Combating the exotic pet trade: Effects of conservation messaging on attitudes, demands, and civic intentions

The exotic pet trade poses a major threat to biodiversity conservation. To combat biodiversity loss, it is essential to reduce demand for exotic pets and engage people in civic actions for wildlife conservation. Although messaging has been extensively used in conservation practice, little is known about how it can influence attitudes and various types of actions pertaining to the exotic pet trade. This study examined the impact of conservation messaging in the context of exotic pet ownership and wildlife entertainment visitation as common practices of the exotic pet trade. We randomly assigned participants in the United States to one of five messaging conditions: biodiversity loss and animal abuse (M1), zoonotic disease risks (M2), illegality (M3), social disapproval (M4), and neutral biological information as a control condition (M5). We found that all conservation messages (M1–M4) significantly decreased people's favorable attitudes toward the exotic pet trade and their desire to visit wildlife entertainment. However, conservation messaging did not influence the desire for exotic pet ownership or intentions to take civic actions. Our findings highlight the potential of conservation messaging for attitude change and demand reduction for wildlife entertainment, but different approaches are necessary for promoting more effortful actions such as exotic pet ownership and civic actions.


| INTRODUCTION
The private keeping of wild animals is widespread and increasingly popular internationally, driving demand for the unsustainable exotic pet trade (Bush et al., 2014;Lockwood et al., 2019;McMillan et al., 2021;Siriwat & Nijman, 2018;Smith et al., 2017;Tapley et al., 2011).Exotic pets refer to wild animals that do not have a long history of captivity and are kept at home or commercial entertainment venues such as roadside zoos, circuses, exotic animal cafés, and a variety of tourist attractions (Bush et al., 2014;Collard, 2020).Although the actual number is unknown, these animals are often taken from their natural habitats and traded in high volume and diversity to satisfy global demand each year (Nijman & Shepherd, 2009;Smith et al., 2017).For example, between 2009 and 2018, the United States imported live, wild-caught animals, totaling more than 8.8 million individuals from globally unregulated terrestrial vertebrate species and 800,000 individuals from species regulated under the United Nations Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES; Watters et al., 2022).Among these imports, 28% of unregulated species and 72% of regulated species were at risk of imminent extinction, indicating the significant conservation challenges posed by the unsustainable global and domestic exotic pet trade (Watters et al., 2022).However, its true extent remains largely undocumented due to challenges in tracing and the rise of nontraditional platforms such as online marketplaces (Indraswari et al., 2020;Nijman et al., 2017;Siriwat & Nijman, 2018;Toomes et al., 2023).
The scale of the exotic pet trade threatens biodiversity conservation and animal welfare (Ashley et al., 2014;Baker et al., 2013;Harris et al., 2017;McMillan et al., 2021), poses risks to public health and safety (Can et al., 2019;Chomel et al., 2007), and undermines the rule of law (Freund et al., 2017;Kitade & Naruse, 2020;Lyons & Natusch, 2011).Captive-bred exotic pets still present conservation challenges because these animals can originate from wild populations due to the increasing exploitation of breeding facilities for illicitly sourcing the animals (Lyons & Natusch, 2011).Consequently, the demand for such captive-bred pets can drive the need for more wild-caught animals.To tackle the illegal or unsustainable trade, substantial efforts have been made globally, including trade restrictions through both national legislations and CITES, law enforcement, incentive programs, and demand reduction interventions including campaigns (Challender & MacMillan, 2014).In particular, there has been a growing literature on demand reduction, providing valuable insights into potential strategies to influencing consumer behaviors (Chaves et al., 2018;Doughty et al., 2021;MacFarlane et al., 2022;Moorhouse et al., 2021;Rizzolo, 2021;TRAF-FIC & The Behavioural Insights Team, 2018;Veríssimo et al., 2020;Wang et al., 2020).
Despite these measures, current approaches have important gaps.In general, demand reduction efforts have primarily focused on discouraging people from owning exotic pets, while paying much less attention to public demand for wildlife entertainment (Veríssimo & Wan, 2019).The exhibition of wild animals for commercial entertainment purposes is a major force driving the exotic pet trade (Andersson et al., 2021).Furthermore, most behavior change interventions have only targeted discrete private actions such as recycling, conserving energy and water, and avoiding unsustainable consumption (Bergquist et al., 2023;Farrow et al., 2017;Mertens et al., 2022;Nisa et al., 2019;Thomas-Walters et al., 2023), yet overlooking individuals' potential contributions to changing broader systems through collective civic actions (e.g., voting, signing a petition, and participating in a demonstration, Naito et al., 2022).Combating the exotic pet trade will require not only reducing demand or transforming policies, trade regulations, and industry standards, but also individuals to engage in civic actions that call for system change (Naito et al., 2022).
In conservation science and practice, messaging has been used widely to provide information and engage people in a variety of conservation actions (Kidd et al., 2019;Nielsen et al., 2021;Veríssimo, 2019).Although the effectiveness of messaging varies across intervention types and target behaviors (Bergquist et al., 2023;Thomas-Walters et al., 2023), messaging can promote positive change in pro-environmental attitudes, potentially with lasting impacts (Echeverri et al., 2017;Nabi et al., 2018;Ranney & Clark, 2016).There is also encouraging evidence of framing (i.e., the way message content is constructed and presented) to motivate individual choices and actions that could benefit wildlife conservation (Eyster et al., 2022;Moorhouse et al., 2017).
While prior research has shown a gap between attitudes and behaviors, where individuals fail to translate their beliefs and aspirations into concrete actions (Dilkes-Hoffman et al., 2019;Guagnano et al., 1995), certain attitudes can lead to pro-environmental intentions and behaviors, especially when external factors provide conditions that eliminate barriers to action or enable action (Ajzen, 1991;Bamberg & Möser, 2007;Klöckner, 2013).The influence of attitudes on pro-environmental behavior holds important implications for wildlife conservation.Attitudinal change plays a vital role in shaping public opinions and facilitating support for conservation policies and behaviors, thereby advancing sustainability goals (Kidd et al., 2019;Leiserowitz et al., 2006;Manfredo et al., 2021;Pietsch & McAllister, 2010).
The purpose of the current study is to better understand both possibilities and limitations of messaging for influencing attitudes and various types of individual actions in order to inform future behavior-change interventions.Here we conducted an online randomized controlled trial to examine the impact of conservation messages on attitude change associated with the exotic pet trade (H1), reducing demand for exotic pet ownership (H2a) and wildlife entertainment visitation (H2b), and promoting a range of civic actions for wildlife conservation (H3).We also investigate how exposure to different conservation messages might affect attitudes toward exotic pet trade and motivate conservation intentions.While the focus of this research is on examining the causal relationships between conservation messages and the outcome measures, it is important to note that we do not explore the interactions among these measures nor suggest a causal progression from attitude change to behavior change.

| Data collection and sampling
A power analysis revealed a minimum of 485 participants in total was required for this study (for an assumed minimum effect size f = 0.15, alpha = .05,power = 0.80, five between-subject conditions and two within-subject measurements, and an assumed correlation among repeated measures of .80;Faul et al., 2009).Thus, we recruited 527 participants via Amazon Mechanical Turk (MTurk) after removing participants who failed attention checks (i.e., a question assessing participants' ability to identify a statement not mentioned in the message they just read) or completed the survey in 100 seconds or less.Participants who were 18 years old or older and living in the United States took the survey in return for US$0.50.For demographics, 59% of the participants were female and 78% were Caucasian.Median annual household income was within the range of $50,000-99,999; 49% of the participants identified as liberals, 20% moderates, and 32% conservatives; and 13% of the participants reported to have owned an exotic pet and 58% have been to a commercial wildlife entertainment venue to see exotic animals.All data were collected anonymously using the Qualtrics survey platform between April 16 and 26, 2021.This study was conducted under the approval of the University of British Columbia Behavioral Research Ethics Board (H20-03797).The survey instrument is presented in SI-1.

| Messaging conditions
To examine the effects of messaging, we developed five types of messages.Of the five conditions, four conservation messages highlighted the negative consequences of the exotic pet trade with regards to biodiversity loss and animal abuse (M1), zoonotic disease risks (M2), illegality (M3), and social disapproval (M4).By emphasizing different aspects of the consequences, our conservation messages aimed to elicit certain emotional, perceptual, cognitive, or behavioral responses among participants (Bargh, 2006).A fifth neutral statement provided simple biological information about wild animals (M5); the effects of conservation frames were assessed against this "control condition."Figure 1 presents the five message frames presented in the survey.
The selection of messaging conditions in our study was guided by prior research that employed similar approaches to evaluate the impact of framing on pro-environmental behaviors (Cialdini, 2003;Echeverri et al., 2017;Moorhouse et al., 2017Moorhouse et al., , 2021)).Biodiversity loss, animal abuse, zoonotic disease risks, and illegality are widely recognized issues in the wildlife trade (Bush et al., 2014;IPBES, 2020).Here we combined biodiversity and animal welfare concerns into one messaging condition because these two issues are closely related in people's minds (Dubois & Fraser, 2013) and in their efforts to protect both biodiversity and the well-being of individual animals (Sekar & Shiller, 2020;Wallach et al., 2018).Additionally, we carefully developed the specific narratives of our conservation messages based on past studies that have identified key determinants of pro-environmental behaviors (van Valkengoed et al., 2022).Factors such as problem awareness, internal attribution (e.g., morality and ascription of responsibility), social norms, as well as contextual factors and habits, all play crucial roles in shaping pro-environmental attitudes, intentions, and behaviors (Bamberg & Möser, 2007;Klöckner, 2013;Onwezen et al., 2013).Given the constraints of an online experiment, we directed our focus toward cultivating awareness of consequences and evoking a sense of responsibility when constructing the narratives for each conservation message.

| Outcome measures
Attitudes toward the exotic pet trade (DV1) Participants were asked to indicate their reactions to images exemplifying the private ownership of exotic pets and wildlife entertainment visitation on a 7-point Likert scale ranging from "1 = horrible" to "7 = great" (see SI-1 for the exact questions).In our analyses, we referred to exotic pet ownership and wildlife entertainment visitation collectively as common practices associated with the exotic pet trade (Andersson et al., 2021).We repeated the question three times based on the three taxonomic groups (i.e., mammals, birds, and reptiles) and used the mean score across the three ratings as a proxy for people's attitudes toward the exotic pet trade.The internal consistency of the three ratings was high (Cronbach's α = .87).

Demand for the exotic pet trade (DV2)
Participants were asked to rate on a 7-point Likert scale of "1=absolutely not desired" to "7 = absolutely desired" on how much they desired to own an exotic pet or visit a wildlife entertainment place in the next few years (see SI-1 for the exact questions).

Intended civic actions (DV3)
Participants were asked to indicate their intentions to take civic actions for wildlife conservation (i.e., "How likely are you to engage in the following actions in the next few years?").We adapted 10 civic actions (Table 1) from previous research as examples of civic engagement (Alisat & Riemer, 2015;Tindall & Robinson, 2017).All responses were measured on a 7-point Likert scale ranging from "1 = extremely unlikely" to "7 = extremely likely."The internal consistency of the 10 ratings was high (Cronbach's α = .95).Thus, we averaged the 10 ratings to derive a mean score as a measure of civic action intentions.

| Procedure
Pre-messaging Before reading the message, participants in all five conditions viewed several images of exotic pets and indicated their attitudes toward the exotic pet trade (DV1).They also indicated the level of desire to own an exotic pet or visit a wildlife entertainment place in the future (DV2).

Messaging
Participants were randomly assigned to one of the five messaging conditions.In each condition, participants read the assigned message and had to pass an attention check to make sure they had read the message.Only those who passed the attention check proceeded to complete the rest of the survey.

Post-messaging
After reading the message, participants in all five conditions viewed the same images and answered the same questions regarding their attitudes toward the exotic pet trade and the desire to own an exotic pet or to visit a commercial wildlife entertainment venue (DV1 and DV2).Additionally, participants indicated their intentions to engage in a range of wildlife conservation civic actions (DV3).Demographic questions about gender, ethnicity, the highest education level attained, political orientation, and household annual income were included at the end of the survey.

| Statistical analyses
All analyses were performed with the statistical programming software R (R version 4.1.3).In all statistical models, we set the significance level at α < .05 with two tails, applied Bonferroni corrections for multiple comparisons, and calculated partial eta (η p 2 ) as effect sizes.We adopted Cohen's benchmark to interpret the magnitude of effect sizes: η p 2 values of .01≤ .06 as small, .06≤ .14 as medium, and ≥.14 as large effects, respectively (Cohen, 1988, pp. 285-288).
We first conducted a two-way mixed-design ANOVA (within-subjects 2 pre-post time Â between-subjects 5 messaging conditions) to examine the extent to which conservation messages changed people's attitudes toward the exotic pet trade (H1).The within-subjects design enables higher power because it compares an individual's response before and after messaging, as would a paired t-test, rather than simply comparing population-wide averages.In case of a significant main effect of messaging, we would subsequently perform post-hoc tests to identify any significant differences between conditions.
Second, we conducted the same two-way mixed-design ANOVA to examine whether conservation messages changed people's desire to own exotic pets or to visit wildlife entertainment venues (H2a for owning and H2b for visiting).In addition, we conducted non-parametric Kruskal-Wallis tests because of the violation of normality assumption in our data (see SI-2).Both statistical tests produced highly similar results, and therefore, we provide our ANOVA results here as they are more statistically robust (Norman, 2010).
Finally, we performed a one-way between-subjects ANOVA to compare the effects of messaging across the five conditions on people's intentions to engage in civic actions for wildlife conservation (H3).Without the benefit of pre-messaging values for the civic actions, this test thus has weaker power.

| Attitudes
The two-way ANOVA for H1 revealed significant main effects of pre-post time (F[1,522] = 396.972,p < .001,η p 2 = .432)and conditions (F[4,522] = 3.808, p = .005,η p 2 = .028).Importantly, there was a significant two-way interaction between time and conditions with a small effect size (F[4, 522] = 6.635, p < .001,η p 2 = .048).Post hoc analyses further showed that participants' favorable attitudes toward the exotic pet trade significantly decreased after messaging in all conservation messaging conditions (M1-M4) by 23%-33%, compared to the control condition (M5) (Figure 2a).No significant interactions were found between any of the four conservation messaging conditions, suggesting that the four conservation messages are roughly equally effective in reducing favorable attitudes toward the exotic pet trade.Means and standard deviations of these measures are provided in SI-3.
T A B L E 1 Measures for intended civic actions for wildlife conservation.

Civic actions for wildlife conservation
Intention to learn about exotic pet trade issues This means that the conservation messages had little impact on reducing demand for exotic pet ownership compared to the control condition.
Because a large proportion of participants indicated a low desire to own exotic pets prior to messaging (i.e., they scored <3 on average on the 7-point Likert scale), this may be a floor effect obscuring the impact of messaging.Thus, we conducted a subgroup analysis, splitting the data into likely exotic pet owners (i.e., those who scored ≥4 on the scale) and unlikely owners (i.e., those who scored <4 on the scale), and performed a two-way mixed-design ANOVA test for each group.However, this additional analysis showed no significant interaction effects in either group, supporting our main finding of conservation messaging for having little influence over people's desire to own exotic pets (see SI-4).This said, it is still possible that the floor effect somewhat biased our result.
The post hoc analyses showed that all conservation messages (M1-M4) significantly reduced people's desire to visit wildlife entertainment venues by 20%-26%, compared to 6% reduction in the control condition (M5) (Figure 2c & SI-3).There were no significant interactions between any of the conservation messaging conditions, suggesting that the four conservation messages were likely to be equally effective in reducing the demand to visit wildlife entertainment.

| Intended civic actions
Lastly, the one-way between-subjects ANOVA revealed no significant differences among the five messaging conditions on people's intention to take civic actions for wildlife conservation (F[4,522] = .756,p = .554,η p 2 = .006)(Figure 2d).Means and standard deviations of the original 10 action items are provided in SI-5, and the one-way ANOVA output of each action is available in SI-6.

| DISCUSSION
The current paper demonstrates both possibilities and limitations of conservation messaging for influencing people's attitudes and behavioral intentions in relation to the exotic pet trade.The results show that conservation messages emphasizing the negative consequences of the exotic pet trade are effective at reducing favorable attitudes toward the exotic pet trade and consumer demand for wildlife entertainment.However, we did not find significant evidence that messaging reduces demand in exotic pet ownership or promotes civic action intentions for wildlife conservation.Such variations in messaging effects may be explained in terms of the nature of actions in question.Although both avoiding wildlife entertainment visitation and exotic pet ownership are classified as private actions (Naito et al., 2023), these two actions are likely different in terms of the level of personal investment.Wildlife entertainment visitation tends to occur more casually and occasionally with relatively low effort, whereas exotic pet ownership entails strong preferences and long-term commitment and relationships with animals.In general, people are unlikely to be moved easily by value-based messages when they hold strong preferences or convictions (John et al., 2009;Sunstein, 2017).Thus, as with other consumers of wildlife as traditional medicine and ornaments (Burgess et al., 2019;Doughty et al., 2019), individuals who strongly favor exotic pet ownership are not likely to be easily influenced by conservation messaging.On the other hand, messaging seems to be effective in changing more casual and spontaneous behaviors, such as wildlife entertainment visitation, because these behaviors are relatively easy and do not require careful deliberation or major adjustments in one's convictions.This finding is consistent with previous studies which showed that the impact of messaging is generally limited to casual private actions such as energy use and recycling (Asensio & Delmas, 2015;Farrow et al., 2017;Luo et al., 2022).This suggests that different approaches are necessary for engaging with existing and potential exotic pet owners who have a strong interest in private keeping of wild animals.Measures such as regulatory reforms, strict law enforcement, and imposing barriers to the purchase and private keeping of wild animals might be more effective in dissuading potential owners of exotic pets.
The current study also shows little impact of conservation messages on motivating people to engage in civic actions, although this part of the study had more limited power (without pre-messaging measures).Yet many wildlife campaigns use conservation messages to solicit donations, petition signatures, and sharing of online content-some of the civic actions we measured in the study.The impact of conservation messaging on civic action may be limited because civic engagement is often intentional by nature and strongly driven by social ties, relational values, and the availability of resources, rather than one's awareness of the environmental issues (Jones, 2006;Tindall & Robinson, 2017).Moreover, connection to other devoted activists and the feeling of achievement are key to active civic engagement (Manzo & Weinstein, 1987).Thus, promoting civic actions will likely require more socially embedded approaches than messaging that help enhance the sense of community and human-nature relationships.
Consistent with a large body of research on attitudes, our results demonstrate that conservation messaging can reduce overall favorable attitudes toward the exotic pet trade at least for a short period of time.This finding is encouraging because attitudes can influence public opinions, preferences, acceptance about certain policies and behaviors, and engagement in pro-environmental actions (Albarracin & Shavitt, 2018;Pietsch & McAllister, 2010;Sockhill et al., 2022).This also has an important implication beyond individual behavior change.If conservation messaging can successfully shift public opinion about the exotic pet trade beyond a tipping point, even with as little as 10% of the population (although the actual percentage depends on specific contexts), it could cause the majority to adopt the minority opinion (Xie et al., 2011).Consequently, there is a possibility that the once widely accepted practices of owning exotic pets and visiting wildlife entertainment venues may no longer be perceived as socially appropriate, similar to the shifts that have occurred in other practices such as public smoking and the use of wild animals in circuses.Such a shift in public sentiments might lead to a large-scale behavior change and the transformation of systems in which the exotic pet trade operates (Naito et al., 2022;Otto et al., 2020).However, the effectiveness of messaging in producing lasting changes in attitudes and behaviors remains unknown.A longitudinal study is required to explore this further.
Unlike our study, Moorhouse et al. (2017) found that legality and zoonotic disease information significantly reduced consumer demand for exotic pets, as opposed to information regarding conservation and welfare impacts.The difference between their study and ours may be partly due to the difference in study designs and participant characteristics.Our study used a pre-post message design across treatment and control conditions to examine the impact of conservation messaging, but Moorhouse et al. (2017) focused solely on post message, between-subject effects without considering pre-existing baselines before reading the message.Moreover, while our study participants were drawn from the general public in the U.S. population (Buhrmester et al., 2011;Levay et al., 2016), Moorhouse et al.'s participants were those who voluntarily visited their "exotic pet match" website after searching for keywords associated with exotic pet ownership, a self-selected group who were interested in owning exotic pets.Our subgroup analysis showed that participants with an interest in owning exotic pets in the future were unaffected by the conservation messages, but the sample size of these participants was small.As such, the subgroup analysis may be underpowered.
Although messages can sometimes backfire to produce the opposite effects (Cialdini, 2003;Miller & Prentice, 2016), none of our conservation messages on average resulted in more favorable attitudes toward the exotic pet trade, higher demand for exotic pet ownership and wildlife entertainment visitation, or lower levels of intentions to engage in civic actions.Past research suggests that information tends to backfire when it highlights the commonness of undesirable behavior, seems unreasonable and manipulative, threatens self-identity, or provokes resentment (Chan et al., 2017;Hummel & Maedche, 2019;Schultz et al., 2018).Backfire effects can undermine efforts to promote sustainable behaviors.Thus, it is crucial that conservation messages avoid conveying information that may be construed as threatening individuals' sense of self or righteousness (White et al., 2019).
Future research might examine the effectiveness of conservation messaging on different segments of the population, including those who have owned exotic pets or have a strong interest in buying one in the future.Such a segmented approach will be particularly useful for designing targeted interventions, given that everyone holds their own views on environmental challenges and reasons for action or inaction (Zhao et al., 2021).We also recommend using actual behavioral measures rather than intentions in future experiments.On average, intentions have been found to explain about 28% of the variance in actual behaviors (Sheeran, 2002).Adding actual behavioral measures will likely provide stronger evidence to support the effectiveness of messaging (Clements et al., 2015).For example, a recent study by Luo et al. (2022) has incorporated actual behavioral measures in a field setting to examine the effectiveness of visual conservation messaging for reducing plastic waste.Importantly, more research is needed to better understand effective approaches to promoting civic actions.Future research might investigate the effectiveness of social-oriented interventions in an experimental design, as civic actions tend to address issues affecting societies and require people to work together.Finally, it is important to note that the generalizability of our findings may be limited to the United States where all of our participants were sampled, especially in the context of civic action which could manifest differently in other societies or countries.Since wild animals are extensively traded for pets in East and Southeast Asia, Europe, and the Middle East as well (Andersson et al., 2021), we recommend that future research may also focus on these key regions.

| CONCLUSION
As the world faces an unprecedented biodiversity loss and ecological crisis, there is a growing recognition that wildlife conservation needs diverse approaches to transformative change.The current paper showed both potentials and limitations of messaging for influencing attitudes about and intentions to act in response to the unsustainable exotic pet trade.So far, conservation campaigns have largely adopted unitary messaging strategies for both demand reduction and calls to civic action.Instead, we suggest that such campaigns use more targeted approaches to behavior change based on different types of actions and also focus on attitude change to promote new norms and meanings of the exotic pet trade.Attending to these differences will likely advance the effectiveness of conservation interventions.
AUTHOR CONTRIBUTIONS RN, JZ, and KM conceptualized the study.RN and JZ designed the study and survey instrument with feedback from KC. RN collected the data and analyzed the data with inputs from JZ. RN created the figures; and led the writing of the manuscript with insightful comments and edits from JZ and KC; and all authors gave final approval for publication.

ACKNOWLEDGMENTS
This study was conducted under the approval of the University of British Columbia Behavioral Research Ethics Board (H20-03797).The authors would like to thank Dr. Alejandra Echeverri for her disciplinary expertise and statistical advice.We also thank members of the CHANS Lab and the Behavioral Sustainability Lab for their helpful feedback on the design of the survey instrument.This work was supported by the Four-Year Doctoral Fellowship Program from the University of British Columbia, the Canada Research Chair Program, andSocial Sciences andHumanities Research Council (Insight Grant #435-2017-1071).

F
I G U R E 1 Message frames in M1-M5, highlighting different aspects in relation to the exotic pet trade.

F
I G U R E 2 Changes in attitudes and behavioral intentions, for five message frames.The four panels: (a) favorable attitudes toward the exotic pet trade, (b) desire to own exotic pets, (c) desire to visit wildlife entertainment venues before and after messaging, and (d) civic action intentions (depicted as box and violin plots to represent the multiple civic actions together in a between-subject design without a premessaging measure).Message frames highlight animal abuse and biodiversity loss (M1), zoonotic disease risks (M2), legality (M3), social disapproval (M4), and neutral biological information as a control condition (M5).Dots represent means, and vertical lines represent error bars at 95% confidence intervals.