Impact of Attributional Style of Behavior Outcome on the Sustainable Development of Residents’ Energy-Saving Behavior: Differences in Policy Responses of Residents

Abstract

In the context of promoting the construction of an energy-saving and environment-friendly society, it is crucial to promote the sustainable development of residents’ energy-saving behaviors, determine how attributional styles of behavior outcome affect residents’ energy-saving behaviors, and examine how different residents respond to intervention policies. Based on the attribution theory, this study constructs a theoretical model of the impact mechanism of the attributional style of behavior outcome on residents’ energy-saving behaviors. Regression analysis and the quadratic response model were used to test all 1254 valid questionnaires. The results show that the residents’ degrees of understanding, support, and implementation of the energy-saving policies presented inconsistent situations. Residents’ degrees of understanding and support of energy-saving policies positively promote the degree of implementation of energy-saving policies. The consistency of understanding and support has a non-linear influence on the degree of implementation of energy-saving policies. Additionally, the public’s understanding, support, and implementation of energy-saving policies have significant differences in the attribution style of behavioral outcomes. Finally, some relevant policy recommendations are put forward.

1. Introduction

In recent years, China’s rapid economic growth has driven the growth of energy consumption to a certain extent. China’s total energy consumption reached 4.98 billion tons of standard coal in 2020, which is 2.11 percent more than the previous year, according to data released by the National Bureau of Statistics [1]. With the rapid development of society, residents’ energy consumption increases rapidly [2]. In 2020, China’s household energy consumption accounted for 12.93% of the total energy consumption, reaching 643.8 million tons of standard coal [3], with an annual growth rate of 4.15% [1], which is the main growth point of China’s energy consumption. As a kind of environmental protection behavior, the selection process of residents’ energy-saving behavior is complex, and there are many influencing factors [4]. The behavior of residents in the terminal link of energy consumption not only directly affects the scale growth of energy consumption and carbon emissions but also is the main driving factor of energy consumption and carbon emissions in industries such as industry, construction, transportation, and the service industry. For the implementation of the “carbon peaking and carbon neutrality” grand strategy, the State Council issued by Action Plan for Carbon Dioxide Peaking Before 2030 in October 2021. This plan points out that it is necessary to promote green and low-carbon national actions, enhance the people’s awareness of saving, advocate a green and low-carbon lifestyle, and emphasize the need to transform the green concept into a conscious action of all people [5]. It can be seen that it is of great significance to reduce residents’ energy consumption and promote the development of residents’ energy-saving behavior [2].

In recent years, the government has introduced a series of policies and measures to encourage residents to save energy, which has played a certain role in promoting energy-saving behavior. For example, in Shanghai, since the implementation of the Shanghai Municipal Household Waste Management Regulations on 1 July 2019, urban residents’ awareness of garbage sorting has been continuously enhanced, and remarkable results have been achieved. Every year, Japan has an event called "Earth Day," during which an increasing number of businesses and landmark buildings turn off their lights to promote environmental protection. The new energy vehicle subsidy policy promotes the public’s purchase behavior for new energy vehicles and so on. Although they have played a certain role in policy guidance, they still do not have systematic long-term incentive mechanisms and innovative institutional design on the whole. At present, there are still some problems with residents’ energy-saving behavior, such as poor implementation, a lack of energy-saving motivation, and a lack of identification of the results of energy-saving behavior [6]. Therefore, it is necessary to analyze the response of residents to energy-saving policies from a deeper level, maximize the intervention role of policies, guide the sustainable development of residents’ energy-saving behavior, and make the society form a good social atmosphere of advocating energy-saving.

Compared to research on the factors that affect the early stages of energy-saving behavior, research on the long-term growth of energy-saving behavior is still in the exploration stage. In the process of sustainable behavior development, the result of behavior has a significant effect on the sustainable implementation of that behavior. The effect is more pronounced when the outcome of behavior is inconsistent with the individual’s expectations. According to psychological attribution theory, the degree of matching between behavioral results and individual expectations makes individuals have positive or negative emotions, thus forming different attribution styles [7]. Positive attribution style increases behavioral confidence, whereas negative attribution style decreases willingness to engage in certain behaviors [8]. If targeted policy intervention is carried out at this stage, the specific environment of behavior will always exist, and the specific environment will gradually develop into a kind of suggestion that can induce residents to continue and automatically adopt energy-saving behavior to a greater extent, and then promote the sustainable development of energy-saving behavior.

This study explores how the attribution style of behavior results affects the long-term development of residents’ energy-saving behavior. This study takes residents’ energy-saving behavior as the research object, takes psychological attribution theory as the theoretical basis, and analyzes the influence of different attribution styles of energy-saving behavior results on the intervention policies of energy-saving behavior from the perspective of behavioral sustainable development. Additionally, this study constructs a policy response model based on behavioral outcome attribution style to further explore the differences in intervention policy responses.

2. Literature Review

Residents’ energy-saving behavior is affected by many complex factors. In relevant studies, most scholars regarded residents’ energy-saving behavior as a specific environmental behavior [9,10,11,12]. Existing research focused on the factors affecting residents’ energy-saving behavior from different perspectives.

By integrating previous research results, this paper divides the influencing factors of individual energy-saving behavior into three categories. First, it includes demographic characteristics [10], such as gender [11], age [12], income [13], and education [14,15,16]. However, relevant studies were controversial, and there was no relatively consistent conclusion. Second, it includes psychological factors, such as energy-saving attitudes, green values, behavioral intentions, and so on. Since individuals are the audiences of public policies, if there is no psychological condition for individual behavior implementation, the energy-saving behavior encouraged by external policies may not last long. Boomsma [17] and Belaid [12] et al. proved that people’s attitudes and ideologies toward energy-saving can promote their energy-saving behavior. Irawan et al. investigated the impact of different values on citizens’ pro-environment behaviors, and relevant results showed that different values had a significant impact on pro-environment behaviors [18]. Finally, external situational factors are involved, including social influences [19,20], tax incentives and subsidies [21], public publicity and price control policies [22,23], etc.

Specifically for psychological attribution, Heider first proposed the concept of attribution in his work [24]. Seligman defines the attribution style [25]. With the development of psychology, many scholars gradually realized that individuals in the process of processing information often show different preferences. The research on attribution is becoming more extensive, mainly involving psychology, education, behavior, and so on. Carson pointed out that only when diversity, stability, and consistency are high, individuals will attribute their behavior to the external environment [26]. Hu and Liu proposed the complete attribution model in the field of organizational behavior, which expanded the attribution theory to a broader research field [27]. Jungmeen et al. confirmed that different attribution styles play a moderating role in adolescents’ perceptual processes and depression [28]. Yang et al. found that employees’ attribution style can directly affect job performance and indirectly influence job performance through the mediating effect of decision style [29]. However, there are few types of research on attribution in energy-saving behavior. Yue et al. found the feedback effect of behavioral outcome attribution modulated perceived behavioral outcomes on energy-saving intention and behavior [30]. Since the attribution process is very consistent with the cognitive and behavioral characteristics of people and the development process of people’s behavior, it is useful to include it in the study of people’s energy-saving behavior.

For external situational factors, relevant studies have shown that policy intervention is an effective institutional measure for residents to adopt energy-saving behaviors [6,31]. In terms of policies to guide residents’ energy-saving behaviors, Peter believed that price policies, tax incentives, and energy subsidy policies had significant incentives, constraints, or inhibitions on residents’ energy-saving behaviors and were conducive to guiding residents’ energy consumption behaviors [32]. In terms of policies and means to promote residents’ energy conservation, Liu et al. believed that price policies must play a full role in promoting residents’ active energy conservation and consumption reduction [33]. Mi et al. found that informational strategies have a positive, promoting effect on residents’ energy-saving behavior, and the post-intervention use of informational strategies is more effective [34]. Chen et al. studied the public’s reaction to the municipal solid waste control policy and emphasized that the government should give more consideration to the attitude and reaction of the “policy recipients” when formulating the control policy [35].

As can be seen from the above literature, related scholars have conducted a lot of research on energy-saving behavior. Many researchers have also indicated the importance of low-carbon and energy-saving policies for regulating residents’ energy-saving behavior and have put forward relevant countermeasures and suggestions. However, the influence of individual behavioral attribution styles on energy-saving behavior is still insufficient. Furthermore, the relationship between psychological factors and policy influence needs to be further explored. Therefore, this study focuses on the sustainable development stage of residents’ energy-saving behavior and introduces the psychological attribution theory. Additionally, we study how the attributional style of behavior outcome affects the sustainable development of residents’ energy-saving behavior and further explore the corresponding differences in intervention policies.

3. Research Hypothesis and Model Construction

3.1. Definition of Residents’ Energy-Saving Behavior

Energy-saving behavior refers to the behavior of an individual to save energy consumption in a certain way. Residents’ energy consumption is the household energy consumption in daily life, mainly including oil, natural gas, electricity, and other related energy consumables. The existing research literature generally divides energy-saving behaviors into two categories [9,36]. One is habitual energy-saving behavior, which refers to the behavior of reducing energy consumption by adjusting daily lifestyles and energy use habits, which needs to be repeated, such as adjusting temperature control equipment and turning off lights on time, and so on. The other type is energy efficiency investment behavior, which refers to the behavior of indirectly achieving an energy-saving effect by purchasing energy-saving equipment or investing in energy-saving technology.

With the continuous improvement of China’s urbanization level, energy, and environmental pressure are increasing. Therefore, the importance of changing residents’ energy use behavior should be highlighted. In the time dimension, residents’ energy-saving behavior can be short-term or long-term. Generally, residents’ energy-saving behaviors in the short term are very random and may show different characteristics. In the long run, individual behavior shows strong plasticity. The energy-saving behavior of residents is affected by many external factors and may also show stability in the long term. Existing research on residential energy-saving behaviors mostly focuses on the influencing factors at the initial formation stage of energy-saving behaviors, and the research on the development process of residents’ energy-saving behaviors is still in the exploratory stage [7].

Based on this, and drawing on previous studies, in this study, residents’ energy-saving behaviors are divided into habit adjustment energy-saving behavior (HAB) and efficiency investment energy-saving behavior (EIB). Focusing on the sustainable development stage of residents’ energy-saving behaviors, this study introduces psychological attribution theory to excavate the influencing mechanism of residents’ energy-saving behaviors based on the consideration of promoting the sustainable development of energy-saving behaviors.

3.2. Construction of Interventional Policy Response Model

The successful realization of environmental protection is largely influenced by the design of relevant policies [6]. Most domestic and foreign researchers stated that policies play an important role in guiding residents to save energy and improve efficiency. At present, the academic community usually divides the types of policies into three categories [6]: coercive policy, economic policy, and information policy. The coercive policy has the characteristics of high execution costs and strong mandates, such as price control and fee collection, which are respectively pre-intervention and post-intervention strategies. Economic policy instruments can provide monetary incentives for energy efficiency reforms, such as tax credits and financial subsidies. Information intervention policy aims to influence residents’ energy-saving behavior by disclosing and publicizing relevant information about conservation, such as through media publicity and activities, and guiding residents to save energy and improve energy efficiency.

In this paper, four policies, namely the economic subsidy policy (ESP, an economic policy), the price control policy and fee collection policy (PCP and FCP, both mandatory policies), and the campaign propaganda policy (CPP, an information intervention policy), are selected as external intervention policies in the development of sustainability of residents’ energy-saving behavior for further research.

Policy intervention is an important factor affecting the sustainable development of energy-saving behaviors, and how to improve the effectiveness of policy to effectively guide residents to implement energy-saving behaviors is particularly important. From the perspective of the public, relevant studies have shown that individuals’ cognition and willingness to support policies will determine their implementation of those policies. If it is not based on a full understanding of the policy, the individual’s sense of identity for the policy, due to endogenous factors (environmental awareness, civic awareness, etc.) cannot be sustained [37,38,39,40]. In addition, individuals’ awareness of the policy and willingness to support it are not completely consistent. Although residents understand a certain policy, if they do not support it, its effectiveness will be greatly reduced [41]. In the public’s attitude toward and reaction to the policy, the implementation of the policy is the most direct reflection of its effectiveness. The cognition and identification of the policy determine whether the individual is willing to comply with the policy, which further affects the effective implementation of the policy. This study supposes that individuals’ attitudes and responses to energy-saving policies are the comprehensive embodiment of understanding, support, and implementation.

Based on the response modes of individual cognition, emotion, and attitude [42], this study defines residents’ degree of understanding of energy-saving policies (UTP) as the degree of individual awareness and understanding of policy contents. The degree of support (STP) indicates the individual’s identity and willingness to support a certain policy. The degree of implementation (ITP) indicates the degree of individual compliance with a certain policy. The relationship between the degree of understanding, the degree of support, and the degree of implementation needs to be further discussed. Therefore, this study proposes the following research hypotheses:

Hypothesis 1 (H1). 

Residents’ degree of understanding of energy-saving policies positively affects the degree of implementation of the policies.

Hypothesis 2 (H2). 

Residents’ degree of support for energy-saving policies positively affects the degree of implementation of the policies.

Hypothesis 3 (H3). 

The consistency of the residents’ degrees of understanding and support of the energy-saving policies has a nonlinear influence on the degrees of implementation of the policies.

3.3. The Introduction of the Attributional Style of Behavior Outcome

According to psychological attribution theory, attribution refers to the process in which individuals summarize causes in various ways according to their behavior and consequences. According to this theory, behavioral results have a significant impact on repetitive behaviors, and the impact is more obvious when the actual behavioral results are inconsistent with individual expectations [43,44]. In the process of attribution, individuals will produce different emotions and show different choice preferences. Seligman defined individuals’ preferences for the interpretation of different types of information as attribution styles and divided them into positive attribution style, intermediate attribution style, and negative attribution style [25]. Chen proved that behavioral attribution styles differ significantly in the depth and pattern of individual information searches [45]. Hu constructed the complete attribution model, which expanded the attribution theory to a wider research field [27]. He unified attribution cognition and attribution outcome and constructed a cognitive model of the attribution process. It consists of four nodes: “consequence-cause-cognition-behavior”. The basic theories are as follows: the impact of the analysis of behavioral causes on cognition (i.e., attributional results); when cognition is reinforced by attributional consequences, cognition will influence new behaviors.

Because of the repeatability of individual behavior, the attribution process for individual behavior is a long-term iterative process. Energy-saving behavior is repetitive, and the attribution process for energy-saving behavior is also a long-term iterative process. In this process, individuals with a positive emotional preference have a strong will to perform the behavior, while a negative emotional preference will make the individual lose the will to perform a certain behavior. If policy intervention is carried out at this stage to strengthen the perception of behavioral results and positive emotions, residents can be incentivized to continue to adopt energy-saving behaviors. Therefore, it is necessary to explore the impact of attributional style on residents’ energy-saving behaviors in the context of sustainable development. Based on previous studies, this study selects the attributional style of behavior outcome for energy-saving behavior to measure residents’ positive, negative, or intermediate emotions when making the attribution. Combined with external policy intervention, this study focused on exploring the different influences of different attribution styles on the responses of residents to energy-saving policies. Further combining with the previous construction of the policy response model, hypothesis H4 is proposed:

Hypothesis 4 (H4). 

Residents’ degrees of understanding, support, and implementation of energy-saving policies have significant differences in the attributional style of behavior outcome.

In summary, this study constructed a policy response model of residents’ energy-saving behavior, combined with an attribution style. Based on the public’s degree of understanding, support, and implementation of the energy-saving policy, this study selected a Price control policy (PCP), Fee Collection Policy (FCP), Economic subsidy policy (ESP), and Campaign publicity policy (CPP) as the external policy interventions, to further explore the differences in response to different energy-saving intervention policies. Additionally, based on the psychological attribution theory, this study focused on the sustainable development process of energy-saving behavior. We also discussed the difference in the response of the attributional style of energy-saving behavior outcomes to the intervention policy in the sustainable development stage of residents’ energy-saving behavior. The theoretical research model of this study is shown in Figure 1.

4. Data

4.1. Questionnaire Design

To test the hypothesis mentioned above, relevant data were obtained through a questionnaire survey (see Supplementary Materials) in this study. Using the information from this study, the existing maturity scale was revised and improved, and the Likert-5 scale was used to evaluate corresponding items. The higher the score, the more respondents agreed with the description of the item.

Demographic variables of respondents included gender, age, education level, and marital status. The corresponding scale was designed by referring to Sun [46]. In this study, two variables, habit adjustment energy-saving behavior (HAB) and efficiency investment energy-saving behavior (EIB), were used to measure residents’ energy-saving behaviors. Both behaviors involve two subitems, which refer to Lind’en [47] for the table. In the question items (such as actively turning off the power when household appliances are not in use), the subjects choose from five scales (the Likert-5 scale) of never, occasionally, generally, often, and every time according to their behavior, corresponding to 1–5 points, respectively. Finally, the scores of the two options with the same dimension are added and averaged.

For the attributional style of the behavioral outcome, based on the research of Peterson [48], two opposite situations (positive and negative) of behavioral outcome were designed in the questionnaire. The first question in each case is designed to lead the respondent. Then, the relevant questions measured the stability, consistency, and internal and external factors of the subjects in different scenarios. The total score in the positive scenario minus the total score in the negative scenario is the final score. When the final score is greater than zero, residents belong to the positive attribution style (PAS); when it is equal to zero, residents belong to the intermediate attribution style (IAS); and when it is less than zero, residents belong to the negative attribution style (NAS). In this study, attribution style was mainly used as a categorical variable.

For policy responses, by referring to the response scale of garbage control policy developed by Chen [35], three dimensions of measurement items were designed, which are respectively “the degree of my understanding of the policy”, “the degree of my support for the policy”, and “the degree of my obedience to or implementation of the policy”, and each dimension involved eight items and four energy-saving policies (PCP, FCP, ESP, and CPP). Three variables representing residents’ understanding degree, support degree, and implementation degree of different energy-saving policies were constructed.

4.2. Data Collection

To ensure the validity of the formal survey scale, this study tested the reliability and validity of the initial scale based on a small preliminary investigation. According to the survey results, the initial scale was optimized by combining the feedback of the respondents and the suggestions of relevant experts in the field.

The questionnaire survey for this study was carried out in Xuzhou, Jiangsu Province, and was mainly conducted online and supplemented by offline research. As Xuzhou is a transportation hub in China, its economic scale also has great advantages, and it is at the forefront of energy conservation and emission reduction, so we believe that this city has a certain representation in our study. By randomly forwarding questionnaires on social platforms, we expanded the research scope and ensured the randomness of the samples. On this basis, we conducted an offline survey in Xuzhou City and randomly distributed paper questionnaires in parks, shopping malls, residents’ buildings, and other crowded places.

Through the network survey, a total of 965 questionnaires were recovered, and 192 invalid questionnaires were eliminated. A total of 773 valid questionnaires were collected online. A total of 587 questionnaires were collected by offline survey, including 481 valid questionnaires. Finally, 1254 valid questionnaires were collected, with an effective recovery rate of 80.80%.

4.3. Common Method Deviation Test

The Harman single-factor test was used to test the common method bias [49]. Six factors with characteristic roots greater than 1 were extracted from the results of the unrotated exploratory factor analysis. The maximum factor variance explanation rate is 28.04%, much less than 40%, so this study does not reveal the serious problem of common method biases.

4.4. Reliability and Validity Test

Reliability refers to the consistency of the results obtained after repeated measurement of the same variable, which can reflect the authenticity of the characteristics of the respondents. Since the measurement of variables in this study was based on multiple items, and it was difficult to conduct a follow-up survey on respondents, Cronbach’s alpha was selected as the test for initial scale reliability, and SPSS 26 software was used for analysis. The Cronbach’s α reliability coefficients of the attributional style of behavior outcome and policy response in the questionnaire are 0.622 and 0.872, respectively, indicating that the energy-saving behavior attributional style questionnaire and policy response questionnaire had an acceptable reliability level.

To test the validity of the data, SPSS (Version 26.0. IBM Corp: Armonk, NY, USA) software was used in this study. According to the test result, the KMO value is 0.819, above 0.5, and the significance of the Bartlett sphericity test is 0.000 < 0.05. Therefore, it can be considered that the questionnaire has good structural validity, and the questionnaire questions can effectively detect the degree of things to be measured. Concerning the content validity of the questionnaire, it was first pretested and then changed based on the results of the pretest. The revision and development of the questionnaire were based on in-depth interviews and literature research regarding existing maturity scales. Additionally, invite professors and experts engaged in energy conservation research for a long time to evaluate and revise it. Therefore, it can be considered that all questionnaires in this study have good content validity.

5. Results and Discussion

5.1. Demographic Analysis

The absolute values of skewness coefficients of each measurement item of each variable in the sample are all less than 2, which means that the scale data is close to the normal distribution, so the sample passes the normality test. This study uses an independent-sample t-test and one-way analysis of variance to analyze the differences in residents’ energy-saving behavior and attribution style in terms of demographic characteristics. The results are shown in

Table 1. Demographic analysis of energy-saving behaviors and attribution style.

	Demographic Variables	Proportion	HAB	EIB	Attribution Style
Gender	Male	48.56%	3.567	3.772	0.860
	Female	51.44%	3.662	3.778	0.780
	F		0.357	0.611	1.832
Age	18–25 years old	16.91%	3.502	3.758	0.420
	26~35 years old	20.49%	3.629	3.722	0.340
	36~45 years old	38.68%	3.635	3.810	1.090
	46~59 years old	22.25%	3.662	3.775	1.090
	Over 60 years old	1.67%	3.575	3.775	1.100
	F		1.143	0.436	7.753 ***
Education	High school or below	12.12%	3.610	3.627	1.180
	Technical secondary school or college degree	24.80%	3.673	3.702	0.870
	Bachelor’s degree	47.29%	3.622	3.829	0.830
	Postgraduate degree (Master or doctoral degree)	15.79%	3.517	3.835	0.450
	F		1.282	3.074 *	3.246 *
Marital status	Unmarried	23.05%	3.493	3.684	0.250
	Married	73.84%	3.659	3.804	0.970
	Divorced	3.11%	3.558	3.791	1.470
	F		4.050 *	2.035	13.902 ***

Note: *** p < 0.001; * p < 0.05. HAB indicates habits adjustment energy-saving behavior and EIB indicates efficiency investment energy-saving behavior.

.

Table 1 shows that efficiency investment and energy-saving behaviors have significant differences in educational level and habit adjustment. Energy-conservation behaviors have significant differences in marital status, and attribution styles have significant differences in age, educational level, and marital status. Specifically, gender had no significant effect on any of the three (HAB, EIB, and attribution style). However, the comparison of average values shows that the implementation level of HAB among women is higher than that of men, which indicates that women are more likely to focus on saving energy daily and choosing an energy-efficient lifestyle. Some scholars [16,50] also believed that women paid more attention to daily energy-saving, mainly considering the differences in social roles and lifestyles between men and women. Energy-saving behaviors show an “inverted U-shaped” trend with increasing age. Residents aged 46 to 59 have the highest level of HAB, and residents aged 36 to 45 have the highest level of EIB. Existing studies have not reached a unified conclusion on the impact of education level on energy-saving behavior. In this study, energy-saving behavior is divided into HAB and EIB, and it is found that residents’ education level significantly affects their efficiency investment and energy-saving behavior. Education level significantly affects residents’ EIB (F = 3.074, p < 0.05). The higher the education level, the higher the implementation level of EIB. As for HAB, residents with technical secondary school or college degrees scored the highest, while residents with postgraduate degrees scored the lowest.

For attribution style, it shows completely significant age differences (F = 7.753, p < 0.001) and marital status (F = 13.902, p < 0.001) and differs significantly in education (F = 3.246, p < 0.05). As can be seen from the results in Table 1, compared with women, men’s attribution style is more positive. In terms of age, the attribution styles of residents aged 36 to 45 are more positive, and the attribution style shows an “inverted U-shape” with the increase of age. It is worth mentioning that residents’ attribution style becomes more negative with the improvement of their education level. Considering marital status, divorced residents have the most positive attribution style, and unmarried residents have the most negative attribution style.

5.2. Analysis of Residents’ Attitudes towards Interventional Policies

Statistics of the residents’ degrees of understanding (UTP), support (STP), and implementation (ITP) of energy-saving policies are shown in

Table 2. Frequency statistics of residents’ degrees of understanding of energy-saving policies.

	Extremely Inconsistent	Relatively Inconsistent	Uncertainty	Relatively Consistent	Extremely Consistent 3
PCP	3.51%	19.06%	8.13%	50.24%	19.06%
FCP	4.94%	24.88%	19.70%	38.20%	12.28%
ESP	5.10%	23.60%	26.00%	31.10%	14.19%
CPP	7.66%	28.63%	30.70%	24.00%	9.01%

Note: “PCP” indicates price control policy, “FCP” indicates fee collection policy, “ESP” indicates Economic subsidy policy, and “CPP” indicates Campaign publicity policy.

,

Table 3. Frequency statistics of the residents’ degrees of support for energy-saving policies.

	Extremely Inconsistent	Relatively Inconsistent	Uncertainty	Relatively Consistent	Extremely Consistent 3
PCP	3.35%	7.97%	20.41%	31.74%	36.52%
FCP	4.07%	8.45%	19.22%	34.37%	33.89%
ESP	0.64%	1.28%	7.34%	42.74%	48.01%
CPP	0.56%	1.59%	11.48%	45.69%	40.67%

Note: “PCP” indicates price control policy, “FCP” indicates fee collection policy, “ESP” indicates Economic subsidy policy, and “CPP” indicates Campaign publicity policy.

and

Table 4. Frequency statistics of residents’ degrees of implementation of energy-saving policies.

	Extremely Inconsistent	Relatively Inconsistent	Uncertainty	Relatively Consistent	Extremely Consistent 3
PCP	0.96%	4.63%	19.14%	50.96%	24.32%
FCP	3.11%	6.86%	20.65%	42.50%	26.87%
ESP	1.12%	4.86%	17.86%	43.22%	32.93%
CPP	0.56%	2.15%	26.87%	46.97%	23.44%

Note: “PCP” indicates price control policy, “FCP” indicates fee collection policy, “ESP” indicates Economic subsidy policy, and “CPP” indicates Campaign publicity policy.

. From the residents’ degrees of understanding of energy-saving policies, more than 20% of the residents almost do not know about energy-saving policies. Among them, the number of people who did not know about the CPP is the largest, reaching 36.29%. The number of people who did not know or did not know much about PCP is the lowest. From the degrees of support for energy-saving policies, most residents support energy-saving policies, among which ESP and CPP have the largest numbers of supporters. From the point of view of the degrees of implementation of energy-saving policies, most residents’ energy-saving policies are well implemented.

To intuitively compare the relationship among the residents’ degrees of understanding, support, and implementation of the energy-saving policies, according to the above analysis data, “extremely inconsistent”, “relatively inconsistent”, “uncertainty”, “relatively consistent”, and “extremely consistent” are valued at 1 to 5, respectively. Then, they are substituted into Formula 1 to calculate the scores of the degrees of understanding, support, and implementation of each type of policy, respectively.

$${\mathrm{S}}_k=\frac{1}{N_k}{{\displaystyle \sum }}_{j=1}^5{r}_{ij}\ast W_{ij} ,$$

(1)

In Formula (1), $k$represents the type of energy-saving policy; $S_k$indicates residents’ degrees of understanding, support, and implementation of type $k$ energy-saving policy; $i$ refers to the item sequence number of energy-saving policy; $j$ represents the evaluation level (one to five) of residents’ degrees of understanding, support, and implementation of energy-saving policies; $N_k$ represents the number of measurement items of type $k$ energy-saving policy; $r_{ij}$is the corresponding score of degrees of understanding, support, and implementation of energy-saving policies when the evaluation level is $j$; $W_{ij}$ is the proportion of the degrees of understanding, support, and implementation of the type $i$ energy-saving policy in the total number of residents when the evaluation level is $j$. The calculation results are shown in

Table 5. Residents’ degrees of understanding, support, and implementation of energy-saving.

	UTP	STP	ITP
PCP	3.59	3.89	3.91
FCP	3.25	3.85	3.83
ESP	3.22	4.46	4.04
CPP	3.10	4.41	3.87
Mean	3.29	4.15	3.91

Note: “UTP” indicates residents’ degree of understanding of energy-saving policies, “STP” indicates residents’ degree of support of energy-saving policies, and “ITP” indicates residents’ degree of implementation of energy-saving policies.

.

Table 5 shows that residents have a low overall understanding of energy-saving policies. According to previous studies, the environmental and civic awareness of Chinese residents still needs to be improved [51]. In contrast, residents’ support for and implementation of policies they already know is higher. Among them, the degree of understanding of CPP is the lowest and that of PCP is the highest; the degree of support of FCP is the lowest, and the degree of support of ESP is the highest. The degree of implementation in FCP is the lowest, and the degree of implementation in ESP is the highest. Intuitively, a high degree of understanding of energy-saving policies does not necessarily correspond to a high degree of implementation, and a high degree of support for a policy often corresponds to a high degree of implementation.

5.3. Quadratic Response Surface Regression Analysis of Intervention Policies

The statistical analysis principle of quadratic response surface regression is based on the idea of multiple regression, using the hierarchical regression analysis method, and using the response surface technology to calculate, process, and interpret the complex regression coefficient results. It can better explore the interaction effects of variables and the influence effects of higher dimensions [52,53]. This method is mainly applied to the research of matching humans and organizations [53,54]. There is also a fitting relationship between residents’ degree of understanding and degree of support of the energy-saving policy, that is, the degree of “consistency” between residents’ understanding and support of the policy. It is similar to the matching connotation in human-organization matching theory. Therefore, this study used the method of quadratic response surface regression analysis to deeply explore the response relationship among residents’ degree of understanding, support, and implementation of energy-saving intervention policies. In this paper, the influence of the consistent degree of understanding and support of residents on the implementation degree of energy-saving policy is further explored.

First of all, a correlation analysis is conducted on the residents’ degree of understanding, support, and implementation of the energy-saving policy. As shown in

Table 6. Correlation analysis.

Price Control Policy				Fee Collection Policy
	UTP	STP	ITP		UTP	STP	ITP
UTP	1			UTP	1
STP	−0.042	1		STP	0.219 **	1
ITP	0.109 **	0.437 **	1	ITP	0.177 **	0.375 **	1
Economic Subsidy Policy				Campaign Publicity Policy
	UTP	STP	ITP		UTP	STP	ITP
UTP	1			UTP	1
STP	0.182 **	1		STP	0.110 **	1
ITP	0.219 **	0.354 **	1	ITP	0.274 **	0.378 **	1

Note: ** p < 0.01, significant correlation.

, there is a significant correlation among residents’ degrees of understanding, support, and implementation of energy-saving policies. Residents’ degree of understanding of energy-saving policies has a positive impact on the degree of implementation, and the degree of support for residents’ energy-saving policies has a positive impact on the degree of implementation, which proves hypotheses H1 and H2.

In the following, the data were processed by the quadratic response surface regression method. Residents’ degree of implementation of energy-saving policies is taken as the dependent variable, and residents’ degree of understanding and degree of support are taken as the independent variables for analysis and modeling. To avoid multicollinearity, the measurement index has been centralized, and the analysis method of the quadratic polynomial regression equation is as follows: In the first step, $x_{1 }$and $x_{2 }$ are put into the regression equation to test the linear relationship with $Y$ (Model 1). In the second step, $x_1^2, x_{2 }^2$ and $x_1{x}_2$ are put into the equation (Model 2) to explore the nonlinear effect of understanding and support on execution and their cross effect. The regression model constructed by quadratic response surface regression analysis is as follows:

$$Y_1={\beta }_0+a_1{x}_1+a_2{x}_2+e,$$

(2)

$$Y_2={\beta }_0+a_1{x}_1+a_2{x}_2+a_3{x}_1^2+a_4{x}_2^2+a_5{x}_1{x}_2+e,$$

(3)

Model 1 is the linear regression of UTP ($x_1$) and STP ($x_2$) on ITP ($Y$). Model 2 is the quadratic response regression model.

Table 7. Results of quadratic response surface regression analysis.

	Price Control Policy (PCP)		Fee Collection Policy (FCP)
	Y		Y
	Model 1	Model 2	Model 1	Model 2
$x_1$	0.104 ***	0.128 ***	0.097 ***	0.100 ***
$x_2$	0.328 ***	0.333 ***	0.320 ***	0.352 ***
$x_1^2$		0.023		−0.028
$x_2^2$		0.003		0.025
$x_1{x}_2$		−0.075 ***		0.092 ***
R²	0.206	0.215	0.149	0.161
F	163.353	69.564	110.452	49.065
	Economic Subsidy Policy (ESP)		Campaign Publicity Policy (CPP)
	Y		Y
	Model 1	Model 2	Model 1	Model 2
$x_1$	0.141 ***	0.150 ***	0.188 ***	0.206 ***
$x_2$	0.443 ***	0.280 ***	0.460 ***	0.481 ***
$x_1^2$		0.070 ***		−0.105 ***
$x_2^2$		−0.186 ***		0.002
$x_1{x}_2$		0.054		−0.037
R²	0.149	0.175	0.196	0.221
F	110.400	54.148	153.764	71.946

Note: *** p < 0.001; $x_{1 }$indicates residents’ degrees of understanding of energy-saving policies; $x_{2 }$indicates residents’ degrees of support for energy-saving policies; Y indicates residents’ degrees of implementation of energy-saving policies.

shows the results of the analysis of intervention policies. From the point of view of the coefficient alone, the coefficient of support ($x_2$) is larger. For PCP, FCP, and CPP, the interaction coefficient of residents’ degree of understanding and support is significant. This indicates that residents’ degree of understanding and support have a significant interaction effect on the degree of implementation of these three kinds of policies.

As can be seen from the results of Table 7, the R2 of Model 2 is significantly improved compared with Model 1. The coefficients of the quadratic term and interaction term in Model 2 are significant, indicating that Model 2 has higher explanatory power. The response analysis was further carried out through the three-dimensional graph. Specifically, MATLABR2021a software is used to draw the three-dimensional diagram of the relationship between residents’ degrees of understanding, support, and implementation of energy-saving policies. As shown in Figure 2, when the consistency of residents’ understanding and support of energy-saving policy is high (the consistency is, $x_1=x_2$), residents’ degree of implementation of the policy is also high, and hypothesis H3 is true.

The knowledge-attitude-behavior model [55] points out that behavior is related to knowledge and belief. Residents’ understanding of policies needs to be supported by relevant knowledge. Residents’ support for policies is an individual attitude and belief, which can predict the implementation of compliance policies. However, residents’ understanding of policies does not translate directly into their implementation. The promotion of energy-saving knowledge should be emphasized. To a certain extent, this indicates that the government should pay attention to the consistency of residents’ understanding and support for energy-saving policies and give consideration to the far-reaching influence of propaganda policies on residents’ long-term intervention in energy-saving behaviors.

5.4. The Difference Analysis of Behavioral Outcome Attribution Style on the Intervention Policy Response

According to the theoretical model established above, individuals summarize causes in various ways according to their behaviors and results, which is the process of attribution. The degree of match between the outcome of the behavior and the expectation will affect the individual’s emotions. Furthermore, attribution style can affect an individual’s mode of information processing and way of thinking, then affect the resident’s subsequent behavior choices. To explore how the attribution style of residents’ energy-saving behavior results influences the effect of energy-saving policy intervention, this study analyzed the difference in the response of the attributional style of behavior outcome to policy intervention.

Based on the questionnaire and classification method of attribution style designed in this study, the subjects were divided into positive, intermediate, and negative attribution style groups. Then, their degrees of understanding, support, and implementation of different policies were calculated, respectively, as shown in

Table 8. The PAS, IAS, and NAS of residents with different attributional styles of behavior outcome.

Class	PAS			IAS			NAS
	UTP	STP	ITP	UTP	STP	ITP	UTP	STP	ITP
PCP	3.595	3.933	3.939	3.450	3.771	3.711	3.656	3.919	4.003
FCP	3.351	3.799	3.819	2.971	3.701	3.514	3.344	3.955	3.995
ESP	3.217	4.274	3.909	2.866	4.282	3.828	3.395	4.642	4.214
CPP	3.100	4.240	4.207	2.965	4.171	4.098	3.163	4.597	4.451
Mean	3.316	4.062	3.969	3.063	3.981	3.788	3.390	4.278	4.166

Note: “PAS” indicates positive attributive style, “IAS” indicates intermediate attribution style, and “NAS” indicates negative attribution style; “PCP” indicates price control policy, “FCP” indicates fee collection policy, “ESP” indicates Economic subsidy policy, and “CPP” indicates Campaign publicity policy.

. On average, residents with a positive attribution style have the highest degree of understanding, support, and implementation of energy-saving policies, while those with an intermediate attribution style have the lowest score. Residents in the three groups of attribution styles have the lowest degree of understanding of the CPP and the lowest degree of implementation of the FCP. In contrast, the three groups of residents have the highest degree of support for ESP and the highest degree of implementation of CPP.

Furthermore, a one-way ANOVA was used to analyze the difference in the response degree of residents’ energy-saving policies in different attribution styles.

Table 9. Difference analysis of UTP on the attributional style of behavior outcome.

Category		UTP
		Mean	Levene Test
			F	Sig.P
Attributional Style	PAS	3.4755	22.874	0.000
	IAS	3.0495
	NAS	3.3247

Note: “UTP” indicates the degree of understanding of energy-saving policies.

shows that UTP has a significant difference (F = 45.991, Sig = 0.000) in the attribution style. Residents with PAS have a better understanding of various energy-saving policies than those with IAS and NAS.

Table 10. Difference analysis of STP on the attributional style of behavior outcome.

Category		STP
		Mean	Levene Test
			F	Sig.
Attributional Style	PAS	4.3107	28.043	0.000
	IAS	3.9975
	NAS	3.9723

Note: “STP” indicates the degree of understanding of energy-saving policies.

shows that STP has significant differences (F = 84.865, Sig = 0.000) in the attribution style. From the mean point of view, residents with PAS have the highest STP, which is significantly higher than those with IAS and NAS.

Table 11. Difference analysis of ITP on the attributional style of behavior outcome.

Category		ITP
		Mean	Levene Test
			F	Sig.
Attributional Style	PAS	4.1771	36.738	0.000
	IAS	3.8001
	NAS	3.7981

Note: “ITP” indicates the degree of understanding of energy-saving policies.

shows that ITP has significant differences in attribution style (F = 103.322, Sig = 0.000). Among them, residents with PAS have the highest degree of implementation of energy-saving policies. Therefore, residents’ degrees of understanding, support, and implementation of energy-saving policies have significant differences in the attributive style of behavior outcome, so hypothesis H4 is valid. At the same time, by combining the results of the three tables, we found that residents with the intermediate attribution style all scored the lowest on the policy response. This may be caused by the insensitivity of residents with an intermediate attribution style to policy changes.

5.5. Quadratic Response Surface Regression Analysis of Intervention Policies under Different Attributional Styles of Behavioral Outcomes

In the process of policy intervention in residents’ energy-saving behavior, the implementation of the policy is the most direct reflection of its effectiveness. To further explore the nonlinear influence of residents’ understanding of and support for energy-saving policies on their implementation under different attribution styles, the quadratic response surface regression analysis model (Formula (3)) with a better fitting effect was adopted in this study to further process and analyze the data. According to the attribution grouping mentioned above, the secondary response regression analysis was conducted for the positive, intermediate, and negative attribution style groups, respectively.

Table 12. Quadratic response surface regression analysis of intervention policies under positive attribution style.

	Y (The Degree of Implementation)
	PCP	FCP	ESP	CPP
$x_1$	−0.039	−0.001	−0.288	−0.600 *
$x_2$	0.298 ***	−0.065 *	1.009 *	−1.607 *
$x_1^2$	0.006	0.032	0.102 ***	0.071 **
$x_2^2$	0.040 ***	0.036	−0.031	0.242 ***
$x_1{x}_2$	0.127 ***	0.808 ***	−0.071	0.058
R²	0.984	0.382	0.218	0.414

Note: *** p < 0.001; ** p < 0.01; * p < 0.05.

,

Table 13. Quadratic response surface regression analysis of intervention policies under intermediate attribution style.

	Y (The Degree of Implementation)
	PCP	FCP	ESP	CPP
$x_1$	−0.461	−0.859 ***	0.112	−0.476 *
$x_2$	0.582 *	0.806 ***	1.801 ***	1.828 ***
$x_1^2$	0.090 *	0.127 ***	0.054 *	0.110 ***
$x_2^2$	−0.022	−0.037	0.151 ***	−0.148 ***
$x_1{x}_2$	−0.027	−0.010	−0.102	−0.044
R²	0.188	0.298	0.143	0.349

Note: *** p < 0.001; * p < 0.05.

and

Table 14. Quadratic response surface regression analysis of intervention policies under negative attribution style.

	Y (The Degree of Implementation)
	PCP	FCP	ESP	CPP
$x_1$	0.0976 ***	−0.329	−0.320	1.583 ***
$x_2$	−0.601 **	−1.189 ***	0.837 **	−0.239
$x_1^2$	−0.108 ***	−0.110 ***	0.057 **	−0.171 ***
$x_2^2$	0.139 ***	0.092 ***	−0.042	0.117 **
$x_1{x}_2$	−0.025	0.300 ***	0.024	−0.095 **
R²	0.251	0.372	0.377	0.319

Note: *** p < 0.001; ** p < 0.01.

show that in the positive attribution style group, residents’ understanding of and support for the policy has a significant nonlinear effect on the degree of implementation. Under the intermediate attribution style, residents’ degrees of understanding and support have no significant positive effect on the degree of execution and even have a negative effect on the degree of execution in terms of ESP and CPP. Under the negative attribution style, residents’ degrees of understanding and support for energy-saving policy have a significant negative effect on the degree of implementation, which is significantly reflected in PCP, FCP, and CPP. This further verifies hypothesis H4 mentioned above.

This paper used the three-dimensional diagram to conduct the response surface analysis. Figure 3, Figure 4 and Figure 5, respectively, show the relationship between residents’ degree of understanding, support, and implementation of energy-saving policies. The X-axis represents residents’ degree of understanding of related policy, the Y-axis represents residents’ degree of support willingness of related policy, and the Z-axis represents residents’ degree of implementation of related policy. Figure 3 shows that in the positive attribution style group, the residents’ response trend to energy conservation policies is obvious, indicating that when the residents’ understanding and support of the policies are consistent. Moreover, the degree of consistency is higher, and the residents’ implementation degree of the policies is higher. In the intermediate attribution style group, residents’ understanding, support, and implementation of the policy fluctuated significantly, as shown in Figure 4. Specifically for the economic subsidy policy, the interaction effect between the residents’ understanding of the policy and the degree of support is not obvious, and the degree of implementation is more significantly affected by the degree of support. In the negative attribution style group, it is obvious that a high degree of understanding and a high degree of support cannot completely bring about a high degree of implementation, especially in the price control policy and campaign publicity policy, as shown in Figure 5. In conclusion, the results of the analysis of the different attribution style groups show that when residents attach personal positive emotions to cognition, it is easier to convert cognition into behavior, while under the influence of negative emotions, residents are less likely to convert cognition into behavior. Kaiser et al. [56] and Richard et al. [57] also pointed out that when individuals attach individual positive emotions cognitively, it is easier to convert cognition into the behavioral intention. Combined with psychological attribution theory, it also means that residents with a positive attribution style will have a higher willingness to implement when they understand the policy with the support of positive emotion, which will further promote the implementation effect of energy-saving policies.

6. Conclusions and Recommendations

Based on the theory of psychological attribution and from the perspective of sustainable behavior development, this study constructed a policy response model of residents’ energy-saving behavior. It further explored the mechanism of residents’ response to energy-saving policies from the perspective of attributional style of behavior outcome and analyzed sample data. Our results lead to the following conclusions:

First, residents’ overall participation in the policy was low, and their understanding of the energy-saving policy was not high. Additionally, residents’ degree of support was significantly higher than their understanding and implementation. Residents’ support for the related policies was inconsistent with their actual implementation. Second, residents’ degrees of understanding and support for energy-saving policies positively promoted the degree of implementation. When residents’ degrees of understanding and support of energy-saving policies are both high, the implementation of the policy is also high. The consistency of understanding and support for the policy had a nonlinear influence on its implementation. Finally, this study verified that residents’ understanding, support, and implementation of energy-saving policies had significant differences in the attribution style of their energy-saving behavior results. Residents in the positive attribution style group had significantly higher understanding, support, and implementation of various energy-saving policies than those in the intermediate and negative attribution style groups. Under different attribution styles, residents’ degrees of understanding and support for policies had different effects on implementation.

Collectively, this paper puts forward recommendations to promote the sustainable development of residents’ energy-saving behavior as follows:

Innovate ways to popularize policies and foster an energy-efficient society. According to the research results, residents’ understanding of the policy is not sufficient, and both understanding and support can positively promote the degree of implementation. The government should strengthen the publicity of energy-saving-related policies through the Internet, television, advertising, and other channels. Give play to the propaganda role of society and make the residents clearly understand the specific regulations and implementation rules of various energy-saving policies. Strengthen the popularization of different energy-saving policies and improve the understanding and support of the residents for the policies to ensure the intervention effect of the policies.

Innovate in the form of policy guidance, set up customized information intervention channels, and attach importance to residents’ responses to energy conservation policies. The government should strengthen its energy-saving information intervention to increase the consistency of residents’ understanding and support for the policy. Through customized information dissemination channels, directional feedback is carried out to guide different groups to pay attention to energy conservation issues. We should make a good return visit to residents with low policy support, find out the root cause of the problem, and fully consider the problem from the perspective of residents. Targeted feedback is provided through customized information dissemination to enhance residents’ understanding of energy conservation and emission reduction issues such as climate warming.

Promoting the positive development of individual attribution and enhancing the sustainability of energy-saving behavior. The government should increase residents’ positive experiences of energy-saving behaviors, stabilize the psychological conditions for implementing energy-saving behaviors and enable residents to repeatedly carry out energy-saving and emission-reduction behaviors. On the one hand, community learning and energy-saving knowledge competitions can be held to inculcate the cognition of energy-saving significance and create an energy-saving community atmosphere. Rewards in the form of points, money, and prizes can be implemented to encourage energy efficiency investment-related energy-saving behaviors. On the other hand, the government can also use the information feedback system to strengthen residents’ perceptions of their energy-saving behavior and increase residents’ positive emotions about energy-saving behavior.

7. Limitations and Future Research Direction

Although this study has certain enlightenment and reference significance for the development of energy-saving behavior, there are still some limitations.

First, this paper is an empirical study based on a questionnaire survey. Due to the influence of subjective factors on respondents, it is difficult to rule out the possibility that there may be a deviation between the data filled in by residents and the real situation. In the future, the scope and number of samples can be expanded as far as possible. In addition, it is possible to conduct field behavior experiments, use dynamic tracking methods, etc.

Second, compared to previous studies, this study only considers the intervention scenarios of four policy tools. With more research and innovative development of energy-saving policies, future studies can further expand the types of policy tools and conduct intervention analyses of different policy combinations.

Finally, human behavior is complex. Based on psychological attribution theory, this paper analyzes the influence of different attribution styles of energy-saving behavior on the intervention policies of energy-saving behavior from the perspective of sustainable development of energy-saving behavior. Future research could extend the existing theoretical framework, such as by integrating values, social learning theory, etc. To sum up, relevant research can be further expanded in the future.