Australia has one of the most unique and biodiverse natural environments in the world. However, due to climate change, habitat destruction, invasive species, pollution, and resource extraction, the state of Australia’s natural environment is rapidly deteriorating (Cresswell et al. 2021). At a global level, the World Economic Forum (2022) has declared biodiversity destruction the third most severe threat humanity will face in the next ten years. National and intergovernmental bodies are developing strategies to mitigate the loss of threatened species and habitats (Australian Government 2022; United Nations 2021). However, mobilisation of broad public support is also essential to prevent and reverse biodiversity loss (Convention on Biological Diversity 2010; Stephens 2023).

Biodiversity loss stems from human activities (Amel et al. 2017; Maxwell et al. 2016). Thus, the solutions also lie in changing our behaviour (Schultz 2011, 2014). Government policy instruments, such as regulations, are often adopted in an effort to influence human behaviour for environmental benefits. However, public opinion can also influence the effectiveness of environmental policies (Burstein 2003; Massingham et al. 2023). For example, in relation to waste reduction, banning or taxing single-use plastics has been implemented by many national and local governments, with the intention of reducing their use (Borg et al. 2022). At the same time, community support or opposition can encourage or discourage decisionmakers from introducing such restrictions on single-use plastics (e.g. Bharadwaj et al. 2021; Macintosh et al. 2020; Murdoch 2010). In order to enable effective biodiversity conservation, it is important to understand how to foster nature conservation behaviours as well as policy support among the wider community.

Supporting nature

The behaviours that support conservation outcomes are typically considered a subset of general ‘pro-environmental’ behaviours (i.e. any behaviour that minimises one’s impact on the environment), as they focus on activities that support the conservation and restoration of the biodiversity of plants and animals (Richardson et al. 2020). While behaviours that support nature conservation could be considered in aggregate, distinctions between different types of behaviours are necessary because the behavioural drivers will likely vary alongside the behavioural characteristics (Church et al. 2023).

According to Stern (2000), general pro-environmental behaviours can be categorised based on the level of involvement to perform the behaviour and visibility of the behaviour. Private-sphere environmentalism involves behaviours such as the purchase, use, and disposal of personal and household products. In contrast, public-sphere environmentalism can be separated into environmental activism, which is active involvement in environmental movements (e.g. petitioning, donating to environmental organisations), and nonactivist behaviours in the public sphere, which represents non-activist support for public policy. Others have also identified similar public vs. private high-level categories: for example, Larson et al. (2015) classified pro-environmental behaviours into lifestyle (i.e. private-sphere) and environmental citizenship (i.e. activism), as well as social environmentalism and land stewardship (the latter only applicable to land owners). Specifically in the conservation area, Selinske et al. (2020) identified six categories of biodiversity behaviours, including public-sphere (advocacy and donating) and private-sphere (consumption, lifestyle, and stewardship), as well as social behaviours (such as discussing environmental issues).

While ‘policy support’ is not always included in categories of environmental behaviour, due to its more passive nature, it is well established that wicked problems require a fundamental societal shift and approach to public policy (Kaufman et al. 2021). For example, in the context of climate change, it has been recognised that behaviour change alone is not sufficient to reduce global emissions without additional structural changes, such as those enabled by public policy (Flagg and Kirchhoff 2018). In many areas of conservation, there has been a shift away from regulatory policies, with increasing emphasis on financial or market-based incentives (Bouwma et al. 2018; Dean et al. 2023). Despite this shift, there has been little exploration of community perceptions and preferences for different types of conservation policies – including those that do and do not affect individual behaviour. Given that policy framing is known to be important for biodiversity conservation (Jokinen et al. 2018), it is important to understand the different ways that people support conservation policy initiatives.

Influencing support for nature

Human behaviour is influenced by a vast array of factors, evidenced by the sheer number of behaviour and behaviour change theories and models (Darnton 2008; Davis et al. 2015). One of the more common, albeit simplistic, models of behaviour change is the knowledge-attitude-behaviour model, which proposes that the more people have a cognitive understanding about an issue, the more likely they are to change their attitudes towards the issue, and ultimately their behaviour (De Vries 2020). The model was originally proposed in the health context (Bettinghaus 1986), however, it has also been applied to environmental contexts (e.g. see Myung 2018; Sousa et al. 2021).

A wide variety of contextual, social, and psychological factors are known to influence human behaviour (Darnton 2008; Davis et al. 2015). However, previous research has consistently found a relationship between cognitive factors (such as knowledge or awareness) and attitudinal factors (such as concern) and engagement in pro-environmental behaviours (Casey and Scott 2006; Loyau and Schmeller 2017; Rhead et al. 2015). Additionally, Arponen and Salomaa (2023) found that the conservation actions with the greatest potential to leverage systemic transformative change were awareness raising strategies that seek to change behaviour and education and training. However, it is important to recognise that awareness raising in their study did not refer to factual knowledge acquisition, but included a variety of outreach and communication methods that foster empowerment and enablement. It is also important to recognise that while some level of cognitive understanding of an environmental problem is necessary, it is not sufficient to achieve behaviour change in isolation (Dean et al. 2016). For example, it has been suggested that the role of cognitive and attitudinal factors in influencing behaviour may be supplemented by the presence of an affective factor – such as feeling connected to the natural environment (Carmi et al. 2015; Liu et al. 2020).

In recent years, there has been increasing evidence suggesting a moderate to strong relationship between feeling connected to nature and engaging in pro-environmental behaviours (Mackay and Schmitt 2019; Whitburn et al. 2020), including nature conservation, or pro-biodiversity, behaviours (Prévot et al. 2018; Richardson et al. 2020). Connection to nature has been conceptualised in multiple ways (Gosling and Williams 2010; Hatty et al. 2020; Mayer and Frantz 2004), but it ultimately refers to the way humans perceive, relate to, and interact with nature (Hatty et al. 2020), beyond what they know about or their attitudes towards the environment.

While many factors influence behaviour, there is an opportunity to determine the extent to which cognitive factors (such as awareness), attitudinal factors (such as concern), and affective factors (such as connection to nature) are associated with different types of nature conservation support – i.e. public-sphere and private-sphere behaviour and policy support. Such an understanding will be of particular value to conservation practitioners and policymakers. For example, cats pose a significant threat to wildlife in Australia (Kearney et al. 2019; Legge et al. 2020). In order to address this problem, there are many actions that could be taken – e.g. cat owners could keep their cats indoors (private-sphere behaviour); members of the community could petition for politicians to introduce wildlife management techniques for feral cats (public-sphere behaviour); or policymakers could introduce laws that prevent domestic cats from roaming the streets – which may receive support or opposition from the community (policy support). While each action addresses the same issue, different factors will affect whether people would be willing to support each of them (van Eeden et al. 2021). Understanding patterns of existing engagement and drivers of each action can inform design of engagement strategies that promote these actions.

For the current research, we posed the following research questions:

  • RQ1. To what extent are awareness, concern, and connection associated with public- and private-sphere nature conservation behaviours and policy support? and.

  • RQ2. How do these factors differ between public-sphere behaviour, private-sphere behaviour, and policy support?

Methods

Research design

Data for this research came from a larger study involving Australian adults (aged 18 years or over) who completed a 10-minute online survey in November and December 2022. Respondents were selected by a research company from their panel of members and emailed an invitation to complete the survey. Soft quotas were applied using population data from the Australia Bureau of Statistics 2021 Census for age group (18–24: 11%, 25–34: 18%, 35–44: 18%, 45–54: 16%, 55–64: 15%, 65–74: 12%, 75+: 10%), gender (female: 51%, male: 49%), and geographic location (Metro: 67%, Regional: 33%; NSW: 32%, Vic: 26%, Qld: 20%, SA: 7%, WA: 10%, Tas: 5%, NT: 5%, ACT: 5%) – for characteristics of the responding sample, see Table 1. Survey data were quality checked and cleaned prior to analysis – e.g. respondents who completed the survey too quickly or answered most Likert-scale questions with the same response were excluded. The final sample included 4,048 respondents. The research was approved by the lead author’s University Human Research Ethics Committee (project #36,212).

Table 1 Descriptive statistics – for scale variables, means and standard deviations (SD) are presented; for categorical variables, column proportions (%) for the responding sample and the population are presented
Full size table

The larger survey from which this study draws its data explored a range of factors to understand biodiversity awareness and concern among the general public in Australia. This included awareness of biodiversity issues, connection to nature, information sources for nature-related issues, perceptions about the state of the environment, concern about biodiversity-related issues, engagement in nature conservation behaviours, opinions about government performance and level of support or opposition for biodiversity policies (Borg et al. 2023). From the larger survey, only relevant variables (including socio-demographic characteristics) were included in the current study and are described further below. For an overview of the full survey structure and the specific questions included in the current study see Supplementary Material, Figure A1 and Table A1.

Dependent measures

Behaviour

Eight categories of nature conservation behaviours were included in the survey. Including a mixture of ‘public-sphere’ behaviours (e.g. ‘advocate for nature’, ‘be a supporter for nature’) and ‘private-sphere’ behaviours (e.g. ‘be a sustainable consumer’, ‘manage pets / garden for nature’). The items were adapted from research conducted by the Australian Conservation Foundation in 2022 by collapsing specific actions (i.e. ‘choose “environmentally friendly” products’; ‘eat less meat’) into high-level behavioural categories (i.e. ‘be a sustainable consumer’). Respondents indicated if they had ever done each type of behaviour, coded to 0 = No and 1 = Yes for analysis.

Policy support

Twelve items measured support or opposition for different types of conservation policies. This included items related to penalty actions (e.g. ‘Introduce tougher fines and stronger laws to stop illegal tree clearing and forest / habitat destruction’); restoration actions (e.g. ‘Restore nature in cities and towns’); and items related to regulation, markets, and consumption (e.g. ‘Develop “eco labels” that allow consumers to make informed choices about products’). These items were developed by the research team in consultation with biodiversity experts. They represent a combination of existing policies and hypothetical future policies which would benefit biodiversity in Australia. Respondents indicated their level of opposition or support on a scale from 1 (strongly oppose) to 5 (strongly support).

Independent measures

Connection

Six items were used to measure connection to nature, adapted from the CN-12, a multidimensional connection with nature instrument (Hatty et al. 2020). Example statements included, ‘Being in nature helps me deal with everyday stress’ and ‘I enjoy spending time in nature’. Respondents were asked to indicate their level of agreement or disagreement with each statement on a scale from 1 (strongly disagree) to 5 (strongly agree).

Concern

Eight statements described biodiversity-related issues which were rated on a scale from 1 (not at all concerned) to 5 (extremely concerned). These statements were adapted from previous research (Fielding et al. 2021) and extended to include issues identified by biodiversity experts. Response options were modified from ‘worried’ to ‘concerned’. Example statements included ‘The possible extinction of native plants and animals’ and ‘Overuse of native species (e.g. over-fishing)’.

Awareness

Finally, participants were asked to indicate whether seven statements about biodiversity-related issues were familiar to them, such as, ‘Most of Australia’s mammals and frogs are found nowhere else in the world’ and ‘Since European settlement, Australia has lost most of its forests’. Respondents indicated if they had heard of each statement previously. Statements were generated by discussion with biodiversity experts and covered issues related to endemicity, ecosystem services, and threats to biodiversity. Responses were coded to 0 = No/not sure and 1 = Yes. For the analyses, a composite score was generated by summing ‘yes’ responses (range: 0–7).

Socio-demographic characteristics

At the group level, there are well-established differences in environmental concern and behaviour between different population sub-groups – e.g. based on age, education, and gender (Gifford and Nilsson 2014). Therefore, several socio-demographic control variables were also included in the current study. These included age (in years), gender (0 = man, 1 = woman), residential area type (0 = metro/suburban, 1 = regional), country of birth (0 = Australia, 1 = other), language spoken at home (0 = English only, 1 = another language), employment status (0 = not working, 1 = working), highest level of education completed (0 = school only, 1 = tertiary qualification), and household income (0 = negative or nil income, 1=$1-$499 per week, 2=$500-$999 per week, 3=$1,000-$1,499 per week, 4=$1,500-$1,999 per week, 5=$2,000-$3,999 per week, 6=$4,000 or more per week).

Analyses

After data screening (see Supplementary Material Table A1), component structure was assessed for connection, concern, behaviour, and policy support. Principal component analyses with Direct Oblimin rotation was used, given that the objective was to explore component structure, with an expectation that the components would be correlated (Tabachnick and Fidell 2019). For all analyses the Kaiser-Meyer-Olkin Measure of Sampling Adequacy was 0.80 or higher and Bartlett’s Test of Sphericity were significant at p < .001.

For connection, a single-component solution accounted for 66.2% of the variance and all items loaded at 0.73 or above. For concern, a single-component solution accounted for 67.1% of the variance and all items loaded at 0.76 or above. Mean scores were calculated for connection (α = 0.90) and concern (α = 0.93) across all items and used in subsequent analyses (see Supplementary Material Table A2 for item factor loadings).

For behaviour, two components were identified with Eigenvalues greater than 1, which accounted for 57.6% of the variance. All items in the rotated solution loaded onto one of the two components above 0.50 with minimal cross-loading. The two components generally aligned with ‘public-sphere’ and ‘private-sphere’ behaviours, as anticipated (see Supplementary Material Table A3). Composite scores were calculated for public-sphere behaviour (α = 0.79) and private-sphere behaviour (α = 0.69) by summing the number of ‘yes’ responses (range: 0–4).

For policy support, two components were initially identified with Eigenvalues greater than 1. However, the results indicated that the vast majority of explained variance was attributed to one component, with little contribution from the second. An additional analysis was conducted forcing a single-component solution which accounted for 56.4% of the variance in policy support, with all items loading above 0.50. The mean score was calculated for policy support (α = 0.92) across all items and used in subsequent analyses (Supplementary Material Table A4 for item loadings for the two- and single-component solutions).

Next, three Hierarchical Multiple Regressions were conducted with public-sphere behaviour, private-sphere behaviour, and policy support respectively as the dependent variables. The key independent variables – connection, concern, and awareness – were added first, followed by the socio-demographic control variables. This allowed us to determine how much of the variance in behaviour (public- and private-sphere) and policy support was explained by the key independent variables alone and how much was explained while controlling for the influence of the socio-demographic control variables (Pallant 2013). Due to the danger of ‘too much power’ associated with large sample sizes, effect sizes and a p-value of < 0.01 were used to guide interpretation of the results (Tabachnick and Fidell 2019). For effect sizes, Cohen’s f2 was used for the model summaries and squared semi-partial correlation (sr2) was used for the independent variables – converted to percentages – to determine the unique contribution of each variable.

Results

The most common behaviour was spending time in nature (88% ‘yes’) and the least common behaviour was advocating for nature (36% ‘yes’; see Fig. 1). Respondents engaged in private-sphere behaviours more than public-sphere behaviours – with 70% or more indicating that they engaged in private-sphere behaviours compared to less than 60% for the public-sphere behaviours.

Fig. 1
figure 1

Frequency distribution – Behaviour, ever engaged in (dichotomous)

Full size image

For policy support, respondents generally supported all 12 policies, with at least 65% selecting ‘support’ or ‘strongly support’ for all items (see Fig. 2). Policies eliciting the highest level of support (i.e. over 80% of respondents selected ‘support’ or ‘strongly support’) involved penalty actions to protect the natural environment, such as increasing fines for importing or smuggling illegal wildlife and increasing fines for high polluters, and restoration actions involving restoring water to wetlands and rivers and restoring nature in cities and towns. Policies with the lowest level of support (70% or less selected ‘support’ or ‘strongly support’) were those that could affect individual behaviours, such as developing eco labels that allow consumers to make informed choices about products and introducing laws to prevent domestic cats roaming the streets.

Fig. 2
figure 2

Frequency distribution – Policy support (Likert)

Note: labels less than 5% have been removed to improve readability

Full size image

On average, respondents engaged in 1.74 (out of 4) public-sphere behaviours and 2.89 (out of 4) private-sphere behaviours and gave an overall policy support rating of 4.07 out of 5 (see Table 1). The dependent measures generally had small to medium correlations with the key independent measures (awareness, connection, and concern), with the largest correlation (r = .61) between concern and policy support (see Supplementary Material Table A5).

Regressions

The full model explained 22.9% of the variance in public-sphere behaviour (Table 2) – an increase from 13.3% after the socio-demographic variables were added. It also explained 30.2% of the variance in private-sphere behaviours – an increase from 23.9%. Finally, the full model explained 43.5% of the variance in policy support, which was almost unchanged from the key independent variable-only model (42.5%).

Table 2 Hierarchical Multiple Linear Regressions: Model summaries (amount of variance explained (R2) and effect size (f2))
Full size table

For public-sphere behaviour, the strongest contributions were from concern (sr2 = 3.7%) and age (younger, sr2 = 3.6%; Table 3). Awareness (sr2 = 1.9%) and connection (sr2 = 0.7%) also made unique contributions to the variance in public-sphere behaviour. In addition, employment (working), education (tertiary qualification), and gender (male) made statistically significant contributions to the model (p < .01); however, the amount of variance explained was quite small (sr2 ≤ 0.5%). For private-sphere behaviour, the greatest contributions were from connection (sr2 = 4.0%) and concern (sr2 = 3.7%), followed by awareness (sr2 = 2.5%), and age (younger, 1.7%). Gender (female) and employment (working) also made significant (p < .01) but minor (sr2 = 0.3%) contributions to the variance in private-sphere behaviour. Finally, for policy support, the greatest contribution was from concern (sr2 = 15.3%). Connection (sr2 = 2.5%), and awareness (sr2 = 2.5%) also made unique contributions to the variance in policy support. Additionally, age (older) and gender (female) made significant (p < .01) but quite small (sr2 < = 0.2%) contributions.

Table 3 Hierarchical Multiple Linear Regressions: Coefficients (standardised Beta (β) and effect size (squared semi-partial correlation, sr2)
Full size table

Discussion

The aims of this research were to explore community engagement in nature conservation behaviours and support for different types of nature conservation policies, and to assess the extent to which awareness of biodiversity issues, concern about biodiversity, and connection to nature were associated with engagement in conservation behaviours and policy support. We found that people reported engaging in private-sphere behaviours more than public-sphere behaviours, and that all nature conservation policies were supported by a clear majority – with the most support for restrictive policies that penalise practices that harm nature. When assessing the drivers of behaviour and support, the full models (including socio-demographic characteristics) explained around one-quarter of the variance in public-sphere behaviour, one-third of the variance in private-sphere behaviour, and two-fifths of the variance in policy support. For all models, the key independent variables – awareness, connection, and concern – were associated with all dependent variables, with some notable patterns of variation, which we discuss below. The socio-demographic characteristics made almost no contribution to policy support, and a sizeable contribution to public and private-sphere behaviour. Below we explore the theoretical and practical implications of these findings.

The findings provide support for the utility of awareness and concern (Casey and Scott 2006; Loyau and Schmeller 2017; Rhead et al. 2015) and for the importance of considering connection as an additional, unique factor (Prévot et al. 2018; Richardson et al. 2020). While these findings replicate work in similar contexts (e.g. Fielding et al. 2021; Sockhill et al. 2022), to our knowledge our study is among the first to explore a combination of public- and private-sphere behaviours alongside policy support in the context of protecting Australia’s biodiversity. In addition, while not a primary focus of our study, the findings regarding engagement in different types of behaviours and support for different policies warrants further discussion.

Given it is unfeasible, and ill advised, to attempt to change all actions that support biodiversity conservation simultaneously, behavioural scientists typically advise undertaking behaviour prioritisation to determine where resources and effort can be deployed most efficiently (Kneebone et al. 2017). Based on our findings, there is more opportunity to increase adoption of public-sphere behaviours (i.e. people engage in fewer private- compared to public-sphere behaviours), particularly those related to advocacy. Such actions are generally considered more ‘impactful’; however, they also have lower behavioural plasticity – i.e. perceived likelihood or ease of adoption (Selinske et al. 2020). For example, while voting for a political candidate based on their nature-related policies may seem relatively straightforward, voting preferences are highly complex and affected by many factors including political ideology, personality traits, and identity – including social and political identity (Kulachai et al. 2023). As such, more research is required to better understand the variety of contextual, social, and psychological drivers and barriers to specific public-sphere behaviours in order to develop effective interventions.

An alternative strategy may be to focus less on individual behaviour (which is difficult to change and influenced by a wider variety of factors not captured in this study), and instead focus on the more passive policy support (which was largely driven by concern). There are diverse policy instruments available to support conservation outcomes, including regulatory instruments, incentive schemes, and suasive approaches that build skills or motivation (Howlett 2019; Pacheco-Vega 2020). It has been suggested that there has been a shift away from restrictive policies, such as regulation, to less restrictive policies, such as incentive schemes, partly in response to neoliberal approaches (Bouwma et al. 2018; Dean et al. 2023).

Given this, it is interesting that our respondents expressed high levels of support for a range of policy initiatives, including restrictive approaches that involve penalties. This may be partly attributed to use of the word ‘illegal’ in describing such policies. For example, previous research has found that environmental policy support is associated with perceptions of effectiveness, intrusiveness, and fairness (Huber et al. 2020). Penalties for illegal activities may be perceived as fair, without being intrusive. In contrast, policies that could affect private-sphere behaviours may be perceived as intrusive, and not necessarily fair – such as those related to cat containment (van Eeden et al. 2021). The relationship between concern and policy support also raises questions about expectations of the ‘self’ vs. ‘others’ (including government policy) in protecting nature. For example, individuals may perceive their own role in addressing nature conservation as less significant (e.g. planting a native garden for wildlife), compared to the role of government policy, which could address larger-scale conservation issues (e.g. stop illegal tree clearing and forest / habitat destruction). Further research exploring the reasons for supporting different biodiversity policies is recommended to better understand these differences.

While the models explained some of the variance in behaviour and policy support, considerable variance was not accounted for, particularly for public-sphere behaviour. This is unsurprising, given that simple behavioural models do not account for the wide variety of contextual, psychological, and social drivers of behaviour (Darnton 2008; Davis et al. 2015). This may also explain the larger contribution of the key independent variables to the variance in policy support, which does not necessarily require an observable action. For example, according to the COM-B model of behaviour change (Michie et al. 2011), behaviour (B) is influenced by capability (C), opportunity (O), and motivation (O). While the current study included several motivation-related factors (awareness, concern, and connection), it did not include factors associated with capability (e.g. skills) and opportunity (e.g. physical environment). Simpler models of behaviour may be better able to explain less involved forms of nature conservation support, reflecting the more straightforward relationship between concern and policy support (i.e. if someone is concerned about something, they would likely support policies that allay their concerns). In contrast, active forms of engagement – particularly public activism – are likely to be influenced by a greater variety of factors, including values, cognitive biases, norms, and the physical environment (Gifford and Nilsson 2014).

Furthermore, while awareness and connection made unique contributions to all models, the size of their contribution varied. Consistent with previous research (Gkargkavouzi et al. 2019; Hatty et al. 2020; Nisbet et al. 2009; Whitburn et al. 2020), connection made the largest contribution to private-sphere behaviour but only a very small contribution to public-sphere behaviour. This suggests that feeling connected to nature is more likely to be associated with engagement in everyday lifestyle activities and less likely to be associated with engagement in public activist-type activities – which are typically driven more by values, ideology, and identity (Molinario et al. 2020; Scopelliti et al. 2018). This discrepancy highlights the importance of differentiating nature conservation behaviours in research and practice and recognising that different constructs – while still relevant – likely exert differing degrees of influence on different types of behaviours.

The greater contribution of the socio-demographic variables to the behaviour models (public-sphere: 9.6%, private-sphere: 6.3%) compared to policy support (1.0%) highlights other potential enabling factors (i.e. age, gender, employment, and education) which could also be leveraged for behaviour change. For example, our findings suggest that younger people are more likely to engage in more public-sphere behaviours and slightly more likely to engage in more private-sphere behaviours. In contrast, older people are (marginally) more likely to support biodiversity-related policies. This difference may reflect younger people’s general dismissal of the government (Martyn and Dimitra 2019), or generational differences in perceived responsibility towards environmental protection (Meis-Harris et al. 2019). In other words, young people’s general mistrust of the government may increase their sense of personal responsibility to take individual action (Church et al. 2023). The negative relationship between age and behaviour also differs from some previous research which finds that older people engage in more pro-environmental behaviours (Gifford and Nilsson 2014; Otto and Kaiser 2014). However, recent research suggests that this relationship is not necessarily linear, where younger generations may be becoming more environmentally aware (Fielding et al. 2023). It is important to highlight that these socio-demographic factors appear over and above any role of concern or awareness about biodiversity per se. Thus, researchers and practitioners should seek to identify the psychological mechanisms and structure factors which underpin the effects in future research.

Implications for practice

Findings from this study have important implications for practitioners and policymakers seeking to engage the wider community in nature conservation. First, when designing engagement strategies, practitioners and policymakers should be clear and specific in their desired outcome. If the intention is for individuals to adopt public-sphere nature conservation behaviours, then an appropriate strategy may involve fostering awareness of the biodiversity crisis, while also stimulating concern about biodiversity-related issues. If the desired outcome is for individuals to change their everyday private-sphere behaviours, then it would still be appropriate to foster awareness and concern, but engagement strategies will likely have more impact if they also foster a sense of connection to nature. In contrast, if the desired outcome is support for biodiversity-related public policy, a more appropriate approach may involve focusing specifically on increasing concern about the biodiversity crisis – while ensuring the policy framing emphasises effectiveness, intrusiveness, and fairness (Huber et al. 2020). It is also important to ensure messaging is balanced between messages that trigger negative sentiment, such as concern and fear, with those that trigger positive sentiment, such as empowerment and inspiration (de Lange et al. 2022).

Alternatively, if practitioners and policymakers have limited resources and are choosing which type of support to target, they may be better off focusing their efforts on encouraging policy support. This is not to say that nature conservation behaviours cannot or should not be changed among the general community. Indeed, behaviour change is fundamental to stop damaging the natural world (Amel et al. 2017). However, many practitioners have limited capacity to utilise a wider suite of behaviour change tools which will be required to address the multitude of social, psychological, and situational barriers to achieve large-scale change (Michie et al. 2011). Instead, practitioners with limited resources, who that rely on simpler forms of engagement, such as communication, may be better spent focusing their efforts on the lower hanging fruit of ‘passive’ policy support.

Limitations

While we were able to establish an association between concern, connection, and awareness and nature conservation behaviour and policy support among a large and nationally representative sample, the methodological limitations associated with this study should be discussed. First, the data were from a self-report survey which are subject to limitations such as randomised responding and extreme responding (Paulhus and Vazire 2007). To alleviate this issue, quality appraisal procedures were followed in an attempt to identify and exclude such respondents. Second, online panels typically use non-probability recruitment methods which can lead to sample bias (Pennay et al. 2018). Third, while sampling quotas were used to ensure that the sample broadly reflected the Australian population on certain characteristics, there are additional forms of response bias that may have impacted our data. For example, the survey introduction stated that the project was exploring biodiversity, nature and the natural environment which may have deterred less ‘pro-environmental’ respondents from participating (non-response bias). Additionally, respondents may have answered certain questions, such as those on behaviour, in line with the perceived ‘socially desirable’ responses (van de Wetering et al. 2022). Responses may also have been influenced by the order of the questions (priming) – e.g. awareness of biodiversity questions were asked before questions about concern.

Additionally, while our behavioural measures provide a useful estimate of engagement in public- and private-sphere nature conservation, they involved high-level categories of behaviours and self-reported data. The use of relatively broad categories of behaviours limits our understanding of the specific actions in which respondents engaged (e.g. ‘Advocate for nature’ included examples such as ‘joining a campaign’ or ‘contacting a member of parliament’). Such broad categories could also interpreted differently by respondents (e.g. ‘join a campaign’ could be interpreted as joining an environmental advocacy’s mailing list, or following a campaign on social media). The drivers and barriers of more specific behaviours would also vary considerably – as already evidenced by our findings for public- vs. private-sphere behaviour. Furthermore, while self-reported data is feasible for large-scale social surveys, it is difficult to control for known issues associated with self-reporting (van de Wetering et al. 2022). For example, although there is a strong relationship between self-reported behaviour and observed behaviour, it has been suggested that the functional difference between the two measures can be substantial (Kormos & Gifford 2014). This is partly due to known issues in self-reports surveys such as recall error and social desirability (Oerke and Bogner 2013; Paulhus and Vazire 2007). As such, future research which explores more specific nature conservation behaviours and that ideally employs more accurate measures of behavioural (such as observation) is recommended.

Finally, our study assessed a limited model of behaviour, focusing on awareness, concern and connection to nature. While our findings indicate that raising awareness or concern about biodiversity may have a role to play in promoting community mobilisation, it is important to note there are diverse factors that influence engagement in nature conservation and policy support. For example, research suggests that awareness of actions and the effectiveness of actions may be more important than knowledge of threats alone (Dean et al. 2018; Kaiser and Fuhrer 2003). Furthermore, social factors such as social connections and social capital may support an individual’s motivation or capacity to engage in conservation behaviours (Church et al. 2023; Dean et al. 2019). It is important to consider all such factors when engaging communities in conservation.

Conclusion

To protect and restore Australia’s unique and biodiverse natural environment, urgent systemic change is required, and community mobilisation is an essential component of such change. To enable community mobilisation, conservation practitioners and policymakers can encourage individuals to support nature conservation by engaging in a range of public- and private-sphere behaviours, or by supporting relevant local, State, and Federal government policies. Our findings suggest that fostering a sense of concern, and increasing awareness of biodiversity-related issues and connection to nature could help the community support nature conservation outcomes. Increasing concern for the biodiversity crisis among the wider population, in particular, could be an effective and efficient strategy to contribute to systemic change via support for nature conservation public policies – particularly those that involve punitive measures for offenders. At face value, public support for biodiversity policies could also be utilised by advocates to encourage public officials to anticipate community enthusiasm for stronger conservation measures.