A study of residents’ intentions to participate in the renovation of older communities under the perspective of urban renewal: evidence from Zhangjiakou, China

ABSTRACT Community renovation involves energy conservation and emission reduction while fostering the process of urban renewal as a sustainable method of building upgrading. However, weak resident participation is the greatest obstacle to its development. This study aims to explore the psychology of the residents’ participation to stimulate their intention to participate. By combining the theory of planned behaviour with the norm activation model theory and government incentives, we collected data from 353 residents using a questionnaire and tested it using structural equation modelling. The findings suggest that government incentives, perceived behavioural control and personal norms all directly influence residents’ intentions to participate, with government incentives being the most significant influencing factor. Therefore, residents’ intentions to participate were driven by both rational and moral factors. Moreover, subjective norms and awareness of consequences had no significant effect on personal norms. Finally, the total mediating effect of subjective norms on personal norms and the moderating effect of low-income cohorts were verified. Thus, this study is a reference for the government and the community to improve the residents’ participatory statutes enhance their sense of belonging and the intention of various groups to participate and provide impetus to the urban renewal process. GRAPHICAL ABSTRACT


Introduction
Following the global economic slowdown in the 1980s, countries faced economic restructuring. Owing to its reform and opening-up policies, China was able to overcome its economic slump by promoting rapid urbanisation throughout the real estate industry Weber 2010). This approach boosted economic development in the short term; however, the concentration of time and the lack of foresight in the early economic layout and local development policies has resulted in a plethora of older communities today (Xu, Xue, and Huang 2022). Up to 170,000 old urban communities have been reported as requiring renovation in various regions, according to information made public by the Chinese government and preliminary statistics from the Ministry of Urban and Rural Development. These communities typically suffer from subpar building performance, inadequate public facilities, unreasonable road traffic planning and more safety hazards, which seriously affects the satisfaction of residents' living needs and the acquisition of a sense of well-being (Wu and He 2005;Bai et al. 2018;Wang et al. 2014). Furthermore, the existence of potential social conflicts also affects the appearance of the city. At the beginning of the 21st century, the concept of urban renewal was widely accepted in China to change the urban landscape through the demolition and renovation of old urban infrastructures and communities as well as to meet the needs of residents for a better living environment (Li, Ran, and Yinlong 2021;Lai, Chau, and Cheung 2018). There are two solutions for the renewal of old communities. The first option is to demolish and redevelop the entire old communities, which can change their aesthetic in the short term. However, demolition and redevelopment can cause the following concerns: (1) resource consumption is high, leading to excessive carbon emissions and threatening the environment (Xu et al. 2019;Shi et al. 2012), and (2) old communities are predominantly located in urban core areas, where land is expensive and demolition and redevelopment costs are high. Consequently, certain redevelopment projects suffer from a shortage of funds in the middle of the project and the project is stalled (Gao, Liu, and Dunford 2014) (3) The sense of identity and belonging of the original community is destroyed, which is not conducive to the formation and continuation of community culture (Chang 2010;Hikichi et al. 2016;Su, Zhao, and Tan 2015;Townshend et al. 2015). Thus, due to the problems associated with demolition and redevelopment, more communities are now choosing to adopt the second option, shifting the renovation method from major demolition and construction to dynamic upgrading and renovation. This both aligns with the sustainable development model and reflects the humanistic values of society.
The smooth implementation of the renovation of old communities requires the wide participation of community residents. Relevant policies actively guide and encourage community residents to participate extensively in the renovation of old communities and have an active influence. At the central level, it is proposed that community renovation should be "bottom-up", involving all residents in all phases of the community renewal project, including design, construction, monitoring, post-renovation management and evaluation and providing feedback. At the local level, Hangzhou, for example, explicitly requires that over two-thirds of the owners in the community reach a consensus on the community renewal and renovation plan prior to the implementation of a renewal project. On the other hand, certain scholars have investigated the importance of resident participation from the perspective of urban renewal. For instance, Elander Ingemar argues that sustainable urban renewal often brings inequality, yet citizen participation can help resolve these problems and controversies (Elander 2022). In summary, community renovation cannot be achieved without the participation of residents.
However, currently, many older communities still do not have resident self-governance organisations, and even more lack an efficient resident participation system, and the overall level of resident participation is low (Huang, Ma, and Song 2021;Guo, Zhou, and Li 2021). Moreover, the majority of the current studies by scholars at home and abroad focus on the physical and spatial environment, renovation methods and the economic benefits of the land brought by the renovation itself in the process of community renewal Zhu, Li, and Feng 2019;Duan et al. 2020), and limited studies focus on the issue of public participation (Imrie, Lees, and Raco 2009;Hauge, Thomsen, and Lofstrom 2013;Sanoff 1989). Furthermore, they predominantly focus on the external environment and the personal attributes of the residents (Brownill and Carpenter 2009;Filner 2001), lacking consideration of the residents own subjective initiative, and the weak participation of residents remains an obstacle to community renewal. This has severely constrained the government's ability to develop community renovation policies and has slowed the growth of the city's economy and image. Therefore, we analysed the psychological mechanisms behind resident participation to elucidate the intention to participate. Compared to participation in demolition and reconstruction, residents' participation in the renovation of older communities is altruistic regarding energy conservation and pollution reduction; residents can have a comfortable environment for themselves after participating in the renovation, which is egoistic. Thus, this study examines the pathways influencing residents' intention to partake in community renovation using the well-adapted theory of planned behaviour and normative activation theory. In addition, considering that the government, as an advocate of community renovation, has a nonnegligible influence on the psychology of resident participation, government incentives were introduced into the conceptual model as an expansion variable.
In this study, we define two concepts. First, the scope of the renovation of older communities: The basic category is predominantly for the renovation and upgrading of municipal supporting infrastructure and the maintenance of public parts such as roofs, external walls, stairs and pavements of buildings in the district. Improve the class is mainly the environment and supporting facilities renovation and construction, building energy-saving renovation within the small area, conditional building retrofitting elevators, etc. The upgrading class is predominantly the construction of supporting public service facilities and their intelligent renovation, such as kindergartens and other educational facilities, elderly service facilities, etc. Second, resident participation: The current model of citizen participation in China is primarily focused on the stage of "delegated power" (Arnstein 1969), where citizens have the right to take ownership of the renovation projects in old communities. Residents can contribute ideas to meet their needs before the renovation, demonstrate cooperation in the act of participation, promote the project and guide the renovation site.
This study explores the influence of residents' internal psychological factors on their intention to participate and validates the mechanisms of egoism and altruism in addition to government incentives. Furthermore, it also broadens the theoretical framework of resident participation. Thus, interested parties may use our findings to establish policies that encourage participation from residents.
Three theoretical and practical goals may be supported as a consequence of this investigation. First, the psychological model of residents' intention to participate is studied in terms of the psychology inherent in their subjective initiative, and the combination of the theory of planned behaviour and the theory of normative behaviour completes the psychological model of residents' intention to participate (PI) . Second, the addition of government incentive factors facilitates a more intuitive observation of how the government affects residents' psychological pathways and provides the government a framework upon which to increase people's psychological participation. Third, this study serves as a practical guide for the parties involved in understanding the psychology of resident participation, assists in mobilising residents more broadly for community reconstruction and offers new ideas for future research on community participation behaviour.

Theory of planned behaviour
The theory of planned behaviour (TPB) is an extension of the Theory of Rational Action, introduced by Ajzen in 1985. It is one of the most popular theories currently used to analyse the relationship between individual intentions and behaviour (Maichum, Parichatnon, and Peng 2016;Ajzen 2011). TPB uses perceived behavioural control (PBC), subjective norms (SN) and attitudes (ATB) to predict behavioural intentions (Ajzen 2011). As individuals become more positive about PBC, SN and ATB, their intention to perform the behaviour strengthens (Luiza Neto et al. 2020). PBC refers to how easy or difficult an individual perceives it will be to perform a particular behaviour, and it reflects the individual's perception of factors that facilitate or hinder the performance of the behaviour (Xu, Ramanathan, and Victor 2018). Thus, PBC is the degree to which an individual has control over the performance of a behaviour. SN is a person's perception of the social pressure exerted on them to perform or not perform the relevant behaviour (Ozen and Mauer 2002). Generally, people perform a behaviour when they evaluate it positively or when they believe it is important or others believe they should perform it (Ozen and Mauer 2002). However, ATT is typically considered to be a mediating element between SN and behaviour; therefore, they are not considered in this study (Zhang, Geng, and Sun 2017). The TPB has been applied in numerous pro-environmental and pro-social studies, for example, using it to investigate the main psychological factors influencing the way Brazilians travel (Luiza Neto et al. 2020) in addition to successfully using it to predict an individual's intention to start a business (Misoska, Dimitrova, and Mrsik 2016;Pejic Bach, Aleksic, and Marjana 2018), consumers' intention to visit green hotels (Verma and Chandra 2018) and to adopt hybrid cars (Wang et al. 2016).

Norm activation model theory
Norm activation model theory (NAM) was presented by Schwartz in 1977(Schwartz 1977. The primary factors that influence behaviour are personal norms (PN), the ascription of responsibility (AR) and awareness of consequences (AC). AC refers to an individual's perception of the possible negative consequences of not performing a behaviour (De Groot and Steg 2009;Steg and De Groot 2010,Chen, 1746-1753. AR refers to an individual's perception that the individual is responsible for the negative consequences of not performing a behaviour (De Groot and Steg 2009;Steg and De Groot 2010), and PN refers to an individual's perception of their responsibilities and duties (Schwartz 1977). NAM was originally developed to explain altruistic behaviour; however, it was expanded to explain pro-social behaviour. Several studies have demonstrated the capabilities of NAM in predicting environmentally friendly behaviours (Chen, 1746(Chen, -1753, such as energy-saving behaviour (Zhang, Wang, and Zhou 2013), public transportation (Nordfjaern and Rundmo 2019) and certain environmental behaviours (Schultz et al. 2005;Nordlund and Garvill 2002). Consequently, NAM is efficient in explaining pro-environmental behaviours.

Integration of planned behaviour theory and norm activation model theory
Resident's participation in the renovation of older communities is pro-environmental regarding reducing energy consumption and carbon emissions, and prosocial regarding improving the overall living environment of the community and promoting urban renewal, compared to their participation in demolition and rebuilding, all of which are driven by altruism. The residents' participation in the renovation provides them with a comfortable environment, which is a pro-ego act, and these acts are driven by egoism. TPB is perceived as a rational choice model by certain academics (Abrahamse et al. 2009). However, it disregards the impact of individual and environmental variations on behaviours (Carrington, Neville, and Whitwell 2010), more egoistically motivated. Therefore, TPB is insufficient for fully explaining the behavioural intention of residents to participate in the renovation of old communities (Han and Hyun 2017;Sang et al. 2020). NAM is more adaptable in describing pro-environmental conduct; however, it falls short of explaining egoistic behaviour and the function of self-motivation (Onwezen, Antonides, and Bartels 2013;Han 2015) because it is more altruistic and does not consider the influence of human rationality. Therefore, the conduct of residents participating in the repair of old communities can be better understood by integrating these two theories. In addition, numerous studies have demonstrated the applicability of the combination of the two theoretical models, such as the prediction of landscape maintenance intensity (Souto 2012) and the study of consumer over-ordering behaviour in restaurants (Yu et al. 2021).

Government incentives
Government incentives (GI) are used as an extension factor in this study: first, local authorities and communities predominantly rely on administrative orders to intervene in community environmental issues, and communities are often passive recipients of administrative guidance from local authorities, lacking positive interaction. This results in the renovation being detached from the community base and failing to meet the real needs of residents and inspire ownership. Furthermore, this may result in the low participation of residents in the renovation process and conflicts between the two sides. According to Heclo (2002), residents are reluctant to partake in public affairs since they are costly and provide limited rewards. Consequently, the government should create conditions that would allow residents to lower their input and partake in public affairs more easily. Ng (2002) considered appropriate compensation as an effective means of promoting resident participation. Foley and Martin (2000) argue that residents will become resentful and less likely to participate in community renovation if the government pursues political objectives while ignoring community concerns.
Second, government incentives have been introduced as an important factor in the promotion of green products and the renovation of housing projects. For instance, (Sheu and Chen 2012) used a threestage game model to analyse the effect of government financial intervention on competition in the supply chains for green products. Government incentives are crucial for purchase intention, according to a study by ) on young consumers' intention to purchase green homes in Shandong province, therefore they influence green consumption. From a residential perspective, (Cho and Eun-You 2020) explore the measures of community housing renovation initiatives, with government measures significantly influencing the process.
Finally, the government is the dominant and driving force in the renovation of old communities (Wu et al. 2021), and the majority of previous studies have focused on a combination of single theory and government incentives Wang et al. 2021) or a combination of planned behaviour theory and to study pro-environmental and pro-social behaviour (Sawitri, Hadiyanto, and Hadi 2015;Elhoushy 2022;Lopes et al. 2019). However, it is not possible to fully observe the pathways of government influence on altruistic and self-interested motivations; therefore, the utility of government incentives cannot be measured. Combining these factors to introduce government incentives into the model makes the study more comprehensive.

Research hypothesis
Certain studies have shown that PBC has a significant positive effect on behavioural intentions (Shi and Long 2022;Wang 2020;Xiao, Song, and You 2020;Zhang et al. 2021). Considering the residents' participation in old community renovation behaviour, when individuals present positive evaluations of their economic level and their knowledge of the functions and effects of old community renovation, they will perceive that it is easy to undertake old community renovation. Subsequently, this generates an intention to participate in old community renovation; therefore, individuals' high level of control over their self-perceptions can generate strong behavioural intentions .
PN is internalised social norms based on emotional awakening, social expectations and self-expectations (Schwartz 1977). Numerous scholars have discovered that PN directly impacts behavioural intentions (Sang et al. 2020;Li and Lu 2019;Sun and Lu 2012;Shin et al. 2018;Jansson 2011) and that when PN is activated personal moral and objective expectations constrain objective subjects to produce pro-environmental behaviour, which will result in a behavioural willingness to transform old communities. Therefore, we created the following two hypotheses: H1. PBC positively influences PI.

H2. PN positively influences PI.
In the norm activation model theory, AC and AR are two prerequisites for activating PN (Schwartz 1977). AC positively impacts PN in the area of green behaviour (Han and Hyun 2017;Sang et al. 2020) and residents' perceptions of negative consequences of not participating in the renovation of older communities subsequently activate PN to generate behavioural intentions (Zhang, Geng, and Sun 2017). In this study, AR refers to the ability of residents to actively bear the negative consequences of participating in the renovation of older communities. In the relationship between AC, AR and PN, many studies have highlighted that AR has a mediating role (Sang et al. 2020). For example, (De Groot and Steg 2009) deduced five studies of pro-environmental behaviour, predominantly in a chain mediation model (Rosenthal and Ho 2020;Gao, Huang, and Zhang 2017). When a strong sense of consequence arises it results in a positive effect of responsibility ascription and thus activates PN. Therefore, we hypothesised the following.
In this study, SN refers to residents' perceptions of the pressures caused by social opinion and objective circumstances that may motivate residents to participate in the renovation of older communities. In the relationship between SN and PN, SN precedes PN. Han et al. (2019) identified a significant role for SN and PN in explaining the decision-making process of passengers toward pro-environmental cruise products and verified a positive causal relationship between the two.
In certain studies, SN is perceived as the weakest factor influencing behavioural intentions (Luiza Neto et al. 2020); therefore, SN is not used as a proximal determinant of intention to participate. SN has been shown to influence PBC. Portnov et al. (2018) demonstrated this finding by investigating the behavioural intentions of Thai consumers.
In their study on the factors influencing residents' travel mode choice, (Sang et al. 2020) concluded that PBC directly influences behavioural intentions, and it also affects the degree of activation of PN. Numerous studies also support this conclusion Wittenberg, Bloebaum, and Matthies 2018). PN can only be positively reinforced when there is a strong self-cue. Therefore, the following hypotheses were proposed. Inevitably, the renovation of older communities is constrained by construction sites and construction cycles, causing long periods of inconvenience to residents. In addition, in the absence of financial incentives, residents typically settle for the status quo, and urban renewal will not be possible. Therefore, appropriate government incentives can be used to change the situation where residents resist and refuse to pay for public services. GI has been shown to have a significant relationship with the PI and PBC , for example, (Xiao, Song, and You 2020) derived a significant relationship between government and farmers' intention to participate in an extended model of planned behaviour theory. Therefore, we hypothesised as follows.
Figure 1 depicts a model of the factors influencing residents' intention to participate in the renovation of old communities, which is based on the above assumptions.

Questionnaire design
This paper uses a questionnaire method for data collection. The questionnaire contains three sections. The first explains the relevant background and concepts, the second contains basic information about the respondents and the third contains questions on potential constructs. These questionnaires were administered to residents of communities undergoing renovation of older communities in addition to those preparing for renovation. The questionnaire was designed using a five-point Likert scale (strongly disagree = 1, disagree = 2, neutral = 3, agree = 4 and strongly agree = 5), and the results were analysed by returning the questionnaires to further investigate the factors that influence residents' intention to participate in the renovation of older communities.
Through literature review, the model's sevenquestion items were adapted from the research scales of numerous scholars. In addition, we performed expert interviews prior to releasing the questionnaire and invited four construction industry experts to assess the questionnaire and make any required changes to confirm the reliability and validity of the scale. The revised questionnaire was used to perform 40 pre-surveys, and the questionnaire was further improved based on the pre-survey findings to create the final questionnaire. The original measurement items were all in English; however, we distributed our questionnaires to people in China. Consequently, we asked two bilingual researchers to translate the questions into Chinese and subsequently to translate the final survey results into English. The specific measurement questions are listed in Table 1.

Data collection
This case study was performed in Zhangjiakou, Hebei Province, China, a community-based community renovation that typically does not affect the normal life of the residents (e.g. no moving away from the original place of residence, affecting shopping, etc.), predominantly in spring and summer. It was selected based on the following factors: First, Zhangjiakou is one of the first pilot cities to be renovated, and in 2021, the city's 15 counties and districts plan to renovate 176 old communities, involving 40,492 households and a renovation area of 3,106,400 square metres. The renovation and construction of old communities is still in progress, making it easy to collect data from the research subjects. Second, the low population density of Zhangjiakou makes it more prudent to renovate old communities on an existing basis to achieve double value-added economic and social benefits than to expand the number of new buildings. Finally, there are many fourth-tier cities similar to Zhangjiakou in China, and an empirical study of them would be beneficial for the replication of results, programme implementation and improvement in similar cities.
The survey area involved six districts (Qiaodong District, Qiaoxi District, Xuanhua District, etc.) and 10 counties (Zhangbei County, Guyuan County, etc.) in Zhangjiakou, Hebei Province. The survey began on March 2022 and ended on 13 April 2022. The following data collection methods were used: First, in online research, 210 questionnaires were distributed by anonymous survey with the help of social software (e.g. WeChat, QQ, etc.) and email. Second, field research, in which the surveyors invited residents of old communities to complete 160 paper questionnaires. A total of 370 questionnaires were distributed in the above two ways, of which seven were invalid (under 1 minute to answer the question or 10 consecutive questions with the same answer), 10 were not returned and 353 were valid, with an effective rate of 95%. According to Kline's (Markus 2012) study, there should be at least 210 valid questionnaires in this survey, and the actual number of valid questionnaires is 353, which satisfied the criterion. The demographic information of the respondents is presented in Table 2.

Reliability and validity tests
The reliability and validity of the data were checked prior to analysis. The software SPSS 26 was used to test the reliability of the questionnaire. The Cronbach's alpha coefficient was used to test the internal consistency of the scale; higher values indicate higher reliability. A coefficient of 0.7 or above for Cronbach's alpha is an acceptable range (Hair et al. 2010), and the measurement results indicate that the reliability and reliability of the data are high. As shown in Table 3, Cronbach's alpha passed the test. The validity tests provided convergent and discriminant validities. Convergent validity indicates the degree of convergence of questions in the same dimension (the degree of relevance of the questions). The discriminant validity refers to the degree of difference between different dimensional constructs. Convergent validity was tested using combined reliability (CR) and average extracted variance (AVE) (Fornell and Larcker 1981). CR estimates above a critical value of 0.7 correspond to acceptable levels of convergent validity (Fornell and Larcker 1981), and an AVE value above 0.5 indicates good convergent validity (Fornell and Larcker 1981). Chin et al. (Chin, Gopal, and Salisbury 1997) suggested a standardised factor-loading threshold of 0.6; the data are acceptable when the standardised factor data in the study are close to or above 0.6. As shown by the calculated values listed in Table 3, the model met these criteria. An AVE square root greater than the inter-concept standardised regression weighting coefficient indicates good discriminant validity (Chin, Gopal, and Salisbury 1997). The discriminant validity analysis is shown in Table 4.  In occupation, senior manager is usually the main leader and sub-main leader of the organisation (e.g. school principal, vice principal, party secretary, etc.), middle manager is the person in charge of specific matters in the organisation (e.g. director in charge of teaching, director in charge of safety, etc.), general staff is usually the person who performs actions in the organisation (e.g. teachers, cleaners in the school) and self-employed person is a freelancer who does not have an agency.

Model goodness-of-fit test
The goodness-of-fit indices of the model were examined using AMOS 26. In this study, seven indicators, chisquare/ (degree of freedom (χ2/df), the goodness-of-fit index (GFI), adjusted goodness-of-fit index (AGFI), relative fit index (RFI), normed fit index (NFI), comparative fit index (CFI) and root mean square error of approximation (RMSEA), were selected to test the model fit. The results showed a χ2/df value of 2.119 < 3; GFI, AGFI, NFI, CFI, IFI and TLI were all acceptable criteria Hu and Bentler 1999), and RMSEA value was 0.056 < 0.08. These results all satisfy the criteria, indicating that the model fit was satisfactory and that the model could be used for further research. The results of the model fitting are shown in Table 5.

Results of the path coefficient test
AMOS 26 was used to test the path coefficients. Table 6 shows the results of the structural model. All eight hypotheses were supported at a significance level of 0.05, and two hypotheses were not supported. The standardised path coefficients indicate the relationship and degree of influence between the constructs in the structural model. Together, PBC, PN and GI explained 33.3% of the variance in the willingness to participate in PI, with PBC (β = 0.218, t = 3.242 and p = ***), PN (β = 0.122, t = 2.179 and p = 0.029*) and GI (β = 0.370, t = 5.693 and p = ***) significantly influencing residents' willingness to participate in older communities, thus

Intermediation effects
The mediating effect allows for the analysis of the process and mechanism of influence of the independent variables on the dependent variable. As shown in Figure 1, there are four mediating variables in this model. The confidence interval test was conducted using the Bootstrap confidence interval method in AMOS 26 (Portnov et al. 2018), with 2000 replicates of the sample, and the mediation effect was tested at a 95% confidence interval. The confidence interval test for bias-corrected in the Bootstrap method indicates a significant mediation effect when the intersection of the upper and lower bounds does not contain 0. Table 7 shows the results of the mediation effect analysis of the model. The four paths "SN→PBC→PN", "AC→AR→PN", "PBC→PN→PI" and "GI→PBC→ PI", with the effect percentages of 100%, 27.61%, 18.76% and 24.85% respectively, of which "SN→PBC→PN" was fully mediated.

Multi-group comparison result
Constancy was tested using continuous nesting constraints. In model A, the measurement weights were made equal to that of the baseline model to examine whether the composition of the measurement model was constant across groups. However, in model B, the structural weights were made equal to that of model A to examine whether the structural weights (i.e. path coefficients) were constant across groups. If the results of model A were not significant (p > 0.05), then the composition of the measurement model would be constant across all groups, and the questionnaire (or scale) designed for the study could be applied to different groups. If the results of model A were significant, then different questionnaires had to be designed for different groups or the composition of the latent variables required further adjustment. If the results of model B were not significant (p > 0.05), then the structural model would be constant across all groups, the same influencing mechanism would be discussed across groups and the factors did not moderate the structure of the model. However, if model B was significant, then the assumed influencing mechanism of the baseline model varied across groups and further discussion of the influencing mechanism across groups would be required. Due to sample size limitations, gender, age, income and education were divided into two groups, each according to data distribution characteristics for comparison to ensure that the sample size of each group was as balanced as possible. Gender was divided into the male group (N = 208) and female group (N = 145), and age was divided into the youth group (18-35 years old, N = 265) and the middle-aged group (35 years old and above, N = 88). Income was divided into the lowincome group (monthly household income of 7000 and below, N = 186) and the high-income group (monthly household income of 7000 and above, N = 167) and education was divided into the low- The c denotes total effect value; a*b denotes the mediating effect value; c' denotes the direct effect value. education group (education at college and below, N = 117) and the high-education group (education at bachelor's degree and above, N = 236). The results are presented in Table 8. The results highlight that the p-values of sex, age, education and whether or not to contact the item grouping in models A and B are over 0.05 (i.e. the sample differences are not significant), and the absolute increments of the fitted indicators are small, indicating that sex, age, education and whether or not to contact the item have good constancy in the measurement model and structural model of residents' intention to participate renovation.
Income groupings have a p-value of below 0.05 in model B, indicating that different income groups produce significant differences in intention to participate, i.e. income plays a moderating role in the model, as shown in Figure 2, where the numbers in brackets next to each pathway are the standardised pathway coefficients for different income groups (i.e. low-income group, high-income group). Table 9 shows the critical ratios of the model path coefficients, and when |CR| > 1.96 (i.e. P < 0.05), the path proved to be significantly different between the clusters. The analysis concludes that income has a moderating effect on the effect of PBC on residents' intention to participate (hypothesis H1), with a standardised path coefficient of 0.586 for the highincome group and 0.052 for the low-income group, nearly 10 times the effect of the high-income group than the low-income group. This demonstrates that the facilitation of PBC on residents' intention to participate is more pronounced under the effect of high income and that when the high-income group feels that their financial ability at their disposal, their intention to renovate is stronger.

Discussion
Based on the TPB and the NAM theory, this study investigates the process by which seven factors, including PBC, GI and PN influence residents' intention to participate in the renovation of older communities. Ten hypotheses were proposed, eight of which are supported. H3 and H6 are not supported.   The results support H1, H9 and H10 in the model, which aligns with the findings of Zhang  and Shin et al. (Shin et al. 2018), according to which both GI and PBC significantly affect intention to participate, with GI being the most important factor affecting residents' intention to participate in the renovation of older communities. Furthermore, GI also indirectly affect residents' intention to participate by influencing PBC. In addition, PBC directly influences intention to participate in addition to indirectly through the activation of PN. Therefore, hypothesis H2 is also supported.
Contrary to the findings of other investigations, hypotheses H3 and H6 were not validated, and SN and AC did not directly affect PN (Sang et al. 2020). This is primarily due to Zhangjiakou's ranking as a fourth-tier Chinese city. Compared to residents of medium and large cities, those of smaller cities are more concerned with the dynamics in and around their communities due to multiple factors such as regional economy and culture and are unable to experience first-hand the functions of the renovated communities under the incipient situation of old community renovation. Consequently, they more cautiously approach promotion and publicity from the outside world and do not realise that they require it. In addition, residents of small cities are not sufficiently aware of the relationship between urban renewal and energy efficiency to professionally perceive the negative impacts of demolition and redevelopment or to understand the actual benefits of urban renewal. Hypothesis H6 is not supported by the fact that residents do not see it as their own fault when they think it is a minor issue, which prevents them from activating PN.
In addition, it is hypothesised that H4, H5, H7 and H8 are supported. Personal norms were significantly influenced by the mediating effects of "SN→PBC→PN" and "AC→AR→PN", which also significantly influenced the intention to participate, which aligns with previous research findings. PN, as the link between altruistic and egoistic models, is driven to altruism by their own moral constraints and their external influences. When PN is activated, residents' sense of moral responsibility increases accordingly and they take the initiative to participate in community renovations.
The PBC influence of high and low-income groups on residents' intentions to partake in the renovation of older communities varied, according to the group analysis. Low-income groups may be limited by economic factors, and even if they are not satisfied with the existing living environment, they are not willing to pay for it, so they do not feel strongly about it and are more reluctant to participate in the renovation. However, the higher income groups may be more willing to pay for the improvement of their existing living environment, positively contributing to the perceived power of participation in renovation and showing a strong desire to renovate older communities. Thus, income has a more pronounced moderating effect in the "PBC→PI" path.

Reflections and recommendations
This study has important implications for policy development for the government. First, the government's role significantly impacts residents' participation intentions. The government can more easily achieve the goal of green, low-carbon and sustainable development by introducing environmentally friendly renovation content and processes in conjunction with green energy-saving technologies. This will also raise the residents' awareness of environmental protection and help them more readily awaken their altruistic mindsets. The government should strengthen publicity by introducing content that addresses the policy, meaning, objectives, content and process and demonstrating successful cases and the effectiveness of the renovation. The form of publicity, self-published media platforms like public websites can be utilised for publicity in addition to the conventional techniques to increase locals' awareness of current social concerns and their sense of responsibility. The implementation of the green and low-carbon concept is strengthened, and the renewal of the supporting facilities around the community is promoted, leading to the renovation of the surface. This both influences the intention to participate directly and enhances the PBC of residents indirectly; moreover, it also stimulates their enthusiasm for participation and the spirit of human ownership. In addition, in certain areas, the government provided financial subsidies ranging from RMB 150/m 2 for the small scale of the initial renovation projects in old communities. The concept of "active guidance and appropriate subsidies" has enhanced residents' awareness of the renovation projects. Furthermore, the creation of specific service organisations and attentive public listening are excellent strategies for encouraging resident participation.
Second, the parties involved must improve the community engagement system and focus on the opinions of the residents. The government's all-inclusive restoration model for community renovation, a livelihood project, may conflict with people's immediate requirements, and there is also a dearth of effective exposure to the most recent government policies. Therefore, it is crucial to establish a sound mechanism for community participation, stimulate the ownership of residents, establish special organisations to publicise national policies in addition to lowcarbon targets, relay the needs of residents, to forge links between the government and residents. Residents' PBC and PN will likely be improved by creating a positive community atmosphere and a bottom-up participation paradigm, which will affect residents' intention to engage.
In addition, group analysis of varied income levels indicated a moderating influence in the route of PBC on participation intention. Low-income groups have a low intention to participate, and appropriate subsidies can stimulate their intention to participate. The current policy on subsidies for community renewal projects is low, predominantly due to the existing urban renewal relying excessively on government financial funds, the lack of motivation for funding, the low intention of social capital investment and the lack of more funds to subsidise residents. This reduces the goodwill and participation of low-income residents. Therefore, combining the financial resources of the government, residents and the community is the most effective method to find a solution. These are all measures at the level of egotism.
Finally, increasing a sense of belonging to the community is a powerful strategy to encourage residents' altruistic psychology. Due to the lack of cultural infrastructure and community cultural activities, the majority of older communities currently do not have a good community culture. Therefore, the rational planning of public space infrastructure and community activities is a priority in community renovation, not only to make the residents feel the significance of the renovation but also to enhance their sense of belonging to the community and thus strengthen the altruistic mentality.

Conclusion
The renovation of old communities alleviates the contradiction between urban construction needs and sustainable development goals and is an effective way to achieve energy saving and emission reduction. Despite an expanding corpus of study, the psychological mechanisms influencing residents' intention to participate in renovation remain to be clarified. This study explores the routes of effect from egoistic and altruistic psychology on residents' inclination to participate in the repair of older communities, spanning both rational and moral reasons, based on the TPB-NAM theory and government incentives. By constructing structural equation models, path analysis, mediating effect analysis and multi-group comparative analysis to find the mechanism of influence between the seven elements and the relationship between the relevant elements. The study demonstrates that (1) GI is the most important driver of intention to participate followed by PBC and PN. (2) The partial mediation effects of the AC→AR→PN, PBC→PN→PI and GI→PBC→PI paths and the total mediation utility of the SN→PBC→PN path were verified. (3) Low-income groups are less likely to participate in the pathway of PBC on intention to participate. These conclusions assist those participating in the refurbishment to identify more effective ways to encourage locals to partake in the project.
However, this study has several limitations. First, although intended behaviour is an effective predictor, it is not identical to the actual behaviour. Therefore, future research on the real behaviour of residents participating in community renovations should be performed. Second, despite the fact that Zhangjiakou is a typical city, the sample source is still geographically constrained and might not be representative of some of the more economically powerful provinces and cities, and regions of different economic levels may produce different research results. In future studies, the number and target population of the survey can be expanded, diversified demographics can be added to reflect the actual transformation market, questionnaires can be implemented in other regions and the subsequent research results can be compared with the present research results, which will be of greater value to interested parties. Finally, there is no research on whether residents' previous participation affects their intention to participate again, which can be added in a posteriori studies.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Notes on contributors
Shuochen Xiao, PhD, is a lecturer whose research focuses on user behavior analysis and intelligent decision-making. She has published more than 15 papers in domestic and international academic journals and conferences, and has participated as a key researcher in projects supported by the National Natural Science Foundation of China and the National Key R&D Program.

Lei
Li is a master's student, participated in 2 projects at the departmental and university levels, 5 papers were indexed by CNKI, one of which was indexed by China Science and Technology Core.
Jing Ma is a master's degree student, participated in 3 provincial and departmental level projects, research direction is cost management. Dan Liu is an MA TESOL student at the University of Liverpool.

Jiahao
Li is a master's degree student.