absence of recent data or their collection The first-stage equation is described as follows:
through rigorous methodologies (Ketchen et al., 2023). Since the 2023
educationic = α0 + α1 Ric + α2 Xic + f(yobic ) + εic (1)
Susenas Modul Hansos was not conducted and the 2020 version omitted

3
L. Endriana et al. Social Sciences & Humanities Open 11 (2025) 101612

where educationic represents the educational attainment of individual i actual cutoff. Fig. 1b shows no significant discontinuity when shifting
from cohort c, indicated by the continuous years of schooling completed. the cutoff from 1972 to 1979, with minor changes, if any, being higher
An individual belongs to a cohort affected by the policy, determined by on the side after the cutoff. This contrasts with Fig. 1a, which shows
the dummy variable Ric (1 if born in 1971 or earlier, 0 if born in 1972 or discontinuity around the actual cutoff, with the left side being higher. A
( ) more complete analysis is discussed in section 4.5. Moreover, adding
later). f yobic is a quadratic function of the individual’s birth year. εic
and subtracting bandwidths are performed considering the
denotes the error term, where to address potential correlations among
bias-variance trade-off, where a smaller bandwidth increases precision
individuals in the same cohort, the standard errors are clustered at the
and reduces specification errors, but also tends to increase the variance
birth cohort level.
of the estimated coefficient due to fewer observations (Cattaneo et al.,
Xic reflects the control variables, including a gender dummy (1 if
2019).
female and 0 if male) and a rural–urban location (1 if rural and 0 if
urban), which are assumed to be predetermined covariates set before the
policy (Cattaneo et al., 2019). These variables help minimize sample 4. Results
variation in RD estimation, but do not affect individuals born before or
after the policy (Lee & Lemieux, 2010; Meyer, 2015). Adding these 4.1. The school year extension impact on educational attainment
covariates also helps in model falsification and validation (Cattaneo
et al., 2019, 2021). Table 1 presents the first-stage regression results, estimating the
The second-stage regression equation estimating the impact of edu­ impact of the school year extension policy on educational attainment.
cation on energy behavior is as follows: Columns 1 and 2 use school years as a proxy for the education level
attained, using a linear or quadratic polynomial of the normalized birth
EBic = β0 + β1 educationic + β2 Xic + f(yobic ) + μic (2) year. Columns 3–6 display the highest level of education completed as
an alternative measure of educational attainment, using a quadratic
where EBic represents energy behavior outcomes, including clean polynomial.
cooking fuel use, energy conservation behavior, and the adoption of The findings show that the reform significantly increased educa­
energy-efficient technology, for individual i from cohort c. Clean cook­ tional attainment. The estimates remain stable across the polynomial
ing fuel use is a dummy variable formed from various fuel options (1 for specification, suggesting the reform increased the schooling duration by
using clean fuel and 0 if not). The classification of clean fuel is based on approximately 0.694–0.701 years. These results are consistent with
the CO2 emission potential during cooking activities, signifying energy prior studies (Masuda & Yudhistira, 2020; Parinduri, 2014, 2017),
with very low emission pollution levels during combustion processes, which found similar estimates ranging from 0.611 to 0.731 years.
including biogas, liquefied petroleum gas (LPG), natural gas, electricity, Furthermore, the policy enhanced the chances of finishing primary and
ethanol, and solar energy (International Energy Agency, 2017; junior high school.
Netherlands Enterprise Agency, 2022). Energy conservation behavior is
a binary variable transformed from a Likert scale, as in the response to
the question “How often do you turn off lights when not in use?” (1 for 4.2. Impact of educational attainment on energy behavior
always or often, and 0 for others). Similarly, the adoption of
energy-efficient technology is a binary variable transformed from re­ Table 2 displays the estimation results of the relationship between
sponses to the question “How often do you consider low electricity usage education and various energy behaviors, using OLS as a baseline model
when purchasing electronic appliances?” (1 for always or often, and and 2SLS as the main estimation approach. The OLS results (first col­
0 for others). More details can be found in Appendix B. The coefficient β1 umn) indicate that an increase in schooling years is positively correlated
indicates the impact of education on energy behavior. If education en­ with pro-environmental energy behaviors at the 1 % significance level.
hances energy behavior, then β1 is expected to be positive for all energy However, these estimates are subject to endogeneity concerns. In
behavior outcomes. contrast, the 2SLS estimates (Columns 2 and 3), which account for the
For model specification validation, this study conducts robustness endogeneity of education using local linear and quadratic polynomials
checks using placebo tests and bandwidth adjustments. The placebo test of the normalized birth year, provide more credible results.
involves shifting the cutoff point (Eggers et al., 2024) and then exam­ Overall, this study finds limited evidence that education significantly
ining whether there is discontinuity, and the impact is contrary to the impacts energy behavior in Indonesia, particularly affecting energy

Fig. 1b. Placebo test – Educational attainment by year of birth (placebo cut­
Fig. 1a. Educational attainment by year of birth (actual cutoff 1972). off 1979).

4
L. Endriana et al. Social Sciences & Humanities Open 11 (2025) 101612

Table 1
Impact of the school year extension policy on the education level attained.
(1) (2) (3) (4) (5) (6)

School years Education level attained

Primary school Junior high Senior high College

school school

Experiencing education reform (1 if born in 1971 or 0.694*** 0.701*** 0.110*** 0.076*** (0.006) 0.001 (0.006) 0.032***
earlier) (0.051) (0.051) (0.004) (0.004)
R-squared 0.140 0.140 0.050 0.112 0.068 0.009
Observations 96,712 96,712 96,712 96,712 96,712 96,712

Normalized year of birth polynomial

Linear ✓ ​ ​ ​ ​ ​
Quadratic ​ ✓ ✓ ✓ ✓ ✓
Gender and rural–urban dummies ✓ ✓ ✓ ✓ ✓ ✓

Note: Standard errors reported in parentheses are clustered at the birth cohort level. ***, **, and * denote statistical significance at the 1 %, 5 %, and 10 % levels,
respectively.

in the next section, where potential mechanisms and contextual expla­

Table 2 nations are discussed.
Impact of educational attainment on energy behavior.
(1) (2) (3)

OLS 2SLS 2SLS 4.3. Heterogeneity analysis of education’s impact on energy behavior
A. Choices of cooking fuel
Clean cooking fuel This study examines the heterogeneity of education’s influence on
Years of schooling 0.020*** 0.002 0.005 energy behavior across urban and rural regions, each exhibiting unique
(0.000) (0.008) (0.008) characteristics (Tayyaba, 2012), particularly regarding education, so­
First-stage F-statistics 190.60 187.01
–
cioeconomic status, energy access, and behavioral patterns. Based on the
Observations 96,712 96,712 96,712
summary statistics (Appendix A), the prevalence of clean cooking fuel
B. Energy conservation use and energy-efficient behavior is notably higher in urban areas than
Turn off the lights when not in use
in rural areas. Meanwhile, the rate of energy conservation behavior
Years of schooling 0.003*** − 0.002 − 0.002
​ (0.000) (0.007) (0.007) appears relatively uniform across both regions, with an average of 84 %.
First-stage F-statistics – 190.60 187.01 In terms of socioeconomic characteristics, urban residents tend to have
Observations 96,712 96,712 96,712 higher levels of education, income, and access to information sources
C. Energy efficiency than rural residents, indicating that residents in urban areas are more
Consider low power when purchasing electronics advanced than their rural counterparts. The descriptive differences
Years of schooling 0.015*** 0.025*** 0.027*** provide initial indications of potential disparities in energy behavior
(0.000) (0.009) (0.009)
across regions.
First-stage F-statistics – 190.60 187.01
Observations 96,712 96,712 96,712 Table 3 displays the estimation results. The first-stage F-statistics
exceed the critical threshold established by Stock & Yogo’s guideline
Normalized year of birth polynomial
Linear ​ ✓ ​ (Stock & Yogo, 2005) in both areas, which shows that the instrument is
Quadratic ​ ​ ✓
Gender and rural–urban dummies ✓ ✓ ✓ Table 3
Note: Standard errors reported in parentheses are clustered at the birth cohort Heterogeneity of energy behavior based on regional subsamples.
level. ***, **, and * denote statistical significance at the 1 %, 5 %, and 10 % (1) (2)
levels, respectively.
Urban Rural

A. Choices of cooking fuel

efficiency behavior, while not substantially influencing clean cooking
Clean cooking fuel
fuel use and energy conservation. This aligns with findings in England Years of schooling 0.018* − 0.005
and Wales by Powdthavee (2021), who employed both sharp and fuzzy (0.011) (0.012)
RDD to examine the impact of education reforms on pro-environmental First-stage F-statistics 69.81 115.51
behaviors, such as switching off lamps and electronic devices when not Observations 42,166 54,546

in use. B. Energy conservation

All first-stage F-statistics exceed the critical threshold of the Stock & Turn off the lights when not in use
Years of schooling − 0.008 0.002
Yogo weak instrument test (Stock & Yogo, 2005), indicating that the
(0.011) (0.009)
2SLS estimates are less likely to experience bias due to the weak in­ First-stage F-statistics 69.81 115.51
strument. The 2SLS estimates in Panel A show that years of schooling Observations 42,166 54,546
have no significant effect on the likelihood of using clean cooking fuels. C. Energy efficiency
Similarly, Panel B shows no significant effect on turning off lights when Consider low power when purchasing electronics
not in use. However, Panel C reveals that education significantly impacts Years of schooling 0.016 0.035***
the adoption of energy-efficient technology. Column 3 shows that an (0.013) (0.012)
First-stage F-statistics 69.81 115.51
additional year of schooling increases the probability of considering
Observations 42,166 54,546
low-electricity consumption when purchasing electronic devices by 2.7
percentage points. Compared with the OLS estimates, the 2SLS co­ Note: Standard errors reported in parentheses are clustered at the birth cohort
efficients are higher, suggesting that the OLS results are downward level. ***, **, and * denote statistical significance at the 1 %, 5 %, and 10 %
levels, respectively. All regressions include a quadratic polynomial of normal­
biased due to omitted variable bias. These findings are further explored
ized birth years and control for gender and rural–urban status.

5
L. Endriana et al. Social Sciences & Humanities Open 11 (2025) 101612

strong. In Panel A, the results show that an additional year of schooling by 6.9 percentage points, with a stronger effect in urban areas (8.4
increases the likelihood of using clean cooking fuels by 1.8 percentage percentage points) than in rural areas (5.9 percentage points). This
points in urban areas, and the effect is statistically significant at the 10 % finding aligns with studies from India and Thailand (Chankrajang &
level. However, the study finds no significant effect in rural areas. Panel Muttarak, 2017; Rahut et al., 2019), highlighting the role of education in
B shows no significant causal relationship between education and expanding access to environmental and energy-related information.
turning off lights in either subgroup. In contrast, Panel C reveals that Column 3 shows no statistically significant correlation between an
education significantly increases the likelihood of adopting additional year of schooling and self-reported energy-saving knowledge.
energy-efficient technologies in rural areas by 3.5 percentage points but The near-zero coefficients in both urban and rural areas indicate that
has no significant effect in urban settings. This aligns with findings from marginal increases in education do not meaningfully raise awareness.
China, indicating that education does not significantly impact the This finding aligns with studies in Europe, showing that education level
pro-environmental attitudes and behaviors of urban residents but is does not significantly influence how well-informed an individual is
positively significant in rural areas (Wang et al., 2022). about environmental issues (Meyer, 2015).

4.4. Potential mechanisms 4.5. Robustness checks

To explore how education affects energy behavior, this section in­ This study ensures the reliability of its findings by conducting
vestigates potential mechanisms based on economic and behavioral robustness checks through a placebo test and bandwidth adjustments, in
theories. The first is the income mechanism, grounded in human capital addition to comparing OLS and 2SLS models with linear and quadratic
theory, which suggests that education enhances productivity and earn­ polynomials (as presented in Table 2). The results, summarized in
ing potential (Becker, 1993). It is expected that better education leads to Table 5, consistently confirm the robustness of the estimated effects of
higher income, thereby increasing affordability for accessing clean en­ education on energy behavior.
ergy resources and energy-efficient equipment. The second is the First, a placebo test was conducted, assuming the school year
knowledge mechanism, which posits that education improves access to extension policy was implemented in 1985. Under this assumption, the
and absorption of information relevant to energy-related decisions fictitious policy would have affected individuals born in 1978 or earlier,
(Grossman, 2006). It is expected that higher education increases shifting the cutoff from the actual birth cohort year of 1972–1979. As
knowledge, thereby encouraging individuals to adopt energy behaviors expected, this falsified policy had no significant effect on educational
that consider the environment. In this study, we use energy-saving attainment (see Appendix C), aligning with Fig. 1b. Consequently, the
knowledge and access to information sources as proxies for knowl­ estimated impact of education on energy behavior also differs from the
edge. The results are presented in Table 4. main findings, further validating the model’s credibility. See Table 5,
Due to the absence of reliable income data, this study uses household column 1, for more details.
expenditure as a proxy. Column 1 shows that an additional year of Second, a bandwidth adjustment was performed to test the sensi­
schooling increases household expenditure by 7.7 percent on average. tivity of the results. The bandwidth was narrowed to include birth co­
The effect is stronger in urban areas, where an additional year of horts from 1962 to 1982 and widened to include birth cohorts from 1958
schooling increases expenditure by 8.7 percent compared to 7.0 percent to 1986. As shown in Table 5 (columns 2 and 3), the estimated impact of
in rural areas. These results support the hypothesis that income serves as education on energy behavior remains consistent in both statistical
a channel through which education influences energy behavior. significance and magnitude, reinforcing the stability and robustness of
Column 2 reports the impact of education on access to information the findings.
sources. An additional year of schooling improves access to information
Table 5
Table 4 Robustness checks.
The effects of educational attainment on income, access to information sources, (1) (2) (3)
and energy-saving knowledge.
Placebo cutoff Narrower Wider
(1) (2) (3) bandwidth bandwidth
Income Access to information Energy-saving A. Choices of cooking fuel
sources knowledge Clean cooking fuel
Years of schooling − 0.004 0.005 0.002
Years of schooling 0.077*** 0.069*** 0.000
(0.011) (0.011) (0.009)
(0.009) (0.017) (0.004)
First-stage F- 115.94 103.74 142.12
First-stage F- 187.01 187.01 187.01
statistics
statistics
Observations 96,712 82,743 109,798
Observations 96,712 96,712 96,712
B. Energy conservation
Urban subsample
Turn off the lights when not in use
Years of schooling 0.087*** 0.084** 0.000
Years of schooling − 0.026*** − 0.006 − 0.002
(0.015) (0.030) (0.005)
(0.009) (0.009) (0.008)
First-stage F- 69.81 69.81 69.81
First-stage F- 115.94 103.74 142.12
statistics
statistics
Observations 42,166 42,166 42,166
Observations 96,712 82,743 109,798

C. Energy efficiency
Rural subsample Consider low power when purchasing electronics
Years of schooling 0.070*** 0.059** 0.001 Years of schooling 0.016 0.026*** 0.027***
(0.011) (0.020) (0.006) (0.012) (0.012) (0.010)
First-stage F- 115.51 115.51 115.51 First-stage F- 115.94 103.74 142.12
statistics statistics
Observations 54,546 54,546 54,546 Observations 96,712 82,743 109,798

Note: Standard errors reported in parentheses are clustered at the birth cohort Note: Standard errors reported in parentheses are clustered at the birth cohort
level. ***, **, and * denote statistical significance at the 1 %, 5 %, and 10 % level. ***, **, and * denote statistical significance at the 1 %, 5 %, and 10 %
levels, respectively. All regressions include a quadratic polynomial of normal­ levels, respectively. All regressions include a quadratic polynomial of normal­
ized birth years and control for gender and rural–urban status. ized birth years and control for gender and rural–urban status.

6
L. Endriana et al. Social Sciences & Humanities Open 11 (2025) 101612

5. Discussion These findings suggest that while education shapes certain aspects of
energy behavior, it is not a standalone driver. Economic constraints,
This study empirically establishes a causal relationship between energy access, market incentives, and public policy likely play decisive
education and energy behavior in Indonesia. The findings reveal that roles. For example, high upfront costs, limited access to clean fuel al­
education influences different types of energy behavior in distinct ways. ternatives, or weak institutional support may inhibit behavior change,
To better understand this variation, beyond the mechanisms and even among informed and educated individuals.
regional heterogeneity that have been discussed, it is important to
highlight other aspects: the nature of each energy behavior, the proxy 6. Conclusion
used in this study, and relevant external factors beyond the boundaries
of the study that may help explain the results. Using the 1978 school year extension policy in Indonesia as a natural
The three types of energy behavior analyzed in this study differ in experiment, this study looks at the causal effect of educational attain­
effort, cost, and technical understanding. Cooking fuel choices often ment on energy behavior. The findings indicate that the influence of
involve trade-offs in cost, convenience, and comfort. Energy conserva­ education on energy behavior is not homogeneous but depends on
tion behavior, in contrast, typically requires no financial investment, behavior features. Education significantly influences the adoption of
tends to lower expenses over time, and is relatively easy for most people energy-efficient technology, particularly in rural areas, but has no sig­
to adopt. Energy efficiency behavior requires greater technical knowl­ nificant impact on clean cooking fuel use, except in urban areas. It also
edge and willingness to make upfront investments for long-term savings. does not significantly influence energy conservation behavior, which is
Education significantly influences the use of clean cooking fuel in generally easier to adopt and does not require financial sacrifice.
urban areas but has little impact in rural areas. This discrepancy likely These results indicate that education plays a more meaningful role in
reflects variations in energy infrastructure, fuel accessibility, and eco­ shaping complex and forward-looking decisions. This influence is
nomic factors. In rural communities, traditional fuels such as firewood especially evident in behaviors that require technical understanding and
are more accessible, cheap, and often cost-free, which may diminish the upfront investment. In contrast, behaviors constrained by accessibility
role of education in influencing clean fuel choices. For many rural and affordability, such as clean cooking use in rural areas or simple
households, economic considerations and affordability take precedence, conservation actions, appear to be less responsive to education. The
particularly among those with lower levels of education and income. varying effects across rural–urban contexts highlight the importance of
The income mechanism supports this interpretation: education is posi­ supporting infrastructure, affordability, energy access, and broader so­
tively correlated with expenditure (used here as a proxy for income), cioeconomic conditions in influencing behavioral outcomes.
especially in urban areas. This aligns with the energy ladder theory, This study offers several policy implications. First, while improving
which suggests that rising income facilitates the transition to modern educational attainment remains important, it is not sufficient by itself to
fuels (Muller & Yan, 2018), and is consistent with findings in India and promote sustainable energy behavior. Complementary measures such as
China that show urban households tend to transition more rapidly to­ clean energy subsidies, behavioral campaigns, and improved access to
ward cleaner energy (Farsi et al., 2007; Pachauri & Jiang, 2008). In clean energy alternatives, particularly in rural areas, are necessary.
addition, persistent challenges such as limited clean cooking fuel Second, integrating environmental and energy issues into school
availability, affordability barriers, and insufficient program outreach curricula may strengthen education’s long-term influence. Third, the
continue to constrain energy adoption. This is particularly evident in study finds that education improves access to information sources,
eastern Indonesia, where many rural communities are concentrated. suggesting that public communication through mass and digital media
This structural limitation suggests that while education can play a cat­ can play a crucial role in encouraging sustainable energy behavior.
alytic role, it is not sufficient on its own. This study, while contributing valuable insights, has several limita­
Another important finding is that education does not significantly tions. First, the proxy used to measure energy-saving knowledge relies
affect energy conservation or energy-saving knowledge. At first glance, on a single survey question and may not fully capture the depth of re­
this evidence may appear contradictory to one of the objectives of ed­ spondents’ understanding. Second, due to limited income data, this
ucation, which includes knowledge (Marzano & Kendall, 2008), and the study uses expenditure as a proxy for income, potentially introducing
initial expectation that education would enhance awareness and sub­ measurement bias. Third, the analysis relies on cross-sectional data from
sequently influence energy behavior. However, we should interpret the Susenas Modul Hansos 2017, limiting the ability to assess long-term
these results cautiously. In this study, turning off lights when not in use or dynamic behavioral changes. Finally, this study estimates a local
represents energy conservation behavior. As shown in Appendix A, 84 % average treatment effect (LATE), capturing the impact of the 1978
of respondents report practicing energy conservation, and 95 % policy only on the subpopulation affected by the policy; thus, the results
correctly recognize that saving electricity contributes to energy con­ should not be generalized to the broader population or other national
servation. This suggests that awareness and basic conservation practices contexts (Angrist & Pischke, 2009). Future research should explore these
are already widespread, even among individuals with lower educational limitations using panel data, existing natural experiments, or experi­
attainment. Therefore, additional years of schooling may not signifi­ mental designs in other developing countries to enrich the literature on
cantly increase awareness or alter commonly adopted behavior. energy behavior.
Finally, education significantly affects energy efficiency behavior,
particularly when considering low electricity consumption in appliance CRediT authorship contribution statement
purchases. This suggests that education plays a more substantial role in
shaping complex, future-oriented decisions that require technical un­ Lilia Endriana: Writing – original draft, Software, Methodology,
derstanding and planning. The effect of education on energy efficiency is Formal analysis, Data curation, Conceptualization. Djoni Hartono:
found to be significant in rural areas but not in urban areas. Several Writing – review & editing, Validation, Supervision, Methodology,
factors may explain this rural–urban gap. Appendix A shows that urban Investigation, Conceptualization. Khoirunurrofik: Writing – review &
residents are more likely to consider low-power appliances than their editing, Validation, Supervision, Investigation, Formal analysis. Irfani
rural counterparts. Coupled with generally higher levels of education in Fithria Ummul Muzayanah: Writing – review & editing, Validation,
urban areas, this suggests a saturation effect, where further increases in Supervision, Investigation.
education result in diminishing marginal gains. In contrast, rural areas
offer greater room for behavioral change. Lower baseline education Ethical statement
levels and lower adoption of energy-efficient technologies imply that
additional schooling can produce a more pronounced effect. Ethical approval is not applicable to this manuscript.

7
L. Endriana et al. Social Sciences & Humanities Open 11 (2025) 101612

Declaration of the use of AI Farsi, M., Filippini, M., & Pachauri, S. (2007). Fuel choices in urban Indian households.
Environment and Development Economics, 12(6), 757–774. https://doi.org/10.1017/
S1355770X07003932
During the preparation of this manuscript, the authors used ChatGPT Flamm, B. (2009). The impacts of environmental knowledge and attitudes on vehicle
to identify unclear statements and enhance readability. After using this ownership and use. Transportation Research Part D: Transport and Environment, 14(4),
tool/service, the authors reviewed and edited the content as necessary 272–279. https://doi.org/10.1016/j.trd.2009.02.003
Frederiks, E. R., Stenner, K., & Hobman, E. V. (2015). The socio-demographic and
and take full responsibility for the content of the published article. psychological predictors of residential energy consumption: A comprehensive
review. Energies, 8(1), 573–609. https://doi.org/10.3390/en8010573
Data availability Gelman, A., & Imbens, G. (2019). Why high-order polynomials should not Be used in
regression discontinuity designs. Journal of Business & Economic Statistics, 37(3),
447–456. https://doi.org/10.1080/07350015.2017.1366909
The datasets generated or analyzed during the current study are Gifford, R., & Nilsson, A. (2014). Personal and social factors that influence pro-
available from the corresponding author upon reasonable request. environmental concern and behaviour: A review. International Journal of Psychology,
49(3), 141–157. https://doi.org/10.1002/ijop.12034
Grossman, M. (2006). Education and non-market outcomes. In E. A. Hanushek, &
Funding F. Welch (Eds.), Handbook of the economics of education (pp. 577–633). https://doi.
org/10.1016/S1574-0692(06)01010-5
Hartono, D., Hastuti, S. H., Balya, A. A., & Pramono, W. (2020). Modern energy
This study is partially supported by Universitas Indonesia under consumption in Indonesia: Assessment for accessibility and affordability. Energy for
grant number NKB-29/UN2.RST/HKP.05.00/2024. Sustainable Development, 57, 57–68. https://doi.org/10.1016/j.esd.2020.05.002
Hartono, D., Yusuf, A. A., Hastuti, S. H., Saputri, N. K., & Syaifudin, N. (2021). Effect of
COVID-19 on energy consumption and carbon dioxide emissions in Indonesia.
Declaration of competing interest Sustainable Production and Consumption, 28, 391–404. https://doi.org/10.1016/j.
spc.2021.06.003
The authors declare that they have no known competing financial IEA. (2017). Energy access outlook 2017. From poverty to prosperity. https://doi.
org/10.1787/9789264285569-en.
interests or personal relationships that could have appeared to influence
IEA DSM Energy Efficiency. (2014). Subtasks of task 24 definitions: Behaviour and
the work reported in this paper. energy. 2. https://userstcp.org/wp-content/uploads/2019/12/3.Task24_Ph
ase1_Subtask1_Definitions_combined.pdf.
Ketchen, D. J., Roccapriore, A. Y., & Connelly, B. L. (2023). Using old data: When is it
Acknowledgments
appropriate? Journal of Management, 49(8), 2541–2548. https://doi.org/10.1177/
01492063231177785
The authors acknowledge the partial support from the Hibah Pub­ Lee, D. S., & Lemieux, T. (2010). Regression discontinuity designs in economics. Journal
likasi Terindeks Internasional (PUTI) Pascasarjana, Universitas Indonesia of Economic Literature, 48, 281–355. https://doi.org/10.1257/jel.48.2.281
IPCC. (2023). In H. Lee, & J. Romero (Eds.), Climate change 2023: Synthesis report.
(Grant Number: NKB-29/UN2.RST/HKP.05.00/2024) for the Contribution of working groups I, II and III to the sixth assessment Report of the
2024–2025 fiscal year. intergovernmental Panel on climate change (the core writing team. IPCC. https://doi.
org/10.59327/IPCC/AR6-9789291691647.
Liao, C., Erbaugh, J. T., Kelly, A. C., & Agrawal, A. (2021). Clean energy transitions and
Appendix A. Supplementary data human well-being outcomes in lower and middle income countries: A systematic
review. Renewable and Sustainable Energy Reviews, 145, Article 111063. https://doi.
Supplementary data to this article can be found online at https://doi. org/10.1016/j.rser.2021.111063. May 2020.
Lopes, M. A. R., Antunes, C. H., & Martins, N. (2012). Energy behaviours as promoters of
org/10.1016/j.ssaho.2025.101612. energy efficiency: A 21st century review. Renewable and Sustainable Energy Reviews,
16(6), 4095–4104. https://doi.org/10.1016/j.rser.2012.03.034
References Marzano, R. J., & Kendall, J. S. (Eds.). (2008). Designing and assessing educational
objectives (1st ed.). Corwin Press.
Masuda, K., & Yudhistira, M. H. (2020). Does education secularize the islamic
Allen, M. R., Dube, O. P., Solecki, W., Aragón-Durand, F., Cramer, W., Humphreys, S.,
population? The effect of years of schooling on religiosity, voting, and pluralism in
Kainuma, M., Kala, J., Mahowald, N., Mulugetta, Y., Perez, R., Wairiu, M., &
Indonesia. World Development, 130, Article 104915. https://doi.org/10.1016/j.
Zickfeld, K. (2018). Framing and context. In Global Warming of 1.5◦ C. An IPCC Special
worlddev.2020.104915
Report on the impacts of global warming of 1.5◦ C above pre-industrial levels and related
Meyer, A. (2015). Does education increase pro-environmental behavior? Evidence from
global greenhouse gas emission pathways, in the context of strengthening the global
Europe. Ecological Economics, 116, 108–121. https://doi.org/10.1016/j.
response to the threat of climate change. Cambridge University Press. https://doi.org/
ecolecon.2015.04.018
10.1017/9781009157940.003
Mills, B., & Schleich, J. (2012). Residential energy-efficient technology adoption, energy
Angrist, J. D., & Pischke, S. (2009). Mostly harmless econometrics: An empiricist’s
conservation, knowledge, and attitudes: An analysis of European countries. Energy
companion. Princeton University Press.
Policy, 49, 616–628. https://doi.org/10.1016/j.enpol.2012.07.008
Becker, G. S. (1993). Human capital: A theoretical and empirical analysis, with special
Mundaca, L., Zhu, X., & Hackenfort, M. (2022). Behavioural insights for sustainable
reference to education (3rd ed.). The University of Chicago Press.
energy use. Energy Policy, 171. https://doi.org/10.1016/j.enpol.2022.113292
BPS. (2021). Berita resmi statistik: Hasil sensus penduduk 2020. In Bps.Go.Id (issue No.7/
Muzayanah, I. F. U., Lean, H. H., Hartono, D., Indraswari, K. D., & Partama, R. (2022).
01/Th.XXIV, 21 januari 2021). https://sensus.bps.go.id/berita_resmi/detail/sp2020
Population density and energy consumption: A study in Indonesian provinces.
/1/hasil-sensus-penduduk-2020.
Heliyon, 8(9), Article e10634. https://doi.org/10.1016/j.heliyon.2022.e10634
Broadstock, D. C., Li, J., & Zhang, D. (2016). Efficiency snakes and energy ladders: A
Nakano, R., Zusman, E., Nugroho, S., Kaswanto, R. L., Arifin, N., Munandar, A.,
(meta-)frontier demand analysis of electricity consumption efficiency in Chinese
Arifin, H. S., Muchtar, M., Gomi, K., & Fujita, T. (2018). Determinants of energy
households. Energy Policy, 91, 383–396. https://doi.org/10.1016/j.
savings in Indonesia: The case of LED lighting in Bogor. Sustainable Cities and Society,
enpol.2016.01.009
42, 184–193. https://doi.org/10.1016/j.scs.2018.06.025. December 2016.
Cattaneo, M. D., Idrobo, N., & Titiunik, R. (2019). A practical introduction to regression
Narasimha Rao, M., & Reddy, B. S. (2007). Variations in energy use by Indian
discontinuity designs. A Practical Introduction to Regression Discontinuity Designs, II.
households: An analysis of micro level data. Energy, 32(2), 143–153. https://doi.org/
https://doi.org/10.1017/9781108684606
10.1016/j.energy.2006.03.012
Cattaneo, M. D., Keele, L., & Titiunik, R. (2021). Covariate adjustment in regression
Netherlands Enterprise Agency. (2022). What is clean cooking?. https://english.rvo.nl/in
discontinuity designs. http://arxiv.org/abs/2110.08410.
formation/what-clean-cooking.
Chankrajang, T., & Muttarak, R. (2017). Green returns to education: Does schooling
Nugroho, S. B., Zusman, E., Nakano, R., Takahashi, K., Koakutsu, K., Kaswanto, R. L.,
contribute to pro-environmental behaviours? Evidence from Thailand. Ecological
Arifin, N., Munandar, A., Arifin, H. S., Muchtar, M., Gomi, K., & Fujita, T. (2017).
Economics, 131, 434–448. https://doi.org/10.1016/j.ecolecon.2016.09.015
The effect of prepaid electricity system on household energy consumption - the case
Chou, J. S., & Yutami, G. A. N. (2014). Smart meter adoption and deployment strategy
of bogor, Indonesia. Procedia Engineering, 198, 642–653. https://doi.org/10.1016/j.
for residential buildings in Indonesia. Applied Energy, 128, 336–349. https://doi.org/
proeng.2017.07.117. September 2016.
10.1016/j.apenergy.2014.04.083, 2014.
Oikonomou, V., Becchis, F., Steg, L., & Russolillo, D. (2009). Energy saving and energy
DiNardo, J., & Lee, D. S. (2011). Program evaluation and research designs. Handbook of
efficiency concepts for policy making. Energy Policy, 37(11), 4787–4796. https://doi.
Labor Economics, 4(PART A), 463–536. https://doi.org/10.1016/S0169-7218(11)
org/10.1016/j.enpol.2009.06.035
00411-4
Pachauri, S., & Jiang, L. (2008). The household energy transition in India and China.
Eggers, A. C., Tuñón, G., & Dafoe, A. (2024). Placebo tests for causal inference. American
Energy Policy, 36(11), 4022–4035. https://doi.org/10.1016/j.enpol.2008.06.016
Journal of Political Science, 68(3), 1106–1121. https://doi.org/10.1111/ajps.12818
Parinduri, R. A. (2014). Do children spend too much time in schools? Evidence from a
Esfandiar, K., Pearce, J., Dowling, R., & Goh, E. (2022). Pro-environmental behaviours in
longer school year in Indonesia. Economics of Education Review, 41, 89–104. https://
protected areas: A systematic literature review and future research directions.
doi.org/10.1016/j.econedurev.2014.05.001
Tourism Management Perspectives, 41. https://doi.org/10.1016/j.tmp.2022.100943.
January 2021.

8
L. Endriana et al. Social Sciences & Humanities Open 11 (2025) 101612

Parinduri, R. A. (2017). Does education improve health? Evidence from Indonesia. Spandagos, C., Baark, E., Ng, T. L., & Yarime, M. (2021). Social influence and economic
Journal of Development Studies, 53(9), 1358–1375. https://doi.org/10.1080/ intervention policies to save energy at home: Critical questions for the new decade
00220388.2016.1228880 and evidence from air-condition use. Renewable and Sustainable Energy Reviews, 143,
Powdthavee, N. (2021). Education and pro-environmental attitudes and behaviours: A Article 110915. https://doi.org/10.1016/j.rser.2021.110915. April.
nonparametric regression discontinuity analysis of a major schooling reform in Steg, L. (2008). Promoting household energy conservation. Energy Policy, 36(12),
England and Wales. Ecological Economics, 181, Article 106931. https://doi.org/ 4449–4453. https://doi.org/10.1016/j.enpol.2008.09.027
10.1016/j.ecolecon.2020.106931. July 2020. Steg, L., & Vlek, C. (2009). Encouraging pro-environmental behavior: An integrative
Pratiwi, S. N., Sakita, Y., Kubota, T., Tanaka, H., Nishiiri, S., Takaguchi, H., & review and research agenda. Journal of Environmental Psychology, 29(3), 309–317.
Surahman, U. (2023). Influences of occupant preferences for living environments on https://doi.org/10.1016/j.jenvp.2008.10.004
household energy consumption: A case study of Indonesia. Energy and Buildings, 300, Stephenson, J., Barton, B., Carrington, G., Gnoth, D., Lawson, R., & Thorsnes, P. (2010).
Article 113636. https://doi.org/10.1016/j.enbuild.2023.113636. October. Energy cultures : A framework for understanding energy behaviours. Energy Policy,
Qiao, W., & Yin, X. (2021). Understanding the impact on energy transition of consumer 38(10), 6120–6129. https://doi.org/10.1016/j.enpol.2010.05.069
behavior and enterprise decisions through evolutionary game analysis. Sustainable Stock, J. H., & Yogo, M. (2005). Testing for weak instruments in linear IV regression. In
Production and Consumption, 28, 231–240. https://doi.org/10.1016/j. D. W. K. Andrews, & J. H. Stock (Eds.), Identification and inference for econometric
spc.2021.04.015 models, essays in Honor of Thomas Rothenberg (pp. 80–108). Chambridge University
Radzi, S. F., & Hassan, M. S. (2021). Energy efficiency and sustainability. In Affordable Press. https://doi.org/10.1017/CBO9780511614491.006.
and clean energy (pp. 392–402). Springer. https://doi.org/10.1007/978-3-319- Tayyaba, S. (2012). Rural-urban gaps in academic achievement, schooling conditions,
95864-4_15. student, and teachers’ characteristics in Pakistan. International Journal of Educational
Rahut, D. B., Ali, A., Mottaleb, K. A., & Aryal, J. P. (2019). Wealth, education and Management, 26(1), 6–26. https://doi.org/10.1108/09513541211194356
cooking-fuel choices among rural households in Pakistan. Energy Strategy Reviews, United Nations. (2019). World population prospects 2019: Highlights. (ST/ESA/SER.A/
24, 236–243. https://doi.org/10.1016/j.esr.2019.03.005. March. 423).
Samarakoon, S., & Parinduri, R. A. (2015). Does education empower women? Evidence United Nations Environment Programme. (2023). Emissions Gap Report 2023: Broken
from Indonesia. World Development, 66, 428–442. https://doi.org/10.1016/j. Record – temperatures hit new highs, yet world fails to cut emissions (again). In
worlddev.2014.09.002 Emission Gap Report 2023: Broken Record - temperatures hit new highs, yet world fails to
Santangelo, A., & Tondeli, S. (2021). Consumer behavior in building energy use. In cut emissions (again). https://doi.org/10.59117/20.500.11822/43922. Nairobi.
W. leal Filho, L. Brandli, T. Wall, A. M. Azul, & A. L. Salvia (Eds.), Affordable and Wang, Q., Niu, G., Gan, X., & Cai, Q. (2022). Green returns to education: Does education
clean energy (pp. 205–215). Springer. https://doi.org/10.18356/ affect pro-environmental attitudes and behaviors in China? PLoS One, 17, 1–20.
9789213618851c011. https://doi.org/10.1371/journal.pone.0263383, 2 February.
Schlindwein, L. F., & Montalvo, C. (2023). Energy citizenship: Accounting for the Wang, S., Xie, Z., & Wu, R. (2020). Examining the effects of education level inequality on
heterogeneity of human behaviours within energy transition. Energy Policy, 180, energy consumption: Evidence from Guangdong Province. Journal of Environmental
Article 113662. https://doi.org/10.1016/j.enpol.2023.113662. June. Management, 269, Article 110761. https://doi.org/10.1016/j.
Simon, H. A. (1955). A behavioral model of rational choice. Quarterly Journal of jenvman.2020.110761. April.
Economics, 69(1), 99–118. Zainudin, N., Idris, N. D. M., & Siwar, C. (2021). Energy conservation: Concept and
Simon, H. A. (1957). Models of man: Social and rational. New York: Wiley. approaches. In Affordable and clean energy (pp. 362–372). Springer. https://doi.org/
Solihat, A., & Syukur, A. (2020). Kebijakan kontroversial menteri daoed joesoef 1978- 10.1007/978-3-319-95864-4_42.
1983 : Perubahan tahun ajaran dan libur bulan ramadhan. Jurnal Pendidikan Sejarah,
9(1), 55–73. https://journal.unj.ac.id/unj/index.php/jps/article/view/15782.