Eliciting diverse perspectives to prioritize community actions for biodiversity conservation

Abstract

Communities have a strong role in protecting biodiversity. In addition to participation in restoration, a range of actions in the public or private sphere may support biodiversity. Despite this, there is a lack of clarity about what actions should be prioritized for behavior change campaigns. We developed and applied a method to prioritize community actions for biodiversity conservation that incorporates an expert‐based assessment of impact and a community‐informed measure of the likelihood of uptake. In stage 1, experts (n = 143) completed a survey that quantified the relative impact of actions based on best–worst scaling of perceived impact. In stage 2, surveyed community members (n = 3200) ranked the likelihood of adopting actions based on the ease or difficulty of performing each action, and the opportunity for change based on the proportion of respondents not yet engaging in each behavior. Experts gave the following actions the highest ranking for impact: voting for the environment (first), participating in restoration in ecological priority areas (second), and purchasing and protecting remnant bushland (third). When considering the disciplinary background and institutional background of experts, voting and participating in restoration activities remained in the upper ranked options. However, there was some divergence between these groups. For example, reducing beef consumption was ranked third by university‐based experts but ranked 28th by experts based in state government. Overall, community members ranked the following behaviors as most likely to be adopted: following quarantine laws (first), reducing plastic use (second), and managing pets (third). Top likelihood ranking of actions was minimally affected by community characteristics (nature relatedness, gender, location). Integrating these findings, the action ranked most favorably for impact, likelihood, and opportunity was participating in restoration. Choosing actions for behavior change campaigns requires consideration of the entire social–ecological system—from social factors that enable or constrain adoption to the ecological impact of actions across relevant social and ecological contexts.

INTRODUCTION

Prioritization is a common practice in conservation planning. Prioritization approaches may focus on ecological or other information, such as species richness, cost‐effectiveness, impact on species persistence, or maintenance of ecosystem services (Cullen, 2013; Wilson et al., 2009). Although there is recognition of the contribution of community actions to biodiversity conservation outcomes (Amel et al., 2017; Cinner, 2018; Nilsson et al., 2020; Reddy et al., 2017), relatively less attention is given to methods for prioritizing community actions and who is involved in the prioritization. We developed and applied a method to prioritize community actions for biodiversity conservation that incorporates an expert‐based assessment of impact and a community‐informed measure of the likelihood of uptake. We also explored whether the characteristics of experts or community members participating in the prioritization affected the findings.

Community action to support biodiversity

Conservation narratives have shifted to better incorporate relationships between people and nature (Mace, 2014). In parallel, there is increasing focus on expanding traditional conservation tools to also explore how we can work effectively with people and communities to support stronger conservation outcomes (Amel et al., 2017; Reddy et al., 2017). This opportunity is echoed in a range of conservation strategies. For example, the Convention on Biological Diversity incorporated social targets conserving nature (Convention on Biological Diversity, 2010). The recent update via the Kunming–Montreal Global Biodiversity Framework emphasizes the role of community actions, including community‐based natural resource management (Stephens, 2023).

Despite an accepted need to promote community actions, there is less clarity about which actions are most important. For many conservation challenges, there are typically multiple actions to consider. For example, a recent study on community action to support the management of wild dogs in peri‐urban regions of the City of Gold Coast (Australia) considered 13 actions within a prioritization framework (Please et al., 2018). When considering communities at a national level—as promoted by global targets—the broad nature of threats to biodiversity means there are hundreds of actions available. Typologies of conservation actions—often referred to as behaviors (Schultz, 2011)—consider 4 main types: land‐based stewardship actions (such as engaging in habitat restoration), civic actions (such as engaging in advocacy or donating to conservation organizations), lifestyle actions (such as reducing consumption of products that impact biodiversity), and social actions (such as encouraging others to adopt biodiversity actions) (Larson et al., 2015). Although it may be tempting to promote all potential actions that support biodiversity, behavior change interventions are most effective when a specific action is targeted (Schultz, 2011). Given this and limited resources (Wilson et al., 2009), there is a need to examine approaches to prioritize community actions for biodiversity at scale.

Prioritizing actions

Schultz (2011) argues that prioritizing actions should involve considering the impact of each action (in this case impact on biodiversity) and behavioral plasticity. Behavioral plasticity is the likelihood that a new action can be adopted by a reasonable proportion of the population (i.e., an action with high plasticity is relatively easy to perform and not yet performed by most people) (Schultz, 2011). Such an approach is increasingly recognized as useful (Boulet et al., 2023; Kneebone et al., 2017). A recent study has applied such an approach to prioritize community actions to protect biodiversity (Selinske et al., 2020; discussed below). We considered different approaches to quantifying impact and behavioral plasticity and examined the influence of sources of expertise to inform these metrics.

Approaches to assessing impact

With regard to assessing impact, in some situations (e.g., with energy use and emissions), there may be empirical evidence available to support quantifying impact (Dietz et al., 2009). However, such data are not always available for a diverse suite of actions (Kneebone et al., 2017), and issues, such as biodiversity, are complex, characterized by dispersed and diverse threats, impacts that are spatially dependent, and action and impact that are often distal in space or time (Selinske et al., 2020). In these circumstances, it is common to use expert elicitation. Many studies—across topics as diverse as water consumption, wild dog management, or biodiversity protection—have utilized expert elicitation surveys to rate the perceived impact of actions with Likert‐style scales from high to low impact (Kneebone et al., 2017; Please et al., 2018; Selinske et al., 2020). These studies typically focus on experts with a particular type of experience but do not routinely scrutinize how expert characteristics may influence prioritization outcomes.

Conservation science draws on a range of disciplinary perspectives. Although emerging from the disciplines of biology and ecology, conservation biology has been referred to as a synthetic multidisciplinary science that incorporates perspectives from disciplines such as environmental and natural resource management, environmental monitoring, biogeography, mathematics, and social sciences (e.g., economics, governance, and behavioral science) (Mace, 2014; Soulé, 1985). Conservation experts may also operate across diverse institutional settings, including universities, governments, nongovernment organizations, and land management groups. Disciplinary background or professional focus may shape perspectives of the social–ecological system (Biedenweg et al., 2020; Guerrero et al., 2021; Levy et al., 2018; Moon et al., 2019; Rothenberger et al., 2020). One factor that may mediate the effects of professional background on ranking impact may relate to the use of different information sources to guide decisions. University‐based experts are more likely to draw on scientific research, emphasizing the credibility and legitimacy of scientific knowledge, whereas practitioners draw more heavily on institutional reports, emphasizing salience and relevance to local settings (Archie et al., 2014; Cash et al., 2003; Pfaeffle et al., 2022; Virk et al., 2023). Diverse types of professionals also draw on personal experience and the experience of colleagues (Archie et al., 2014; Cook et al., 2010; Fabian et al., 2019; Fazey et al., 2006; Pullin et al., 2004; Virk et al., 2023; Young & Van Aarde, 2011). This raises the question of how different types or levels of professional experiences may shape ratings of impact in a prioritization assessment. We explored how expert characteristics, including disciplinary background and organization, influence conservation impact ratings.

Approaches to assessing behavioral plasticity

The most common metrics used to assess behavioral plasticity are the likelihood of an action being adopted and the opportunity for change, indicated by whether a reasonable proportion of the population are not yet performing the action (Schultz, 2011). Some studies quantify likelihood and opportunity metrics by asking experts to provide estimates (Selinske et al., 2020). However, professionals with expertise in assessing the impacts of action on biodiversity may not always have corresponding expertise in whether community members are willing or able to adopt specific behaviors. As such, directly surveying community members can provide more direct evidence of likelihood and opportunity.

Adoption of behaviors is shaped by an individual's capacity, motivation, and opportunity to perform them (Michie et al., 2011). Because generating motivation and opportunity is often the role of interventions, assessments of likelihood commonly focus on capacity. Some studies assess likelihood by quantifying capacity as the perceived difficulty or ease of performing a behavior (Kneebone et al., 2017). Although comparisons have been made between community members and water professionals (Kneebone et al., 2020), it is not clear how the characteristics of community groups themselves may influence ratings of likelihood and prioritization outcomes. For example, it is well established that social factors, such as nature relatedness and demographics, influence willingness to engage in nature protection behaviors (Church et al., 2023; Fielding et al., 2023; Sockhill et al., 2022). Although it would be expected that individuals with greater nature relatedness may perceive less difficulty involved in nature protection behaviors, it is not clear whether this would also result in a divergence of behavioral rankings between those with higher or lower nature relatedness. We extended previous prioritization studies (Selinske et al., 2020) by using community perceptions rather than experts to quantify impact and likelihood and to examine how community characteristics influence prioritization outcomes.

Prioritizing community actions

We examined a method to prioritize community actions to support biodiversity conservation across Australia. Australia provides an important setting to explore these issues. More than 1700 Australian species and species groups are at risk of extinction (Dielenberg et al., 2023). The Australian Biodiversity Strategy argues that all Australians must take responsibility for biodiversity conservation and has a target of increasing public participation in conservation (Natural Resource Management Ministerial Council, 2010). Although volunteering, citizen science, and land covenants are suggested as examples, no specific actions are prioritized. Similarly, Australia's 2021–2031 Threatened Species Strategy emphasizes the importance of community leadership and engagement (Australian Government, 2021). Suggested opportunities for community contribution include responsible pet ownership, planting wildlife gardens, and citizen science, but no specific recommendations or specific targets for actions are provided, and priority actions have not been explicitly identified. We aimed to address this gap. In addition, by examining how expert and community characteristics contribute to prioritization outcomes, we sought to generate insights for those applying such methods across diverse contexts. Within this context, we conducted surveys of experts and community members to examine: expert rankings of the impact of multiple biodiversity actions and the influence of expert characteristics on impact rankings and community ratings of likelihood (ease or difficulty of uptake) and opportunity (rates of existing uptake) to estimate behavioral plasticity and the influence of community characteristics on likelihood rankings.

METHODS

We used data from 2 Australian surveys. In the first, we examined conservation experts’ perception of the impact of community conservation actions. In the second, we recruited community members to estimate the likelihood and opportunity. Ethical approval of the study design was provided by the University of Queensland Human Research Ethics Committee for both the expert survey (approval 2019000252) and community survey (approval 2019001168).

Developing a list of actions

There are diverse lists of potential community actions for conservation (Appendix S1) (Borg et al., 2024; Fielding et al., 2023; Selinske et al., 2020). A literature review and conversations with conservation experts (senior researchers within a biodiversity conservation research center at a large Australian university) were conducted to generate an initial list of actions that have the potential to generate benefits for biodiversity. Based on behavioral typologies (Larson et al., 2015; Stern, 2000), we aimed to ensure that this list incorporated a range of behaviors, including land‐based stewardship behaviors (activities or events that seek to manage or restore habitat); lifestyle behaviors (actions taken within the household and everyday life, such as restraining pets or choosing consumer products with low environmental impact); civic behaviors (persuading governments or decision makers to strengthen conservation policies through actions such as signing petitions, voting, and writing letters); and social behaviors (encouraging others to support biodiversity conservation by adopting land‐based, lifestyle, or civic behaviors).

Our initial list had over 200 individual actions. We reduced the number of actions to ensure a manageable length for the expert and community surveys while providing adequate breadth. We ensured key threats (i.e., habitat loss and degradation, climate change, invasive species, pollution, and resource extraction [Natural Resource Management Ministerial Council, 2010]) were represented by at least one action. We also identified overlaps and potential groups of actions and selected key actions from the different behavioral types. For example, many different actions can contribute to climate mitigation (Wynes & Nicholas, 2017). Because the aim of this exercise was not to prioritize climate actions per se, we integrated climate mitigation actions into a single item. We also ensured that the activities included were not restricted to those who were landholders, but also included actions that most Australians had the opportunity to adopt. During this reduction process, we continued consultation with experts (senior researchers at The University of Queensland's Centre for Biodiversity and Conservation Science) to ensure actions of high‐potential impact were not excluded and our assumptions about grouping behaviors were reliable. This process resulted in a list of 28 behaviors (Appendix S2).

Expert recruitment

Diverse samples are thought to generate more accurate judgments (Sutherland & Burgman, 2015). Given this, we aimed to recruit a diverse sample of conservation experts that was not restricted by experience or qualifications. We aimed to recruit a sample sufficiently large to permit examining subgroups (minimum n = 100).

From May 2019 to October 2019, conservation experts with experience working in an Australian context were invited to participate in our survey via multiple strategies. A.D. shared invitations to participate with professional groups such as the Ecological Society of Australia, the Centre for Biodiversity and Conservation Science, and natural resource management organizations. Representatives from these groups then distributed this information throughout their networks (e.g., via newsletters). Participants who received these invitations were also invited to share the link within their networks. Targeting our recruitment strategy on distribution via these groups (and individuals within them) means that we were not able to calculate the total number of professionals invited or a specific response rate. Invitations to participate stated that it was focused on actions that people can do to support “biodiversity conservation in Australia.”

Ranking of actions in the expert survey

Respondents ranked impact through best–worst scaling, a choice‐based prioritization method (full survey in Appendix S3). Many expert elicitation processes are designed for questions that involve estimating numerical quantities or probabilities (Hemming et al., 2018; Speirs‐Bridge et al., 2010). These methods have been applied to behavioral prioritization by inviting experts to rate perceived impact on a scale of high to low impact (Linklater et al., 2019; Please et al., 2018; Selinske et al., 2020). However, a lack of precise units for impact can create ambiguity in interpreting terms, which may contribute to ambiguity in responses (Carey & Burgman, 2008; Hardy & Ford, 2014). Another challenge with using rating scales for prioritization is when respondents rate items consistently high or low with little distinction between options. Choice‐based approaches can overcome these limitations when the goal is prioritization rather than providing numerical estimates (Louviere et al., 2013). The final list of actions was presented according to a balanced incomplete block design identified by Cochrane and Cox (1992) (plan 11.39). Participants were presented with 36 blocks (i.e., sets of actions); each block contained 7 actions. Over the survey, each action appeared 9 times (r = 9), and each possible pair of actions co‐occurred twice (λ = 2). For each block of 7 actions, participants were asked “If everyone in Australia adopted just one of the following behaviors, which would have the strongest direct benefit for biodiversity?” After selecting the highest impact behavior, they were then asked to select the behavior with the least direct benefit. Presentation of both block order and order of actions within each block was randomized.

Expert characteristics

To allow us to assess how certain expert characteristics influence impact ratings, we asked experts to provide information on their organization, discipline, experience, subjective confidence, and demographics. For their organizations, response options included university, state government, federal government, local government, natural resource management organization, scientific organization, nongovernment organization (NGO), environmental consultancy, and open‐ended option (multiple responses allowed). For analysis, we created 3 groups: state government, universities, and others (NGOs, local government, consulting, other, and no involvement with state government or universities). Although some respondents classed as state government or universities also had other affiliations (e.g., NGOs), any respondents involved with both universities and state government were not included in this comparison. Experts within state governments were not combined with other governments because state governments have the primary responsibility for environmental protection in Australia (McGrath, 2012).

For discipline, response options included environmental science, environmental management, ecology, biology, zoology, botany, earth sciences, policy and governance, economics, and open‐ended option (multiple responses allowed). For descriptive analyses, we created 3 groups: environmental science and management; ecology and biology, zoology, or botany; and other. Experts whose disciplinary focus cut across these groups were excluded from descriptive analyses.

For experience, participants were asked to indicate how many years of experience in the areas of biodiversity conservation and management (to the nearest year) they had.

For subjective confidence, participants were asked to rate their confidence. We included this item so as to consider the potential influence of overconfidence (Speirs‐Bridge et al., 2010). Rather than ask experts to estimate confidence for each rating (as typically done with interval estimates), we used a single question to rate their overall confidence in responding to the best–worst scaling questions (1 = extremely confident; 2 = moderately confident; 3 = slightly confident; 4 = slightly unconfident; 5 = moderately unconfident; and 6 = extremely unconfident). We collected the following demographic data: age, gender, and educational attainment.

Community participants and procedure

Participants were recruited via a social research company online panel (PureProfile). We aimed to recruit 3200 adults residing in the Australian state of New South Wales, representative of age and gender. New South Wales is the most populous state in Australia (over 8 million); two thirds of the population live in the Greater Sydney region (New South Wales Government, 2023). Panel members were invited to participate and offered standard compensation (points and entry into prize draw). Rather than focusing on response rates, we set a recruitment target (n = 3200). The company then invited panel members to participate, which continued until the target was met. Participants provided online consent. The 15‐min, online survey was administered during October 2019 (Appendix S4).

Survey questions to quantify the likelihood and opportunity

The specific wording of each action was modified to suit the community audience (Appendix S2). For actions that were dependent on having a private garden or pet (e.g., keep pets inside, choose native plants for garden), a “not applicable” option was available. To reduce the survey length, some similar items were combined or removed. For example, purchasing land for protection versus purchasing land for restoration would have similar levels of difficulty, so these were combined into one item. Some experts raised concerns about the suitability of the broader public trapping invasive species. As such, this item was removed from the community survey. This left a total of 26 behaviors in the community survey.

For each action, we quantified likelihood by asking participants to rate the ease or difficulty of performing each action (10‐point scale from 1 = very difficult to 10 = very easy) and opportunity for change by asking participants whether they had performed the action never, once, a few times, or many times.

Sociodemographic data were collected, including age and gender. Postcode was used to identify the degree of remoteness based on the Australian Standard Geographical Classification—Remoteness Area (ASGC‐RA) criteria (ABS, 2002). This classifies regions based on distance from major urban centers: major cities, inner regional, outer regional, remote, and very remote. For comparison, we split the sample into residents of major cities (hereafter urban residents) versus residents from other areas. Nature relatedness was quantified using the mean of 3 items from the brief nature relatedness scale (NR‐6) (Nisbet & Zelenski, 2013), as used by Massingham et al. (2019) (response options from 1 = strongly disagree to 6 = strongly agree). To identify response patterns that indicate a lack of attention, one item was reworded to be reverse scored (“I never think about how my actions affect the environment.”).

Data analyses

To generate impact scores for each action, we followed the recommendations of Louviere and colleagues (2013). For each action, a count score was generated for both the number of times it was selected as the highest impact and the number of times selected as the lowest impact. The best–worst scaling score for each action was then calculated by subtracting the lowest impact count from the highest impact count with IBM SPSS Statistics 27. The final best–worst scaling score was used as an indicator of impact—the higher the score, the higher the impact (Louviere et al., 2013). In our study, possible scores ranged from 9 to −9. To rank impact, mean best–worst scaling scores were generated for each action and then ranked. Differences between subgroups are presented as descriptive data showing ranking across subgroups.

To calculate the likelihood, the mean likelihood of each action was calculated using IBM SPSS Statistics 27. Mean likelihood scores were then ranked for the whole group and for each subgroup of interest. Opportunity was calculated as the percentage of the sample that reported they had never done this behavior. Behaviors where at least 50% of respondents indicated that they have never performed the behavior were considered high opportunity because there is more room for change in these behaviors.

The prioritization matrix was generated by plotting impact scores against likelihood with plotting markers to distinguish high‐ and low‐opportunity actions. However, matrix‐style solutions may be influenced by methodological choices. For example, placing crosshairs at statistical midpoints versus scale midpoints may influence which behaviors fall into prioritization segments (Sever, 2015). For comparison, we generated 2 matrix figures ranked scores with the crosshairs placed at median (i.e., midpoint of the rankings) and raw scores with crosshairs placed at scale midpoints (i.e., 0 for impact and 5.5 for likelihood).

RESULTS

Expert characteristics

Overall, 143 experts completed the survey (mean [SD] age: 46.7 years [12.8]; females: 53.1% of 143 respondents) (Table 1). Almost two thirds (62.7%) had a postgraduate qualification (38% PhD, 19% master's degree, 6% graduate certificate). On average, people reported 17.3 years of experience in conservation management (range: 1–50, median: 16 years); 71% had more than 10 years’ experience. Participants represented a diversity of disciplines, including environmental science and management (55.9%, n = 80), ecology (45.5%, n = 65), and biology, zoology, or botany (27.3%, n = 39) (correlations in Appendix S5).

TABLE 1.

Profile of experts in biodiversity conservation completing expert elicitation survey in which they ranked community actions for perceived impact.

Expert characteristics	Sample
	n	%
Gender
Female	76	53.1
Male	65	45.5
Prefer not to say	2	1.4
Age	Mean: 47.0 years	Range: 18–75
Education and qualifications
PhD	55	37.9
Masters	27	18.6
Bachelor degree	53	36.5
Experience in biodiversity conservation and management
>20 years	42	29.4
11–20 years	49	34.3
1–10 years	52	36.4
Disciplinary background
Environmental science or management	80	55.9
Ecology	65	45.5
Biology, zoology, or botany	39	27.3
Policy and governance	23	16.1

Expert ranking impact

The highest ranking action was “vote for strong conservation policies,” followed by “participating in bush restoration in an ecological priority area” and “purchasing and protecting a small parcel of land.” Apart from voting, most of the highest impact actions were land‐based stewardship actions. The lifestyle action with the greatest impact was “make serious impact to reduce carbon footprint” (Figure 1). The actions with the lowest impact ranking were reducing the use of household chemicals, participating in beach cleanups, and choosing local plants for the garden (Figure 1).

FIGURE 1.

Conservation expert perceptions about impact of conservation actions in rank order from high to low based on mean scores generated by best–worse scaling (error bars, standard deviation).

Influence of expert characteristics on impact ratings

Of the 10 highest ranked behaviors overall, 7 remained ranked in the top 10 across groups with different organizational backgrounds (Table 2). However, there were some notable differences. The largest difference within all expert comparisons related to beef consumption. Respondents based in universities were more likely to rate minimizing beef consumption highly (third), compared to the ranking allocated by state government respondents (28th) and experts based in other organizations (24th). Similarly, university‐based respondents were also more likely to rate advocacy‐related activities as higher impact than those from other organizations. For example, writing a letter to your local member of parliament was ranked ninth by university‐based experts and 21st by state government employees, and donating money to an advocacy organization was ranked 11th by university professionals and 25th by state government experts.

TABLE 2.

Rank of expert ratings of impact of community actions on biodiversity from high (1) to low (28) impact for the full sample and subgroups based on expert experience, confidence, organization, and discipline.

		Years experience			Confidence			Organization			Discipline
Action	Full sample	1–10	11–20	>20	None	Slight	Moderate to extremely	University	State government	Other	Environmental science or management	Ecology or biology	Other
Vote for strong conservation policies	1	1	1	2	1	1	1	1	1	1	1	1	1
Bush restoration of ecological priority area	2	5	3	1	4	2	2	5	2	2	3	3	4
Purchase and protect remnant bushland	3	4	2	3	2	3	3	2	5	3	5	2	3
Purchase and restore degraded land	4	3	4	8	3	6	5	4	8	4	6	4	7
Reduce carbon footprint	5	2	7	7	5	4	6	6	3	7	2	9	2
Participate in bush restoration in local area	6	6	5	4	6	5	4	8	4	5	7	5	8
Participate in riparian protection activities	7	10	6	5	8	7	7	15	6	6	4	6	9
Do not allow pets to escape into the wild	8	7	8	6	7	8	9	10	9	8	8	7	5
Encourage others to adopt conservation action	9	8	9	10	15	9	8	16	7	9	11	16	6
Sign up to covenant on private land	10	12	14	11	9	12	14	12	14	11	13	8	15
Green infrastructure in home	11	11	12	14	13	16	10	17	13	10	10	14	14
Trap invasive mammals	12	18	11	12	19	11	11	21	12	13	16	11	11
Make serious effort to reduce plastic use	13	9	16	17	11	15	13	13	11	18	9	23	10
Follow quarantine laws	14	22	10	9	14	13	12	23	10	21	12	19	13
Donate to national land trust organization	15	16	15	13	10	14	15	7	15	12	15	13	20
Keep pets restrained at night	16	15	17	15	18	10	18	19	16	15	14	18	12
Let local member of parliament know that you support biodiversity protection	17	17	13	22	17	17	16	9	21	25	25	10	19
Sustainable seafood guide	18	21	18	16	12	19	22	14	18	22	17	17	21
Artificial habitat for wildlife in yard	19	14	20	20	21	18	17	18	19	14	18	12	16
Insect and bird attracting plants for garden	20	20	22	21	23	20	21	20	17	16	20	20	17
Minimize beef consumption	21	13	21	28	16	22	26	3	28	24	19	15	28
Donate to conservation advocacy group	22	19	25	23	20	23	24	11	25	23	22	21	23
Monitor species in terrestrial areas	23	24	19	18	26	21	19	28	20	17	21	22	22
Monitor health of waterways and coasts	24	26	24	19	25	24	20	26	22	19	23	24	18
Purchase only forest stewardship council certified wood and paper	25	23	23	24	22	25	25	22	23	26	26	26	27
Choose local plants for garden	26	25	26	25	27	26	23	24	26	20	28	25	25
Participate in beach cleanup events	27	27	27	27	24	27	27	27	24	28	27	27	26
Minimize use of household chemicals	28	28	28	26	28	28	28	25	27	27	24	28	24

Note: Numbers reflect the order in which actions were ranked, and colors show variation from high (blue) to low (orange) conservation impact.

Of the top 10 ranked behaviors overall, 8 of these remained ranked in the top 10 across groups with different disciplinary backgrounds (Table 2). Those with a disciplinary background in ecology or biology rated advocacy actions as higher impact, where donating money to an advocacy organization was ranked 10th for experts with training in biology or ecology but 25th by experts with training in environmental science and management.

To assess the influence of experience on rankings, experts were grouped into terciles: 1–10 years’ experience (36.4%, n = 52), 11–20 years’ experience (34.3%, n = 49), and more than 20 years’ experience (29.4%, n = 42). Looking across these terciles, the top 9 behaviors overall fell into the top 10 among all experience subgroups (Table 2). The most experienced subgroup (>20 years) was the only group to not rank voting as number one, instead selecting participating in bush restoration in an ecological priority area. The largest difference in rankings related to reducing beef consumption, which was ranked 13th for those with 1–10 years’ experience but 21st and 28th for those with more experience.

The mean (SD) score for subjective confidence was 2.99 (1.19). For comparison, we grouped experts into 3 groups: not confident (confidence of 4, 5, or 6; 30.1%, n = 43), slightly confident (confidence of 3; 26.6%, n = 38), and moderately to extremely confident (confidence of 1 or 2; 43.4%, n = 62). Only 6 participants (4.2%) rated themselves as extremely confident. Confidence had minimal impact on ratings for the top 10 behaviors (Table 2). Specifically, the top 9 behaviors were in the top 10 among groups with all levels of confidence. The largest difference in rankings related to dietary choices, where not confident respondents rated reducing beef consumption and using a sustainable seafood guide as higher impact (by up to 10 ranked places) than more confident respondents.

Community participant characteristics

There were 3220 completed community member surveys. A range of age groups were represented (13%: 18–24 years; 18%: 25–34 years; 17%: 35–44 years; 18%: 45–54 years; 15%: 55–64 years; 18%: 65 years and over), and half were male (49.8%). Just over one third had completed tertiary education (39.6%). Just over two thirds reported living in a major city (69.3%). Mean nature‐relatedness scores were 4.48 (SD: 1.04, range: 1–6).

Community ranking likelihood

The behaviors rated as easiest to adopt were following quarantine laws, making a serious effort to reduce plastic use, and keeping pets restrained at night (Figure 2). The behaviors rated as the most difficult were purchasing and protecting or restoring a parcel of land, donating to a national land trust organization (“an organization that buys land to protect and restore nature”), and installing green infrastructure in the home (“reducing hard surfaces and installing special gardens to reduce stormwater pollution”).

FIGURE 2.

Ranking of likelihood of adoption stewardship behaviors by community members (n = 3200) based on mean scores for ease or difficulty of adoption.

Influence of community characteristics on likelihood rankings

Characteristics of community respondents (nature relatedness, gender, and remoteness) had minimal impact on upper rankings and lower rankings: all of the top 4 ranked actions remained in the top 4 places for all subgroups assessed (Appendix S6). Similarly, the bottom ranked 2 actions were ranked the lowest by all subgroups.

However, there were a number of notable differences. The largest difference was observed between urban participants and those who lived in regional or rural areas. Specifically, reducing carbon footprint by driving less was ranked as more difficult (by 10 places) by regional or rural participants. However, regional participants ranked artificial habitat for wildlife in yard as easier (by 8 places). When looking at gender, the strongest ranked differences involved minimizing beef consumption and driving less. Specifically, women rated minimizing beef consumption as easier than men (by 7 ranked places), whereas men rated driving less as easier than women (by 7 ranked places).

Opportunity

Actions with the highest opportunity (i.e., those with the lowest rates of uptake) were purchasing and protecting land (79%) and donating to national land trust organizations (70%) (Appendix S7). Behaviors with the lowest opportunity were keeping pets restrained at night (4%) and reducing plastic use (5%). Certain actions were not applicable to some respondents, including those who do not have pets (31.4%), do not have a yard for habitat (20.1%), or do not eat seafood (15.1%).

Impact–likelihood matrix

The impact–likelihood matrix showed actions ranked highly for both impact and likelihood, with the highest priority for action located in the top right quadrant, including engaging in restoration activities, voting for a party that protects biodiversity, ensuring pets do not escape into the wild, encouraging others to take action to protect biodiversity, protecting one's own land, and following quarantine laws. When examining opportunity (i.e., which actions have not yet been adopted by the majority of the population), the only actions approaching the upper‐right quadrant related to engaging in restoration practices (in ecological priority areas [Figure 3a,b], local areas, or riparian areas [Figure 3b]).

FIGURE 3.

Impact–likelihood matrix of potential priority actions for community stewardship showing: (a) ranked impact and likelihood scores (dotted cross at the midpoint of rankings) and (b) mean impact and likelihood values (dotted cross placed at scale midpoints) (blue circles, actions with high opportunity [i.e., at least 50% of community respondents report having never performed this behavior]).

DISCUSSION

Community involvement in biodiversity conservation contributes to international and local conservation goals. Nonetheless, methods to prioritize community actions for biodiversity conservation have received little attention. We integrated expert and community perspectives on diverse community actions and explored the degree to which this method is influenced by the characteristics of these experts and community members. The types of behaviors that could be considered for prioritization were diverse, including participating in restoration activities, behaviors related to following laws about quarantine or pets, voting, and encouraging others to contribute to biodiversity protection. Overall, the characteristics of experts and community members had minimal influence on rankings, but important distinctions have implications for motivating communities across large scales to adopt conservation actions.

Legislative and policy change as the most important pathway to impact

Voting was the top‐ranked action for impact. This prioritization of government‐level change suggests that experts view legislative and policy change as more important for conservation than other types of community actions. At the time of the study, there had been much debate about Australia's primary environmental protection laws—The Environment Protection and Biodiversity Conservation (EPBC) Act—and the need to strengthen these laws (Dielenberg et al., 2023; Driscoll et al., 2019; Lindenmayer & Burnett, 2022; Ward et al., 2019). A strong element of these debates argued that changing government would generate political impetus for stronger environmental protection. It is possible these debates strengthened expert perceptions about the importance and potential impact of voting patterns. As discussed by Griffin and Tversky (1992), this may be an example of people being more sensitive to the strength of evidence (described as the extremeness of an evidence) rather than the weight of evidence (the credibility of the evidence). It is also possible that experts view voting as a simple action and pathway to change. However, we caution against this possible assumption and suggest that encouraging people to vote for conservation is not as simple as it may appear. First, voting for conservation is difficult to operationalize, and many individuals may vote trusting that their preferred party has an environmental policy but not scrutinize the quality of this policy or whether investments are made to implement and enforce the policy. This aligns with our ratings of opportunity, where only 37% of respondents reported that they never did this. If the item were rephrased to “vote for party X,” then the difficulty of engaging in this may have been rated very differently. Second, from a behavioral perspective, voting is a complex behavior and is influenced by multiple factors including partisanship, issues raised during electoral campaigns, and leadership appraisal (Bean & McAllister, 2009; Carson et al., 2020; Colvin & Jotzo, 2021; Rekker, 2022). Many organizations invest substantial amounts to influence voting patterns, making this a challenging behavior to shift. Also, given that voting is an infrequent behavior, it creates a limited window for impact.

Influence of type of experience versus years of experience

Although expert rankings were consistent for voting and land‐based stewardship, there were notable divergences for other types of actions. The largest points of difference related to reducing beef consumption, which was ranked highly (third) by university‐based experts but last (28th) by experts in state government. This raises the question of how different types of experts draw on experience or information to make their judgments. Although all professionals draw on personal experience for decision‐making (Fabian et al., 2019), academics are more likely to draw on scientific research, whereas managers and government staff are more likely to rely on experience and institutional reports with a localized focus (Cook et al., 2010; Virk et al., 2023; Young & Van Aarde, 2011). Within this context, university researchers may draw on scientific literature about grazing impacts on habitat loss, pollution, and carbon emissions (Godfray et al., 2018; McAlpine et al., 2009; Seabrook et al., 2006; Skidmore et al., 2021). However, because the impacts of beef consumption on biodiversity are more spatially and temporally distal, the capacity to directly observe the benefits of dietary changes is more challenging. As a result, experts drawing on personal experience or localized reports may consider these actions as having limited impact. This divergence was not detected for in situ stewardship actions (e.g., participating in tree planting), which may generate outcomes that are directly observable within short time frames (Ryan et al., 2001). The other set of behaviors exhibiting differences between experts were advocacy‐related actions, which were ranked more highly by university‐based experts than government experts. When considering the impact of these actions, government‐based experts may draw on experiences of decision‐making processes, their understanding of what shapes policy development and political decisions, and the limitations of advocacy actions. We observed smaller differences across subgroups related to years of experience and subjective confidence. There is much research on confidence and expert ratings (Hemming et al., 2018, 2020; Speirs‐Bridge et al., 2010), but most of this focuses on quantifying confidence in interval judgments, as opposed to the choice‐based parameters used in this study. Overall, these findings highlight the influence of the type of professional experience, rather than the quantity of experience, and the importance of developing integrated knowledge systems for conservation decisions (Cash et al., 2003).

Considering constraints to action

Much of the discourse about promoting community conservation actions emphasizes motivation stemming from connection to nature (Baird et al., 2022; Clayton et al., 2017; Soga et al., 2016; Whitburn et al., 2019). However, the capacity and opportunity to perform the action are also necessary, especially for more difficult actions (Attari et al., 2011; Michie et al., 2011). Three types of constraints to action have been theorized: intrapersonal constraints relate to cognitive or emotional perspectives about action suitability; interpersonal constraints relate to the influence of others; and structural constraints involve contextual barriers such as financial costs or physical opportunity to act (Crawford et al., 1991; Gage & Thapa, 2011; Godbey et al., 2010; Moghimehfar & Halpenny, 2016). The difference in rankings for driving less between urban and regional respondents is likely to reflect a structural constraint, where individuals in regional areas have greater car dependency and limited access to alternatives (Carroll et al., 2021). Gender differences in rankings for driving less may reflect a mix of structural, intrapersonal, and interpersonal barriers, where women may have greater parenting responsibilities and greater safety concerns about alternatives such as public transport or cycling (Carroll et al., 2020; Craig & van Tienoven, 2019; Kawgan‐Kagan, 2020; Ouali et al., 2020). The gender difference in rankings of reducing beef consumption is likely to reflect an intrapersonal barrier for men. Men have a stronger attachment to eating meat and are more reluctant to consider vegetarian dietary options (Graça et al., 2015; Rosenfeld, 2020; Rothgerber, 2013). Overall, these differences reflect the influence of diverse types of barriers to conservation action and that experiences of constraints may vary across social groups. The presence of barriers to many higher impact actions raises the question of potential trade‐offs. Should conservation programs focus on strengthening the uptake of easy actions or target a higher impact action for change (Attari et al., 2011)? In general, behavior change programs recommend a multistage approach, including (but not limited to) choosing a behavior that has a reasonable impact on the desired outcome; identifying barriers to engaging in this behavior, especially structural barriers; and developing a program that decreases the barriers to action (such as strategies for making the action easier or providing some form of incentive) (Schultz, 2014). As suggested by Attari et al. (2011), we encourage conservation programs to extend beyond the low‐hanging fruit and consider meaningful ways to strengthen motivation, capacity, and opportunities to engage in higher impact actions.

Strengthening participation in habitat restoration

The only actions that involved high impact, likelihood, and opportunity were engaging in land‐based activities such as participating in restoration and land management. Habitat loss is a key threatening process driving the loss of biodiversity both in Australia and globally (Brooks et al., 2002; Legge et al., 2023; Ward et al., 2019). The United Nations Decade on Ecosystem Restoration has positioned restoration as a vital activity to tackle biodiversity loss (Aronson et al., 2020). Parallel to this, there is an increasing emphasis on the social dimensions of restoration. It is suggested that engaging local communities in restoration activities, from planning to implementation, can ensure that restoration initiatives deliver positive outcomes for social and ecological systems (Tedesco et al., 2023). In Australia, diverse programs enable participation in restoration, including national programs that support networks of landholders or volunteers (Cary & Webb, 2001) and community organizations focusing on local environments (Gooch & Warburton, 2009). It is important to recognize that increasing the size of volunteer cohorts requires additional investment in training and management. Volunteering is becoming more episodic, with overall rates of volunteering on the decline (Brudney & Meijs, 2009; Pagès et al., 2018; Snyder & Omoto, 2008). Given this, it has been suggested that ensuring suitable management and care of existing volunteers is just as important as the recruitment of new volunteers (Brudney & Meijs, 2009).

Limitations

Our study has a number of limitations. One challenge was the extensive number of potential actions to protect biodiversity. Although it was necessary to reduce our list to a practical size, it is possible that other potentially important actions were excluded. Our actions were framed to specifically focus on biodiversity, but a range of other broader actions may also contribute to positive outcomes for biodiversity. For example, engaging in civic actions to reduce corruption and promote transparency in government decision‐making is likely to generate a range of societal benefits, including benefits for biodiversity (Driscoll et al., 2018). With regard to ranking actions based on impact, there are additional contextual factors that warrant further exploration, such as the number of people performing the behavior. Many actions, such as reducing emissions, have the greatest effect when a greater proportion of the population engages in them. In contrast, actions such as participating in restoration do not require such high participation rates to generate benefits. Another factor that we were not able to consider is who is performing the behavior and the context in which it is performed. Actions such as restoration may benefit from promoting participation in individuals with certain skills, whereas other actions such as civic engagement may be more powerful when politically influential groups are more active. We looked at actions applicable across a national scale, which meant that descriptors of actions were deliberately broad. Applying these findings in a specific context may require tailoring to the local situation (e.g., “plant local species” may become “plant species X”). Although we focused on a national scale, we recognize that the experience of experts may encompass different scales, from the highly localized scale to international scales. Our questions asked participants to focus on Australia, but it is possible that other types of expert experiences also shaped their perceptions of impact. One important factor we did not consider in our prioritization was the programmatic costs of promoting and enabling certain actions (Selinske et al., 2022). Another limitation relates to the use of self‐reported scales for assessing opportunity and ease or difficulty of performing the action. Although research suggests moderate relationships between self‐reported behavior and actual behavior (Kormos & Gifford, 2014), it is possible that social desirability bias or salience of positive examples of action (rather than inaction) results in overestimates of frequency. With regard to assessing estimating likelihood, it is possible that participant ratings of ease or difficulty were influenced by factors such as experience with the action or psychological factors, such as hope or perceptions about personal control (Cleveland et al., 2020; Dean & Wilson, 2023). Lack of experience may also shape perceptions. For example, simplified wording to describe private land covenants (e.g., “make a commitment to protect wildlife on your land”) may have resulted in high perceptions of ease.

Globally, there is increasing recognition of the importance of working with people and communities to engage them as active contributors to biodiversity protection. While acknowledging that these global strategies target broad goals, we argue that it is vital to provide specific calls to action and guidance on selecting these. Without guidance, the ambiguity of recommendations to protect nature may result in a situation where individuals select the most familiar or easy actions that may generate minimal ecological change. Choosing actions for behavior change and engagement campaigns requires consideration of the entire social–ecological system—from social factors that enable or constrain adoption to the ecological impact of actions across relevant social and ecological contexts. Our findings demonstrate how insights from experts and communities can identify priority targets for engaging communities in conservation.

Supporting information

Supplementary Materials

Dean, A. J. , Fielding, K. S. , Smith, L. D. G. , Church, E. K. , & Wilson, K. A. (2025). Eliciting diverse perspectives to prioritize community actions for biodiversity conservation. Conservation Biology, 39, e14372. 10.1111/cobi.14372