Quantifying the Influence of Emotions on Management Acceptability for White-Tailed Deer (Odocoileus virginianus)

Abstract

Emotions pervade human-wildlife relationships across social identities and cultures. Yet research on how emotions influence the cognitive processing of wildlife encounters remains sparse. In this study, we quantify the role of anticipated emotions in processing hypothetical encounters with white-tailed deer (Odocoileus virginianus). In 2021, we surveyed Indiana residents about deer and deer management (n = 1.806). Under four hypothetical deer encounters, we estimated the structural relationships among respondents’ general attitudes toward deer, mutualism wildlife beliefs, scenario-specific emotions, and scenario-specific lethal control acceptability. Emotions mediated 14% of the effect of general attitudes on lethal control acceptability when encountering a fawn and completely mediated this effect when encountering a diseased deer. Our findings suggest that emotions work together with cognitions to process stimuli in a human-wildlife encounter and make a normative decision. Accounting for emotions in decision-making will help practitioners develop more effective and socially accepted approaches to wildlife conservation and management.

Keywords:

• Acceptability
• cognitive hierarchy
• Country or Region
• emotions
• Human dimensions of fisheries and wildlife
• Human-Wildlife Interactions
• North America
• Review Areas
• structural equation modeling
• Topic Area
• white-tailed deer
• wildlife management

Introduction

Human emotions infuse our relationships with nature and wildlife, at different times evoking solitude and spirituality (Long, More, and Averill 2006), awe and enjoyment (Hicks and Stewart 2020; McIntosh and Wright 2017), or fear and frustration (Hill 2015; Jacobs et al. 2014). Emotions also escalate social conflicts over nature and wildlife, like that between environmental preservation versus use (Buijs and Lawrence 2013; Madden and McQuinn 2014). Perhaps no medium has been as influential in popularizing this conflict as Walt Disney’s Bambi. From the death of young Bambi’s mother to the fiery hunt for Bambi as an adult buck, the film became an emotional appeal against hunting and humanity’s domination over nature in favor of nurturing an idyllic and vulnerable forest (Hastings 1996). This elicited backlash from hunters who were outraged by the film’s depiction of their practices and produced an enduring conflict pitting supporters of ethical hunting against the “excessive sentimentality” of those supporting animal rights, humaneness, and wilderness preservation (Lutts 1992, 161). Although the pro- versus anti-hunting debate can be classified as a clash of value orientations (Teel and Manfredo 2010), emotions typically exacerbate social animosities through opposing derogatory slurs like “Bambi-killer” and “Bambi-lover” (Cartmill 1993; Hastings 1996).

Yet much more complexity underlies human emotions toward wildlife. These emotions shift with place (i.e., of encounter or attachment; Halpenny (2010; Jacobs and Vaske 2019), power relations (González-Hidalgo and Zografos 2020), spiritual or cultural meanings (Blumenthal 1990; Gogoi 2018), and the type of animal encountered (Nyhus 2016). Wildlife such as deer, which initially seem harmless, can elicit feelings of fear, anxiety, and anger depending on the person and their deer-related experiences (Stinchcomb, Ma, and Nyssa 2022).

Despite the pervasiveness of emotions in society and popular culture, emotions remain largely absent from research on human behaviors toward and preferences for wildlife, particularly in the North American context (Jacobs 2012; Jacobs and Vaske 2019). Scholars attribute this in part to an institutional perception that human emotions are irrational, weak, and reactive, to and to challenges in measuring emotions, and to the novelty of studying emotions in the human dimensions of wildlife field (Hicks 2017; Manfredo 2008; Stinchcomb, Ma, and Nyssa 2022). Indeed, emotional expression has historically been marginalized from Western thought, exiled into feminine and private spaces, and placed in stark opposition to masculine and publicly acceptable rationality (Anderson and Smith 2001; Batavia et al. 2021). Contemporary understandings recognize the fluidity of emotions across gendered and public/private boundaries and their centrality to social-political life (Thrift 2004). As such, the role of emotions in how we ‘rationalize’ experiences with and within nature warrants empirical investigation across disciplines. Here, we aim to advance emotional research in the human dimensions of wildlife by quantifying the influence of emotions on people’s cognitive processing of encounters with white-tailed deer (Odocoileus virginianus).

Cognition and Emotion in the Human Dimensions of Wildlife

Research in the human dimensions of wildlife (HDW) has widely used social-psychological theories to measure the relationships among human beliefs, attitudes, and behaviors toward wildlife. These theories typically describe a hierarchy of human cognition, where values lie at the foundation of cognitive processing and are operationalized through basic beliefs, which then influence more specific attitudes and norms, which finally shape behavioral intentions or outcomes (Fishbein and Ajzen 1975; Homer and Kahle 1988; Whittaker, Vaske, and Manfredo 2006). Moving up this hierarchy from values to behaviors, the constructs become less abstract, so that specific attitudes better predict behavioral outcomes than do more fundamental values. This specificity principle (Fishbein and Ajzen 1975) remains critical to cognitive-hierarchy research because it recognizes the contextual dependence of human-wildlife interactions and provides criteria for how quantitative research should specify values, attitudes, and behaviors to uncover statistical relationships among them (Sponarski, Vaske, and Bath 2015; Whittaker, Vaske, and Manfredo 2006).

Despite its prominence in HDW research, the cognitive hierarchy typically explains only half of the variation in normative evaluations of management actions (Sponarski, Vaske, and Bath 2015; Vaske, Roemer, and Taylor 2013). Recent research suggests that human emotions are influential in processing human-wildlife interactions and may partly account for the remaining variability (Drake et al. 2020; Gogoi 2018; Hicks and Stewart 2020; Sponarski, Vaske, and Bath 2015; Straka, Miller, and Jacobs 2020). Here, we quantify the role of emotions within a modified cognitive hierarchy framework. While cognitive hierarchy models typically examine how beliefs and attitudes affect behaviors or behavioral intentions, we instead assess their effect on normative judgements about wildlife-management actions, and how emotions modify that effect. Recent work employed similar models to analyze the role of emotions in hypothetical encounters with wolves (Jacobs et al. 2014; Landon et al. 2020) and coyotes (Sponarski, Vaske, and Bath 2015). These studies measured emotional responses and normative judgements about management in several encounter scenarios, which could then be structurally associated with basic beliefs about and general attitudes toward the predators. Sponarski, Vaske, and Bath (2015) found that emotions mediated the effects of coyote-related beliefs about and attitudes toward the acceptability of lethal control in all but the most severe scenario of a coyote snarling, in which underlying attitudes toward coyotes most strongly predicted the acceptability of lethal control to respondents.

Still, research on emotions toward wildlife remains sparse and primarily focused on carnivore species (Drake et al. 2020; Jacobs et al. 2014; Slagle, Bruskotter, and Wilson 2012; Sponarski, Vaske, and Bath 2015). Few studies quantify emotions related to non-carnivores, such as ungulates. Members of the Cervidae family—including elk and deer species—persist in high densities across much of the United States and impact human communities through damage to agricultural crops or hobby gardens, vehicle collisions, and disease risk (DeNicola et al. 2000). Depending on the context, ungulates can elicit complex emotions, ranging from negative feelings of frustration or anxiety (Hill 2015; Stinchcomb, Ma, and Nyssa 2022) to positive feelings of awe, joy, or cultural connection (Blumenthal 1990; Hicks 2017). Yet, emotions toward these taxa have been neither quantified nor accounted for by wildlife management agencies (Manfredo 2008).

In Indiana, white-tailed deer (hereafter “deer”) populations proliferated after their reintroduction in 1934, and now occur at high densities across the state (Brown and Parker 1997). To control deer populations –and balance their impacts and benefits—deer managers in Indiana widely use lethal control measures like licensed hunting and targeted culling (Swihart et al. 2020). Yet mutualism values for wildlife—which consider wildlife as part of one’s social community, worthy of care, and deserving of similar rights to humans—are expected to be increasing, especially in urbanized areas of the state, and affecting public trust in management decisions (Dietsch et al. 2019; Manfredo, Teel, and Henry 2009; Schroeder et al. 2021).

However, before 2021, assessments of social perceptions of deer and deer management in Indiana had only considered hunters and farmers. An exploratory qualitative study in 2021 found emotions to be highly influential on perceptions of deer and deer management, regardless of stakeholder identity (Stinchcomb, Ma, and Nyssa 2022). Emotions expressed toward deer depended on the age or sex of the animal, its condition or behavior, and the person’s prior experiences and values, motivations, or beliefs related to deer and deer management (Stinchcomb, Ma, and Nyssa 2022). In this case, emotions and cognitions did not act separately when processing experiences with deer but rather were intertwined in a complex cognitive system.

Conceptual Framework

In this study, we examine how anticipated emotions within hypothetical human-deer encounters influence the acceptability of lethal control among residents of Indiana. Our modified cognitive hierarchy model includes mutualism beliefs about wildlife, general attitudes toward deer, and scenario-specific emotional dispositions and judgements about lethal control acceptability. We define beliefs broadly as statements or ideas that an individual takes as truths, even if they are not true (following Vaske and Manfredo 2012). Four sets of wildlife-related beliefs (i.e., appropriate use, hunting, social affiliation, and care) “orient” underlying values to wildlife onto domination or mutualism dimensions (Landon et al., 2019; Manfredo and Dayer 2004; Whittaker, Vaske, and Manfredo 2006). Previous scholarship demonstrates that wildlife value orientations fundamentally explain individual variation in attitudes toward wildlife and its management (Dietsch et al. 2019; Teel and Manfredo 2010).

Attitudes are defined as “a mental state reflected by cognitive (e.g., beliefs) and affective (e.g., emotions) components” (Sponarski, Vaske, and Bath 2015, 240). More generally, attitudes comprise a directional evaluation (i.e., positive or negative) of some object that can change with situational and experiential factors (Eagly and Chaiken 1993; Whittaker, Vaske, and Manfredo 2006). For example, attitudes toward deer may vary depending on where one typically encounters them (e.g., natural area vs. private property), what the deer is doing (e.g., resting vs. eating), and the cohort of the deer (e.g., a large buck vs. a young fawn). Attitudes toward wildlife management are typically measured as the acceptability of specific actions to control a wildlife population. Following Bruskotter et al. (2009, 121), we conceptualize management acceptability as “a judgement or decision regarding the appropriateness of a particular action or policy.” Such normative judgements about wildlife management are influenced most significantly by other cognitive factors—such as one’s beliefs about a species’ impacts and general attitudes toward that species—and to a lesser extent by one’s social identity or affiliations (Bruskotter et al. 2009).

Generally, emotions consist of physiological, cognitive, and behavioral responses to external stimuli and influence many cognitive functions from motivation and memory to decision-making (Izard 2007; Jacobs 2012; Lerner et al. 2015; Tyng et al. 2017). Emotions interact continuously and dynamically with cognitions like values, beliefs, and attitudes to produce an individualized evaluation or judgment about a stimulus (Izard 2009; Lewis 2005). For instance, an individual who values wildlife existence may exhibit a stronger emotional response to lethal control measures than would an individual who remains distanced from wildlife or values personal property over wildlife protection.

How an individual processes an emotional experience depends on emotional personality traits that reflect one’s identity, are always present, and remain relatively stable over time (Izard 2007; Jacobs and Vaske 2019; Sponarski, Vaske, and Bath 2015). Individuals use these traits as criteria to appraise a stimulus and judge its emotional relevance (Jacobs, Vaske, and Roemer 2012; Sponarski, Vaske, and Bath 2015). Emotional traits typically evoke strong memories about or personal experiences with wildlife (Vaske, Roemer, and Taylor 2013) and predict behavioral decisions or normative judgments, such as those about the acceptability of management (Jacobs and Vaske 2019). How these traits vary with the species or cohort of animal remains unknown (Jacobs, Vaske, and Roemer 2012).

Emotional responses likely represent an interaction between enduring personality traits and rapid situational evaluations (Izard 2009). However, delineating between emotional traits and momentary responses remains difficult for quantitative social science and was beyond the scope of this work. In this paper, we measure Indiana residents’ anticipated emotions in response to hypothetical deer encounters using discrete and cross-culturally recognized emotions such as fear, joy, anger, or anxiety (Izard 2007; Jacobs and Vaske 2019; Sponarski, Vaske, and Bath 2015). We define an anticipated emotion as the cognitive representation of a future emotional state, i.e., how one expects to feel in future situations or about future outcomes (Brown and McConnell 2011; Loewenstein and Schkade 1999). Rather than measuring the emotional component of an attitude toward deer (Deonna and Teroni 2015; Russell 1980)—the positive feeling associated with, e.g., “deer are cute,” or the negative feeling associated with, e.g., “deer are a nuisance”—we measure an anticipated emotion in the (hypothetical) moment, which can regulate attitudes or behaviors toward a specific object like the deer that is encountered (Brown and McConnell 2011; Halperin et al. 2013). Residents’ attitudes toward deer are solicited prior to introducing the scenarios but likely remain cognitively present, so we focus on the mediating effect of emotions felt in each scenario on forming judgements about the acceptability of lethal deer control (Vaske et al. 2021). We recognize the ongoing debate within the psychology and broader social science literature regarding whether attitude or emotion come first in relation to normative judgements, behaviors, and behavioral intentions (Allen, Machleit, and Kleine 1992; Eagly, Mladinic, and Otto 1994; Forgas 2003; Van Den Hooff, Schouten, and Simonovski 2012; Vaske et al. 2021). The structure of this relationship can vary with situational contexts (Allen, Machleit, and Kleine 1992; Halperin et al. 2013) and potentially with the wildlife species encountered (Jacobs, Vaske, and Roemer 2012; Vaske et al. 2021). Comparing between different roles of attitudes and emotions, however, was beyond the scope of this study.

We examine the degree to which anticipated emotions mediate the relationship among beliefs, attitudes, and management acceptability, and how this mediation changes with the type of deer encountered in each scenario. Based on the specificity principle, we expect mutualism beliefs and general attitudes to influence scenario-specific anticipated emotions. These emotions and cognitions should, together but differentially, influence the acceptability of lethal control in each deer encounter scenario (Figure 1). Specifically, we hypothesize that:

Figure 1. Hypothesized structural relationships among general attitudes toward deer, mutualism wildlife beliefs, scenario-specific anticipated emotions, and scenario-specific lethal control acceptability. Φ: covariance between latent variables.

1. Compared to people who hold negative attitudes toward deer, people who hold positive attitudes toward deer are, overall, less accepting of lethal control of deer regardless of the encounter scenario.

2. Compared to people who do not agree with mutualism beliefs, people who agree with mutualism beliefs about wildlife are, overall, less accepting of lethal control of deer regardless of the encounter scenario.

3. Anticipated emotions mediate the relationship between general attitudes and the acceptability of lethal deer control and the relationship between mutualism beliefs and lethal control acceptability

4. Anticipated emotions will show the strongest mediating effects when encountering a fawn or a diseased deer

Methods

Data Collection

We measured the perceptions, values, attitudes, beliefs, and emotions related to deer and deer management among Indiana residents using a state-wide survey during June–August 2021 (Stinchcomb et al. 2022). Our study was approved for human subjects research by the Institutional Review Board of Purdue University, protocol number 1902021653. We used a 2 × 4 stratified design to randomly sample 6,000 residents. The higher-order stratum divided residents into 3,000 customers of the Indiana Department of Fish & Wildlife (DFW) and 3,000 non-customers (non-DFW). Customers of the Indiana DFW include anyone who has purchased a resident hunting, fishing, or trapping license. We further divided DFW and non-DFW strata into four landscape types—forest, farmland, developed area, and “integration”—from each of which we randomly sampled 750 tax-parcel addresses. The “integration” sub-sample consisted of tax parcels within 6.4 × 6.4 km (4 × 4 mi) grids where ecological data were collected by our colleagues from Purdue University. We conducted sampling in ArcGIS Pro (ESRI 2022) using land cover, tax parcel, and address data from IndianaMap, the Indiana Department of Local Government and Finance, and the DFW. We re-sampled addresses until no duplicates, blanks, and public or corporate owners remained.

We followed Dillman’s (2014) tailored-design method for survey dissemination. Via postal mail, each resident received a pre-notification postcard followed a week later by a printed survey packet. Printed survey packets included a link to take the survey online in the enclosed cover letter. Two weeks after first contact, we mailed a reminder postcard to residents who had not yet responded via mail or online. Over the next month, we sent two additional survey packets to non-respondents (two weeks apart), for a total of three survey mailings.

We compared key demographic characteristics of survey respondents with the corresponding census data for Indiana to check for non-response bias. Based on Chi-squared tests, respondent demographics were significantly different from proportions expected at the state level (Stinchcomb et al. 2022). Since we intended to test a baseline model describing the role of emotions in human-deer encounters, we pooled our data for analyses. We caution against generalizing our results to the entire population.

Measurement

Scenario Design

We used four hypothetical encounters with deer to assess how respondents’ anticipated emotions influence the acceptability of deer management in specific scenarios (Table 1). The encounters were based on recent qualitative work (Stinchcomb, Ma, and Nyssa 2022) showing that emotions expressed by Indiana residents typically depend on the age and/or sex of the deer encountered (e.g., buck or fawn) and its behavior (e.g., foraging, appearing sick or emaciated). We thus selected two cohorts of deer—a large buck and a fawn—and two behaviors—eating the nearest plants and looking diseased—for our encounter scenarios. We kept the behavior between the two cohorts constant and kept the cohort between the two behaviors constant (Table 1) to minimize variation that could elicit additional emotional responses. Because deer typically do not evoke emotions as predators do in many human-wildlife situations, our goal in selecting and analyzing these scenarios was to initially assess what types of deer—with “type” including cohort, behavior, or condition—arouse emotional responses from the average person. We do not statistically compare the individual effects of each deer type and situation on anticipated emotions; rather, we qualitatively examine across scenarios how deer type influences the magnitude, direction, and significance of the effects of attitudes, beliefs, and anticipated emotions on lethal control acceptability.

Table 1. Wording and expected effects of four hypothetical encounters with different white-tailed deer types, i.e., cohorts, behaviors, or conditions.

Deer encounter scenario	Exact wording “While walking on your property or in your neighborhood…”	Expected emotional response	Expected effect of anticipated emotions on lethal control acceptability	Literature cited
1: Adult deer eating nearest plants	“…an adult deer appears and stops on the path in front of you, then begins eating the nearest plants.”	Frustration, if residents experience frequent damage to plants on their property. Neutrality otherwise.	Negative in direction, but insignificant due to neutrality of behavior.	Jacobs, Vaske, and Roemer (2012)
				Stinchcomb, Ma, and Nyssa (2022)
2: Large buck looks your way	“…a large buck (male deer) appears and stops on the path in front of you, then looks your way.”	Anxiety or fear from some residents, awe or joy from others.	Not different from zero due to mixed emotions across sample.	Kellert (2012)
				McIntosh and Wright (2017)
				StinchcombMa, and Nyssa (2022)
3: Fawn looks your way	“…a fawn (baby deer) appears and stops on the path in front of you, then looks your way.”	Joy and other positive emotions for most residents.	Negative in direction and significant.	Buckley (2022)
				Stinchcomb, Ma, and Nyssa (2022)
4: Adult deer looking diseased	“…an adult deer appears and stops on the path in front of you, looking diseased.”	Sad or upsetting emotions for most residents.	Positive in direction and significant.	Buckley (2022)
				Stinchcomb, Ma, and Nyssa (2022)

Model Variables

The structural model comprises 19 observed items, organized into three latent variables and one manifest variable: mutualism beliefs about wildlife (latent, 7 items), general attitudes toward deer (latent, 4 items), scenario-specific anticipated emotions (latent, 7 items per scenario), and the scenario-specific acceptability of lethal deer control (manifest, 1 item per scenario). Anticipated emotions and management preferences were measured in each of four hypothetical encounters with deer. Mutualism beliefs and general attitudes were taken from questions elsewhere in the survey and thus did not vary with scenario context. We provide our complete questionnaire in Supplemental Online Material (Methods SOM 01).

Using standard language in the HDW field, mutualism beliefs about wildlife were: “To what extent do you disagree or agree with the following statements about wildlife in general: (i) Animals should have rights similar to the rights of humans; (ii) I view all living things as part of one big family; (iii) I feel a strong emotional bond with animals; (iv) I care about animals as much as I do about people; (v) We should strive for a world where humans and wildlife can live side by side without fear; (vi) I value the sense of companionship I receive from animals; (vii) Wildlife are like my family and I want to protect them”(Manfredo, Teel, and Henry 2009; Teel and Manfredo 2010; Whittaker, Vaske, and Manfredo 2006). We measured belief items on a 5-point Likert-type scale, recoded −2 to 2 for analysis where −2 = “Strongly Disagree” and 2 = “Strongly Agree” (Methods SOM 01). Disagreement with these items aligns with traditionalist or distanced value orientations while agreement aligns with mutualist or pluralist value orientations (Landon et al. 2020). We could have chosen to analyze domination beliefs, which we would expect to have a positive association with lethal control acceptability. However, we were interested in mutualism belief items due to their rise across the U.S.

Following Sponarski, Vaske, and Bath (2015), our four general attitudes were: “In general, do you think of deer as: (i) bad/good; (ii) dangerous/harmless; (iii) detrimental/beneficial; (iv) nuisance/asset?” We measured each attitude on a 5-point bipolar scale, recoded from −2 to 2 for analysis. For instance, responses to “In general, do you think of deer as dangerous/harmless,” were recoded as (−2) very dangerous; (−1) slightly dangerous; (0) neither dangerous nor harmless; (1) slightly harmless; (2) very harmless (Methods SOM 01). The same scale applies to the remaining attitudes, replacing the adjective.

Anticipated emotions were based on each hypothetical scenario of encountering deer. We measured emotions on 5-point semantic differential scales. For example, given Scenario 2, “While walking on your property or in your neighborhood, a large buck (male deer) appears and stops on the path in front of you, then looks your way,” the respondent was asked, “to what extent would you feel the following: (i) nervous/calm; (ii) unexcited/excited; (iii) upset/pleased; (iv) tense/relaxed; (v) scared/not scared; (vi) sad/joyful; (vii) alert/not alert?” Items were recoded from −2 (e.g., upset) to 2 (e.g., pleased) for analysis. We asked respondents to report their anticipated emotions in each of the four scenarios.

Our dependent variable, acceptability of lethal management, was measured directly in the survey, and thus represents a manifest, rather than latent, variable.

In each scenario, we asked the respondent, “Under this scenario, how unacceptable or acceptable would it be [for authorities] to use the following management actions: (i) lethal control; (ii) nonlethal control; (iii) advise & monitor; (iv) do nothing/no management” (Methods SOM 01). Responses were measured on a 5-point Likert-type scale recoded −2 to 2 for analysis where −2 = “Very Unacceptable” and 2 = “Very Acceptable.” For this paper, we analyze how scenarios change the acceptability of lethal control, because it is the primary method of deer management in Indiana.

Analysis

We took a structural equation modeling approach to examine how emotions influence the relationship among mutualism beliefs, deer-related attitudes, and deer management preferences (Figure 1). We used confirmatory factor analysis to ensure that our observed indicators measured the appropriate, and distinct, latent variables. We then evaluated the internal consistency of the latent variables (mutualism beliefs, general attitudes, and anticipated emotions) using Cronbach’s alpha (Cronbach 1951). Once we established reliability of the measurement models, we estimated the structural model (Figure 1) four times, once per scenario, using scenario-specific data for the anticipated emotions and lethal control variables. For each scenario, we evaluated overall model fit using a variety of criteria established by Hu and Bentler (1999), including chi-square ($\Delta {\mathrm{\chi }}^2,$ ${\mathrm{\chi }}^2/df$), Root Mean Square Error of Approximation (RMSEA; acceptable value .05–.08), comparative fit index (CFI; acceptable value > 0.90), and Tucker-Lewis index (TLI; acceptable value > 0.90).

To test for the mediating role of emotions, we examined the direct, indirect, and total effects of general attitudes and mutualism beliefs on lethal control acceptability. In STATA IC v 16.1 (StataCorp 2019), we used the sem package and its estat teffects command to output direct, indirect, and total effects for each model pathway and their significance. We then calculated the proportion of effect mediated by emotions as the indirect effect through emotions divided by the total effect for each pathway. Although we do not statistically evaluate the individual effects of deer cohort and behavior on model results, we qualitatively assess how the specific deer encountered in each scenario affects the mediating influence of anticipated emotions on lethal control acceptability.

Only 14 out of 1,806 responses showed missing values across our model variables. Because this represents a small proportion, we used the Maximum Likelihood Estimator (MLE) with missing values method for all model estimations. We report standardized coefficients to compare model fit and estimates across scenarios. We conducted all estimations and analyses using the sem package in STATA IC v 16.1 (StataCorp 2019).

Results

We received 1,806 completed surveys. An additional 500 were undeliverable, deceased, or otherwise ineligible, for an overall response rate of 33%. Respondents identified as 76% men and 23% women, mostly White/Caucasian (92%), and primarily rural (43%) or urban (25%) residents. The average respondent was 60 years old and had resided in Indiana for 51 years at the time of the survey. For a complete profile of survey respondents, please see Table 1 in Stinchcomb et al. (2022).

Across our sample, respondents generally felt positively toward white-tailed deer, describing them as mostly good, harmless, beneficial, and an asset as opposed to bad, dangerous, detrimental, and a nuisance (Table 2). Respondents showed mixed agreement and disagreement on mutualism beliefs with mean values ranging from −0.51 [animal rights] to 0.69 [companionship] (Table 2). Respondents exhibited mostly positive anticipated emotions when encountering an adult deer eating plants (scenario 1) and a fawn (scenario 3), increased alertness when encountering a large buck (scenario 2), and mostly negative emotions when encountering a diseased deer (scenario 4). Lethal control was generally rated as unacceptable across the first three scenarios, with the least acceptable rating occurring when encountering a fawn (mean = −1.34). The average lethal control rating switched to acceptable when encountering a diseased deer (mean = 0.92; Table 2).

Table 2. Summary, factor loadings, and reliability of structural equation model variables.

Variable	Description	Mean	Std Dev	Standardized factor loading	Item-total correlation	Alpha if item deleted	Cronbach’s alpha
General Attitudes							0.87
In general, do you think of deer as:
ATT_BG	Bad/Good	1.02	1.02	0.70	0.66	0.85
ATT_DH	Dangerous/Harmless	0.67	1.14	0.71	0.67	0.85
ATT_DB	Detrimental/Beneficial	0.58	1.12	0.88	0.79	0.80
ATT_NA	Nuisance/Asset	0.54	1.22	0.85	0.76	0.81
Mutualism Beliefs							0.85
WVO2	Animals should have rights similar to the rights of humans	−0.51	1.27	0.62	0.56	0.83
WVO4	I view all living things as part of one big family	0.35	1.26	0.75	0.66	0.82
WVO6	I feel a strong emotional bond with animals	0.27	1.15	0.61	0.58	0.83
WVO8	I care about animals as much as I do about people	−0.13	1.29	0.62	0.58	0.83
WVO11	We should strive for a world where humans and wildlife can live side by side without fear	0.36	1.25	0.59	0.53	0.84
WVO12	I value the sense of companionship I receive from animals	0.69	1.09	0.65	0.62	0.83
WVO13	Wildlife are like my family and I want to protect them	−0.04	1.21	0.84	0.73	0.81
Emotional Dispositions
Scenario 1	An adult deer appears and stops on the path in front of you, then begins eating the nearest plants						0.83
	In this scenario to what extent would you feel…
s1_EM1	Nervous/Calm	0.99	1.12	0.70	0.62	0.79
s1_EM2	Unexcited/Excited	0.82	1.06	0.34	0.35	0.83
s1_EM3	Upset/Pleased	0.95	1.07	0.74	0.71	0.78
s1_EM4	Tense/Relaxed	0.89	1.11	0.90	0.78	0.77
s1_EM5	Scared/Not scared	1.10	1.10	0.87	0.73	0.78
s1_EM6	Sad/Joyful	1.00	0.99	0.73	0.72	0.78
s1_EM7	Alert/Not Alert	−0.12	1.37	0.25	0.22	0.87
Scenario 2	A large buck (male deer) appears and stops on the path in front of you, then looks your way						0.86
	In this scenario to what extent would you feel…
s2_EM1	Nervous/Calm	0.42	1.33	0.77	0.74	0.83
s2_EM2	Unexcited/Excited	1.04	0.98	0.44	0.42	0.87
s2_EM3	Upset/Pleased	0.88	1.10	0.78	0.76	0.83
s2_EM4	Tense/Relaxed	0.43	1.24	0.86	0.81	0.82
s2_EM5	Scared/Not scared	0.72	1.21	0.92	0.81	0.82
s2_EM6	Sad/Joyful	0.94	0.99	0.72	0.70	0.84
s2_EM7	Alert/Not Alert	−0.51	1.42	0.36	0.32	0.90
Scenario 3	A fawn (baby deer) appears and stops on the path in front of you, then looks your way						0.85
	In this scenario to what extent would you feel…
s3_EM1	Nervous/Calm	1.27	1.05	0.78	0.72	0.82
s3_EM2	Unexcited/Excited	1.21	0.98	0.55	0.54	0.84
s3_EM3	Upset/Pleased	1.29	0.97	0.82	0.79	0.81
s3_EM4	Tense/Relaxed	1.17	1.06	0.90	0.81	0.80
s3_EM5	Scared/Not scared	1.32	1.00	0.90	0.78	0.81
s3_EM6	Sad/Joyful	1.27	0.96	0.76	0.75	0.82
s3_EM7	Alert/Not Alert	−0.07	1.47	0.23	0.21	0.91
Scenario 4	An adult deer appears and stops on the path in front of you, looking diseased						0.81
	In this scenario to what extent would you feel…
s4_EM1	Nervous/Calm	−0.26	1.29	0.83	0.62	0.76
s4_EM2	Unexcited/Excited	−0.03	1.10	0.13	0.21	0.83
s4_EM3	Upset/Pleased	−0.87	1.01	0.52	0.62	0.77
s4_EM4	Tense/Relaxed	−0.31	1.12	0.94	0.73	0.74
s4_EM5	Scared/Not scared	0.13	1.22	0.76	0.59	0.77
s4_EM6	Sad/Joyful	−1.05	1.05	0.40	0.54	0.78
s4_EM7	Alert/Not Alert	−0.84	1.15	0.46	0.49	0.79
Management Option
s1_mg1	Scenario 1: How acceptable is lethal control?	−0.70	1.47
s2_mg1	Scenario 2: How acceptable is lethal control?	−0.62	1.51
s3_mg1	Scenario 3: How acceptable is lethal control?	−1.34	1.20
s4_mg1	Scenario 4: How acceptable is lethal control?	0.92	1.41

Confirmatory factor analysis empirically verified the reliability of indicators for our latent constructs (i.e., general attitudes, mutualism beliefs, and anticipated emotions) and the discriminant validity among those constructs. All measurement indicators loaded onto their associated latent variable with factor loadings $\ge$ 0.61 except for the fifth mutualism belief (WVO11) which showed a factor loading of 0.59. Anticipated emotions Unexcited/Excited and Alert/Not Alert showed factor loadings between 0.13 and 0.55 for most deer encounter scenarios (Table 2). We removed these two anticipated emotion items because they showed higher alpha-if-item-removed statistics than the overall Cronbach’s alpha for three out of four emotions scales (Table 2).

Overall, reliabilities of items measuring general attitudes toward deer and mutualism beliefs were high (Cronbach’s alpha = 0.87, 0.85 respectively). Reliabilities for the five remaining anticipated emotions in each of the four encounter scenarios were also high (Table 2). Post-estimation tests confirmed the convergent and discriminant validity of each measurement model. Across the four scenarios, the structural models fit the data with RMSEA values $\le$ 0.08, CFI values > 0.89, and TLI values $\ge$ 0.87. Although these values are slightly lower than the combinations recommended by Hu and Bentler (1999), the values indicate an adequate fit.

Emotions Interact with Deer Type to Influence Lethal Control Acceptability

Here, we discuss the direct effects of attitudes and mutualism beliefs on lethal control acceptability, then the mediating effect of emotions across scenarios. Direct effects are noted by $\lambda$ in parentheses. Indirect effects (noted by ${\lambda }_{\mathit{ind}}$) are mentioned in paragraphs three and four, to demonstrate how emotions mediate the effects of attitudes on lethal control acceptability.

Our first two hypotheses, that people who hold positive attitudes toward deer (H1) and agree with mutualism beliefs (H2) are overall less accepting of lethal control, generally held across all four scenarios. In the fourth scenario, encountering a deer that looks diseased, the direct effect of positive attitudes toward deer on acceptance of lethal control did not differ from zero ($\lambda$= 0.017, p = 0.73; Figure 2d). Mutualism beliefs consistently showed a significant and negative effect on lethal control acceptability (−0.419 < $\lambda$< −0.201, all p < 0.001; Figure 2a–d).

Figure 2. Structural equation model results for each deer encounter scenario. ϕ: covariance between latent variables. Stars depict the statistical significance of each effect. *: p < 0.05; **: p < 0.001. Indirect effect (IE) is the path from attitudes/mutualism beliefs to emotions multiplied by the path from emotions to lethal control, e.g., in (a): indirect effect (attitudes) = 0.461*0.063 = 0.029. Significance of each IE was derived from STATA v 16. Proportion of effect mediated by emotions (PM) is calculated as the IE of general attitudes or mutualism beliefs over their total effect on lethal control acceptability, in absolute value, e.g., in (a): total effect (attitudes) = indirect + direct effect, e.g., 0.029 + −0.396 = −0.367. |PM| = 0.029/|−0.367| = 0.079. Global fit of each model is assessed by Root mean squared Error of Approximation (0.05 < RMSEA < 0.08), comparative fit index (CFI > 0.90), and Tucker-Lewis index (TLI > 0.90). scenario 1: RMSEA = 0.066, CLI = 0.938; TLI = 0.923; scenario 2: RMSEA = 0.074, CLI = 0.927, TLI = 0.910; scenario 3: RMSEA = 0.076, CLI = 0.926, TLI = 0.909; scenario 4: RMSEA = 0.081, CLI = 0.896; TLI = 0.871.

We expected anticipated emotions to mediate the relationships between general attitudes and lethal control acceptability and between mutualism beliefs and lethal control acceptability to some degree in all scenarios (H3). Our data produced partial support for this hypothesis that differed across scenarios. Anticipated emotions did not significantly mediate the effect of general attitudes on acceptability of lethal control when encountering either an adult deer eating the nearest plants or a large buck (Figure 2a,b). The role of anticipated emotions became significant when encountering a fawn, mediating 14% of the effect of attitudes on lethal control acceptability (Figure 2c). Emotional mediation of the attitudes-lethal control pathway remained significant and reached nearly 150% when encountering a diseased deer (Figure 2d). Anticipated emotions significantly mediated the effect of mutualism beliefs on the acceptability of lethal control only when encountering a diseased deer (5% mediated, Figure 2d).

As hypothesized, the relationship among general attitudes, anticipated emotions, and lethal control acceptability depends on the deer cohort encountered (H4). Our data show that the direct effect of anticipated emotions on lethal control acceptability becomes statistically significant when encountering a fawn ($\lambda$ = −0.096, p = 0.02), but not when encountering a buck ($\lambda$ = 0.012, p = 0.77). Moreover, the indirect effect of general attitudes on lethal control acceptability (through anticipated emotions) increases in magnitude and becomes significant when encountering a fawn, while the total effect of general attitudes decreases (${\lambda }_{\mathit{ind}}$ = −0.040, p = 0.02). About 14% of the effect of attitudes is thus subsumed (or mediated) by anticipated emotions in the fawn scenario, compared to 2% in the large buck scenario (Figure 2b,c). Together, these results indicate that the influence of anticipated emotions on residents’ preferences for deer management changes depending on the cohort of deer encountered and becomes much more prominent when encountering a young fawn, as we initially hypothesized (H4).

The behavior displayed by a deer also differentially affected the relationships in our structural model (H4). When encountering a deer eating the nearest plants, anticipated emotions did not affect the acceptability of lethal control ($\lambda$ = 0.063, p = 0.243). Instead, positive general attitudes ($\lambda$ = −0.396, p < 0.001) and mutualism beliefs ($\lambda$ = −0.356, p < 0.001) directly and significantly decreased the acceptability of lethal control (Figure 2a). Indirect effects in this encounter were not statistically significant. Conversely, when encountering a diseased deer, the effect of general attitudes on the acceptability of lethal control became fully mediated by anticipated emotions, with the indirect effect becoming significant and the direct effect becoming insignificant (${\lambda }_{\mathit{ind}}$ = −0.054, p < 0.001; $\lambda$ = 0.017, p = 0.73; Figure 2d). In the diseased deer encounter, anticipated emotions showed their largest and most significant effect on lethal control among all four scenarios ($\lambda$ = −0.211, p < 0.001; Figure 2a–d). The direct effect of mutualism beliefs on lethal control remained strong and highly significant, with more mutualist beliefs decreasing respondents’ acceptance of lethal control for a diseased deer ($\lambda$ = −0.347, p < 0.001; Figure 2d). Encountering a diseased deer was the only scenario in which mutualism beliefs showed a significant indirect effect on lethal control acceptability through anticipated emotions (${\lambda }_{\mathit{ind}}$ = 0.016, p = 0.04; Figure 2d).

Discussion

Our study aimed to assess the role that anticipated emotions play in processing hypothetical encounters with an ungulate species, and how those emotions vary with the type of animal encountered, i.e., its cohort, behavior, or condition. As hypothesized, anticipated emotions influenced the relationship between general attitudes and lethal control acceptability in two hypothetical deer encounters: encountering a fawn and encountering a diseased deer. Encountering a fawn initiated a significant mediating role of emotions, suggesting that fawns elicit shared feelings of joy, pleasure, and calm that influence most residents’ processing of an encounter with this cohort. Large bucks, on the other hand, may elicit mixed feelings of awe or majesty from some residents and anxiety or fear from others (Table 1), leading to an insignificant effect of emotions in the buck scenario. Because mutualism beliefs consistently decreased the acceptability of lethal control across scenarios, our data suggest that the “Bambi” effect is driven more by anticipated emotions than by underlying beliefs. As such, the popularized clash over Bambi and what it represents may be one of emotional elicitation whereby seeing a fawn produces positive emotions that counteract the idea of killing that animal. We found that strong anticipated emotions coincide with strong mutualism beliefs among urban residents (Discussion SOM 01), but qualitative work demonstrates that farmers and rural landowners also express feelings of joy or awe toward fawns over other deer cohorts (Stinchcomb, Ma, and Nyssa 2022). It remains unclear whether individualized experiences or a shared awe of Bambi drive this emotional elicitation. Future studies should examine how anticipated emotions differ with wildlife cohorts across taxa, and across human social groups.

Anticipated emotions spiked when respondents imagined encountering a diseased deer, mediating the entire effect of general attitudes on lethal control acceptability, and about 5% of the effect of mutualism beliefs. Supported by previous qualitative findings, this implies that reflecting on the suffering of deer can shift normative judgments about deer management toward an overwhelming acceptance of lethal control measures like culling (Stinchcomb, Ma, and Nyssa 2022). Alternatively, this spike in emotions could result from a fear of disease transmission to other deer or to humans. In hypothetical encounters with wolves and coyotes, the fear of being attacked drove a stronger effect of emotions on lethal control acceptability (Sponarski, Vaske, and Bath 2015; Vaske, Roemer, and Taylor 2013; Jacobs et al. 2014). Concerns about animal welfare—which shift the subject of emotional projection from oneself to the animal—have been expressed toward diseased ungulates (Stinchcomb, Ma, and Nyssa 2022), captive predators (Cohen 2013), and the humaneness of wildlife management (Slagle et al. 2017), but not toward healthy predators encountered in the wild. Ungulates and predators thus elicit similar emotional dispositions under certain scenarios, but emotions depend critically on the condition of the animal and the context of encounter.

In contrast, emotions showed no effect on normative judgments about how to manage an adult deer eating the nearest plants. This lack of effect could be due to how we presented this scenario. We wanted to capture a generalized interaction that could apply to everyone in our state-wide sample. Thus, for an urban/suburban resident “eating the nearest plants” might mean the grass on the side of the neighborhood road or their own ornamentals; while for a farmer walking their property, those plants could be their soybean crops. Our data might have elicited a larger effect if we had specified “eating the nearest crops, seedlings, or ornamentals,” as deer damage to these plants tend to elicit frustration compared to a seemingly harmless deer grazing on the roadside.

Several recent studies analyzed the roles of emotions and cognitions simultaneously and argued for the importance of studying them together (Hill 2015). We move beyond this argument to suggest that emotions should be considered a component of the cognitive system (see also LeDoux and Brown 2017). Conceptualizing the mind as complex and adaptive allows us to account for dynamic feedbacks among cognitions, emotions, personal experiences, and external forces including social, political, and ecological stimuli (Holling 2001; Jochum et al. 2014). These feedbacks oppose the linearity of a cognitive hierarchy model and help to explain when, where, and how different emotions shift attitudes toward wildlife, even when underlying values remain similar (Jacobs and Vaske 2019; Stinchcomb, Ma, and Nyssa 2022). Differential roles of cognitions and emotions, then, are attributed to the bundle of stimuli being processed (LeDoux and Brown 2017). As shown in our study, while an adult deer eating plants may produce minimal emotional stimuli, a fawn or a diseased deer stimulate complex emotions that shape people’s normative judgements about acceptable management. Moreover, reflecting on emotional memories, or emotional imagery of wildlife (Straka, Miller, and Jacobs 2020), can shift one’s cognitions about wildlife and their management from a negative to a positive attitude or from unacceptability of lethal management to acceptability.

Our work, along with several other studies on human emotions toward wildlife and the environment (e.g., Buijs and Lawrence 2013; McIntosh and Wright 2017; Nightingale 2011), clearly demonstrate that emotions are not confined to private, feminized spaces but, rather, felt across various constituent groups to comparable intensities. Scholarship in political ecology reveals how people’s understandings of the social-ecological world are embodied, subjective, and mediated by feeling (Thien 2005) and interact with conflicting subjectivities and power relations (González-Hidalgo and Zografos 2020). Such emotional conflicts over bodies, spaces, and ways of being underlie most social struggles over the use, access to, and control of wildlife and nature (Sultana 2011). Engaging with affective political ecology (Singh 2018) and emotional geography (Anderson and Smith 2001) can yield a more holistic understanding of how human emotions shape human-environmental interactions, and, more specifically, the social acceptance of wildlife conservation and management. As such, our work contributes to the ongoing scholarly shift away from a duality of emotion versus reason and highlights the importance of bringing alternative perspectives about human emotion into conventional approaches to wildlife management.

Through participatory processes, researchers have found emotions and empathy—when their expression is safely supported—to be critical for understanding one’s interrelations with others and co-producing conservation strategies (Tremblay and Harris 2018). Empathy for wildlife, specifically, elevates individuals’ sense of moral obligation to nature and increases support for conservation (Ghasemi and Kyle 2021). When faced with perceived threats, injustices, or unmet needs, emotions also motivate people to organize and engage in political or environmental activism (Buijs and Lawrence 2013). Indeed, the word “emotion” stems etymologically from Latin emovere “to move out, agitate” and 16^th century French émotion “a (social) moving, stirring, agitation” (Online Etymology Dictionary 2021). To emote is thus to organize feeling, thinking, and doing together (Anderson and Smith 2001). The blending of emotion and reason, considered a condition for compassionate conservation, allows discursive processes, such as collaborative decision-making, to elucidate mutual dependencies, vulnerabilities, and interrelations (Batavia et al. 2021). As Batavia et al. (2021, 4) describe, when we understand compassion to mean passively suffering with, it reveals how we can “beget concern for all living beings…by recognizing that they, too, are embedded within and susceptible to the world.”

As our findings confirm, emotions are part of the cognitive process, influencing a person’s reflections on an experience and their attitudes toward conservation and management possibilities. In the hypothetical encounter with a diseased deer, reflecting on the deer’s suffering and sickness produced a shared emotional response for the suffering to end and, by implication, not spread to others. Opening institutional practice to the idea of thinking and feeling simultaneously will ‘enliven’ collective capacities to co-manage resources (Nightingale 2011; Vasile 2019) and imagine ways of ‘becoming with’ others, both human and nonhuman (Haraway 2008; Singh 2018).

Enlivening management with emotions, rather than repressing them, will elucidate the drivers of social conflict over deer and means of expanding public engagement in deer management. Social conflicts are often embedded in long histories of contention and power imbalances, over which each new dispute intensifies group emotions (Madden and McQuinn 2014) and motivates further escalation of conflict (Buijs and Lawrence 2013). Moreover, such inter-group conflicts foster mistrust in management when group needs remain unaddressed or not reflected in management decisions (Tremblay and Harris 2018). Although discussing emotional experiences can be uncomfortable (Madden and McQuinn 2014), doing so can uncover shared relations and moralities to wildlife (Lute and Gore 2014). A recent qualitative study with Indiana residents showed that hunters, rural residents, and urban residents all expressed a sense of responsibility to steward deer populations, and while they disagreed over the morality of hunting, they agreed on general objectives for deer management (Stinchcomb, Ma, and Nyssa 2022). As Lute and Gore (2014) suggest, developing shared morals like stewardship can facilitate compromise on wildlife management goals among diverse stakeholders and help narrow divides between groups with opposing value orientations, such as hunters and animal rights advocates (Patterson, Montag, and Williams 2003). Moreover, rousing emotions of place and compassion stimulates people to assess their beliefs about an issue and focus clearly on the environmental goals of a decision (Buijs and Lawrence 2013; Wilson 2008). It remains crucial for stakeholders to collectively define the meaning of each shared emotion, moral, or value and how they apply to wildlife management goals (Patterson, Montag, and Williams 2003; Slagle et al. 2019).

As such, our research suggests that wildlife management and conservation will benefit from using participatory, value-based, and goal-oriented approaches that foster iterative collaborations with diverse constituents (Slagle et al. 2019). Doing so will require a deeper integration of social science into each stage of the planning and decision-making processes (Niemiec et al. 2021). Acknowledging and allowing for the expression of emotions, experiences, and power relations among constituents plays a crucial role in defining management problems and objectives by ensuring that all relevant voices are represented, and their commonalities and differences can be safely negotiated (Brugnach and Ingram 2012). Yet in North America, integrating social science and participatory planning into wildlife management still faces substantial barriers including institutional culture, dominant ontological assumptions, and limited social science capacity (Niemiec et al. 2021).

Before concluding, we caution against generalizing our results to a broader population. Our sample was significantly more Caucasian, male, well-educated, wealthy, and rural residing than expected from Indiana census data (see Stinchcomb et al. 2022, Table 1). Our sampling strategy contributed to these differences because it deliberately targeted rural properties, urban areas—most of which contain universities—and existing customers of the Indiana DFW. Group-level heterogeneity can affect the direction and significance of SEM model estimates with pooled data, but few accessible methods exist to remedy the effects of unobserved heterogeneity for covariance-based models (Becker et al., 2013) and software is currently limited to comparing two groups at a time. Although our study aimed to present a baseline analysis of emotions in deer-human interactions, we conducted a supplemental comparison of scenario models between urban and rural residents (Discussion SOM 01). We found that mutualism beliefs about wildlife and emotional mediation of the attitude-lethal control pathway were stronger among urban residents whereas rural residents showed consistently strong effects of general attitudes on lethal control acceptability. Attitudes may thus be more proximate than emotions to assessing the acceptability of deer management among rural residents. Social conflicts over wildlife-related beliefs and emotions will likely be amplified between urban and rural constituencies (Gimpel et al. 2020; Heberlein and Ericsson 2005; Patterson, Montag, and Williams 2003). We encourage future research to explore emotional conflicts at the urban-rural interface. Future research should also examine similarities and differences between Indiana DFW customers (44% of sample) and non-customers (56% of our sample), men (73%) and women (22%), or deer hunters (67%) and non-deer-hunters (37%) because these comparisons may be of interest to the state wildlife management agency. We have provided a baseline analysis necessary to carry out these comparisons.

Our study demonstrates that emotions are a critical part of human cognitions about wildlife. Even when people hold strong beliefs about wildlife or hunting, deer-related emotions can shift attitudes toward management approaches. In the case of white-tailed deer, expressing care for Bambi does not necessitate opposition to hunting. Rather, the “Bambi effect” can be felt by anyone encountering a fawn or, more powerfully, a suffering deer. Here, empathy—“the ability to perceive, understand and care about the experiences of another” human or non-human being (Wharton et al. 2019, 158)– results from the interaction among cognitions, emotions, and situational contexts. Just as emotions can mediate the relationship between attitudes and normative judgements, emotions themselves are mediated by our relationships with each other and with the spaces we inhabit. The experience of situated emotions therefore differs for different genders, cultures, and their intersections (Batavia et al. 2021), which contests notions of homogeneity within, for example, rural or urban communities (Panelli, Little, and Kraack 2004). Similarly, emotions expressed toward one species of wildlife differ not only by the type of animal encountered and its behavior, but also by when, where, and by whom it is encountered. It remains crucial to acknowledge the situational dependency of human-wildlife interactions and assess local landscapes (both social and ecological) prior to developing management interventions (Zimmermann et al. 2021), as interactions can change with variation in people, their cognitions and emotions, wildlife, landcover, and socio-political leanings. Continuing to involve emotions in research on human-wildlife interactions, and conversing with political-emotional geographies, will help to elucidate when, where, and how emotions influence public preferences for wildlife populations and management or conservation outcomes.

Additional information

Funding

This paper is a contribution of the Integrated Deer Management Project, a collaborative research effort between Purdue University and the Indiana Department of Natural Resources Division of Fish & Wildlife. This work was supported by the Indiana Department of Natural Resources under Grant W-48-R-02. A previous version of this manuscript appears in the first author’s dissertation from May 2022. The authors thank the Human Dimensions Lab and its undergraduate research assistants for graciously assisting with survey dissemination and data entry.