https://orcid.org/0000-0003-4510-1748
Figures
Abstract
This paper builds on a model of individual and collective climate action on the neighbourhood level recently presented by Klöckner et al. [1]. In this model, types of local climate action were empirically categorized (diet, travel, protest, other climate actions), and it was found that both individual and collective intentions contribute to self-reported climate actions in these categories and that collective intentions were weaker than individual intentions. Based on these findings, the current paper proposes a theoretically derived intervention strategy based on neighbourhood events. These events comprised hands-on work on contextualized climate action, experiential learning, and creative and disruptive communication techniques, aiming at strengthening the collective motivation to act against climate change in the neighbourhoods. The interventions were implemented in nine European neighbourhoods, and we were able to collect some first data on their potential in a survey in seven of the nine neighbourhoods. In total, 46 respondents answered the survey both before and after the interventions, 13 of whom participated in at least one of the intervention events. With this explorative small sample, we find indications that the interventions might be successful in increasing the perceived social norms in the neighbourhoods, the identification with the neighbourhoods, and decreasing perceived barriers to action. Smaller positive effects seem to occur for collective intentions and collective efficacy, and behaviour change. The individual factors appear to be mostly unaffected by the interventions, with potentially some improvement in individual efficacy. Overall, this pilot study points to the potential of neighbourhood-based climate interventions as a new methodology for activating a path to climate action underutilized in current campaigns. The preliminary findings we present here help generate studies to test them under more robust conditions and present a methodology for innovative intervention design.
Citation: Klöckner CA, Löfström E, Brenner-Fliesser M, Winkler C, Kofler V, De Gregorio E, et al. (2025) Can neighbourhood interventions strengthen collective climate action? PLOS Clim 4(12): e0000571. https://doi.org/10.1371/journal.pclm.0000571
Editor: Fanli Jia, Seton Hall University, UNITED STATES OF AMERICA
Received: February 26, 2025; Accepted: December 2, 2025; Published: December 18, 2025
Copyright: © 2025 Klöckner et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The data these analyses are based on can be accessed here: https://zenodo.org/records/10992143.
Funding: This study has been conducted in the JPI Climate project “CLEAN Cultures”, which is funded in JPI Climate joint transnational call SOLSTICE (Enabling Societal Transformation in the Face of Climate Change). Project numbers: Norwegian Research Council 321315, funding achieved by CAK; Austrian Research Promotion Agency 879545, funding achieved by MBF; funding by the Italian Ministry of Education, University, and Research, funding acquired by GC; Academy of Finland, funding decision number 338128. Funding acquired by LS. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Fast and deep changes on all levels of society are imperative if the target of limiting to anything close to 1.5 degrees should remain realistically achievable [1]. While many of these transition decisions need to be made and implemented on the intergovernmental level, they also include components of citizen action and behaviour change, as mere technological transitions will not be able to get the world on the right track towards a sustainable future [2]. This might be diet choices or energy use, investment decisions in efficiency, sufficiency decisions (reducing consumption), or political action. Consequently, individual climate action and its drivers have attracted substantial interest by researchers from the more individual-focused social sciences like psychology [3–6] or behavioural economics [7]. At the same time, this perspective has been criticized for blaming the individual for the failures of the higher-order socio-techno-political systems [8–10]. In this paper, we take a recently published model joining individual and collective drivers of climate actions as a starting point [11]. Klöckner et al. conclude that the potential of collective climate action is underutilized and that an approach strengthening collective action-taking might overcome the limits of individual action. Based on these hypotheses, we designed an experience-based series of neighbourhood interventions to strengthen the capacity of neighbourhoods to take climate action. This paper proposes an approach for developing such interventions and explorative results of an evaluation in several small pilots. We consider this paper primarily as a starting point, presenting our ideas, stimulating discussion, and providing a basis for further exploration in future studies. In the remainder of the introduction, we first turn to a brief review of research on individual and collective drivers of climate action, before introducing the model combining both individual and collective drivers. Then, we will propose an intervention concept to strengthen collective climate action taking on the neighbourhood level and exemplify this with interventions that we developed for nine European neighbourhoods. In the final section of the paper, we present a preliminary evaluation of these pilots and discuss the implications for further research and practice.
Individual vs. collective drivers of climate action
Drivers and barriers of individual climate action have been extensively studied, often using popular psychological decision models like the Theory of Planned Behaviour [12], the Norm-Activation Model [13], or the Value-Belief-Norm Theory [14] as an inspiration. More recently, there have been attempts to integrate some of these theories into more comprehensive models [15,16]. These models identify factors as individual intentions to act (the will to make an effort to implement a climate action in the near future), attitudes towards different climate actions (a general evaluation of the behavioural alternatives), perceived efficacy or behavioural control (the ease of implementing different climate action), social norms (the influence of other people’s expectations and actions), personal norms and values (moral considerations about the behaviours in question) as important drivers of an individual’s climate-related behaviour. Other authors have explored the important role of a personal environmental identity as a driver of climate action across different situational contexts [17–19]. Clayton [18] defined environmental identity as “one part of the way in which people form their self-concept: a sense of connection to some part of the non-human natural environment, based on history, emotional attachment, and/or similarity” (page 45).
More recently, however, the focus has changed away from analysing and understanding individual actions to studying collective actions in the environmental domain [20–22]. In this perspective, the assumption is that demanding actions, which require sacrifices, are more easily taken by people if they take them as part of a group they identify with, and where the group derives an advantage that outweighs the individual losses. This then makes social identity (identification with a specific group and willingness to contribute to the success of that group) a central variable. In the Social Identity Model of Pro-Environmental Action (SIMPEA) [22], the key assumptions are that the assessment of environmental crises may lead to personal and collective emotional responses. If the in-group norms are in favour of action, collective efficacy (“we can do this together”) is experienced, and people socially identify with the group, collective action is likely. Furthermore, Barth et al. [21] argue that collective climate action is triggered by collective intentions to act jointly in the same way as individual actions are by individual intentions.
The preceding study presenting a model of individual and collective climate action on the neighbourhood level
In a previously published paper, we developed a model combining individual and collective motivation of climate action and tested it in a survey in nine European neighbourhoods [11]. In our analyses in that paper, we focused on the neighbourhood as a social unit, as it can be easily identified for interventions, and as people usually spend a significant time of their days within the area [23] and have common activities with neighbours [24], as well as that people often identify with the place they live [25]. We found that individual and collective intentions are relevant drivers of climate action in four domains (diet, travel, protest, and general climate behaviour, including, for example, energy use, less food waste, or consumption reduction). However, in the current conditions, the impact by individual intentions on climate action exceeded about three-fold the influence by collective intentions controlling for socio-structural differences between people. Individual intentions were mostly driven by feelings of individual efficacy, whereas collective intentions were driven by social norms, collective efficacy, and the perceived social capital in the neighbourhood. In our model in the previous paper, we included social capital, since it captures well the capacity of a neighbourhood to act together and constitutes the socio-economic context in which neighbourhood action may unfold [26–28]. We included this variable as a potential moderator (but we were unable to test its effect due to the small sample size).
Transformative learning, experiential learning, and disruptive interventions
Taking our analysis of nine neighbourhoods and the existing climate actions (and their drivers and barriers) presented in the previous section as a starting point, we then designed an intervention strategy to strengthen the collective pathway on stimulating climate action. From our previous work in the neighbourhoods, we concluded that there may be a big potential for neighbourhood action, but that the neighbours were stuck in an individualist approach, lacking collective reflection about potential action pathways. We therefore developed three contextualized intervention events in each neighbourhood, which included components of transformative and experiential learning to stimulate a mind shift and a questioning of the status quo, and disruptive communication to disrupt habitualized ways of living and create emotional engagement. In this section, we will elaborate further about the concepts. We also included activities in our interventions that were designed to engage as many representatives of the neighbourhood as possible in fun social activities. The activities were designed to be low-threshold and to appeal to a wider audience than just the already climate-motivated. They included elements of knowledge dissemination about climate change and the results of a preliminary analysis of the neighbourhoods we conducted based on document studies and expert interviews with key actors in the neighbourhoods prior to our interventions [29]).
For the design of our interventions, we built upon the Transformative Learning Theory [30]. The theory assumes that people can change their perspectives and beliefs through critical reflection, which then might lead to a personal transformation (hence, transformative learning). The theory was conceived to understand how people learn and grow when confronted with challenging situations and disorienting dilemmas (both of which apply well to climate change actions). The theory assumes that people have stable frames of reference, which consist of their habits of mind, viewpoints, and established cognitive response patterns. These stable frames make people non-responsive to new situations unless they are encountering what the theory refers to as disorienting dilemmas, hence experiences that challenge an individual’s frames of reference, strong enough to cause discomfort. Consequently, frames of reference might be questioned, and critical reflection starts. This stage of critical reflection is essential for transforming the reference frames. Finally, critical reflection can lead to the acknowledgement of new reference frames. This then constitutes the transformation. For our design of the intervention events, this means that we were aiming to create situations in which such disorienting dilemmas regarding climate change and (in)action on the neighbourhood level become tangible. We then aimed to encourage critical reflection, while creating arenas for open dialogue and exchange between neighbours and other stakeholders where relevant. We aimed at empowering participants in our interventions to take meaningful action and establish learning communities, where knowledge and experiences are shared between peers.
As such, the concept of our interventions shows a certain overlap with experiential learning [31]. Kolb describes experiential learning as a circular process where a concrete (often physical) experience of something leads to critical reflective observation, which triggers an abstract conceptualization, which then is broken down again into pragmatic active experimentation, which then again leads to the next cycle starting with contextually rich concrete experiences [32]. For our intervention planning, this means that we aimed to create episodes of hands-on experiences, creating an emotional connection to the topic, which were then iterated with periods of reflection and abstraction, before breaking it down again to pragmatic, actionable assumptions, which lead to new experiences.
Finally, we also utilized the idea of disruptive communication [33–35] to create moments in the events where the disorienting dilemmas as outlined above were provoked. Disruptive communication, as we understand it, is communication that is designed to trigger emotional reactions by interfering with people’s established way of living (following the frames of reference mentioned above). These disruptive communication elements often have a creative or art-related element, which makes them more stimulating and easier to tolerate than blunt and brutal interferences with people’s lives. Following this idea, we implemented creative disruptive elements to energize the participants in our events.
Based on these theoretical considerations, we designed intervention campaigns in nine neighbourhoods. We then piloted them and explored their potential effects. Fig 1 shows the overall structure of our activities, and which paper is based on which dataset. After the brief introduction of the nine neighbourhoods in the next section, we present the interventions in the following section in detail.
Selection and short description of the neighbourhoods (This section is paraphrased from the previous paper [11] and its supplementary information, please refer to that paper for a more comprehensive description.)
The development of the neighbourhood interventions was part of a larger project on neighbourhood climate action (www.cleancultures.org). The participating neighbourhoods were selected based on a list of pragmatic criteria: (a) they were located in the countries of the respective researchers involved in the project so that investigations could take place without language barriers and that the researchers were familiar with the cultural context, (b) they should together represent a good distribution of rural and urban areas, as well as a spread of socio-economic conditions, and (c) first contacts with the municipalities in these areas had already been established. This allowed us to start with an already existing level of trust between the research team and key local actors. The selection of neighbourhoods was conducted with a focus on diversity of living conditions, not representativity for each country, let alone Europe [11]. Three Austrian (Admont, Jakomini, Eggenberg), two Italian (Marco Simone, Santa Maria), two Norwegian (Driva, Myrsletta), and two Finnish (Pyhäntä, Simo) neighbourhoods were included in the project. The nine neighbourhoods are very different in size, composition of inhabitants, and main challenges.
Admont is a village with around 5,000 inhabitants. Tourism is an important driver of the local economy. Demographically, Admont shows a shrinking population and higher proportions of older people. Local discussions are not strongly centred around climate change, but rather livelihood, environment, and sustainability. One highly discussed issue is (sustainable) mobility, since Admont is poorly connected to public transport. In general, climate change is not really seen as a big problem in this neighbourhood.
The multi-cultural district of Jakomini plays an important role in Graz not only because of its central location, but also because of its local institutions, remarkable buildings, and numerous events. Over 30,000 people live in an area of 4.06 km². Jakomini is the hottest district in Graz, due to its high density of buildings and comparably little green space. One long-lasting challenge is the unsatisfying traffic situation with too much car traffic. Another challenge is the creation of more public green space. Climate change is not necessarily seen as something that needs to be tackled at the neighbourhood level.
Eggenberg is a middle-class district in the city of Graz. Currently, there is much construction/densification in Eggenberg. Eggenberg has only a few social open spaces (sports field, playground), but a lot of green space. The population of Eggenberg (21,000) is marked by a share of around one-third of people coming from other countries. The two main challenges in the district are 1) how to provide living space and infrastructure for a growing population, and 2) how to improve mobility concepts. Climate change, in general, does not seem to be a big subject of discussion in the district.
Marco Simone (15,000 inhabitants) is a relatively new settlement in the city of Guidonia close to Rome, mostly composed of single or semi-detached houses, and it is populated by middle-class families mostly commuting to Rome on a daily basis. The environmental situation is affected by the effects of the big city (pollution, car traffic). The dominant themes mainly concern two areas. The first is the management of urban and industrial waste. A second relevant issue is that of mobility and transport. Climate change does not emerge as a very salient theme among the residents interviewed.
Santa Maria (10.000 inhabitants) is a neighbourhood in the outer edge of the municipality of Macomer in the northwest of Sardinia. The surrounding area is characterized by a rural economy, and it is also an area of strong tourist attraction. The industrial activities have faced an important crisis over the last two decades. The area is also subject to a shrinking population. The main climate-related challenge is the increased frequency and severity of wildfires during summer.
Driva (400 inhabitants) is a settlement in the municipality of Oppdal in the Norwegian mountains. Driva is a popular resort for vacations, especially during the wintertime. Collective transport does not connect Driva well, but the nearest town, Oppdal, is accessible via train and bus. Driva’s population is increasing and is expected to continue to grow. The need for and protection of the local, more traditional farms is also highlighted. It is a close-knit society where people know each other and often do business together or otherwise cooperate with each other.
Myrsletta (500 inhabitants) is a neighbourhood located 2 km from the centre of the town of Ski. The neighbourhood consists of town and semi-detached houses. Many inhabitants are families with children, but also some seniors. The area is expected to grow quickly due to a recently improved connection via commuter trains to Oslo. Private car ownership dominates, and people do not use public transport much. Local sharing of tools in a neighbourhood Facebook group has become popular over the last few years.
Pyhäntä is a small municipality in the Northern Ostrobothnia area. The municipality covers an area of 847.5 square kilometres. Pyhäntä has a population of 1,631. Its population density is 1.8 inhabitants/ km2. The municipality has succeeded in turning the trend of population loss during the last few years, and the economy of the community has remained stable. Pyhäntä is often referred to as “the most industrialised municipality in Finland”. Use of land for windmills is one important way of compensating for the losses from the phase-out of peat production.
Simo (3,000 inhabitants) is a municipality by the Baltic Sea. As a whole, Simo is a sparsely populated area with a population density of 2.0 inhabitants/ land-km2. The population of Simo has been decreasing over the last years, and the share of people over 65 years has been increasing. Simo is characterized as a forerunner area of Lapland in wind power, and it is the largest wind power producer in the area. Even though the population of Simo is small, the large land area of the municipality means that not everyone knows everyone.
Contextualized interventions to strengthen neighbourhood climate action
Based on the findings from our initial study [11], the analyses of the neighbourhoods, and the theoretical approach to transformative and experiential learning outlined in the previous sections, each country team designed a series of contextualized neighbourhood interventions, all including elements of creative stimulation and disruptive communication. We aimed to implement three events in each of the nine neighbourhoods, specifically designed for the local neighbourhood conditions. The theoretical framework for these interventions was developed together in the transnational research team, and intervention ideas were discussed together and probed on their compatibility with the design principles outlined above. The local research teams then decided specifically about which interventions to develop further and implemented them, hiring professional actors, musicians, artists, etc., where necessary. Table 1 displays the dates of the interventions. Unfortunately, a survey after the intervention series could not be implemented in the Finnish neighbourhoods, so the exploratory assessment of the impact of our interventions is based on results from only seven of our nine initial neighbourhoods. We present the Finnish interventions, nonetheless, in the following paragraphs.
Table 2 presents an overview of the interventions implemented. A detailed description of the interventions per neighbourhood can be found in the supplementary materials. The guiding principles of the intervention design were to combine art or spiritual experiences with discussions and reflections about climate change and actions against it on the local level. The events were organized to build and strengthen the local communities and show that people in the neighbourhood are engaged in climate action and that collective agency can make a difference. The interventions were adapted to the locally available cultural and social resources and challenges identified in the neighbourhood analysis. By doing this, we could capitalise on existing social relations in the neighbourhood. A disadvantage of this approach is that we were not able to evaluate specific intervention elements across different contexts, and that the intervention concepts differed substantially between the neighbourhoods.
As can be seen in Table 1 and Fig 1, the implementation of the survey waves and interventions varied substantially in timing between the neighbourhoods. This was due to practical restrictions. In some neighbourhoods, we took advantage of events being implemented anyway in the neighbourhoods (e.g., outdoor cinema, concerts, markets, festivals) to attract a larger audience. This required us to adhere to the scheduling of these events. Furthermore, the financing for the Italian arm of the study was received much later than for the other countries; thus, the whole implementation was delayed here. This asynchronous implementation has implications for the comparability of the intervention effects: As global events (both climate-related and not) may have interfered, we cannot compare effects between our pilots. Furthermore, there were different time lags between interventions and the second wave of the survey, which also has implications for how strong effects can be expected when we were able to implement the surveys. However, as this is an explorative study, we consider these potential interferences as tolerable.
An explorative evaluation of the effect of our intervention pilots
After implementing a series of interventions as pilots in all nine neighbourhoods as described above, we aimed to assess the potential effects of such interventions on neighbourhood members. To do this, we build on the model developed in our previous paper [11] and track the effects on variables that were identified as having a significant impact for local climate action in that paper (see Table 3). We assess potential effects of our interventions by comparing survey answers of participants of our intervention events with non-participants from the same neighbourhood before and after the interventions. We repeated parts of the survey that our previous paper, which was based on the same neighbourhoods, but excluded sections of the survey not relevant for the analyses described below to reduce survey length in the second round. Table 3 presents an overview of our hypotheses and a justification for each of them. In addition to the factors derived from the theoretical overview presented above and the results of the previous paper, we also tested the potential of changes in perceived barriers towards climate action in the neighbourhood and whether people perceive manifestations of climate change in the neighbourhood.
Methods
Ethical clearance
This study was approved by the following committees for research ethics: Norwegian Agency for Shared Services in Education and Research, SIKT (Ref.nr. 121957, approved on 22/04/2022), Roma Tre Ethics Commission (approved in commission meeting 15/02/2022). For the Austrian sample, an ethical clearance was not required for an anonymous paper-and-pencil survey as per the ethical procedures of Joanneum Research. Participants gave informed consent to participate after being informed about their rights at the beginning of the surveys by proceeding to the first survey page. No minors or people unable to give consent were included in the empirical evaluation of the interventions, although children were allowed to participate in the intervention activities themselves. Recruitment for the study was open in Austria from 01/07/2022–15/09/2022, in Italy from 01/09/2022–20/05/2023, and in Norway from 03/06/2022–30/09/2022 (periods of first wave data collection). The second wave of data was collected in Austria from 28/03/2024–24/04/2024, in Italy from 18/04/2024–28/07/2024, and in Norway from 14/09/2024–26/09/2024. Data from the Finnish case is not analysed in this exploratory impact evaluation as the post-intervention data collection could not be implemented.
Evaluation methodology
To assess whether variables identified as impactful on neighbourhood climate action in Klöckner et al. [11] changed as expected (see Table 3), we replicated the sections of the survey that measured these variables and calculated the scores in the same way as described in [11].
To be able to track changes within the same person, participants were asked in the first survey to generate an individual code based on, for example, the second letter of the maiden name of their mother, the third letter of their father’s first name, the second digit in the mother’s birthdate, etc. In the second round of the survey, after the interventions were finished, the same code generation instrument was used, and answers were matched based on these codes. For 46 participants, matching codes in both surveys could be identified. Data in the first round before the interventions was collected between 03/06/2022 and 20/05/2023 (see [11] for more details), the second round of data collection was implemented between 01/06/2024 and 30/09/2024.
In the survey, we measured the following variables (see [11] for full details, see Table 4 for reliability scores, which are, on average, rather stable across the waves):
Self-reported behaviour was measured by providing people with a list of 19 behaviours, asking which of them they were already doing. We followed the same method of aggregating these into four categories as in the first paper: (a) Diet (3 items: eating a diet low on animal products, eating vegetarian, eating vegan as three steps of increasing difficulty). (b) Travel behaviour (7 items, e.g., replacing short-distance car trips by walking or cycling, avoiding short flights, carpooling, etc.). (c) Protest behaviour (3 items, e.g., sign a petition, contact politicians). (d) general pro-climate action (6 items, e.g., reducing food waste, reducing energy use, buying second hand, etc.). Afterwards, the scores were standardized to the average percentage of people doing the behaviours within each domain.
Individual and collective intention were measured with one item each (“I personally intend to contribute to local climate actions in the neighbourhood within the next year” and “We in the neighbourhood intend to take local climate action together within the next year”). Individual and collective efficacy were measured by three items each (e.g., “I think that I personally can manage to permanently lower my personal CO2-emissions” and “I am capable to make a small but important contribution towards a climate-neutral society together with other people in the neighbourhood”). Attitudes were measured by three items (e.g., “To act together against climate change in our neighbourhood would be good”). Social norms were measured by two items (e.g., “Most people in the neighbourhood expect me to take action against climate change”). Social capital was measured by four items (e.g., “We in the neighbourhood all draw in the same direction”).
Identification with the neighbourhood was measured with four items (e.g., “I am very attached to the neighbourhood”). Environmental identity was measured with one item (“Acting pro-environmentally is an important part of who I am”). Perceived manifestation of climate change in the neighbourhood was measured with one item (“Have you perceived changes in your local area that you think are connected to climate change?”). Finally, barriers against climate action were measured by four items (e.g., “It is difficult in the neighbourhood to reach an agreement between the neighbours on what to do against climate change”). All items were answered on a five-point Likert scale (with the exception of climate change perception, which had five answers from “definitely not” to “yes, definitely”).
Sample description (This section is partly paraphrased from the previous paper [11] and its supplementary information, please refer to that paper for a more comprehensive description.)
In the seven neighbourhoods participating in this longitudinal explorative study, we received a total of 884 answers in the first round before the interventions were started (excluding the two Finnish neighbourhoods included in the previous paper [11]). In the second wave, we received a total of 161 answers. However, for only 46 of them we were able to match the first and the second survey, so these form the basis of our analyses presented in this paper. Data collection varied between the neighbourhoods based on what the local research teams assessed as being the potentially most successful approach. In Austria, data were collected with paper-and-pencil questionnaires distributed in the selected neighbourhoods by mail as well as with an online questionnaire, also available in English and Turkish. In Norway, the researchers hired local adolescents to go from door to door in the neighbourhoods to distribute an invitation letter with the link to an online version of the questionnaire. In Italy, a similar approach was chosen.
Of these 46 participants, 13 indicated having visited at least one of the intervention events, which were briefly described to them in the second survey. We did not expect that the remaining 33 participants who did not indicate having attended at least one of the events were affected indirectly by the events happening in their neighbourhoods. Thirty-three of the participants were from the Austrian neighbourhoods, 12 from the Italian neighborhoods, and one from one of the Norwegian neighbourhoods. Twenty-six were identifying as females (56.5%), 19 as males (41.3%), and one as divers (2.2%). The age distribution is rather balanced with 10 in the age group 18–34 years, 14 in the group 35–49 years, 18 in the 50–65 years group, and four being older. Most of the participants have higher education (25 with a college or university degree), and most place themselves in the higher groups in social status (34 participants on level 7 or higher of 10, where 10 indicates the highest social status). The participants in the intervention were not different from the non-participants in distribution of genders (chi2 = 4.27, df = 2, p = .118), age groups (chi2 = 2.81, df = 3, p = .422), or social status (chi2 = 8.51, df = 5, p = .130). However, they had a higher educational status (chi2 = 9.16, df = 3, p = .027) with almost exclusively highly educated people. The two groups also did not differ in the variables tested below before the interventions (t-tests, all p > .05). We decided not to treat our data to even out differences in distribution across our neighbourhoods and education, as the dataset is very small and such techniques would have overemphasized single answers. Furthermore, this is an explorative study meant to stimulate a discussion rather than giving final results; thus, we decided to keep the dataset as it is, being aware that our explorative results presented in the next section are highly influenced by the included respondents and their distribution to different neighbourhoods and educational status.
Analysis strategy
As our sample size is extremely small, we base our exploration on a combination of three criteria, being aware that sample size is very small for the more traditional statistical tests: (1) We inspected the plots of the effects for the two groups visually and checked if the observed effects appear to be in the expected direction and of a visual size big enough to justify the assumption that the hypothesis is met. (2) We used traditional significance testing (here, univariate tests in conjunction with MANOVA multivariate tests) to test if the difference in differences before-after intervention between participants and non-participants is significantly different from 0. As the sample is small and we had directed hypotheses (see Table 3), we used one-sided testing. However, the sample size we have and the uneven group size result in the fact that we only have enough power to detect large effects. (3) Thus, we also inspected the effect sizes of the differences in differences between the two groups. Using the rules-of-thumb as for example outlined in Adams and Conway [43], we interpret values of eta2 below.01 as small effects, values around.06 as medium-sized effects, and values around or above.14 as large effects. Our preliminary conclusions presented in the discussion are based on a combination of these criteria.
Explorative evaluation results
Changes in climate mitigation behaviour
The first analysis conducted was to compare the self-reported frequency of climate change-related behaviour on the individual level before and after the neighbourhood interventions, compared between the participants who attended at least one of the three neighbourhood events in the respective neighbourhood during the project and those who did not. All 95% confidence intervals between the measurements overlap. Thus, no statistically significant effects could be detected. A visual inspection of the effects shows that in three out of four behavioural categories, the participants of the neighbourhood events increase their self-reported behaviour frequency, whereas the non-participants remain stable or reduce slightly (see Fig 2).
A one-way MANOVA test with Pillai’s Trace did not reveal a statistically significant multivariate effect of event participation on z-standardized before-after behavioural differences across all four domains (F(4, 34)=.779, p = .546, effect size eta2 = .084). Among the univariate results, differences between event participants and non-participants in changes of diet behaviour are the closest to statistical significance (F(1, 37)=1.774, p = .191, eta2 = .046), whereas the differences for travel (F(1, 37)=.009, p = .924, eta2 = .000), protest (F(1, 37)=.040, p = .843, eta2 = .001) and general climate behaviour (F(1, 37)=.430, p = .516, eta2 = .011) are smaller.
Changes in individual drivers of behaviour
In the next step, the effects of event participation on individual intentions to act against climate change and three drivers of that intention (attitudes, individual efficacy, and environmental identity) were tested. Visual inspection shows no changes in intentions in both groups, as well as small effects in the other variables (see Fig 3). Please be aware that the effect on attitudes is opposed to what was expected.
This is confirmed by a one-way MANOVA test with Pillai’s Trace, which did not reveal a statistically significant multivariate effect of event participation on z-standardized before-after differences in these factors (F(4, 41)=1.957, p = .119, effect size eta2 = .160), but the multivariate effect is larger for behaviours. Among the univariate results, differences between event participants and non-participants in changes of individual efficacy are the closest to statistical significance (F(1, 44)=2.160, p = .144, eta2 = .048), whereas the differences for attitudes (F(1, 44)=1.789, p = .184, eta2 = .040), individual intention (F(1, 44)=.024, p = .879, eta2 = .001) and environmental identity (F(1, 44)=.341, p = .562, eta2 = .008) are smaller.
Changes in collective drivers of behaviour
In contrast to the individual drivers of climate change behaviour, the collective drivers show stronger effects in a visual inspection (see Fig 4). In all four cases, the difference between participants and non-participants points to an improvement in the collective driver after the events relative to the development of the non-participants. However, 95% confidence intervals overlap also here.
Furthermore, a one-way MANOVA test with Pillai’s Trace did not reveal a statistically significant multivariate effect of event participation on z-standardized before-after differences in these factors (F(4, 40)=1.387, p = .256, effect size eta2 = .122), and the multivariate effect is a bit smaller than for the individual factors. Among the univariate results, differences between event participants and non-participants in changes of social norms are the closest to statistical significance (F(1, 43)=2.977, p = .092, eta2 = .065). Also the difference for identification with the neighbourhood are close to statistical significance (F(1, 43)=2.810, p = .096, eta2 = .063), whereas the differences for collective intention (F(1, 43)=1.358, p = .250, eta2 = .031) and collective efficacy (F(1, 43)=2.552, p = .117, eta2 = .056) are smaller. Effect sizes in the collective drivers are higher than for the individual drivers, and two of the univariate results are statistically significant for one-sided testing (as positive effects were expected).
Other drivers of behaviour
Finally, an inspection of effects on perceived barriers to climate action in the neighbourhood, local indications of climate change in the neighbourhood, and perceived social capital in the neighbourhood indicates that the only effect seems to show in that barriers are perceived as less high (see Fig 5).
Also here, the one-way MANOVA test with Pillai’s Trace did not reveal a statistically significant multivariate effect of event participation on z-standardized before-after differences in these factors (F(3, 31)=1.258, p = .301, effect size eta2 = .084). Among the univariate results, differences between event participants and non-participants in changes of barrier perception are the closest to statistical significance (F(1, 43)=3.862, p = .056, eta2 = .082), whereas the differences for climate change perception in the neighbourhood (F(1, 43)=.000, p = .996, eta2 = .000) and social capital (F(1, 43)=.135, p = .715, eta2 = .003) are non-existent.
Table 5 below presents an overview of the results of our explorative hypothesis testing.
Discussion
The exploration of the potential effects of our pilot implementations points mostly in the expected direction. For the self-reported behaviours, all four results are in line with our expectation: We see a medium-sized effect for changes in diet behaviour, and a small effect for general climate behaviour. Against our expectation, the effect on protesting behaviour is very small, which seems to indicate that protesting behaviour is per se uncommon in the studied neighbourhoods and that the neighbourhood action may have rather opened for local action than protesting against the authorities (which often were involved in the interventions). Effects on travel behaviour were as expected extremely small, which indicated that travel choices are mostly determined by external conditions rather than internal motivational factors [39,44].
As we expected, the effects of the neighbourhood interventions on the individual factors seem to be small or absent. Individual environmental identity and individual intentions appear to be mostly unaffected by the interventions, which may underline that these processes still mostly happen on the individual level, unaffected by the interventions aiming at strengthening the collective capacity to act. We found indications for a small to medium-sized effect for increased individual efficacy, though, which might indicate that the interventions might have indirectly strengthened the participants’ perceived individual efficacy, also by making the effects of one’s actions in the larger context more salient in line with the Theory of Planned Behaviour [12]. Looking back at our theoretical framework for creating reflection and learning experiences, this might be an outcome of experiential learning, which provides hands-on experiences with action-taking [31,32]. However, we also found an unexpected small to medium-sized effect on attitudes to act: Participants of the intervention events seem to have less positive attitudes to individual climate action after the event than before, which might indicate that especially very motivated people adjust their attitude levels to the lower average level of the group [45].
As expected, the strongest effects were observed for collective factors, here in particular social norms and identification with the neighbourhood (both achieve one-sided significance in spite of the small sample size). Thus, we see an indication that our intervention events might make social norms salient and creating a stronger identification with the neighbourhoods. Effects of collective intention and efficacy point in the same direction, but they are weaker than expected. This is in line with what we intended to achieve in the design of our interventions. With elements of disruptive communication (for example, the improvisation theatre group literally interrupting the mayor on the opening of the festival, or the clown breaking into the presentation of project results), we potentially created moments of irritation or stopping up, which, according to the theory behind disruptive communication, opens space for reflection [33]. As the series of events were designed to question accepted assumptions in line with transformative learning theory [30] about climate action in the neighbourhood (especially false consensus and pluralistic ignorance effects that lead to underestimating other neighbours’ engagement [46]), they may have succeeded in making salient, that the collective level of engagement is higher than expected.
Finally, among the other factors, climate change perception in the neighbourhoods and perceived social capital were apparently not affected by the interventions, whereas perceived barriers to action seem to have been reduced, probably because the interventions focused on concrete climate actions in the neighbourhoods and the barriers were phrased in terms of social barriers as part of the experiential learning approach. Also, this effect was significant in one-sided testing and of medium effect size.
Limitations of the study
The main limitation of the presented study is obviously the very small sample size, which prevents us from following conventional significance testing of the effects and puts a large degree of uncertainty on the presented results. Therefore, we propose to treat them rather as indications generating hypotheses for future research than conclusive results. However, we consider the combination of three criteria (visual inspection, significance testing, and inspection of effect sizes) as reasonably robust, in conjunction with the overall pattern of results further underlines our narrative and warrants further research. It is possible that what we found has been caused by just random variations in people’s assessments and the (self-) selection of participants.
A second limitation is that participation in the events is, of course, not randomly assigned. We organized the events in the neighbourhoods and invited all neighbours to participate, but far from all did. Thus, the group of participants was strongly self-selected, which indicates that they might have been particularly receptive to the interventions. Along these lines, it is worth noticing that the participants in the events did not differ from non-participants in the tested variables, but had a higher educational level. Maybe this form of events appeals less to people with lower education than to those with higher education, which needs further exploration, as it would mean that specific groups of people might not feel invited to these events. However, self-selection should have worked in the direction of climate-interested people (and thereby already more active people) being more motivated to attend the events, which would rather under- than overestimate the effects of the intervention. Another dimension of self-selection is more critical, though: It is likely that people who are interested in social interaction with neighbours and are identified with the neighbourhood are more likely to participate in the events than other people. These might also be people with higher education. This might lead to an overestimation of the effects on collective factors, as people less connected to the social environment of the neighbourhood may respond less positively.
A third limitation is that the intervention concepts differed substantially between the countries based on the resources and networks available to each national research group. Also, the timing was very different between the neighbourhoods for practical reasons as outlined above. This means that it is impossible to draw conclusions on which elements in our intervention concept of transformative, experiential learning with elements of disruption were effective, and which timings are optimal. Further, more systematic research on these aspects (e.g., implementing more comparable intervention concepts across several neighbourhoods in a synchronized way) is necessary.
Finally, the presented explorative results are based on seven European neighbourhoods (as the two interventions in Finland could not be analysed). These are diverse on many criteria, but nonetheless a very particular selection. In addition, response rates in the Italian neighbourhoods were higher than in the Austrian and Norwegian neighborhoods, which means that the results are strongly impacted by the conditions in the two Italian neighbourhoods than in the other five. It is hard to say what impacted the very different number of responses per neighbourhood in the surveys. First of all, the neighbourhoods are of very different sizes. Even if the absolute numbers of responses in Norway were small in wave 1 (see [11]), the response rate was still higher than in the larger neighbourhoods. However, with the expected dropout rates, this number was reduced even further. Different recruitment strategies for the second wave might also have affected the response rates (as personal approaches, as used in Italy, were apparently more successful in motivating participants to take the second wave survey). In hindsight, a combination of quantitative and qualitative evaluation approaches might have alleviated some of these issues by producing more contextualized information, also from small samples.
Conclusion
In this paper, we describe the development of an intervention strategy to stimulate collective climate action in neighbourhoods, based on a model of local climate action [11], and transformative learning theory [30], experiential learning [31,32], and disruptive communication [33]. We implemented pilots of this intervention strategy in nine diverse European neighbourhoods and found some indication that our neighbourhood interventions might yield promising effects. With all caution that the small sample warrants, the results seem to indicate that neighbourhood interventions as a means of stimulating collective climate action should be studied further as they may be a promising alternative to over-individualized action appeals. Further research is necessary to consolidate the presented effects, but we consider the results interesting enough to justify more research in this new arena of climate communication and action.
Supporting information
S1 Text. Detailed description of the interventions per neighbourhood.
https://doi.org/10.1371/journal.pclm.0000571.s001
(DOCX)
S1 Checklist. Inclusivity in global research.
https://doi.org/10.1371/journal.pclm.0000571.s002
(DOCX)
References
- 1.
Masson-Delmotte V, et al. Global Warming of 1.5°C. An IPCC Special Report on the impacts of global warming of 1.5°C above pre-industrial levels and related global greenhouse gas emission pathways, in the context of strengthening the global response to the threat of climate change, sustainable development, and efforts to eradicate poverty. Geneva, Switzerland: World Meteorological Organization; 2019.
- 2. Shwom R, Lorenzen JA. Changing household consumption to address climate change: social scientific insights and challenges. WIREs Climate Change. 2012;3(5):379–95.
- 3.
Klöckner CA. The psychology of pro-environmental communication – going beyond standard information strategies. London: Palgrave Macmillan; 2015.
- 4. Bamberg S, Möser G. Twenty years after Hines, Hungerford, and Tomera: A new meta-analysis of psycho-social determinants of pro-environmental behaviour. Journal of Environmental Psychology. 2007;27(1):14–25.
- 5.
Swim J, et al. Psychology and global climate change: Addressing a multi-faceted phenomenon and set of challenges. A report by the American Psychological Association’s task force on the interface between psychology and global climate change. 2009.
- 6. Whitmarsh L, Poortinga W, Capstick S. Behaviour change to address climate change. Curr Opin Psychol. 2021;42:76–81. pmid:33991862
- 7. Reichl J, et al. The drivers of individual climate actions in Europe. Global Environmental Change. 2021;71:102390.
- 8. Evans D. Blaming the consumer – once again: the social and material contexts of everyday food waste practices in some English households. Critical Public Health. 2011;21(4):429–40.
- 9. Holm L. Blaming the consumer: On the free choice of consumers and the decline in food quality in Denmark. Critical Public Health. 2003;13(2):139–54.
- 10. Shove E. Beyond the ABC: Climate Change Policy and Theories of Social Change. Environ Plan A. 2010;42(6):1273–85.
- 11. Klöckner CA, Brenner-Fliesser M, Carrus G, De Gregorio E, Erica L, Luketina R, et al. Climate actions on the neighbourhood level—Individual, collective, cultural, and socio-structural factors. PLOS Clim. 2024;3(11):e0000424.
- 12. Ajzen I. The theory of planned behavior. Organizational behavior and human decision processes. 1991;50(2):179–211.
- 13.
Schwartz SH, Howard JA. A normative decision-making model of altruism. In: Rushton JP, Sorrentino RM, editors. Altruism and helping behavior. Hillsdale: Lawrence Erlbaum; 1981. p. 189–211.
- 14. Stern PC. New Environmental Theories: Toward a Coherent Theory of Environmentally Significant Behavior. Journal of Social Issues. 2000;56(3):407–24.
- 15. Klöckner CA. A comprehensive model of the psychology of environmental behaviour—A meta-analysis. Global Environmental Change. 2013;23(5):1028–38.
- 16. Klöckner CA, Blöbaum A. A comprehensive action determination model: Toward a broader understanding of ecological behaviour using the example of travel mode choice. Journal of Environmental Psychology. 2010;30(4):574–86.
- 17. Carrus G, Nenci AM, Caddeo P. The role of ethnic identity and perceived ethnic norms in the purchase of ethnical food products. Appetite. 2009;52(1):65–71. pmid:18775756
- 18.
Clayton S. Environmental identity: A conceptual and an operational definition. Identity and the natural environment: The psychological significance of nature. 2003. p. 45–65.
- 19.
Masson T, et al. Identity Processes and individual factors in Energy Decisions – Two comprehensive Meta-Analyses. in Public report in the H2020 project ECHOES. Project number 727470. Trondheim: NTNU; 2017.
- 20. Bamberg S, Rees J, Seebauer S. Collective climate action: Determinants of participation intention in community-based pro-environmental initiatives. Journal of Environmental Psychology. 2015;43:155–65.
- 21. Barth M, Masson T, Fritsche I, Fielding K, Smith JR. Collective responses to global challenges: The social psychology of pro-environmental action. Journal of Environmental Psychology. 2021;74:101562.
- 22. Fritsche I, Barth M, Jugert P, Masson T, Reese G. A social identity model of pro-environmental action (SIMPEA). Psychol Rev. 2018;125(2):245–69. pmid:29265852
- 23. Lewis S, Zapata E. How Much Time Do We Spend in Neighborhoods? Preprint available at SSRN 5180441, 2025: https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5180441
- 24. Crow G, Allan G, Summers M. Neither Busybodies nor Nobodies: Managing Proximity and Distance in Neighbourly Relations. Sociology. 2002;36(1):127–45.
- 25.
Carrus G, et al. Place attachment, community identification, and pro-environmental engagement. Place attachment. Advances in theory, methods and applications. 2014. p. 154–64.
- 26.
Adger WN. Social capital, collective action, and adaptation to climate change. Der klimawandel: Sozialwissenschaftliche perspektiven. 2010. p. 327–45.
- 27.
Aldrich DP, Page-Tan CM, Paul CJ. Social capital and climate change adaptation, in Oxford research encyclopedia of climate science. 2016.
- 28. Carmen E, Fazey I, Ross H, Bedinger M, Smith FM, Prager K, et al. Building community resilience in a context of climate change: The role of social capital. Ambio. 2022;51(6):1371–87. pmid:35015248
- 29.
Klöckner CA, et al. Neighbourhood Case Analysis, in Reports from the CleanCultures Project. Trondheim, Norway: Norwegian University of Science and Technology; 2023. p. 79.
- 30. Mezirow J. An overview on transformative learning. Lifelong learning. 2008:40–54.
- 31.
Kolb DA. Experiential learning: Experience as the source of learning and development. FT press: 2014.
- 32. Morris TH. Experiential learning – a systematic review and revision of Kolb’s model. Interactive Learning Environments. 2019;28(8):1064–77.
- 33.
Klöckner CA, Löfström E. Disruptive Environmental Communication. Springer Nature; 2022.
- 34. Löfström E, Klöckner CA, Nesvold IH. Nature in Your Face - Disruptive Climate Change Communication and Eco-Visualization as Part of a Garden-Based Learning Approach Involving Primary School Children and Teachers in Co-creating the Future. Front Psychol. 2020;11:568068. pmid:33391083
- 35. Löfström E, Richter I, Nesvold IH. Disruptive communication as a means to engage children in solving environmental challenges: A case study on plastic pollution. Frontiers in Psychology. 2021:3599.
- 36. Ojala M. Hope and climate-change engagement from a psychological perspective. Curr Opin Psychol. 2023;49:101514. pmid:36502586
- 37. Diaz JOP. Active Hope as a Catalyst for Mental and Psychosocial Health in Climate-Related Disasters. PSYCH. 2024;15(05):634–44.
- 38. Stern PC. New Environmental Theories: Toward a Coherent Theory of Environmentally Significant Behavior. Journal of Social Issues. 2000;56(3):407–24.
- 39. Javaid A, Creutzig F, Bamberg S. Determinants of low-carbon transport mode adoption: systematic review of reviews. Environ Res Lett. 2020;15(10):103002.
- 40. Petty RE, Wegener DT, Fabrigar LR. Attitudes and attitude change. Annual Review of Psychology. 1997;48(1):609–47.
- 41. Whitmarsh L, O’Neill S. Green identity, green living? The role of pro-environmental self-identity in determining consistency across diverse pro-environmental behaviours. Journal of Environmental Psychology. 2010;30(3):305–14.
- 42. Crona B, Wutich A, Brewis A, Gartin M. Perceptions of climate change: Linking local and global perceptions through a cultural knowledge approach. Climatic Change. 2013;119(2):519–31.
- 43.
Adams MA, Conway TL. Eta squared. Encyclopedia of quality of life and well-being research. Springer; 2021. p. 1–2.
- 44. Scheiner J, Holz-Rau C. Travel mode choice: affected by objective or subjective determinants?. Transportation. 2007;34(4):487–511.
- 45. Binder DM, Bourgeois MJ. Direct and indirect effects of group discussion on consensus. Social Influence. 2006;1(4):249–64.
- 46. Leviston Z, Walker I, Morwinski S. Your opinion on climate change might not be as common as you think. Nature Clim Change. 2012;3(4):334–7.