About Us Advertise Careers Contact Us Privacy Policy Community Guidelines
Locating… 13°C English WhatsApp Alerts Email Alerts
BREAKING NEWS
IMF follow-up programme would steady Sri Lanka as debt repayments loom, analysts say - Reuters Saqib frustrates Sri Lanka as Pakistan extends lead to 123 in second cricket test - The Washington Post Saqib frustrates Sri Lanka as Pakistan extends lead to 123 in second cricket test - The Daily Gazette PLD Space Investor Selects MIURA 5 to Launch Demo Mission U.S. Ambassador Leads Pledge of Service for 34 New Peace Corps Volunteers in Sri Lanka - U.S. Embassy in Sri Lanka (.gov) APOD: 2026 November 18 – An M-class Solar Flare Sierra Snow Season Spins Up 5th Match, Group A, Asia Cup Rising Stars (50 Overs) at Doha, Nov 21 2026 - ESPN Sri Lanka slips to 74-3 after Pakistan takes massive 227-run lead in 2nd test - Oskaloosa Herald Sri Lanka slips to 74-3 after Pakistan takes massive 227-run lead in 2nd test - The News-Enterprise
SRI LANKA NEWS

Multidimensional tolerance towards Asian elephants varies with demographic and spatial correlates among rural communities in Sri Lanka - Nature

Lanka Hit News • Jul 4, 2026 · 4mo ago • 67 min read
Share:

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Scientific Reports volume 16 , Article number: 30801 ( 2026 ) Cite this article

Human–wildlife conflict (HWC) threatens both human communities and many wildlife populations. We investigated the sociodemographic and spatial factors shaping tolerance towards elephants in rural Sri Lanka, where human–elephant conflict is prevalent. We surveyed residents near three protected areas (PAs: Minneriya National Park, Kaudulla National Park, Hurulu EcoPark) frequently encountering elephants. We examined how perceived risks, social norms, and individual agency interact with sociodemographic variables to shape tolerance. We tested whether the tolerance-related construct structure published in a previous study from another region of Sri Lanka replicated in our data, using confirmatory factor analysis on the same survey instrument. Three of the five tolerance-related constructs replicated with acceptable reliability (community shooting norms, utilitarian wildlife value orientations, perceived agency) while two did not, reflecting culturally specific differences in how sacredness of wild elephants is interpreted and near-universal disagreement that elephants should be near homes or croplands. Tolerance-related responses varied significantly despite the respondents’ ethnoreligious homogeneity. Gender effects appeared in two distinct constructs but with different magnitudes: men reported somewhat less permissive community-shooting norms but substantially lower perceived agency in their elephant interactions. Recent change in elephant-encounter frequency emerged as the most pervasive predictor, with increasing encounters associated with weaker emotional bonds, weaker coexistence belief, and more utilitarian wildlife value orientations. Distance to the nearest PA boundary was associated with weaker coexistence belief—the first explicit individual-level link between spatial setting and a tolerance-related response in this dataset. Analyses revealed a coherent gradient in affective and behavioral responses across PAs (Kaudulla = high-tolerance pole; Minneriya = opposite pole; Hurulu intermediate), while abstract value orientations and perceived agency were spatially uniform. These findings highlight the complex interplay of sociodemographic, experiential, and spatial factors in shaping tolerance-related responses and demonstrate that what varies across the landscape is how people feel about and act toward elephants, not how they fundamentally value wildlife. The integration of theory-anchored measurement with spatial analysis provides a more nuanced understanding of tolerance and can inform targeted interventions to mitigate HWC and promote coexistence.

Human–wildlife conflict (HWC, also referred to as “negative human–wildlife interactions”) poses a significant threat to the long-term survival of wildlife populations and the wellbeing of human communities that interact with them 1 . The costs of HWC to people include destruction or damage of property, lost livelihoods, disease transmission, forgone time or income from avoiding or managing wildlife, and compromised physical safety resulting from animal attacks 2 , 3 , 4 . The actual and perceived hazards imposed by HWC are intensifying globally as human–wildlife interactions increase with expanding human activity 5 . Sustainable coexistence (defined as a dynamic state in which humans and wildlife co-adapt to shared landscapes under governance systems that allow wildlife persistence at tolerable risk levels) is elusive and context-dependent, rarely achieved by eliminating conflict entirely 6 , 7 . Individual tolerance towards wildlife is shaped by a balance between perceived benefits, such as cultural meaning or revenue from tourism, and associated costs 8 , 9 , 10 , further influenced by a spectrum of positive to negative personal experiences 11 , 12 . Identifying factors that shape variation in tolerance is crucial for developing effective HWC mitigation strategies and fostering coexistence. Here, we use the term “tolerance” to refer to a multidimensional constellation of psychological and experiential responses that shape how individuals perceive, evaluate, and accept the presence of wildlife in shared landscapes. Therefore, characterizing tolerance requires an understanding of complex psychological factors that shape attitudes and behaviors. Rather than treating tolerance as a single dichotomous or continuous variable (e.g., tolerant vs. intolerant), we conceptualize tolerance as emerging from multiple, interrelated dimensions—including perceptions of risk, evaluations of coexistence and management, instrumental assessments of costs and benefits, and perceived social norms—that together characterize how people live with wildlife. In this framing, tolerance represents an individual-level psychological condition that can enable, but does not guarantee, human–wildlife coexistence at broader social and ecological scales 13 .

Tolerance is a learned behavior, and various proposed models explain its underpinnings 14 . Hazard acceptance models describe the balance of perceived risks and benefits associated with living alongside rather unpredictable (“hazardous”) wildlife to explain tolerance 15 , influenced by trust, affect, and perceived control over risks. However, these models may omit other significant drivers of behavior. Notably, social and cultural norms—shared beliefs about how people should or actually behave—can contribute to wildlife tolerance 16 . The theory of planned behavior (TPB) posits that tolerance is influenced by attitudes and beliefs, perceived social norms, and behavioral control 17 , 18 . Rather than testing these frameworks formally, in this study, we draw on them as interpretive reference points for understanding how tolerance-related responses towards wildlife co-vary in contexts where behavioral intentions are difficult to measure directly. Importantly, while theories such as hazard acceptance models and TPB provide useful conceptual guidance, both are difficult to test directly in HWC contexts where strong cultural taboos, legal restrictions, and social desirability bias limit truthful reporting of behavioral intentions related to lethal control. Established frameworks such as TPB and hazard acceptance models are therefore used as interpretive reference points rather than as confirmatory models.

Elephants are frequently represented in the HWC literature 1 , 19 , 20 , perhaps reflecting their propensity to engage in negative interactions with humans and/or their impacts on human livelihoods and safety 21 , 22 , 23 . Human interactions with Asian elephants ( Elephas maximus ) are multi-faceted and represent a continuum between acceptance stemming from cultural and economic value to intolerance resulting from conflict. Across Asia, elephants are culturally and/or religiously important, holding value that has persisted in some regions for centuries 24 , 25 . More recently, some local communities have generated significant revenue from wildlife tourism centered around wild elephants 26 , 27 . However, dense human populations and habitat fragmentation across Asia have caused intensifying human–elephant conflict (HEC), particularly in rural areas around small-holder farms 28 , 29 , 30 . Perceptions of elephants in local communities are complicated and may show unexpected variation in tolerance levels 31 . Individual variation in tolerance is shaped by perceived risk, attitudes, beliefs, and social norms—among other factors—and so characterizing this variation holds promise in developing opportunities for human–elephant coexistence (HECx) 32 , 33 , 34 , 35 , 36 , 37 , 38 .

Understanding the dynamic process of individual tolerance is critical in places like Sri Lanka, where HEC jeopardizes both the survival of the endangered Asian elephant and human wellbeing. The island hosts a globally significant elephant population 39 , 40 , 41 believed to be approximately 7400 individuals according to the Department of Wildlife Conservation, although this assessment is unpublished and unverified. Though traditional agricultural practices once fostered cohabitation for centuries 42 , tolerance is waning. Human-induced elephant mortality is common 42 , 43 , with up to 400 elephants dying from HEC annually in recent years 44 , 45 . Expanding, unsustainable agriculture practices have led to the destruction of natural habitats, creating human-dominated areas that exacerbate the conflict, and even so, Asian elephants have a broad ecological niche and can navigate anthropogenic landscapes 46 , 47 . Spatial variability of HEC is influenced by the climate, physical environment, and anthropogenic activity 48 , 49 , 50 , 51 , 52 , 53 , 54 , so localized mitigation is essential. While various sociodemographic factors have been proposed that drive HEC 44 , 55 , understanding how perceived risk, attitudes, beliefs, and social norms interact to determine tolerance is poorly understood. de Silva, et al. 11 found that tolerance of farming communities in southern Sri Lanka towards elephants was associated with community-wide fear and worry, causing widespread negative feelings about cohabitating with elephants. The authors of that study advocated for collective HEC interventions to limit emotional contagion of fear. However, tolerance of communities in other areas of Sri Lanka may demonstrate distinct patterns resulting from unique regional contexts 43 , 50 , 56 . For instance, communities that benefit from wildlife tourism may be more willing to tolerate elephants 56 , unlike farmers surveyed by de Silva, et al. 11 , who reported no tangible tourism benefits. Furthermore, while frequent negative experiences with elephants may promulgate negative feelings towards elephants, more positive sociocultural values may outweigh these feelings to foster tolerance 57 , 58 , 59 , 60 . Thus, studies that spatially and psychologically characterize the multidimensional nature of tolerance are essential for developing effective HECx strategies.

To address this need, we surveyed perceptions towards free-ranging elephants among stakeholders in a rural community in Sri Lanka’s dry zone. Our analytical approach had two complementary aims. First, we tested whether the construct structure published by de Silva, et al. 11 for a southern Sri Lankan agricultural sample replicates in our north-central Sri Lankan sample, using confirmatory factor analysis on the same survey instrument. Because tolerance-related constructs are known to be context-dependent, we expected partial rather than full replication and treated the comparison as informative regardless of outcome. Second, for the constructs that did replicate, we examined how sociodemographic, experiential, and spatial factors are associated with variation in tolerance-related responses, expecting predictors such as gender, age, education, livelihood, income, and exposure to elephants to shape this variation. We also explored spatial patterns in these responses across the study landscape, expecting proximity to protected areas (PAs) to be associated with more positive responses given the role of PAs in fostering wildlife tourism. By integrating a theory-anchored measurement model with spatial visualization, this study offers a transferable framework for characterizing how multiple, partially independent dimensions of wildlife tolerance vary across landscapes and populations, with implications for HEC mitigation in Sri Lanka and HECx strategies more broadly.

We conducted household surveys across rural communities adjacent to three protected areas in Sri Lanka’s north-central dry zone: Minneriya National Park, Kaudulla National Park, and Hurulu EcoPark (Fig. 1 ). We surveyed 151 respondents (68.2% female, 31.1% male, 0.7% other/prefer not to disclose), whose age was 49.7 ± 1.2 yrs. Respondents represented a relatively homogeneous rural population with varied engagement in agriculture (see Appendix B for full demographic summary). Only 12.6% of respondent households were engaged in the tourism sector; of those not involved, 38.5% had interest in working in tourism. Respondents involved in tourism reported higher annual household income (740,211 ± 154,498 LKR) than those that were not (402,758 ± 28,184 LKR) (Welch’s t (19.2) = − 2.15, P = 0.045; n tourism = 19, n non−tourism = 132).

Map of study area, centered around Hurulu EcoPark (HEP), Kaudulla National Park (KNP), and Minneriya National Park (MNP); protected areas are shaded. Black triangles indicate the main entrance of each protected area (park). Individual respondent locations are not shown to preserve respondent anonymity. Map was constructed via QGIS 3.26.1 Buenos Aires ( https://qgis.org ). Spatial datasets for protected areas were downloaded from the World Conservation Monitoring Centre (UNEP-WCMC, 2025); roadway and land use data were obtained from the Urban Development Authority of Sri Lanka in 2022.

Respondents reported frequent elephant sightings; 58.9% reported seeing them daily, with 70.9% of respondents seeing them at least weekly and 75.5% at least monthly. Frequent and widespread elephant presence occurred across seasons and land-use types (Appendix B ). 65.6% of respondents believed that the elephant population in the area increased over the previous 5 years; only 15.2% reported a decreasing population. Similarly, 68.5% of respondents reported that their encounters with elephants had increased in frequency compared to 5 years prior, compared to 21.2% who reported a decreased frequency. Still, 70.8% of respondents stated they wanted the elephant population in the area to increase over the next 5 years; only 18.8% thought it should decrease.

Additionally, 81.5% of respondents indicated property (including to crops) damage from elephants in the past 5 years. For these respondents, the cost of the damage accounted for 10.6% ± 1.5% (162,564 ± 23,073 LKR, median = 75,000 LKR) of a household’s annual income. Of these respondents, 28 (22.8%) applied for government compensation, and 3 of those 28 applicants (10.7%) received it. The most common reason cited for not applying (71.1%) was distrust that the government would actually provide compensation. Given the small subgroup size of compensation recipients, this proportion is reported descriptively and we did not pursue statistical inference on this subgroup.

Nine respondents (6.0%) reported having lost family members to wild elephants, while 80 respondents (53.0%) had lost a neighbor or friend in the area to an elephant. Of the 89 respondents who had lost family or neighbor(s), 63 (70.7%) wanted elephants in the area to increase over the next 5 years; this proportion was 63.8% (40 of 62) among respondents who had not lost family or neighbors. Given the small absolute counts (particularly the nine respondents who lost family members), this comparison is presented descriptively and we did not pursue statistical inference on the subgroup of respondents who lost family members. Five respondents (3.3%) indicated that at least one member of their household experienced an injury or health problem from an elephant—these included face, hand, and leg injuries.

Across respondents, perceived accuracy was highest for information coming from family and community members and lowest for institutional and expert sources, including government agencies, media, NGOs, and researchers (Appendix B ).

Information about respondent experiences and attitudes towards other wildlife is in Appendix B .

For clarity, the following results describe patterns in individual questionnaire items belonging to the constructs tested via confirmatory factor analysis (see “ Data analysis ” section). Figure 2 shows the agreement rates towards statements on values, beliefs, norms, and risk perceptions (“ Values, attitudes, norms, and perceptions ” section). Most respondents demonstrated strong values related to wildlife, expressing agreement with a statement that reflected emotional bonds with animals (57.6% agreement) and disagreement that it is acceptable to kill wildlife, even if animals pose a threat to lives (88.1% disagreement) or property (90.1% disagreement) (Fig. 2 a). However, a large percentage of respondents were neutral about whether the needs of humans take priority over wildlife protection (49.0%) or if wildlife are primarily for people to use (39.1%).

Respondent agreement with statements related to ( a ) values of wildlife, ( b ) subjective norms about elephants, ( c ) existence values and beliefs, and ( d ) risk perceptions of living with elephants. Calculated percentages exclude non-responses and “don’t know” responses.

Even so, there were prevailing trends that reflected a desire for elephants to be away from homesteads and/or agriculture (Fig. 2 b). For example, 97.3% agreed with the statement that “elephants should not be near our homes.” Additionally, 92.1% found it unacceptable for elephants to be near agricultural lands, with 71.3% agreeing that authorities should remove problematic elephants. Only 32.7% agreed they could protect their crops against elephants (Table 1 ).

There was a prevailing trend to value elephants now and for the future (92.7% agreement)—this did not appear to be related to perceived sanctity of elephants (only 17.9% agreed that wild elephants are sacred, despite 91.4% agreeing captive elephants in temples are sacred) (Fig. 2 c). The perceived value of wildlife also contrasted against the relatively rare notion that their households cancohabitatewith elephants (35.1% agreement). Notably, there was worry over the risks of living alongside elephants (Fig. 2 d): 90.7% of respondents were worried over being injured by an elephant, 81.3% were concerned over house and property damage, 80.0% worried about crop damage, and 61.7% were concerned over losing stored paddy to elephants.

Norms and beliefs around killing elephants indicated that respondents were opposed to lethal forms of elephant management (Fig. S1 ). 91.4% disagreed that they had the ability to kill an elephant, and only 26.2% agreed that the decision to shoot at an elephant was their own. Very few respondents (1.3%) reported they would shoot at an elephant near their agricultural lands. Only 25.5% agreed that people in their community shoot at elephants, and few thought people in their family (5.3%) or community (25.5%) thought that shooting at elephants was acceptable.

Feelings about living around elephants were generally negative. Most commonly, respondents indicated negative or strongly negative feelings towards each of these statements: (1) “Which describes your feelings towards elephants living in this area?” (48.3%), (2) “To have elephants in this area is for you:” (56.7%), and (3) “Living with elephants is for you:” (68.2%) (Fig. 3 ). However, respondents also shared their feelings towards other wildlife using the same statements; attitudes towards elephants were significantly less negative compared to other wildlife for each statement: first statement (χ 2 (2) = 23.635, P < 0.001), second (χ 2 (2) = 11.096, P = 0.004), and third (χ 2 (2) = 8.298, P = 0.016).

Respondent feelings towards each of three statements for elephants and other wildlife. For ease of comparison, the responses “strongly negative” and “negative” were combined, and “strongly positive” and “positive” were combined. For each statement, feelings towards elephants were significantly less negative than those for other wildlife (see main text for statistics).

Of de Silva et al.’s five tolerance-related constructs, 11 three (V1 community shooting norms, V4 utilitarian wildlife value orientations, V5 perceived agency) replicated with acceptable reliability in our confirmatory factor analysis and are the focus of the analyses below (Table 1 ); the two non-replicating constructs (V2 and V3) are omitted for reasons described in the Methods. Following the CFA-validated measurement model, we modeled six tolerance-related response outcomes—three composite scores (V1, community shooting norms; V4, utilitarian wildlife value orientations; V5, perceived agency) and three substantively important single items (Q5, emotional bond with animals; Q9, belief that the household can coexist with elephants; Q10, acceptability of shooting at elephants as a deterrent)—against a set of demographic, experiential, and spatial predictors (Table 2 ). Linear models with backward stepwise selection at α = 0.05 were used for all six outcomes; an ordinal-regression sensitivity check (proportional-odds logistic regression) was performed on the three single-item outcomes (Q5, Q9, Q10) and yielded substantively identical conclusions to the linear models. We retain the linear-model presentation here for parallel reporting across all six outcomes.

V1 (community shooting norms) was significantly associated with respondent gender and age: respondents identifying as male held less permissive norms about shooting elephants than those identifying as female (β = − 0.27, P = 0.015), and older respondents held more permissive norms (β = 0.008 per year, P = 0.020; R 2 = 0.07).

V4 (utilitarian wildlife value orientations) was associated only with change in elephant-encounter frequency. Respondents who reported that encounters had increased over the past 5 years held significantly more utilitarian wildlife value orientations than those reporting steady encounters (β = − 0.42, P = 0.004; R 2 = 0.06); respondents reporting decreasing encounters did not differ significantly from the steady group.

V5 (perceived agency) showed the largest gender effect of any model in the analysis: male respondents reported substantially lower perceived agency in their interactions with elephants than female respondents (β = − 0.46, P < 0.001; R 2 = 0.18). No other predictor was retained in this model.

Among the single-item outcomes, Q5 (emotional bond with animals) was associated with change in encounter frequency: respondents who reported either decreasing or increasing encounters over the past 5 years reported weaker emotional bonds with animals than respondents reporting steady encounters (decreasing: β = − 0.54, P = 0.022; increasing: β = − 0.62, P = 0.003; R 2 = 0.06). Q9 (coexistence belief) was associated with both encounter change and spatial proximity to the nearest PA boundary: respondents who reported increasing encounters had marginally weaker coexistence beliefs (β = − 0.48, P = 0.07), and respondents farther from the nearest PA boundary had significantly weaker coexistence beliefs (β = − 8.5 × 10 −5 per meter, P < 0.001; R 2 = 0.16). Q10 (acceptability of shooting at elephants) showed near-universal disagreement (98.7% disagreed) and no sociodemographic, experiential, or spatial predictor was retained in the model ( R 2 = 0.00).

Three substantive patterns emerge across the six models. First, gender effects appear in two distinct constructs (V1 community shooting norms and V5 perceived agency) but with different magnitudes: men report somewhat less permissive community-shooting norms and substantially lower perceived agency in their interactions with elephants. Second, recent change in elephant-encounter frequency emerges as the most pervasive predictor, appearing in V4, Q5, and Q9; in each case, respondents reporting increasing encounters held less tolerance-related responses than respondents reporting steady encounters. Third, distance to the nearest PA boundary now appears as a significant predictor in one of the six models (Q9 coexistence belief), providing an explicit link between the spatial setting and an individual-level tolerance-related response in this dataset.

Throughout this section, P BH denotes a Benjamini–Hochberg false discovery rate-adjusted P value across the 26-test omnibus family (14 closest-protected-area tests plus 12 tolerance-zone tests; see “ Data analysis ” section), and P Bonf denotes a Bonferroni-adjusted P value for a pairwise post-hoc comparison within a significant omnibus.

At the construct level, we tested whether the three composite scores (V1 community shooting norms, V4 utilitarian wildlife value orientations, and V5 perceived agency) differed across respondents grouped by their closest protected area (PA) using Kruskal–Wallis tests. V1 scores varied significantly across PAs (H(2) = 9.78, P BH = 0.028, ε 2 = 0.07); respondents nearest Kaudulla held less permissive norms about shooting elephants than respondents nearest to either Hurulu ( P Bonf = 0.010) or Minneriya ( P Bonf = 0.018), while Minneriya and Hurulu did not differ. Neither V4 utilitarian wildlife value orientations nor V5 perceived agency showed significant spatial variation (both P BH = 0.92), suggesting that abstract value orientations and perceived agency to act on elephants are spatially uniform across the study area, while behavioral norms about shooting are not.

At the item level, all three feeling-toward-elephants statements showed strong and consistent spatial variation across PAs (Cramér’s V = 0.33–0.35; all P BH < 0.001). Respondents nearest Minneriya reported the most negative feelings toward elephants living in their area (χ 2 (8) = 35.5, P BH < 0.001) and toward having elephants present (χ 2 (8) = 32.6, P BH < 0.001), differing significantly from both Hurulu ( P Bonf = 0.009 and 0.018, respectively) and Kaudulla (both P Bonf < 0.001). For feelings about living with elephants, respondents nearest Kaudulla reported the most positive responses (χ 2 (8) = 37.8, P BH < 0.001), differing significantly from both Hurulu ( P Bonf = 0.035) and Minneriya ( P Bonf < 0.001).

Three additional item-level outcomes also varied spatially. Respondents nearest Kaudulla reported less community shooting of elephants (χ 2 (8) = 27.2, P BH = 0.003, Cramér’s V = 0.31), stronger emotional bonds with animals (χ 2 (6) = 16.1, P BH = 0.038, Cramér’s V = 0.23), and stronger belief that their household could cohabit with elephants (χ 2 (8) = 20.4, P BH = 0.029, Cramér’s V = 0.26) than respondents nearest the other two PAs (post-hoc P Bonf < 0.05 in each case).

Experiential exposure to elephants also varied across PAs: respondents nearest Minneriya were significantly more likely to report that elephant-encounter frequency had changed in the preceding 5 years (χ 2 (4) = 18.9, P BH = 0.004, Cramér’s V = 0.25) than respondents nearest Kaudulla ( P Bonf = 0.007). Reports of sighting frequency ( P BH = 0.072), property or crop damage from elephants ( P BH = 0.20), and family or neighbor loss or injury ( P BH = 0.68) did not differ significantly by PA.

Together, these results describe a coherent spatial gradient. Respondents nearest Kaudulla expressed the highest tolerance-related responses across affective, behavioral, and normative dimensions and reported the lowest exposure to elephants; respondents nearest Minneriya occupied the opposite end of this gradient; respondents nearest Hurulu were typically intermediate and rarely differed significantly from either end. Importantly, this spatial structure appears in affective and behavioral responses but not in abstract value orientations (V4) or perceived agency (V5), suggesting that what varies across the landscape is how respondents feel about and act toward elephants, not how they fundamentally value wildlife or perceive their own behavioral control.

The spatial distribution of V1 (community shooting norms), V4 (utilitarian wildlife value orientations), and V5 (perceived agency) is visualized in Fig. 4 . To examine whether sociodemographic and experiential variables were associated with discrete tolerance zones, we binned the inverse-distance-weighted interpolated values for each composite into six tolerance zones (sextiles) and tested for association between four predictors (income quintiles, change in elephant-encounter frequency, property or crop damage from elephants, and closest PA) and each set of zones. After BH-FDR correction across the full 26-test omnibus family, only one tolerance-zone association reached significance: closest PA versus V1 (community shooting norms) zones (χ 2 (10) = 40.8, P BH < 0.001, Cramér’s V = 0.37). Pairwise Bonferroni-corrected post-hoc tests indicated that respondents nearest Kaudulla fell into different V1 zones than respondents nearest Minneriya ( P Bonf < 0.001) or Hurulu ( P Bonf < 0.001), while Minneriya and Hurulu did not differ. Income (quintiles), encounter change, and property or crop damage were not significantly associated with V1, V4, or V5 zones after correction; nor were any of these predictors significantly associated with V4 or V5 zones. The spatial structure of community shooting norms therefore aligns with the closest-PA effect observed at the construct level (“ Social and experiential predictors of perceptions and risks ” section), while value orientations and perceived agency remain spatially uniform across the study area.

Geospatial distributions of the three CFA-anchored composite scores represented as raster images, coded from low (dark blue) to high (yellow) values of the composite: V1 community shooting norms (left panel; higher values = less permissive of shooting), V4 utilitarian wildlife value orientations (middle panel; higher values = less utilitarian), and V5 perceived agency (right panel; higher values = lower perceived agency to shoot or harm elephants). Composite scores were derived from a confirmatory factor analysis of the structural model of de Silva et al. 11 ; the two non-replicating constructs (V2 sacredness/emotional bonds/confidence and V3 normative beliefs) are omitted (see “ Data analysis ” section). Shaded polygons indicate protected areas: Hurulu Ecopark, Kaudulla National Park, and Minneriya National Park, with black triangles indicating main entrances to these areas. Inverse-distance-weighted (IDW) interpolation parameters and projection are as described in “ Data analysis ” section.

Our study revealed complex patterns of attitudes, perceptions, and experience that interacted to form distinct social and spatial landscapes of tolerance towards elephants in this region in Sri Lanka’s dry zone. Indeed, the opinions of the respondents who participated in the survey appeared at times contradictory, perhaps reflecting the complicated struggle between a cultural heritage in Sri Lanka that values elephants and the modern reality of humans and elephants overlapping spaces more than ever before 61 . For example, these communities reported extreme intensities of conflict with elephants that incur high costs, and yet the majority also indicated belief that elephant populations in the area should increase over the coming years. Notwithstanding elephants appearing to enjoy better standing compared to other wildlife in the minds of the respondents, our analysis also suggests this dissonance may be explained by various psychosocial and demographic factors. It is important to emphasize that this affective preference for elephants over other wildlife—reflected in less-negative feelings toward elephants relative to other species—may stem from elephants’ distinctive cultural and religious standing in Sri Lanka and should not be equated with behavioral tolerance. Affective preference (liking) and behavioral tolerance (willingness to coexist without retaliation) are related but separable dimensions of human–wildlife relations: an individual may express positive feelings toward an animal yet still engage in defensive or lethal behaviors, just as someone with negative feelings may refrain from harming wildlife. The descriptive pattern reported in “ Descriptive patterns of tolerance-related response items ” section should therefore be interpreted as evidence of more positive affect toward elephants relative to other wildlife, not as direct evidence that respondents are more tolerant of elephants.

Our results support the notion that the intangible costs of HWC (e.g., fear, perceived risk) may be more important in driving tolerance for wildlife. Of the five tolerance-related constructs described by de Silva et al. 11 (their Table 1 ), three replicated with acceptable reliability in our sample—V1 community shooting norms, V4 utilitarian wildlife value orientations, and V5 perceived agency—while two (V2 sacredness/emotional bonds/confidence and V3 normative beliefs about elephant proximity) did not. Our analyses focus on six outcomes: the three replicating constructs plus three substantively important single items (Q5 emotional bond with animals, Q9 coexistence beliefs, Q10 acceptability of shooting). Each captures a distinct psychological or experiential dimension of tolerance, and all six vary systematically with respondents’ demographics (e.g., gender, age) and recent experience with elephants. We also found spatial correlates of these perceptions that contributed to rather localized (in)tolerance “hotspots” in the region, underscoring the importance of site-specific approaches to HEC mitigation 34 , 60 . Here, tolerance zones should be interpreted as descriptive summaries of spatially clustered individual responses, rather than as spatial entities that shape perceptions; they reflect where similar experiences and perceptions co-occur across the landscape. In other words, mapping tolerance zones does not imply that tolerance itself is spatially produced, but rather that social–psychological responses to elephants are unevenly distributed across space due to localized experiences, exposures, and livelihoods. Further, where tolerance has often been operationalized as a single attitudinal outcome or scalar measure, our findings support a more nuanced view in which tolerance is expressed through multiple, partially independent response domains that capture distinct psychological and experiential dimensions of living with elephants.

Because we used a similar survey instrument, we can draw direct comparisons to the previous study from a different region of Sri Lanka by de Silva et al. 11 . The partial replication of their five-factor construct structure—three constructs replicated with acceptable reliability in our north-central sample while two did not—is itself consistent with regional heterogeneity in tolerance-related constructs within Sri Lanka. Even within our relatively homogeneous respondent pool (predominantly Sinhalese Buddhist adults living in the same north-central dry zone), only one tolerance-zone association survived BH-FDR correction (closest PA × V1 zones), so very localized factors—rather than broad regional structural differences—appear to drive the spatial clustering of perceptions at the scale of this study. What broader factors at the central vs. southern Sri Lanka scale account for the differences in construct replication remains an open question; future cross-regional studies that hold survey instrument constant while varying study location should help to untangle which sociocultural, economic, or environmental factors explain regional differences in how tolerance is structured.

The apparent affective preference for elephants over other wildlife carries broader implications for human–wildlife coexistence in Sri Lanka. Several non-elephant species (e.g., wild boar, peafowl) have been declared agricultural pests, and improvised explosive devices (locally known as “hakka pattas”) set primarily for wild boar regularly injure or kill elephants as a non-target consequence. The relative acceptance that elephants currently enjoy in our study area therefore does not extend to wildlife more broadly, and the management challenge of human–elephant conflict cannot be separated from the wider landscape of human–wildlife conflict across the country. Interventions that successfully reduce HEC without also addressing the broader HWC context may displace rather than resolve the underlying conflict.

We anchored our measurement model to the construct structure published by de Silva et al. 11 , following the protocol Kyle et al. 62 recommend for measuring latent variables in conservation research. Confirmatory factor analysis of de Silva et al.’s five-factor structure on our data showed acceptable comparative fit (CFI = 0.92) but poor approximation (RMSEA = 0.14; SRMR = 0.14) along with inadmissible solutions for two factors (Heywood cases for items in V1 norms and V4 utilitarian orientations) and inadequate reliability for two further factors (V2 sacredness/bonds/confidence, McDonald’s ω = 0.23; V3 normative beliefs, ω = 0.24). Three of the five constructs (V1 community shooting norms, V4 utilitarian wildlife value orientations, V5 perceived agency) replicated with reliability values commonly considered acceptable for short scales in cross-cultural research; following de Silva et al.’s own methodological convention for constructs that show acceptable reliability but do not admit a clean latent-variable solution, we computed composite scores as the mean of reverse-coded items for these three constructs and treated them as theoretically distinct correlates of tolerance-related responses rather than as a unified “tolerance” construct. Items belonging to the non-replicating constructs were retained either as descriptive findings (where ceiling effects dominated; Q11 and Q12) or as individual observed outcomes in the regression and item-level chi-square analyses (Q5, Q9, Q10). Accordingly, references to hazard acceptance models and the theory of planned behavior in this manuscript are intended to provide interpretive context for these response patterns rather than to represent formal tests of these frameworks. The use of CFA followed by composite scoring for retained constructs is consistent with recommended practice for latent-variable analysis in conservation social science 63 . Testing TPB in human–elephant conflict settings presents well-recognized challenges, particularly where cultural taboos and legal prohibitions constrain open discussion of behaviors such as shooting or killing elephants. As noted by de Silva, et al. 11 , such conditions often result in limited response variance and heightened social desirability bias, undermining the ability to evaluate planned behavior or behavioral intentions directly. While sophisticated methods exist to address these issues, they typically require substantially larger sample sizes or alternative survey designs beyond the scope of the present study. Rather than attempting to test TPB or hazard acceptance models formally, this study focuses on characterizing how tolerance-relevant perceptions, risks, norms, and evaluations co-occur and vary across individuals and space under these constraints. In this sense, TPB and hazard acceptance models serve as interpretive frameworks for understanding response patterns, not as predictive models evaluated through confirmatory hypothesis testing.

By modeling each retained construct and substantively important single item as a separate outcome, we were able to identify how different demographic and experiential factors are associated with distinct psychological and perceptual dimensions of tolerance towards elephants (Table 2 ). Even within our relatively homogeneous sample—comprised entirely of Sinhalese Buddhist adults living in the same rural area of north central Sri Lanka—we found significant individual variation in tolerance-related responses. Gender appeared in two of the six regression models but with different magnitudes: male respondents reported somewhat less permissive norms about community shooting of elephants (V1), and substantially lower perceived agency to act on elephants (V5), than female respondents. The strong gender effect on perceived agency likely reflects the gendered division of labor in this region, where men are more often involved in crop protection and direct interactions with elephants. Although often unrecognized, women often bear a disproportionate burden of the negative impacts that result from HWC 64 , 65 , 66 , 67 , 68 , and women may feel the costs more significantly from stronger laws that restrict use of PAs and/or perceived risk from living alongside wildlife 63 , 64 . Older respondents may have experienced a longer history of negative interactions with elephants and/or a bigger shift in the nature of these interactions over time, negatively impacting perceived opportunity for HECx 31 . Additionally, younger respondents may have lower costs and/or larger benefits associated with elephant presence 65 . Similarly, respondents involved in agriculture and/or those with lower education levels (and thus more limited economic opportunity) may reflect increased vulnerability to crop damage and financial losses, leading to less favorable perceptions 44 , 66 . Higher income earners were more likely to exhibit positive sentiments towards HECx, while those who reported an increased frequency of elephant encounters or experienced crop damage held less favorable attitudes, highlighting the significant impact of direct negative experiences on tolerance levels 67 . Interestingly, our analyses identified contrasting results between V1 and V2 (perceptions about HWC management) in relation to gender and education. In contrast to their attitudes towards HECx, female respondents and those with collegiate- or university-level education had more positive perceptions about HWC management. Other studies have demonstrated women may be more willing and/or motivated to address HWC 68 , 69 , 70 , 71 , and it is possible that higher education is associated with greater awareness of, or trust in, HWC management strategies. Changing frequencies of elephant encounters (both increasing and decreasing) had a negative association with perceptions about the efficacy of wildlife management, suggesting that increased interactions with wildlife may erode trust in mitigation efforts and/or decreased wildlife sightings may indicate wildlife protection strategies are inappropriately administered. These findings underscore the complex and sometimes contradictory nature of the factors influencing tolerance towards wildlife.

Recent change in elephant-encounter frequency emerged as the most pervasive predictor across our six regression models, appearing as a significant driver of utilitarian wildlife value orientations (V4), emotional bond with animals (Q5), and coexistence belief (Q9). In each case, respondents reporting that their encounters with elephants had increased over the past 5 years held less positive tolerance-related responses than those reporting steady encounters. This pattern is consistent with the broader literature linking direct experience with wildlife to attitudes towards them 78 , 79 , 80 . While urban communities in Sri Lanka (i.e., those that live outside of elephant ranges) tend to exhibit more sympathetic attitudes towards elephant conservation 72 , rural communities may display more resentment. Lower education levels, which often correlate with fewer economic opportunities, were associated with perceived social norms favoring lethal control of elephants 73 , 74 . Given that male respondents in this area are more likely to be farmers regularly involved in crop protection, it is perhaps unsurprising that they also reported more prevalent community norms about shooting elephants and more negative perceptions towards HWC management. Individual and community perceptions around shooting elephants deserve further investigation; respondents may have been hesitant to discuss this activity, as shooting elephants is illegal in Sri Lanka. Despite this, respondents appeared to place strong societal or cultural value on elephants (Fig. 2 c) that may counteract tendencies towards lethal control. Ultimately, the balance between protecting elephants and resorting to lethal measures depends on a sustainable level of tolerance, shaped by a myriad of sociodemographic factors.

Our study also provides a novel spatial perspective on tolerance towards elephants, revealing significant geographic variation in perceptions across the study area. Respondents closer to Kaudulla National Park exhibited more positive tolerance-related response patterns (including more positive feelings towards elephants) compared to those near Hurulu Ecopark and Minneriya National Park. Respondents nearest Minneriya also more frequently reported that encounters with elephants had changed in the recent past, an experiential difference that may help explain the more negative feelings towards elephants and the lower confidence in human–elephant coexistence (HECx) that respondents in this area expressed. A complementary zone-level analysis based on inverse-distance-weighted (IDW) surfaces of perceptions and risks related to HECx (V1) further supported this pattern: respondents in zones with the highest interpolated V1 scores were disproportionately drawn from areas nearest Kaudulla, while respondents in the lowest-V1 zones were disproportionately drawn from areas nearest Minneriya (χ 2 (10) = 40.8, P BH < 0.001, Cramér’s V = 0.37). This spatial heterogeneity in tolerance levels highlights the importance of considering the geographic context of human–wildlife interactions, an aspect that, while increasingly recognized, is not always explicitly addressed in studies with a psychosocial lens toward HWC 59 , 75 . Research has demonstrated the influence of spatial factors such as proximity to PAs, habitat suitability, and landscape composition on HWC and attitudes towards wildlife e.g., 53 , 76 , 77 , 78 , 79 , 80 , 81 , 82 , 83 . Our approach integrates exploratory spatial visualization with a latent variable framework to describe how empirically derived response domains vary across a heterogeneous social–ecological landscape. Rather than testing spatial mechanisms or modeling spatial dependence, this analysis is intended to identify spatial patterning and co-location of social responses that may help generate hypotheses for future, more explicitly spatially structured analyses 84 . Still, our integrative approach offers a more nuanced understanding of spatial heterogeneity of tolerance: zones of higher interpolated V1 scores were not distributed evenly across the study area but instead clustered nearest Kaudulla National Park, suggesting that spatially structured experiences and perceptions, rather than individual-level attributes alone, may be most relevant for targeting interventions to promote HECx. Future research could build upon this framework by examining the influence of finer-scale habitat features, land-use patterns, and the effectiveness of site-specific mitigation strategies on the spatial distribution and multidimensional nature of wildlife tolerance.

The spatial analyses presented here are descriptive rather than inferential; we did not test for spatial autocorrelation, explicitly model spatial dependence, or validate spatial predictions of social responses. As such, the spatial patterns observed should not be interpreted as evidence of spatial drivers or mechanisms of tolerance, but rather as exploratory visualizations of how perceptions are distributed across the study landscape. Recent advances in spatial social science and conservation psychology demonstrate more rigorous approaches for spatializing social and psychological variables, including methods that explicitly model spatial dependence, quantify uncertainty, account for response bias, and generate spatial predictions [e.g., Carroll, et al. 32 ; Hull, et al. 84 ; Manfredo, et al. 59 ; Vasudev and Goswami 34 ]. These approaches offer promising avenues for extending the present work, particularly by linking spatial patterns of tolerance to underlying landscape features, governance contexts, or exposure gradients in a confirmatory framework.

The question remains how to most effectively apply the results of this study (and those that are similar) to mitigate HEC and promote HECx. The answer to this question is complicated by frequent disagreement between so-called “expert” and “non-expert” stakeholder groups about solutions to HEC 74 . For example, most respondents in our study (who may also be considered non-experts because they do not work professionally in Asian elephant conservation) expressed disbelief that they could cohabit with elephants on their land and that authorities should remove (i.e., translocate) problematic elephants. However, various studies unequivocally suggest that translocating elephants involved in HEC is ineffective and actually may escalate conflict 93 , 94 , 95 . Other strategies like switching agricultural activities (e.g., planting crops that are unpalatable to elephants) often enjoy support among conservationists 85 , 86 but may be unfavored by local communities 56 . Discrepancy between experts and non-experts may be further reinforced due to lack of trust in scientific authorities, government officials, and mainstream media sources as they relate to information about elephants, as evidenced by the results of this study. Therefore, while outreach efforts should be evidence-based, they should be supported by local community stakeholders. Notably, many respondents to our survey indicated an eagerness to be involved in local tourism activities, offering potential alternative livelihoods that are compatible with higher levels of tolerance for elephants. Among the sample population in this study, respondents already involved in tourism earned significantly higher household incomes and had more positive perceptions about HECx. Further, areas of positive intensity were concentrated near the entrance gates at the nearby PAs for most of the latent variables that contributed to tolerance for elephants, suggesting a potential spatial association between tourism-related activity and tolerance, although distance to PA boundaries was used analytically to avoid conflating entrance access with uneven spatial exposure. Tourism-based interventions should be implemented (perhaps in tandem with other strategies) in ways that are locally based and integrate characteristic attitudes, beliefs, and experiences prevalent in a specific community.

Several limitations should be noted when interpreting these results. First, although the survey items asking respondents what they wanted to happen to elephant numbers (e.g., Q4.9, Q5.9 in the questionnaire) were phrased in Sinhala as preferences rather than expectations, we cannot fully exclude the possibility that some respondents interpreted these items as predictions of what they expected to occur. The counter-intuitive pattern in which the majority of respondents who had lost family or neighbors still preferred an increase in elephant numbers should therefore be interpreted with this potential interpretive ambiguity in mind. Second, distances to PA boundaries were calculated as Euclidean (straight-line) distances; actual human and elephant movements follow paths shaped by topography, vegetation, infrastructure, and behavioral preferences, and may diverge meaningfully from Euclidean distance. We retained Euclidean distance because the road-network and elephant-corridor data needed to construct a defensible friction surface are not available at sufficient resolution for the study area, and because elephants do not follow human road networks; nevertheless, this is an acknowledged simplification. Third, the distribution of respondents across the three nearest PAs was uneven (84 closest to Minneriya, 43 to Hurulu, and 24 to Kaudulla), reflecting local population density and field-access constraints rather than selection bias. The smaller Kaudulla and Hurulu samples may have reduced statistical power for some between-PA comparisons, and corresponding results should be interpreted accordingly. Fourth, several descriptive comparisons (e.g., reported preferences among the nine respondents who had lost family members to elephants, or the three respondents who received compensation) involve very small subgroups and are reported descriptively rather than inferentially. Finally, the spatial analyses reported here are descriptive and exploratory: the IDW interpolation does not test or model spatial mechanisms, and the resulting tolerance zones should be read as visualizations of where similar individual responses cluster across the landscape rather than as evidence of spatial drivers of tolerance.

Our study demonstrates the complexity of HECx, revealing a landscape of tolerance shaped by intricate interactions between demographics, psychosocial factors, individual experiences, and spatial variables. These interactions highlight the limitations of applying a single framework (e.g., hazard acceptance model, TPB) in isolation and demonstrate the value of an approach such as ours in integrating sociodemographic constructs with spatial analysis to provide a more holistic understanding of tolerance. While respondents expressed positive sentiments towards elephants, a prevailing skepticism in the possibility of coexistence reveals that perceived risks, negative experiences, and prevailing social norms can outweigh these positive attitudes. Specifically, the strong influence of perceived pressure against HWC management and community norms favoring lethal control, especially among older and less educated respondents, aligns with theories of social influence on behavioral intentions 8 , 87 , 88 . This suggests that interventions must address not only individual attitudes but also the broader social context that shapes tolerance. The spatial heterogeneity we observed, with tolerance varying significantly across relatively short distances, emphasizes the importance of tailoring HEC mitigation strategies to local conditions and community-specific needs, incorporating an understanding of how social norms and beliefs are spatially patterned. For instance, given that communities closer to Kaudulla National Park exhibited higher tolerance (e.g., from tourism revenue or other site-specific factors), promoting sustainable, localized economic benefits derived from elephant presence and based in the local cultural context could foster greater acceptance 89 , 90 . Such strategies, however, must be coupled with efforts to mitigate the tangible costs of HEC 3 , 4 , 91 , such as crop damage and property loss, which were significant drivers of intolerance in other areas. The challenge is particularly pressing given the endangered status of Asian elephants, the severe consequences of HEC for both humans and elephants, and the increasing amount and intensity of overlap between human activity and wildlife habitats. This study contributes to a growing body of literature demonstrating that effective human–wildlife coexistence requires a nuanced understanding of the interplay between individual psychology, social dynamics, and spatial ecology. By integrating these perspectives, research can inform the development of more targeted and ultimately more successful conservation interventions, not only for elephants in Sri Lanka, but also for a wide range of HWC scenarios globally.

The target population for this study consisted of adult residents (≥ 18 years old) living and/or working in rural communities within approximately 20 km of at least one of three PAs: Minneriya National Park, Kaudulla National Park, and Hurulu EcoPark (Fig. 1 ). Minneriya and Kaudulla host the largest elephant gathering in Sri Lanka 103 , 104 , and elephants regularly move between these parks, traversing landscapes dominated by agriculture and tourism. Minneriya and Kaudulla are managed by the Department of Wildlife Conservation (DWC), whereas Hurulu is overseen by the Forestry Department (FD) of Sri Lanka. Surveys were conducted between January and April 2024, a period characterized by seasonal flooding in Minneriya and Kaudulla and heightened elephant movement outside PAs, resulting in elevated rates of human–elephant conflict 42 , 43 .

The sampled communities are characterized by mixed livelihoods dominated by small-holder agriculture, with variable engagement in tourism and other non-agricultural occupations. Respondents were limited to community members who lived and/or worked within the study area. The unit of analysis was the individual respondent, although several survey items pertained to household-level experiences (e.g., crop damage, property loss, and injuries caused by elephants). To minimize non-independence, participation was restricted to a maximum of one respondent per household or workplace, and respondents were instructed to answer household-level questions based on shared household experiences. Respondents were reached by vehicle or on foot, and all eligible individuals (≥ 18 years old) were invited to participate voluntarily under informed consent.

While a precise census of the total population within the study area is not available, the surveyed respondents span multiple administrative divisions that comprise several thousand residents in Sri Lanka’s north-central dry zone. Our sample of 151 respondents therefore represents a subset of individuals living in areas with regular exposure to elephants, rather than a statistically representative sample of the entire regional population. Face-to-face interviews were conducted in Sinhala by three native Sinhala-speaking enumerators working in teams of two or three, and all respondents were assured of confidentiality and their right to withdraw at any time. The methodology received ethical clearance from the Postgraduate Institute of Science, University of Peradeniya (CEC_PGIS_2024_02); all research was performed in accordance with relevant regulations and informed consent was obtained from all participants.

We utilized an adapted subset of the instrument designed by de Silva, et al. 11 . Our questionnaire consisted of 82 close-ended questions spanning six sections: (1) respondent demographics, (2) livelihoods, (3) agriculture activity, (4) experiences with elephants, (5) experiences with other wildlife, and (6) social psychological variables (values, beliefs, norms, and risk perceptions). The complete questionnaire is provided in Appendix ( A ) “ Respondent demographics ” and “ Household livelihoods ” sections consisted of categorical and free-response questions about demographics, household composition, transportation, phone access, education, and occupation. Section “ Agriculture/livestock ” was completed by households engaged in agriculture and included questions about cultivation practices, harvest seasonality, and crop protection. Section “ Experiences with elephants ” pertained to experiences with elephants, including: frequency, seasonality, and location of sightings/encounters; experience with HEC, including deterrents; and knowledge about elephants. Section Experiences with other wildlife pertained to experience with other wildlife—including peacocks, primates, boar, deer, and leopards—mirroring the questions from the previous section; as this section was outside the scope of the study, we present results from “ Experiences with other wildlife ” section in Appendix ( B ) Finally, “ Values, attitudes, norms, and perceptions ” section included items to survey social-psychological concepts related to values, attitudes, norms, and perceptions related to HEC and HECx; each item consisted of a statement for which respondents indicated agreement or perceptions using a five-point Likert scale.

All analyses were conducted in R version 4.4.0 92 , with data organization occurring via the tidyverse package 93 . Non-responses (including “I don’t know” responses) were excluded from all summary calculations. In the results section, descriptive statistics are represented by mean values accompanied by standard errors (mean ± SE), where appropriate.

We tested the structural model published by de Silva et al. 11 (their Table 1 ) on our data using confirmatory factor analysis (CFA) implemented in the lavaan package 94 . Indicators were treated as ordered-categorical Likert responses and the model was estimated using weighted least squares with mean and variance adjustment (WLSMV) on the polychoric correlation matrix, with the 29 missing responses imputed in advance via predictive mean matching using the mice package. Item-to-factor assignments followed de Silva et al.’s Table 1 : V1 community shooting norms (items Q13, Q14, Q15); V2 sacredness, emotional bonds, and confidence (Q5, Q7, Q8, Q9, Q19); V3 normative beliefs about elephant proximity (Q10, Q11, Q12); V4 utilitarian wildlife value orientations (Q2, Q3, Q4); and V5 perceived agency (Q17, Q18, Q20). No cross-loadings or error covariances were specified a priori. We report robust scaled fit indices (CFI, TLI, RMSEA with 90% confidence interval, SRMR), standardized loadings, inter-factor correlations, and reliability per factor (Cronbach’s α including ordinal-categorical α, McDonald’s ω, and average variance extracted). The CFA showed acceptable comparative fit (CFI = 0.92; TLI = 0.90) but poor approximation (RMSEA = 0.14, 90% CI 0.13–0.16; SRMR = 0.14) along with inadmissible solutions (Heywood cases) for items in V1 and V4. Reliability values were: V1 (ordinal α = 0.70, ω = 0.51, AVE = 0.60); V2 (ordinal α = 0.35, ω = 0.23, AVE = 0.21; unacceptable); V3 (ordinal α = 0.40, ω = 0.24, AVE = 0.33; unacceptable); V4 (ordinal α = 0.77, ω = 0.74, AVE = 0.69); V5 (ordinal α = 0.72, ω = 0.58, AVE = 0.52). Three of de Silva et al.’s five constructs (V1, V4, V5) replicated with reliability values commonly considered acceptable for short scales in cross-cultural research 95 , 96 ; two (V2 and V3) did not, for substantively interpretable reasons described in the Discussion.

Following de Silva et al.’s 11 own methodological convention for constructs that show acceptable reliability but do not admit a clean latent-variable solution, we computed composite scores as the mean of reverse-coded items for the three retained constructs (hereafter V1_score, V4_score, V5_score) and used these composite scores as the primary measurement model in all subsequent analyses. Items belonging to the non-replicating constructs (Q5, Q7, Q8, Q9, Q10, Q11, Q12, Q19) were retained either as descriptive findings (where ceiling effects dominated; Q11 and Q12) or as individual observed outcomes in the regression analyses (Q5, Q9, Q10) and the item-level chi-square analyses in “ Spatial patterns of tolerance towards elephants ” section (Q5, Q9, Q10, Q15, Q25–Q27). The exploratory factor analysis (EFA) we performed on the same data and described below is retained as a robustness check; an overall Kaiser–Meyer–Olkin (KMO) statistic of 0.78 and Bartlett’s test of sphericity (χ 2 (325) = 1583.475, P < 0.001) indicated that the correlation matrix was suitable for factor analysis, and the EFA yielded four data-driven factors explaining 46.6% of variance. However, in accordance with the protocol recommended by Kyle et al. 62 for measuring latent variables in conservation research—identify a theoretical framework first, operationally define the constructs, then link them to observable measures through theory-based instrumentation—we use the CFA-anchored composites rather than the EFA factor scores as the primary measurement model in all subsequent analyses.

To address inflation of Type I error across the family of statistical tests reported in “ Spatial patterns of tolerance towards elephants ” section, we applied the Benjamini–Hochberg false discovery rate (FDR) procedure to the full 26-test omnibus family—14 closest-protected-area (PA) tests (chi-square and Kruskal–Wallis examining differences across the three closest PAs) plus 12 tolerance-zone tests (four predictors × three composite zones derived from the inverse-distance-weighted surfaces; see below). Within each significant omnibus, pairwise post-hoc comparisons among the three PAs were Bonferroni-corrected with adjusted α = 0.0167. BH-adjusted P values are denoted P BH and Bonferroni-adjusted P values are denoted P Bonf in “ Spatial patterns of tolerance towards elephants ” section and in Table S2 . The full omnibus and post-hoc results are provided in Table S2 of the Supplementary Material.

We initially conducted exploratory factor analysis (EFA) on these same items using parallel analysis and oblimin rotation; the EFA yielded four data-driven factors but produced loadings that mixed theoretically distinct construct types (e.g., affective and risk-perception items loading on the same factor), prompting the shift to the confirmatory analysis described above. Full EFA results are provided in Table S1 ; per-item Kaiser–Meyer–Olkin (KMO) measures of sampling adequacy from the EFA diagnostic stage are reported in Table S3 .

Because tolerance was conceptualized as a multidimensional construct reflected across multiple response domains, we did not specify a single dependent variable representing tolerance. Instead, we modeled each empirically derived response domain as a separate outcome, allowing us to examine how different non-psychological factors (e.g., demographics, experiences, spatial context) are associated with distinct dimensions of tolerance-related responses. Following the CFA-validated measurement model, we modeled six tolerance-related response outcomes [three CFA-anchored composite scores (V1 community shooting norms; V4 utilitarian wildlife value orientations; V5 perceived agency) and three substantively important single items (Q5 emotional bond with animals; Q9 belief that the household can cohabitate with elephants; Q10 acceptability of shooting at elephants as a deterrent)] as separate dependent variables in linear models against a common set of demographic, experiential, and spatial predictors. Visual inspection of normalized quantile–quantile plots indicated approximate normality of the composite scores and acceptable residual distributions for the linear-model treatment of the ordinal single items; an ordinal-regression sensitivity check (proportional-odds logistic regression) on Q5, Q9, and Q10 yielded substantively identical conclusions to the linear models and is reported alongside in Table 2 . For each outcome, we constructed a full model consisting of the additive effects of each of the following fixed n = 128, or 9.4% of responses for the modelling dataset) were imputed using the mice package as described above 97 . In a stepwise fashion, we excluded non-significant ( P > 0.050) effects using a modified F test, resulting in a final model for each of the latent variables. We assessed collinearity among variables in the final models using variance inflation factors (VIF) via the olsrr package 98 ; all variables in each final model resulted in VIF < 5, indicating acceptable collinearity between variables 99 .

Using location data, we mapped each survey point in QGIS 3.26.1 using EPSG:32,644 (WGS 84/UTM Zone 44 N) with park boundary shapefiles from the UN Environment Programme World Conservation Monitoring Centre 100 . Respondent locations were assigned to the nearest PA (Minneriya, Kaudulla, or Hurulu) using Euclidean distance from each household to park boundaries. Distance to PA boundaries was used rather than distance to PA entrances because proximity to a boundary more consistently reflects exposure to elephants across the landscape, whereas respondents grouped by a single entrance can vary widely in actual distance from that entrance, introducing additional spatial heterogeneity due to spatial aggregation choices 101 , 102 . The survey itself was not spatially stratified by park; rather, the spatial grouping was determined post hoc based on location. Of the 151 respondents, 84 were closest to Minneriya, 43 to Hurulu, and 24 to Kaudulla. To assess how responses varied across the landscape, we conducted a chi-square test for homogeneity, comparing the frequency of responses across the five-point Likert scale by closest park with Haldane-Anscombe corrections 103 , 104 and significant results interpreted with Bonferroni exploratory analyses 105 , 106 using an adjusted = 0.010 threshold 107 , 108 .

To visualize spatial trends for each CFA-anchored composite score (V1, V4, V5), we interpolated survey response value points using inverse distance weighting (IDW) 101 , 102 . We selected IDW because our analytical framework groups respondents by proximity to the nearest protected area, and IDW captures the distance-decay assumption underlying that framework: neighboring observations exert greater influence on local interpolated values than distant observations. Alternative methods such as ordinary kriging or thin-plate splines were considered but are designed to model spatial autocorrelation or smoothing across space, neither of which is the central concern of our distance-based tolerance-zone framework. IDW was therefore implemented in QGIS, generating rasters with a resolution of 75 × 75 grid cells, which were subsequently vectorized to create “tolerance zones” in QGIS. Here, tolerance zones refer to areas where individual respondents with similar tolerance-related response profiles are spatially co-located, based on mapped factor scores, rather than discrete spatial units that exert causal influence on attitudes. The raster extent spanned 80.636°–80.997° E and 7.858°–8.226° N, resulting in a pixel size of approximately 0.00488° × 0.00491° (~ 0.53 km × ~ 0.55 km). This resolution was selected to visualize community-scale variation in tolerance rather than fine scale household-level differences while balancing smooth visualization with over-smoothing.

The anonymized dataset used in this study is available from the corresponding author upon reasonable request.

Nyhus, P. J. Human–wildlife conflict and coexistence. Annu. Rev. Environ. Resour. 41 , 143–171. https://doi.org/10.1146/annurev-environ-110615-085634 (2016).

Woodroffe, R., Thirgood, S. & Rabinowitz, A. People and Wildlife: Conflict or Coexistence? 497 (Cambridge University Press, 2005).

Barua, M., Bhagwat, S. A. & Jadhav, S. The hidden dimensions of human-wildlife conflict: Health impacts, opportunity and transaction costs. Biol. Conserv. 157 , 309–316. https://doi.org/10.1016/j.biocon.2012.07.014 (2013).

Braczkowski, A. R. et al. The unequal burden of human-wildlife conflict. Commun. Biology . 6 , 182. https://doi.org/10.1038/s42003-023-04493-y (2023).

IUCN. IUCN SSC Guidelines on Human-wildlife Conflict and Coexistence 1 edn (IUCN, 2023).

Pooley, S., Bhatia, S. & Vasava, A. Rethinking the study of human–wildlife coexistence. Conserv. Biol. 35 , 784–793. https://doi.org/10.1111/cobi.13653 (2020).

Article PubMed PubMed Central Google Scholar

Carter, N. H. & Linnell, J. D. C. Co-adaptation is key to coexisting with large carnivores. Trends Ecol. Evol. 31 , P575–578. https://doi.org/10.1016/j.tree.2016.05.006 (2016).

Kansky, R., Kidd, M. & Knight, A. T. A wildlife tolerance model and case study for understanding human wildlife conflicts. Biol. Conserv. 201 , 137–145. https://doi.org/10.1016/j.biocon.2016.07.002 (2016).

Roy, A., Dash, S. K. & Sathyakumar, S. A combination of cultural values and economic benefits promote tolerance towards large mammals in a hotspot of human-wildlife conflicts in eastern India. Hum. Ecol. 50 , 321–329. https://doi.org/10.1007/s10745-022-00306-8 (2022).

Saif, O., Kansky, R., Palash, A., Kidd, M. & Knight, A. T. Costs of coexistence: Understanding the drivers of tolerance towards Asian elephants Elephas maximus in rural Bangladesh. Oryx 54 , 603–611. https://doi.org/10.1017/S0030605318001072 (2019).

de Silva, S. et al. Experiences and emotional responses of farming communities living with Asian elephants in southern Sri Lanka. Trees Forests People . 14 , 100441. https://doi.org/10.1016/j.tfp.2023.100441 (2023).

Thekaekara, T., Bhagwat, S. A. & Thornton, T. F. Coexistence and culture: Understanding human diversity and tolerance in human-elephant interactions. Front. Conserv. Sci. 2 , 735929. https://doi.org/10.3389/fcosc.2021.735929 (2021).

Knox, J., Ruppert, K., Frank, B., Sponarski, C. C. & Glikman, J. A. Usage, definition, and measurement of coexistence, tolerance and acceptance in wildlife conservation research in Africa. Ambio 50 , 301–313. https://doi.org/10.1007/s13280-020-01352-6 (2021).

Slagle, K. & Bruskotter, J. T. Human–Wildlife Interactions: Turning Conflict into Coexistence Conservation Biology (eds Frank, B., Glikman, J. A. & Marchini, S.) Ch. 5, 85–106 (Cambridge University Press, 2019).

Bruskotter, J. T. & Wilson, R. S. Determining where the wild things will be: Using psychological theory to find tolerance for large carnivores. Conserv. Lett. 7 , 158–165. https://doi.org/10.1111/conl.12072 (2014).

Manfredo, M. J. et al. The changing sociocultural context of wildlife conservation. Conserv. Biol. 34 , 1549–1559. https://doi.org/10.1111/cobi.13493 (2020).

Ajzen, I. & Fishbein, M. Attitude–behavior relations: A theoretical analysis and review of empirical research. Psychol. Bull. 84 , 888–918. https://doi.org/10.1037/0033-2909.84.5.888 (1977).

Kraus, S. J. Attitudes and the prediction of behavior: A meta-analysis of the empirical literature. Personal. Soc. Psychol. Bull. 21 , 58–75. https://doi.org/10.1177/0146167295211007 (1995).

Zimmermann, A. & Baruah, M. Human-Wildlife Conflict: Global Policy and Perception Insights ; https://www.worldbank.org/en/programs/global-wildlife-program/brief/human-wildlife-conflict-global-policy-and-perception-insights#2 (2023).

Abas, A., Rahman, A. H. A., Md Fauzi, T. A. H. T. & Yusof, A. H. M. A bibliometric review of global research on the human-wildlife conflicts. Front. Environ. Sci. 12 , 1517218. https://doi.org/10.3389/fenvs.2024.1517218 (2025).

Linkie, M., Dinata, Y. & Nofrianto, A. & Leader-Williams, N. Patterns and perceptions of wildlife crop raiding in and around Kerinci Seblat National Park, Sumatra. Anim. Conserv. 10 , 127–135. https://doi.org/10.1111/j.1469-1795.2006.00083.x (2007).

Gillingham, S. & Lee, P. C. People and protected areas: A study of local perceptions of wildlife crop-damage conflict in an area bordering the Selous Game Reserve, Tanzania. Oryx 37 , 316–325. https://doi.org/10.1017/S0030605303000577 (2003).

Songhurst, A. Probing the complexities of actual and perceived levels of human-elephant conflict in the Okavango. Botsw. Divers. . 15 , 890. https://doi.org/10.3390/d15080890 (2023).

Locke, P. Explorations in ethnoelephantology: Social, historical, and ecological intersections between Asian elephants and humans. Environ. Society: Adv. Res. 4 , 79–97. https://doi.org/10.3167/ares.2013.040106 (2013).

Sukumar, R. The Story of Asia’s Elephants (The Marg Foundation, 2011).

Curtin, S. What makes for memorable wildlife encounters? Revelations from ‘serious’ wildlife tourists. J. Ecotourism . 9 , 149–168. https://doi.org/10.1080/14724040903071969 (2010).

Laws, E., Scott, N., Font, X., & Koldowski, J. The elephant tourism business. Oxfordshire, UK: CAB International. 293 (2021).

Guarnieri, M. et al. Effects of climate, land use, and human population change on human–elephant conflict risk in Africa and Asia. Proc. Natl. Acad. Sci. 121 , e2312569121. https://doi.org/10.1073/pnas.2312569121 (2024).

Perera, B. M. A. O. The human-elephant conflict: A review of current status and mitigation methods. Gajah 30 , 41–52 (2009).

Xu, H., Jiang, L. & Liu, Y. Mapping the potential distribution of Asian elephants: Implications for conservation and human–elephant conflict mitigation in South and Southeast Asia. Ecol. Inf. 80 , 102518. https://doi.org/10.1016/j.ecoinf.2024.102518 (2024).

van de Water, A. & Matteson, K. Human-elephant conflict in western Thailand: Socio-economic drivers and potential mitigation strategies. PLoS One . 13 , e0194736. https://doi.org/10.1371/journal.pone.0194736 (2018).

Article CAS PubMed PubMed Central Google Scholar

Carroll, S. L. et al. A spatially explicit assesment of factors shaping attitudes toward African elephant conservation. Conserv. Biol. 39 , e14408. https://doi.org/10.1111/cobi.14408 (2025).

de Pinho, J. R., Grilo, C., Boone, R. B., Galvin, K. A. & Snodgrass, J. G. Influence of aesthetic appreciation of wildlife species on attitudes towards their conservation in Kenyan agropastoralist communities. PLoS One . 9 , e88842. https://doi.org/10.1371/journal.pone.0088842 (2014).

Article CAS PubMed PubMed Central ADS Google Scholar

Vasudev, D. & Goswami, V. R. A Bayesian hierarchical approach to quantifying stakeholder attitudes toward conservation in the presence of reporting error. Conserv. Biol. 34 , 299–532. https://doi.org/10.1111/cobi.13392 (2020).

Vogel, S. M. et al. Identifying sustainable coexistence potential by integrating willingness-to-coexist with habitat suitability assessments. Biol. Conserv. 279 , 109935. https://doi.org/10.1016/j.biocon.2023.109935 (2023).

Bhatia, S., Redpath, S. M., Suryawanshi, K. & Mishra, C. Beyond conflict: Exploring the spectrum of human–wildlife interactions and their underlying mechanisms. Oryx 54 , 621–628. https://doi.org/10.1017/S003060531800159X (2020).

Glikman, J. A. et al. Coexisting with different human-wildlife coexistence perspectives. Front. Conserv. Sci. 2 , 703174. https://doi.org/10.3389/fcosc.2021.703174 (2021).

Carter, N. H. & Linnell, J. D. C. Building a resilient coexistence with wildlife in a more crowded world. PNAS Nexus . 2 , pgad030. https://doi.org/10.1093/pnasnexus/pgad030 (2023).

Williams, C. et al. Elephas Maximus (The IUCN Red List of Threatened Species, 2020).

Hendavirtharana, W., Dissanayake, S., Silva, M. & Santiapillai, C. The survey of elephants in Sri Lanka. Gajah 12 , 1–30 (1994).

Dissanayake, S. R. B. et al. The First National Survey of Elephants in Sri Lanka (Ringling Bros. Center for the Study of Asian Elephant AT Rajarata University of Sri Lanka, Mihintale, Sri Lanka, 2012).

LaDue, C. A., Eranda, I., Jayasinghe, C. & Vandercone, R. P. G. Mortality patterns of Asian elephants in a region of human–elephant conflict. J. Wildl. Manag. 85 , 794–802. https://doi.org/10.1002/jwmg.22012 (2021).

Gunawansa, T. D., Perera, K., Apan, A. & Hettiarachchi, N. K. The human-elephant conflict in Sri Lanka: History and present status. Biodivers. Conserv. 32 , 3025–3052. https://doi.org/10.1007/s10531-023-02650-7 (2023).

Köpke, S. et al. Factors driving human-elephant conflict: Statistical assessment of vulnerability and implications for wildlife conflict management in Sri Lanka. Biodivers. Conserv. 33 , 3075–3101. https://doi.org/10.1007/s10531-024-02903-z (2024).

Fernando, P. & Pilapitiya, S. National Action Plan for the Mitigation of Human-Elephant Conflict Vol. 59 (Battaramulla, 2020).

de la Torre, J. A. et al. There will be conflict—agricultural landscapes are prime, rather than marginal, habitats for Asian elephants. Anim. Conserv. 24 , 720–732. https://doi.org/10.1111/acv.12668 (2021).

Fernando, P. & Pastorini, J. Whither the science in wildlife management? Anim. Conserv. 24 , 735–737. https://doi.org/10.1111/acv.12743 (2021).

Fernando, P. et al. Perceptions and patterns of human-elephant conflict in old and new settlements in Sri Lanka: Insights for mitigation and management. Biodivers. Conserv. 14 , 2465–2481. https://doi.org/10.1007/s10531-004-0216-z (2005).

Fernando, P., De Silva, M. K. C. R., Jayasinghe, L. K. A., Janaka, H. K. & Pastorini, J. First country-wide survey of the endangered Asian elephant: Towards better conservation and management in Sri Lanka. Oryx 55 , 46–55. https://doi.org/10.1017/S0030605318001254 (2021).

Anuradha, J. M. P. N., Fujimura, M., Inaoka, T. & Sakai, N. The role of agricultural land use pattern dynamics on elephant habitat depletion and human-elephant conflict in Sri Lanka. Sustainability 11 , 2818. https://doi.org/10.3390/su11102818 (2019).

Santiapillai, C. et al. An assessment of the human-elephant conflict in Sri Lanka. Ceylon J. Sci. (Biological Sciences) . 39 , 21–33 (2010).

Premarathne, N. M. K. C., DIssanayake, S. T. M., de Silva, S., Weerathunga, U. S. & Kumara, T. V. P. Impact of changed rainfall patterns due to climate change and usage of available weather information by communities who face human elephant conflict (HEC) in Udawalawe, Sri Lanka. Proc. Int. Conf. Clim. Change . 1 , 40–50. https://doi.org/10.17501/iccc.2017.1105 (2017).

LaDue, C. A., Farinelli, S. M., Eranda, I., Jayasinghe, C. & Vandercone, R. P. G. The influence of habitat changes on elephant mortality associated with human–elephant conflict: Identifying areas of concern in the north central dry zone of Sri Lanka. Sustainability 13 , 13707. https://doi.org/10.3390/su132413707 (2021).

Jayakody, S. et al. Maxent modeling for predicting the potential distribution of human-elephant conflict risk in Sri Lanka. Appl. Geogr. 173 , 103447. https://doi.org/10.1016/j.apgeog.2024.103447 (2024).

Köpke, S. et al. Human-elephant conflict in the Sri Lankan dry zone: Investigating social and geographical drivers through field-based methods. GeoJournal . 88 , 5153–5172 (2023). https://doi.org/10.1007/s10708-023-10913-7

Suresh, K., Wilson, C., Quayle, A., Managi, S. & Khanal, U. Are farmers willing to accept compensation from tourism revenue for elephant crop damage and coexistence support? Evidence from Sri Lanka. Ecol. Econ. 224 , 108300. https://doi.org/10.1016/j.ecolecon.2024.108300 (2024).

Banerjee, S., Nayak, D. & Sinha, A. Adivasi (Tea Tribe) worldviews of living close to wild Asian elephants in Assam, India. Conserv. Biol. 38 , e14397. https://doi.org/10.1111/cobi.14397 (2024).

Brenner, L. J. & Metcalf, E. C. Beyond the tolerance/intolerance dichotomy: Incorporating attitudes and acceptability into a robust definition of social tolerance of wildlife. Hum. Dimens. Wildl. 25 , 259–267. https://doi.org/10.1080/10871209.2019.1702741 (2020).

Manfredo, M. J., Berl, R. E. W., Teel, T. L. & Bruskotter, J. T. Bringing social values to wildlife conservation decisions. Front. Ecol. Environ. 19 , 355–362. https://doi.org/10.1002/fee.2356 (2021).

Yeshey, Ford, R. M., Keenan, R. J. & Nitschke, C. R. Religious beliefs and wildlife value orientations influence tolerance of wildlife impacts in Bhutan. Hum. Dimens. Wildl. 29 , 300–318. https://doi.org/10.1080/10871209.2023.2234389 (2024).

Santiapillai, C. & Wijeyamohan, S. Conflict, Negotiation, and Coexistence: Rethinking Human-elephant Relations in South Asia (eds Locke, P. & Buckingham, J.) Ch. 10, 229-241 (Oxford University Press, 2016)

Kyle, G., Landon, A., Vaske, J. & Wallen, K. Tools for assessing the psychometric adequacy of latent variables in conservation research. Conserv. Biol. 34 , 1353–1363. https://doi.org/10.1111/cobi.13625 (2020).

Gore, M. L. & Kahler, J. S. Gendered risk perceptions associated with human-wildlife conflict: Implications for participatory conservation. PLoS ONE . 7 , e32901. https://doi.org/10.1371/journal.pone.0032901 (2012).

Mariki, S. B. Social impacts of protected areas on gender in West Kilimanjaro, Tanzania. Open. J. Social Sci. 4 , 220–235. https://doi.org/10.4236/jss.2016.43028 (2016).

Dickman, A. J. Complexities of conflict: The importance of considering social factors for effectively resolving human–wildlife conflict. Anim. Conserv. 13 , 458–466. https://doi.org/10.1111/j.1469-1795.2010.00368.x (2010).

Nyirenda, V. R., Nkhata, B. A., Tembo, O. & Siamundele, S. Elephant crop damage: Subsistence farmers’ social vulnerability, livelihood sustainability and elephant conservation. Sustainability 10 , 3572. https://doi.org/10.3390/su10103572 (2018).

Barua, M., Tamuly, J. & Ahmed, R. A. Mutiny or clear sailing? Examining the role of the Asian elephant as a flagship species. Hum. Dimensions Wildl. 15 , 145–160. https://doi.org/10.1080/10871200903536176 (2010).

Treves, A., Wallace, R. B., Naughton-Treves, L. & Morales, A. Co-managing human–wildlife conflicts: A review. Hum. Dimens. Wildl. 11 , 383–396. https://doi.org/10.1080/10871200600984265 (2007).

Ogra, M. Attitudes toward resolution of human–wildlife conflict among forest-dependent agriculturalists near Rajaji National Park, India. Hum. Ecol. 37 , 161–177. https://doi.org/10.1007/s10745-009-9222-9 (2009).

Yang, N., Zhang, E. & Chen, M. Attitudes towards wild animal conservation: A comparative study of the Yi and Mosuo in China. Int. J. Biodivers. Sci. Ecosyst. Serv. Manage. 6 , 61–67. https://doi.org/10.1080/21513732.2010.509630 (2010).

Almuna, R., Cortés, J., de los Ángeles Medina, M. & Vargas, S. P. We are silently paving the way toward human–wildlife coexistence: The role of women in the rural landscapes of southern Andes. Front. Conserv. Sci. 3 , 1006006. https://doi.org/10.3389/fcosc.2022.1006006 (2022).

Bandara, R. & Tisdell, C. Comparison of rural and urban attitudes to the conservation of Asian elephants in Sri Lanka: Empirical evidence. Biol. Conserv. 110 , 327–342 (2003).

Sampson, C. et al. Rural and urban views on elephants, conservation and poaching. Oryx 56 , 609–616. https://doi.org/10.1017/S0030605321000156 (2022).

Cabral de Mel, S. J. et al. Attitudes towards causes of and solutions to conflict between humans and Asian elephants. Conserv. Sci. Pract. 6 (11), e13238 https://doi.org/10.1111/csp2.13238 (2024).

Puri, M., Johannsen, K. L., Goode, K. O. & Pienaar, E. F. Addressing the challenge of wildlife conservation in urban landscapes by increasing human tolerance for wildlife. People Nat. 6 , 1116–1129. https://doi.org/10.1002/pan3.10604 (2024).

Karanth, K. K., Naughton-Treves, L., DeFries, R. & Gopalaswamy, A. M. Living with wildlife and mitigating conflicts around three Indian protected areas. Environ. Manag. 52 , 1320–1332. https://doi.org/10.1007/s00267-013-0162-1 (2013).

Fortin, D. et al. Quantitative spatial ecology to promote human-wildlife coexistence: a tool for integrated landscape management. Front. Sustainable Food Syst. 4 , 600363. https://doi.org/10.3389/fsufs.2020.600363 (2020).

Kansky, R., Kidd, M. & Fischer, J. Understanding drivers of human tolerance towards mammals in a mixed-use transfrontier conservation area in southern Africa. Biol. Conserv. 254 , 108947. https://doi.org/10.1016/j.biocon.2020.108947 (2021).

Badhe, Y. P. & Jaybhaye, R. G. Habitat suitability area analysis for leopard to mitigate human-wildlife conflict in Junnar Forest Division of Pune Forest Circle. Appl. Ecol. Environ. Sci. 9 , 524–532. https://doi.org/10.12691/aees-9-5-2 (2021).

McCann, N. P. et al. Integrating socioecological suitability with human–­wildlife conflict risk: Case study for translocation of a large ungulate. J. Appl. Ecol. 58 , 2810–2820. https://doi.org/10.1111/1365-2664.14021 (2021).

Thant, Z. M. et al. Factors influencing the habitat suitability of wild Asian elephants and their implications for human–elephant conflict in Myanmar. Global Ecol. Conserv. 43 , e02468 (2023).

Boudreau, M. R. et al. Using habitat suitability and landscape connectivity in the spatial prioritization of public outreach and management during carnivore recolonization. J. Appl. Ecol. 59 , 757–767. https://doi.org/10.1111/1365-2664.14090 (2022).

Sanare, J. E. et al. A socio-ecological approach to understanding how land use challenges human-elephant coexistence in northern Tanzania. Diversity 14 , 513. https://doi.org/10.3390/d14070513 (2022).

Hull, V. et al. Living with wildlife: A review of advances in social-ecological analysis across landscapes. Landscape Ecol. 38 , 4385–4402. https://doi.org/10.1007/s10980-023-01778-9 (2023).

Gross, E. M., Drouet-Hoguet, N., Subedi, N. & Gross, J. The potential of medicinal and aromatic plants (MAPs) to reduce crop damages by Asian elephants ( Elephas maximus ). Crop Prot. 100 , 29–37. https://doi.org/10.1016/j.cropro.2017.06.002 (2017).

Ly, C. T. et al. A pilot study of cultivating non-preferred crops to mitigate human-elephant conflict in the buffer zone of Yok Don National Park, Vietnam. Gajah 51 , 4–9 (2020).

McCleery, R. A. Improving attitudinal frameworks to predict behaviors in human–wildlife conflicts. Soc. Nat. Resour. 22 , 353–368. https://doi.org/10.1080/08941920802064414 (2009).

Jochum, K. A., Kliskey, A. A., Hundertmark, K. J. & Aless, L. Integrating complexity in the management of human-wildlife encounters. Glob. Environ. Change . 26 , 73–86. https://doi.org/10.1016/j.gloenvcha.2014.03.011 (2014).

Berkes, F. Community-based conservation in a globalized world. Proc. Natl. Acad. Sci. 104 , 15188–15193. https://doi.org/10.1073/pnas.0702098104 (2007).

Article PubMed PubMed Central ADS Google Scholar

Waylen, K. A., Fischer, A., McGowan, P. J. K., Thirgood, S. J. & Milner-Gulland, E. J. Effect of local cultural context on the success of community-based conservation interventions. Conserv. Biol. 24 , 1119–1129. https://doi.org/10.1111/j.1523-1739.2010.01446.x (2010).

Karanth, K. K., Gopalaswamy, A. M., Prasad, P. K. & Dasgupta, S. Patterns of human-wildlife conflicts and compensation: Insights from Western Ghats protected areas. Biol. Conserv. 166 , 175–185. https://doi.org/10.1016/j.biocon.2013.06.027 (2013).

R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, Vienna, Austria, 2024).

Wickham, H. et al. Welcome to the tidyverse. J. Open. Source Softw. 4 , 1686. https://doi.org/10.21105/joss.01686 (2019).

Rosseel, Y. lavaan: An R package for structural equation modeling. J. Stat. Softw. 48 , 1–36. https://doi.org/10.18637/jss.v048.i02 (2012).

Taber, K. S. The use of Cronbach’s alpha when developing and reporting research instruments in science education. Res. Sci. Educ. 48 , 1273–1296. https://doi.org/10.1007/s11165-016-9602-2 (2018).

Cortina, J. M. What is coefficient alpha? An examination of theory and applications. J. Appl. Psychol. 78 , 98–104. https://doi.org/10.1037/0021-9010.78.1.98 (1993).

van Buuren, S. & Groothuis-Oudshoorn, K. mice: Multivariate imputation by chained equations in R. J. Stat. Softw. 45 , 1–67. https://doi.org/10.18637/jss.v045.i03 (2011).

olsrr: Tools for Building OLS Regression Models v. 0.6.1. (2024).

Vittinghoff, E., Glidden, D. V., Shiboski, S. C. & McCulloch, C. E. Statistics for Biology and Health 509 (Springer Science+Business Media, LLC, New York, 2012).

UNEP-WCMC and IUCN. Protected Planet: The World Database on Protected Areas (WDPA). Cambridge, UK: UNEP-WCMC and IUCN. www.protectedplanet.net (2024).

Fotheringham, A. S. & Wong, D. W. The modifiable areal unit problem in multivariate statistical analysis. Environ. Plann. A . 23 , 1025–1044 (1991).

Kwan, M. P. The uncertain geographic context problem. Ann. Assoc. Am. Geogr. 102 , 958–968. https://doi.org/10.1080/00045608.2012.687349 (2012).

Haldane, J. B. S. The mean and variation of χ 2 , when used as a test of homogeneity, when expectations are small. Biometrika 31 , 346–355 (1940).

Anscombe, F. J. On estimating binomial response relations. Biometrika 43 , 461–464. https://doi.org/10.2307/2332926 (1956).

Sharpe, D. Your chi-square test is statistically significant: Now what? Pract. Assess. Res. Eval. 20 (2015).

Beasley, T. M. & Schumacker, R. E. Multiple regression approach to analyzing contingency tables: Post hoc and planned comparison procedures. J. Experimental Educ. 64 , 79–93. https://doi.org/10.1080/00220973.1995.9943797 (1995).

Lee, S. & Lee, D. K. What is the proper way to apply the multiple comparison test? Korean J. Anesthesiology . 71 , 353–360. https://doi.org/10.4097/kja.d.18.00242 (2018).

Dunn, O. J. Multiple comparisons using rank sums. Technometrics 6 , 241–252 (1964).

We appreciate the participation of all respondents, whose voices we aimed to amplify. Additionally, we acknowledge the logistical support provided by Pathmasiri Rajapaksha and Upali Weerasinghe. We thank the Department of Wildlife Conservation for permission to carry out work in this region (WL/3/2/30/2023). This research was funded by Oklahoma City Zoo and Botanical Garden’s Round Up for Conservation program.

This research was funded by Oklahoma City Zoo and Botanical Garden’s Round Up for Conservation program.

Sri Lanka Elephant Project, Dambulla, Sri Lanka

Chase A. LaDue, K. H. Darshika L. Jayarathna, P. D. P. J. Jayawardena, A. G. S. Malsha Warnapura, Rebecca J. Snyder & Rajnish P. G. Vandercone

Department of Conservation and Science, Oklahoma City Zoo and Botanical Garden, 2000 Remington Place, Oklahoma City, OK, 73111, USA

Chase A. LaDue, Emily A. Geest & Rebecca J. Snyder

Post Graduate Institute of Science, University of Peradeniya, Peradeniya, Sri Lanka

Department of Zoology, University of Peradeniya, Peradeniya, Sri Lanka

Search author on: PubMed Google Scholar

Chase A. LaDue: Conceptualization, methodology, formal analysis, data curation, writing—original draft, writing—review & editing, visualization, supervision, project administration. K.H. Darshika L. Jayarathna: Investigation, writing—review & editing. P.D.P.J. Jayawardena: Investigation, writing—review & editing. A.G.S. Malsha Warnapura: Investigation, writing—review & editing. Emily A. Geest: Formal analysis, writing—review & editing, visualization. Rebecca J. Snyder: Conceptualization, methodology, writing—review & editing, funding acquisition. Rajnish P.G. Vandercone: Conceptualization, methodology, writing—review & editing, supervision, project administration.

The authors declare no competing interests.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Below is the link to the electronic supplementary material.

Open Access This article is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License, which permits any non-commercial use, sharing, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if you modified the licensed material. You do not have permission under this licence to share adapted material derived from this article or parts of it. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by-nc-nd/4.0/ .

LaDue, C.A., Jayarathna, K.H.D.L., Jayawardena, P.D.P.J. et al. Multidimensional tolerance towards Asian elephants varies with demographic and spatial correlates among rural communities in Sri Lanka. Sci Rep 16 , 30801 (2026). https://doi.org/10.1038/s41598-026-60472-3

Version of record : 05 October 2026

DOI : https://doi.org/10.1038/s41598-026-60472-3

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

&copy; 2026 Springer Nature Limited

Join the Conversation