Introduction

An increasingly urban human population, anticipated to reach 6.6 billion by 2050, poses significant environmental challenges1, including profound concerns around environmental (in)justice2. Rapid urban growth has contributed to the loss of plant and animal species, alongside their critical habitats3,4, and underscored the need to recognize and act upon these losses5,6. Research has consequently moved towards relational frameworks that not only aim to understand the drivers of people’s connections with nature7 but also to confront structural inequities that limit access to nature among low-income and racially minoritized communities8. Simple, everyday experiences that readily occur throughout urban landscapes can be an effective basis for strengthening environmental stewardship9 and reconceptualizing human-nature relationships through a lens of multispecies justice10. However, urban contexts—especially those replacing temperate prairie and other grassland ecosystems—are frequently undervalued11,12, despite their potential to offer co-benefits that foster connections to nature among diverse audiences13. These settings carry great potential to shift the goals of nature conservation beyond a traditional paradigm of nature as distinct from humans toward an inclusive framework of “people and nature” founded on interdisciplinarity, resilience, and engagement of diverse audiences14,15,16. This relational turn in scientific approaches also integrates most people within broader environments, as opposed to keeping them as separate entities, to build more just and equitable pathways for addressing sustainability challenges17,18,19.

Birds are an ideal taxon to understand the dynamic intersections between people and nature that occur in urban and suburban landscapes20,21,22. People regularly interact with and observe birds, and their resulting experiences may range from positive (e.g., feeling more connected to wildlife) to negative (e.g., conflict due to the damage of personal property). Positive interactions with bird communities, in particular, have the potential to increase a person’s likelihood to engage in wildlife-friendly landscaping practices and other pro-environmental behaviors (PEBs) such as voting in favor of conservation initiatives7,23,24. These outcomes offer support for adaptive and transformative management strategies that seek to facilitate more sustainable and just human-nature relationships. Urban sustainability research has been largely conducted in more developed, educated, industrialized, rich, and democratic countries and cultures25, yet even within these regions, there are well-documented disparities in urban centers such as uneven access to vegetation and amenities among disadvantaged populations26,27. As such, we face a risk of developing inequitable conservation strategies that overlook marginalized voices and misrepresent complex social-ecological dynamics. A more pluralistic representation of how diverse audiences interact with their local environments is urgently needed to more effectively and equitably address environmental problems such as landscape alterations and climate change28,29. In this paper, we therefore empirically integrated psychological and ecological data to build more comprehensive knowledge of human-bird interactions within racially and ethnically diverse communities.

Due to various birds’ life histories, different species display observably diverse functional traits that impact the locations and ways in which people interact with these organisms (and vice versa)30,31. The functional traits of birds can include behaviors (e.g., vocalizations for communication, foraging behaviors) and physical characteristics (e.g., plumage colors, body size)32. Past work has argued that larger size, more colorful plumage, more frequent visits to bird feeders, and migratory behaviors all increase the popularity of a bird species, according to Google search trends among people in North America33. These results may indicate how functional traits impact a person’s ability to detect urban birds34. Bird species with vibrant plumage, striking ornamentation (e.g., long crests or tails), and melodic songs are generally more appealing to people35,36. However, there are mixed results for how urbanization impacts the fitness of birds with aesthetically pleasing traits37. For example, previous research has suggested that bird species that are more successful in urban environments often exhibit black, gray, or blue plumage and tend to be more ornamented compared to birds in less urbanized areas38. While people in urban areas may appreciate birds with blue plumage, gray and black plumage is unlikely to be favored36. Such contradictions raise questions about the potential mismatch between human perception and traits that enhance urban fitness, as well as how assessments of these traits shape attitudes towards bird management. The presence (or lack thereof) of diverse urban bird species can also be harbingers of broader issues of injustice related to class, race, and access to nature due to their reliance on suitable habitat39,40. Given that more affluent areas tend to have higher bird diversity as well as more extensive vegetation and greater food resources available for birds41,42, human-bird interactions in underserved communities may affect how residents interact with local birds. These differences warrant research attention to understand the interrelationships between socioeconomic factors and perceived traits of bird populations in urban settings.

Public greenspaces (e.g., parks, nature preserves) have been extensively studied given the benefits they provide for biodiversity conservation43,44,45 and their potential to facilitate positive human-nature relationships46,47. However, private lands are gaining recognition as important places for conservation initiatives such as advancing climate resiliency48, fostering stewardship49, maintaining sustainable food webs50, and indeed supporting diverse bird functional traits51. For example, previous research in the Midwestern USA highlighted the importance of private yards for birds during the winter season because they rivaled public greenspaces in levels of bird species diversity and offered resources such as supplemental bird feeding52. However, these authors only examined the public–private divide for bird populations and did not assess the co-benefits provided for both human and non-human species. Further research is therefore needed to document the affordances of public and private land for bird communities, and in turn, better understand and account for the conditions that give rise to positive human-bird interactions in urban settings5,53.

The diversity of bird species is an important indicator for understanding the quality of environmental conditions54, and the ways in which people value natural spaces55. Biological resource conditions provide a broad context that shapes human behavior56, and perceptions of bird diversity30,46. These human-nature relationships, in turn, may give rise to human perceptions, particularly perceived charisma. Although nascent, studies of wildlife charisma have often evaluated the factors that influence willingness to donate to conservation, such as the characteristics of birds that engender support56, and psychological drivers of people when making avian conservation-donation decisions57. These studies have focused on identifying the species or traits that act as proxies for charisma58,59 rather than developing measurement scales. Lorimer60 defined non-human charisma as the characteristics of a non-human entity that determine its perception by humans and subsequent evaluation. As an extension, we define charisma in birds as a broad set of characteristics that are perceived and collectively evaluated by humans. Our conceptualization of bird charisma is not focused on individual functional traits or their appeal in isolation, but instead, from an emergent, holistic lens such as human-perceived beauty and impressiveness. Building on Lorimer60, charisma can be measured using three dimensions: (1) ecological charisma (i.e., the ease of an individual detecting a species), (2) aesthetic charisma (i.e., the aesthetic characteristics of a species, which trigger an immediate emotional response), and (3) corporeal charisma (i.e., visceral responses triggered by repeated interactions with a species). However, routine interactions between people and common urban species may blur the differences between aesthetic and corporeal charisma.

In addition to understanding the role of traits in shaping how people perceive and interact with birds33, numerous psychological theories can provide guidance for research on the factors that motivate human behaviors that can help to sustain urban bird populations. The Theory of Planned Behavior (TPB) is one of the most prominent, and it offers valuable insights into how attitudes, subjective norms, and perceived behavioral control (PBC) collectively shape behavioral intentions and ultimately actions61. More specifically, attitudes refer to an individual’s positive or negative evaluation towards a specific object, including a person, place, or behavior62. Subjective norms indicate the perceived social pressure to engage or not engage in a behavior63. PBC refers to the perceived ease or difficulty of performing a behavior and considers both external and internal factors61. Theoretically similar to PBC is the idea of self-efficacy, which refers to an individual’s internal confidence in their ability to successfully perform a behavior, instead of considering if a behavior can influence an external outcome64. Although similar, self-efficacy has been shown to have a stronger predictive capacity than PBC65 and can streamline practical implications that arise from interventions66. The TPB has been adapted to environmental management contexts to provide insight into behavioral antecedents67,68, although ecological data are rarely incorporated despite their potential to offer a more complete representation of how an individual’s context shapes their decision-making69,70.

Environmental applications of the TPB have advanced our understanding of the factors driving behaviors aimed at directly or indirectly reducing environmental degradation, also known as pro-environmental behaviors (PEBs)71,72,73. A substantial body of research has centered on PEBs, recognizing the critical role of human behavior in advancing agendas that support resource management70,74. These actions are performed to address various environmental issues relevant to urban contexts, including urban biodiversity loss75,76, natural area conservation77, and pollution78. A diverse range of approaches has been adopted in the conceptualization and measurement of PEB, including scales focused on evaluating human behaviors that promote biodiversity79,80. On the one hand, past work successfully utilized a unidimensional scale for PEBs under the assumption that all measured actions require similar levels of effort and are uniformly motivated by environmental concern81. However, other studies have become sensitive to the multiple dimensions of PEB, such as actions that occur in either a “public sphere” (e.g., voting to support a policy which benefits the local environment) or “private sphere” (e.g., planting native vegetation near your household)82,83. Multidimensional scales have the benefit of evaluating behaviors that differ in their required effort, are driven by different ecological or psychological factors, and/or are context dependent73.

Norms are societal standards of behavior and ethics that have played a substantive role in previous research focused on understanding why people make behavioral decisions84,85. While subjective norms in the TPB highlight social pressures influencing behavior, the norm activation model argues intentions to perform a behavior are principally motivated by moral normative concerns86,87. More specifically, this latter model posits that moral obligations defined as beliefs about the standard practices of reference groups (e.g., friends, relatives, neighbors) compel behaviors88. These internalized social norms are especially relevant in an urban context because they drive yard management, albeit to different degrees in front and backyards89. Previous research has also demonstrated the utility of social normative beliefs in a range of other resource management contexts that provide insight into the portability of this concept90. For example, results from a systematic review suggested that social norm interventions effectively changed behavior in instances of energy usage, recycling, water conservation, and pesticide use91. However, social norms have received far less attention in understanding behaviors directly related to urban birds, though the few studies in this vein have focused on bird conservation in rural landscapes and the consumptive practices of Indigenous communities92,93. The non-consumptive use of bird populations by people in urban centers, therefore, remains to be studied more extensively, particularly regarding the role of social normative beliefs in shaping human behaviors that impact bird populations.

Our research aimed to explicitly integrate knowledge from the social and natural sciences in support of multispecies justice8, and bird conservation initiatives94. We also provided more comprehensive knowledge by introducing a new a focus on the reasons why underserved residents in racially diverse neighborhoods engaged in public and private pro-environmental actions, as well as evaluating bird communities in the same context across scales that varied in space (i.e., on public versus private lands) and time (i.e., summer versus winter seasons) throughout two neighboring cities in the Midwestern USA. We hypothesized that (1) the seasonal number of bird species present near an individual’s household would positively predict perceived functional diversity (H1–H2); (2) perceived functional diversity would positively predict ecological charisma and aesthetic charisma (H3–H4); (3) perceived functional diversity, ecological charisma, and aesthetic charisma would positively predict attitudes towards management (H5–H7); (4) attitudes towards management would positively predict self-efficacy and social normative beliefs (H8–H9); and (5) self-efficacy, attitudes towards management, and normative beliefs would positively predict both public and private PEBs (H10–H15; see Fig. 1).

Fig. 1: Hypothesized model of the theoretical relationships among factors expected to predict private and public pro-environmental behaviors beneficial to bird communities.
figure 1

Latent variables are depicted in ovals, and single-item indicators are in rectangles, including survey items with abbreviations described in Table 3. All hypotheses (H1–H15) are shown beside arrows and expected to be positive (denoted as +).

Full size image

Results

Bird species richness

Across the 465 point counts performed during the winter, 6387 individual birds were detected, including 47 different species. During the 465 point counts performed during the summer, 10,524 individual birds were detected, including 86 species. In total, 16,911 individuals, including 100 different species, were detected across the two seasons52. Although the bird point count surveys were performed prior to selecting Census Block Groups for the self-administered survey of residents, all households were within 1000 m of a point count. The bird species richness layer generated from the point counts resulted in estimations of households being exposed to spanning 9–17 species in the winter and 16–26 species in the summer (see Fig. 2).

Fig. 2: Bird species richness across Champaign-Urbana, IL, USA.
figure 2

A The estimated number of species present in winter 2022. B The estimated number of species present in summer 2022.

Full size image

Survey sample

Survey respondents were purposefully sampled from the most racially and ethnically diverse Census blocks in Champaign-Urbana, IL, USA (see Fig. 3). The gender distribution of respondents was 47.6% female, 44.9% male, 3.2% other, and 4.3% preferred not to answer (see Table 1). About half (51.1%) identified as White, while just under one quarter identified as Black/African American (21.8%). The remaining proportion was identified as Asian (12.8%), other (11.2%), and/or American Indian/Native Alaskan (3.2%). Additionally, 13.8% of our sample identified as Hispanic, Latino, or of Spanish origin when asked about their ethnicity. The sample of respondents included in this study was well educated, with just under half (47.6%) reporting a 4-year college degree or higher. The youngest respondent was 18 years old, and the oldest respondent was 93 years old, with the average age being 44.5 years (SD = 15.2). Respondents did not report extensive previous experience, given an average of 11.8 (SD = 49.4) days traveled in 2022 to watch birds and a total of 8.8 (SD = 16.8) years of experience. Respondent’s birding skills were “lower than average” with a self-evaluated average of 2.3 (SD = 0.96) out of 5.0 when compared to the average person. Respondents reported relying on social media (53%) and friends and family (46%) most often to learn about information related to bird conservation.

Fig. 3: The 24 U.S. Census block groups of residents surveyed in Champaign-Urbana, IL, USA.
figure 3

In February–May 2023, neighborhoods within each of the 24 Census block groups shown were surveyed. Census block groups numbered 1–12 are in Champaign, whereas 13–24 are in Urbana.

Full size image
Table 1 Socio-demographic characteristics of survey respondents (n = 210)
Full size table

We found that slightly more than half of our sample intended to engage in private sphere PEBs to make their yard more desirable for birds through planting native vegetation (57%) or providing bird houses, feeders, or baths (55%; see Table 2). For public sphere PEBs, approximately two-thirds of respondents (64%) were willing to vote to support a policy that affected the local environment, whereas significantly fewer respondents intended to volunteer (29%) or participate in an environmental group (19%). Respondents generally agreed with statements that measured functional diversity (M = 3.53, SD = 0.77; see Table 3), ecological charisma (M = 3.95, SD = 0.77), aesthetic charisma (M = 3.65, SD = 0.77), and self-efficacy (M = 3.68, SD = 0.75), but had negative attitudes towards bird management strategies (M = 2.27, SD = 1.05). Respondents also held social normative beliefs that other residents rarely attempted to protect local birds (M = 2.58, SD = 0.70).

Table 2 Percentage of residents intending to perform pro-environment behaviors in the next 12 months, along with not applicable response rates
Full size table
Table 3 Mean values and factor loadings for latent variables
Full size table

Structural equation modeling results

Fit indices from the measurement model showed that the model fit the sample data well (χ2 = 469.09; df = 215; root mean square error (RMSEA) = 0.076; comparative fit index (CFI) = 0.90; standardized root mean square residual (SRMR) = 0.59). All scales were reliable, given Cronbach alpha (α) coefficients ranging from 0.78 to 0.87 and composite reliability scores ranging from 0.80 to 0.89. All factor loadings were ≥0.40 with the exception of a survey item that measured attitudes towards management. This item (“Wildlife that harms native bird species should be removed”) was dropped from the final model due to a low factor loading score (λ = 0.34), which was below the 0.40 threshold established in previous research95.

Our structural model also showed good model fit (χ2 = 537.10; df = 307; RMSEA = 0.06; CFI = 0.92; SRMR = 0.08). Our hypotheses were largely supported (see Fig. 4), indicating summer bird richness positively predicted perceived functional diversity (β = 0.31; t-stat = 3.385). Perceived functional diversity positively predicted both ecological charisma (β = 0.60; t-stat = 6.485) and aesthetic charisma (β = 0.76; t-stat = 7.495). Additionally, aesthetic charisma positively predicted management attitudes (β = 0.62; t-stat = 4.077). Management attitudes positively predicted social normative beliefs (β = 0.28; t-stat = 3.248), and self-efficacy (β = 0.79; t-stat = 6.632). Moreover, self-efficacy (β = 0.43; t-stat = 2.164) and social normative beliefs (β = 0.16; t-stat = 1.969) positively predicted private PEB, while social normative beliefs positively predicted public PEB (β = 0.20; t-stat = 2.774).

Fig. 4: Statistical results from a latent variable path model.
figure 4

Solid lines indicate significant paths, whereas dotted lines indicate non-significance.

Full size image

Discussion

The growing global urban population introduces deep-seated social-ecological challenges due to rapidly advancing land-use change4,96 and the underrepresentation of diverse voices in the environmental planning process8,25. Given the importance of garnering public support for the success of urban conservation initiatives97,98, we tested a series of hypotheses to understand the role of both combined social and ecological factors in explaining public and private PEB using latent variable modeling techniques. We advanced scientific knowledge of the factors that can energize PEBs beneficial to urban bird communities in prairie ecosystems by merging insights from the norm activation model84,87, TPB61, and self-efficacy theory64. Our results revealed that: (1) actual species richness had a weak positive relationship with perceived functional bird diversity; (2) perceptions of functional diversity moderately explained positive variation in perceptions of charisma; (3) aesthetic charisma emerged as the only significant positive driver of attitudes towards management; (4) attitudes towards management significantly and positively correlated with self-efficacy and normative beliefs; and (5) normative beliefs were the sole significant positive driver of public PEB. These findings aim to contribute to a burgeoning literature on the role of social-ecological factors in urban sustainability.

We observed that the number of species present around respondents’ homes was an effective, albeit weak and positive, predictor of perceived functional diversity. This result aligns with past work suggesting that perceptions of biodiversity are minimally99,100 or even negatively55 correlated with actual biodiversity. While we did not directly measure perceived biodiversity, beliefs about bird functional diversity served as a proxy. We focused respondents’ attention on bird species richness at the neighborhood level with the understanding that personal interactions with birds near a home and/or in gardens may have been a more powerful basis for shaping perception than the total species present in a broader context. The potential disconnect between perceived and actual biodiversity23,46 may have been further thwarted by the complex process of translating benefits provided by birds to human communities through a combination of traits embodied by different species101. We observed that only summer bird richness significantly and positively correlated with perceived functional diversity. Because human-bird interactions were more likely to occur in the late spring (i.e., a migratory stopover period) and early summer (i.e., onset of nesting by resident birds)102, it stands to reason that this path emerged as the sole predictor of perceived functional diversity (H1). For decision-makers working in urban settings, our result suggests that educational campaigns to increase bird biodiversity may not be fully understood by residents without raising awareness of what avian diversity entails and the associated benefits of bird populations that change across seasons103,104.

Urban residents do not solely evaluate the individual traits of birds, but they also form perceptions based on the broad suite of characteristics displayed by birds known as charisma60. Birds are uniquely charismatic to humans because many species utilize striking visual and/or auditory signaling, making them easily detectable and aesthetically appealing36,105. We tested the formative role of birds’ underlying functional traits as a theoretical antecedent to ecological and aesthetic charisma and observed moderate statistical support for our hypotheses (H3–H4)33,58,59. We therefore expanded a nascent but growing literature on the root causes of charisma in species106 and developed a scale to measure the three dimensions posited by Lorimer60, although only two were statistically distinguishable. However, future efforts to measure charisma should consider that many urban birds are easily detected by people, so ecological charisma may not be a universally accurate depiction of city-dwellers’ relationships with birds. Ecological charisma may be more relevant for measuring the differences in human relationships with various taxa in urban ecosystems. Additionally, aesthetic charisma that is shaped by positive or negative evaluations across various trait combinations shows promise in explaining why there are differences in residents’ support for management actions (H7).

In line with the TPB, attitudes towards management practices strongly and positively predicted self-efficacy, which we considered to be interchangeable with the concept of PBC (H8)61. More specifically, our results showed that respondents largely agreed with the survey items that measured self-efficacy, and that this construct was the most powerful driver of PEB in the private sphere, in line with previous research35. These findings suggest there is a promising foundation to motivate and galvanize current PEBs because residents believe their actions will make a difference; however, this relationship will require continued attention—particularly among low-income and minoritized communities—so management agencies should continue to empower residents by demonstrating how their actions can directly impact bird conservation85,107. Our results also aligned with the norm activation model that theorizes attitudes should positively predict social normative beliefs87 (H9); however, we also tested the effects of attitudes on behavior to extend the TPB (H10–H11) and observed a surprisingly non-significant relationship that could be due to our focus on intentions rather than actual behavior57. Future research on observational (rather than reported or intended) behavior is urgently needed, given the disconnect that has consistently emerged between self-reported data and on-ground activity108,109. On average, residents did not believe that others were regularly taking steps to protect local birds, especially in their own neighborhood, which is an alternative explanation for the non-significant relationship between attitudes and behavioral intentions. These findings reveal a valuable opportunity for managers and educators to engage with local communities and showcase how residents are already contributing to the improvement of private greenspaces that may otherwise go unnoticed (e.g., planting native species on their property). By shifting normative beliefs about landscaping practices through increased awareness, more residents will be likely to engage in private PEBs110.

This study evaluated two dimensions of PEB, which were effectively explained by two antecedents. Social normative beliefs positively predicted both private and public PEBs (H14–H15), which signals the potential impact of communication among urban residents who discuss the steps they are taking to improve both their yards and local natural areas. The importance of communication is further evidenced by the reported information sources that were most frequently used by respondents to learn about conservation issues. Given that information sources can influence perceptions of inclusivity in management decisions111, strengthening community information networks can help to fortify existing relationships and build new linkages between people to increase PEB. Self-efficacy positively predicted only private PEBs (H10), which suggests that an individual’s belief in their capacity to perform behaviors beneficial to local birds is spatially and contextually dependent89. This reaffirms the theoretical importance of using a multidimensional scale, as residents are influenced by distinct psychological factors when performing public or private PEBs67,73. Overall, we found strong interest in initiatives to modify private greenspaces to benefit local birds, but aside from voting, little interest in improving public greenspaces. This finding suggests that campaigns targeting private greenspace initiatives will be more likely to gain support among the diverse audiences that participated in this study and underscores the importance of prioritizing action-based behaviors in response to their potential impacts and prevalence80. The ecological impacts from urban residents engaged in our research on bird populations are difficult to assess given the scope of our research. However, any actions creating bird-friendly backyards or regenerating foraging and breeding habitats may have positive conservation outcomes through increasing suitable habitat area, as this is a well-known limiting factor of bird population size112.

Several areas for future research can be drawn from our study findings. First, our sampling procedures purposively targeted Census blocks within Champaign-Urbana that had the highest proportion of non-White residents to build knowledge of these communities that are underrepresented in urban sustainability research. Our sample did not match exactly with the target population; however, multiple non-White residents expressed excitement about the opportunity to participate in this study, because they had not previously been asked to provide input on natural resource management topics. Our use of the drop-off-pick-up method113 was also advantageous because it resulted in a relatively high response rate. Second, we observed significant but low explanatory power of perceived functional diversity. This could be due to the role of knowledge in affecting how attuned respondents were to on-ground biodiversity with respect to the number of species and their functional diversity55. Finally, our measurement approach was theoretically informed but also imperfect. We performed exploratory analyses and carefully evaluated the performance of our survey scales, which led to our PEB sale including fewer survey items than expected and our measure of charisma being reflected by two rather than three originally hypothesized dimensions. The quest to develop valid and place-based measures of PEB should continue108, along with the careful evaluation of how to adapt behavioral antecedents.

Interdisciplinary research that considers the complex interplay between social and ecological dynamics is essential but still underrepresented in urban sustainability initiatives despite growing recognition of the co-benefits of these landscapes for human and non-human species13. We advance efforts to achieve multispecies justice by measuring on-ground diversity that spans spatial and temporal scales and use this knowledge to explain how people perceive the functional diversity of bird communities. Further, we offer new insights on how birds are perceived by diverse communities that have been previously underrepresented in the development of resource management strategies and are faced with inequities such as limited access to safe greenspaces114 that include biologically diverse ecosystems115,116. The multiple psychological drivers of PEB are also explored to provide insight on how to encourage behavior change that will benefit urban bird populations. From our results, we posit that self-efficacy is a powerful catalyst for energizing residents’ bird-friendly activities in their own yards, while beliefs about the behaviors of other people are also instrumental in increasing behaviors performed closer to home and in broader locales57. We also establish a new survey scale to measure the positive charisma of birds and show that these perceptions increase attitudes towards bird management. Indeed, birds are viewed positively by urban residents engaged in this research, thus positioning them as a key motivating factor for advancing sustainability transitions that are inclusive of diverse human communities in an increasingly urbanized world.

Methods

Study context

We conducted this research with residents living throughout Champaign (population ~88,000) and Urbana (population ~38,000), which are two neighboring cities that share a border in east central Illinois, USA. The cities of Champaign and Urbana share the main campus of the University of Illinois system, which serves as one of the top employers for residents and enrolls more than 55,000 students annually. Despite the strong and historic linkages between the two cities, differences in greenspace and bird communities persist due to distinct patterns in urbanization, greenspace management, and city governance priorities52. Moreover, as in many other American cities, there is residential segregation in Champaign-Urbana today that can be traced back to historic redlining. In the 1950s, racially restrictive covenants were written into new subdivision deeds in Champaign County that prohibited the sale, lease, or habitation by “persons not of the Caucasian race” (ref. 117, p. 21). This discriminatory practice forced non-White residents, largely African Americans, by law to live in what was later referred to as the “North End,” positioned on the outskirts of Champaign. The results of redlining were still salient at the time this research was conducted, in that much of the racial and ethnic diversity in our neighboring cities was located around northern Champaign and the University of Illinois campus, according to the 2020 US Census data. Current knowledge of bird diversity can also be affected by historic race-based zoning policies at a national scale, given that bird sampling in redlined neighborhoods has decreased over the past two decades115.

Bird data collection and point count design

To measure the species richness experienced by local residents, we used already collected and published bird community data from Champaign and Urbana (see ref. 52 for a detailed sampling protocol). Briefly, we generated 10-min, 50-m distance point counts in the winter and summer of 2022. We selected 80 locations in both public (e.g., parks, natural areas, cemeteries) and private (i.e., residential yards) greenspaces. Within Champaign and its immediate surroundings, 41 points were selected: 16 in public greenspaces and 25 in private yards. Similarly, within and around Urbana, 39 points were chosen: 23 in public parks and 16 in private yards. Due to the size variation of public greenspaces, we increased the number of point count locations based on their size, leading to a total of 93 point count locations (see Fig. 5), each of which was visited and surveyed five times in both the winter (January–February) and summer (June–July), resulting in a total of 930 point counts across the entire study area. Utilizing the kriging tool in ArcGIS software, the species richness at each of the 93 point locations was then interpolated across the study area to create a species richness layer for each season. Previous work in Champaign-Urbana by Pollock et al.52 demonstrated that species richness and three different metrics of diversity were positively correlated and yielded no significant differences across two different seasons. This justified our use of species richness as opposed to other diversity indices. Furthermore, relative species presence and individual numbers were less relevant given our focus on community perceptions of species richness. This work was permitted by the institutional research board and ethics permits from the University of Illinois at Urbana-Champaign.

Fig. 5: Bird point count locations throughout the study area, including Champaign and Urbana, IL, USA.
figure 5

Bird point counts were conducted in 93 locations across 80 public and private greenspaces in and around the neighboring cities of Champaign and Urbana. Each point count location was surveyed five times during the winter and five times in the summer of 2022. Orange triangles indicate point count locations in private greenspaces (n = 41), while blue circles represent point count in public greenspaces and natural areas (n = 52).

Full size image

Resident survey data collection and sampling design

We also surveyed the most racially and ethnically diverse areas in both Champaign and Urbana, IL, in response to the ongoing inequities of non-White individuals’ access to greenspace and biodiversity27. To select these areas, we linked race and ethnicity information from the 2020 USA National Census to each Census block group, which is the smallest geographical unit for which sample data is published. From this information, we created a proportion of non-White and Hispanic residents living in each block group. We then overlayed property data and removed block groups with less than 100 residential addresses. From the remaining block groups, the 12 in each city with the highest proportion of non-White and Hispanic residents were selected (n = 24 block groups; Fig. 3). In each, residential addresses were randomly selected. These addresses served as starting points to survey all residences on a given street within each block group. In three Census block groups, the selected addresses were on streets with few residential addresses or completely inaccessible apartments, leading to a third street being randomly selected to meet our goals for a minimum sample size.

We collected survey data February–May 2023 using a drop-off-pick-up method113 in both Champaign and Urbana, IL, USA. Questionnaires were packaged in orange plastic bags with personalized correspondence that included a hand-signed cover letter and envelope to seal the completed questionnaire. The included cover letter contained information about the benefits of participating, potential impact of the project, and an incentive to win a $50 Amazon gift card raffle to encourage participation, given decreasing response rates, particularly in rural contexts68. Two survey administrators of different genders surveyed all households to further reduce non-response bias118. Between the hours of 10:30 a.m. and 4:00 p.m. on both Saturday and Sunday, we attempted to make face-to-face contact and verbally communicated with an adult at each residence over the course of two separate weekends. If we were successful in contacting the resident, and they agreed to participate, we left the survey materials with them. We asked the resident to place the completed and sealed questionnaire in the provided orange bag and hang it on their doorknob for retrieval at a later, pre-designated date. Questionnaire retrieval took place Monday–Thursday following the initial delivery from 4:00–6:00 p.m. Up to three attempts were made to collect the completed survey questionnaires, and reasons for refusal were collected from contact logs. The most common reason why questionnaires were not returned was the limited time. In total, we contacted 438 residents during the survey process and received 210 completed questionnaires, resulting in a response rate of 47.9%. We reached our minimum acceptable number of 200 survey respondents, given median sample sizes in previous studies in disciplines such as education and psychology that employ structural equation modeling (see Kline119 for an overview). Survey respondents gave written informed consent, as part of human subjects’ research approval provided by the University of Illinois at Urbana-Champaign.

Survey measures

We asked residents about their intention to perform nine different PEBs that were beneficial to local bird populations in the next 12 months, following previous research24. To enhance the reliability and validity of our original scale, we refined the survey items based on three criteria: (1) type of PEB (i.e., public or private sphere); (2) removal of items that were marked non-applicable greater than 25% of total; and (3) identifying items with similar levels of effort. This resulted in two scales that measured public sphere PEBs with three questions and private sphere PEBs with two questions. The questionnaire also included questions to measure three behavioral antecedents including management attitudes, social normative beliefs, and self-efficacy. Management attitudes were measured using five items from ref. 120 and social normative beliefs were measured with three survey items and adapted from ref. 55. Minor changes to the wording were made to adapt these scales from the context of psychological drivers of angler behavior to bird behavior. Lastly, a self-efficacy scale was adapted from the context of modifying urban landscapes121 to our context on bird management and consisted of three survey items. All survey items analyzed for this study are listed in Tables 2 and 3.

We measured residents’ perceptions of local bird species, particularly the three dimensions of charisma, following ref. 60. We ran an exploratory factor analysis to assess whether our survey items reflected the hypothesized three-dimensional structure of charisma. Principal axis factoring was used to correct for measurement error, and varimax rotation was used to minimize correlation among the latent variables. Our results revealed two distinct factors: ecological charisma measured by three survey items, and aesthetic and corporal charisma consolidated into a single factor measured by six survey items. The modified two-factor solution accounted for 68% of the total variance and resulted in reliable scales that measured ecological charisma and the combined aesthetic charisma. To measure perceived functional diversity, we drew from past research to identify birds’ functional traits relevant to city-dwelling residents30.

Modeling process

We used a two-step structural regression modeling process to test our hypotheses122. We evaluated factor loadings for each survey item to verify that all values exceeded the 0.40 threshold established in previous research95. We also assessed the normality of our data by estimating skewness and kurtosis for each latent variable. All skewness and kurtosis values were within acceptable ranges of normal data95, and thus provided further support for our sample size being considered acceptable119. Model fit was assessed using a χ2 value; however, given this indicator’s sensitivity to sample sizes larger than 200, other fit statistics were also referenced, including the RMSEA ≤ 0.07, CFI ≥ 0.90, and SRMR ≤ 0.07. We used RStudio Version 4.3.1 “Beagle Scout” for our statistical analysis123.

We assessed missing data patterns across our latent variables. A total of 8.1% of the items measuring latent variables were missing or marked as “not applicable” in the survey questionnaire. As a result, we took steps to understand whether the missing items were missing at random, missing completely at random (MCAR), or missing not at random. Little’s124 test indicated the survey items were MCAR (p > 0.05). Given this finding, the full-information maximum likelihood method was applied to all scales to account for our missing data, and a maximum likelihood estimation procedure was used for subsequent analysis.