Introduction

Biodiversity and its related aesthetic experiences

As more than half of the world’s population now lives in cities, the global population has shifted to predominantly urban, and it is projected to become more urban in the future (Sanderson et al. 2018; United Nations, D. o. E. a. S. A., Population Division, 2018). Although urbanization brings both opportunities for economic advancement and social mobility, the expansion of cities also presents significant global environmental challenges, such as habitat loss, diminished habitat quality, fragmentation, and dwindling connection to nature (Liu et al. 2016; Wei et al. 2022). These ramifications are the main drivers of biodiversity loss in multiple taxa, including birds, amphibians, arthropods, and fungi (Bainard et al. 2011; Hamer and Parris, 2011; Herrmann et al. 2012; McIntyre et al. 2001; Shochat et al. 2010). To alleviate the challenges of urbanization and its impact on biodiversity, researchers advocate for enhancing and conserving biodiversity within cities (Langhans et al. 2023). Besides helping to tackle the biodiversity-extinction crisis, urban biodiversity conservation can provide ecosystem services that contribute to improving human health and well-being and enhancing connectedness to nature, particularly by improving access to urban nature (Cox et al. 2017; De Bell et al. 2018; Shanahan et al. 2015; Soanes et al. 2023).

Among various biodiversity ecosystem services, aesthetic value is one of the notable benefits that humans can directly perceive (Tribot et al. 2018). It is also believed to contribute to quality of life, health, or vitality by delivering inspiration, harmony, and peace (Millennium Ecosystem Assessment, 2005). In the research field regarding the aesthetic perception of landscapes, ecological value, and biodiversity, aesthetics is defined as “the enjoyment and pleasure felt through the observation of environmental scenery” (Swaffield and McWilliam, 2013; Tribot et al. 2018). According to Chatterjee and Vartanian (2016), aesthetics encompasses a wide range of elements, including emotional reactions, sensory experiences, and even meaning and knowledge, which can influence decision-making in humans. Therefore, aesthetics is a key determinant in how individuals connect with, experience, and react to the surrounding spaces in urban contexts (Gobster et al. 2007). The aesthetics of urban spaces can evoke emotions, establish the atmosphere for social interactions, and even influence an individual’s sense of belonging (Wang and Yu, 2018). Beautifully designed public spaces, vibrant street art, well-maintained green areas, and thoughtfully planned infrastructure can create a sense of harmony and engagement for residents and visitors alike (Mansor et al. 2012). Within urban settings, the coexistence of diverse plant species and animals, often found near water bodies like lakes or rivers and in urban green spaces, elicits a sense of harmony and tranquility (Szlavecz et al. 2011). This phenomenon fosters people’s profound connection with the surrounding natural world (Lepczyk et al. 2017; Nisbet et al. 2020), thus generating an environment that facilitates emotional bonds and nurtures a profound sense of belonging (Berger et al. 2022). Although aesthetic experiences of nature and wildlife species are also positively linked to environmental ethics, pro-environmental intentions and behaviors (de Pinho et al. 2014; Pereira and Forster, 2015; Tribot et al. 2018; Wang and Yu, 2018), how aesthetic experiences of biodiversity are associated with the belief in biodiversity loss seems to be missing. Thus, it is necessary to address this knowledge gap by examining the association between biodiversity aesthetics and biodiversity-related beliefs among people, especially urban residents who are often perceptually disconnected from nature (Nguyen, Le, et al. 2023; Nguyen, Nguyen, et al. 2023).

In urban settings, plants and pets are two common sources of information that can represent the natural environment, specifically biodiversity. Diverse plant environments are crucial for the well-being of the environment and human communities (Sandifer et al. 2015). According to Janeczko et al. (2020), urban environments with lush greenery can positively affect people’s physical and psychological well-being. These environments mitigate allergies, reduce mortality rates, and promote general wellness (Aerts et al. 2018). Urban vegetation can provide refreshment and relaxation, similar to the benefits observed in wooded settings. In addition, exposure to natural settings helps reduce stress and cognitive strain and improves learning (Vella-Brodrick and Gilowska, 2022). As for pets in urban areas, traditional options like dogs and cats remain popular today (Konstantinova et al. 2021). However, other species, like birds and arthropods, are increasingly acknowledged for their capacity to acclimate to urban environments, thereby establishing a unique niche for themselves (Faeth et al. 2011). Global urbanization prompts a reconsideration of pets’ importance to humans; in densely populated urban areas, where loneliness and alienation are prevalent due to factors such as social isolation, lack of community cohesion, and limited interpersonal connections, pets as companions gain increased importance (Wong et al. 2019). Beyond mere companionship, pets can offer emotional support and contribute to the mental well-being of humans (Fudge, 2014). For instance, interactions with dogs can induce positive emotions, including heightened levels of the “happy hormone” endorphins (Hasin et al. 2018). Given these impacts of plant and pet diversity, they are expected to contribute to the humans’ mental processes and behaviors in relation to the natural environment.

Factors underpinning biodiversity loss belief

Multiple theories and frameworks, such as the Theory of Planned Behavior, Wildlife Value Orientation, Value-Attitude-Behavior Model, etc., posit that beliefs are key factors that shape individuals’ cognitive processes and behaviors in relation to the environment (Ajzen, 1985, 1991; Clayton and Myers, 2015; De Leeuw et al. 2015; Johnson et al. 2021; Lou et al. 2020; Manfredo et al. 2009; Nguyen, Nguyen, et al. 2023; O’Connor et al. 1999; Vaske and Donnelly, 1999). Belief in biodiversity loss can safeguard against extinction denialism and facilitate pro-conservation thinking and behaviors (Lees et al. 2020; Menzel and Bögeholz, 2010; Nguyen, Nguyen, et al. 2023). The theory of Wildlife Value Orientation suggests that values are considered basic patterns of the thought process and shared by members of a social group, so they can profoundly influence attitudes, inform beliefs, and guide behavior toward specific issues, including environmental sustainability (Manfredo et al. 2021; Manfredo et al. 2016; Manfredo et al. 2009). For instance, people with a domination value orientation are more likely to prioritize economic interests and private property rights. In contrast, those embracing mutualism value orientation tend to prioritize habitat protection and equal treatment of interest groups (Manfredo et al. 2016).

However, while these theories and frameworks help explain the relationship between biodiversity conservation’s values, beliefs, and behaviors, they face limitations in reasoning how interactions between humans and their surrounding nature can influence their values, attitudes, and beliefs. One typical theory that can help explain this aspect is the Biophilia Hypothesis, proposed by the biologist Edward Osborne Wilson (Kellert and Wilson, 1995; Wilson, 1986). The Hypothesis suggests that humans have an innate emotional affiliation to other living organisms. Because humans evolved in natural environments, they developed an intrinsic bond with the flora and fauna around them, possibly reinforced by genetic information passed down through generations. This inherent connection to nature influences how individuals perceive and interact with their environment, including urban landscapes (Wilson, 2007). However, the Hypothesis considering Biophilia as a human universal, as part of “ultimate human nature,” has been critiqued for its lack of variability (Woods and Knuth, 2023), which is not suitable for reasoning dynamic socio-psychological issues. Therefore, the Biophilia Hypothesis alone is inadequate for explaining the issue investigated in this study: the relationships between biodiversity interactions, aesthetic values, and biodiversity loss beliefs.

In the current study, the Mindsponge Theory, a psychological and social theory of minds developed based on the evidence from brain and life sciences, was employed as the theoretical foundation for reasoning the psychological process behind the studied relationships for several reasons (Vuong, 2023). Specifically, we drew on two key assumptions from this theory.

Firstly, humans’ beliefs are influenced by multiple elements ranging from socio-psychological to biophysical factors, as indicated by the theories and frameworks mentioned above. Being founded on the information-processing perspective, the Mindsponge Theory can help handle, elaborate, and complement various types of psychological, socio-cultural, and biophysical concepts by bringing them to the same standard: the information-processing scheme. Metaphysically, Davies and Gregersen (2014) suggest that under the information-processing explanatory scheme, the conceptual hierarchy of information is expressed as Information → Laws of physics → Matter, instead of viewing information from the orthodox view, Mathematics → Physics → Information. This approach has enabled the study of various phenomena in evolutionary biology and brain sciences through the information-processing lens (Davies and Gregersen, 2014; Dyson, 1999; T. Li et al. 2022).

Secondly, the Mindsponge Theory helps explain the mental variability of people according to different interaction scenarios with the surrounding environment through the updating mechanism. Recent advances in cognitive neuroscience suggest that the brain’s perception and belief formations are affected by the same fundamental mechanism, which can be reflected through the Bayesian approach (De Lange et al. 2018; Fletcher and Frith, 2009; Seth and Friston, 2016). In particular, how a person perceives something at a basic sensory level relies on their prior knowledge of it (i.e., one’s existing beliefs). Simultaneously, the degree to which one revises these beliefs is influenced by how each experience (i.e., perceptions) contributes to them. The dynamics of the Mindsponge Theory align with this mechanism well as one of its fundamental characteristics is the updating mechanism (i.e., the information absorption and multi-filtering processes), so it enables us to explain how urban residents’ interactions with plant and pet biodiversity can be associated with aesthetic experiences and biodiversity loss beliefs. More details of the Mindsponge Theory are presented in the Theoretical Foundation subsection below. Notably, the updating characteristic of the Mindsponge Theory is also aligned with Toomey and Domroese (2013) conservation behavior feedback model, which stipulates how a person’s conservation attitudes and behaviors can be reinforced through the personal experience of conducting a conservation behavior (i.e., participation in citizen science).

Given the points indicated above, the current study aimed to use the Mindsponge Theory for the following objective:

  • Explore how the diversity of pets and in-house plants is associated with urban residents’ belief in the occurrence and significance of biodiversity loss in different scenarios of aesthetic experiences (positive and negative aesthetic experiences at home due to plants/animals).

The Bayesian Mindsponge Framework (BMF) analytics was utilized on a dataset of 535 Vietnamese urban residents across 35 cities in Vietnam to validate the reasoning (Nguyen, 2021; Nguyen, La, et al. 2022; Vuong et al. 2022).

Vietnam serves as a suitable sampling location for addressing the current research objective due to several reasons. Situated in the Indo-Burma region, one of the world’s most biologically significant and threatened hotspots, the country boasts high biodiversity with over 13,200 plant species and approximately 10,000 animal species (Fauna and Flora International, 2021). However, Vietnam’s natural environment and biodiversity are at risk due to the increasing illegal wildlife trade, which is mainly driven by the growing demand in urban areas (Ngoc and Wyatt, 2013; Shairp et al. 2016). In addition, Vietnam is undergoing rapid urbanization, characterized by the quick expansion of urban areas to support its increasing population. This urban sprawl often prioritizes economic development and infrastructure growth over environmental conservation, leading to cities lacking green spaces and natural habitats (Fan et al. 2019). Subsequently, this trend might create a disconnection between urban residents and nature, posing significant risks to public health, well-being, and environmental sustainability (Mai and Zhang, 2016).

For these reasons, understanding factors that can help improve urban residents’ beliefs in the occurrence and significance of biodiversity loss would help build an eco-surplus culture that can contribute to addressing the biodiversity loss crisis from the demand side, shifting the urban planning focus to nurture urban biodiversity, and improving finance for green spaces in Vietnam (Nguyen, 2022; Nguyen and Jones, 2022a, 2022b). Moreover, as the lack of green space is a global issue, disproportionately affecting low-income and racial minority communities in cities worldwide (Schell et al. 2020), insights from Vietnam can be valuable references for other developing countries with similar urbanization characteristics.

Methodology

Theoretical foundation

Mindsponge Theory, a theory of the human mind’s information processing, was used as the theoretical foundation of this study to examine how access to pet and plant diversity affects urban humans’ belief in biodiversity loss in two aesthetic conditions (Vuong, 2023). The theory was developed from the mindsponge mechanism, a socio-psychological framework of the value absorption and ejection of the human mind, and induced by the most recent findings from the brain and life sciences (Vuong and Napier, 2015). The theory is grounded on the information-processing approach to studying the human mind. This approach posits that information serves as the fundamental basis for constructing physical reality, hence facilitating the exploration of intricate phenomena that need a multidisciplinary understanding (Davies and Gregersen, 2014). Numerous studies have employed the theory as a foundational framework to explore socio-psychological phenomena, encompassing the domains of environmental and conservation psychology (Asamoah et al. 2023; Cheng et al. 2023; Huang et al. 2023; Jiang et al. 2022; Khuc, Dang, et al. 2023; Khuc, Tran, et al. 2023; Kumar et al. 2022; Nguyen, Duong, et al. 2024; Nguyen and Jones, 2022a, 2022b; Santirocchi et al. 2023; Tanemura et al. 2022).

The human mind and environment are considered the two main spectrums of the study. The theory views the human mind as an information collection-cum-processor (a system that collects and processes information simultaneously), which helps explain the way people perceive, think, believe, behave, and establish social constructs (Vuong, 2023). Meanwhile, the environment is conceptually a broader information-processing system (e.g., the Earth system, the socio-economic system, etc.) encompassing the human mind. The mind processor’s inputs are information that is absorbed from the surrounding environment or stored in the memory, while its outputs are things that drive human thinking processes and behaviors, such as value systems, perceptions, thoughts (including creative ideas), feelings, etc. The outputs and perceived feedback from the environment will become inputs for subsequent information processes (Nguyen, Jin, et al. 2022; Vuong, 2022). The mindset, buffer zone, and multi-filtering system constitute the mind. The theory defines mindset as a collection of highly trusted information (core values or beliefs) within the human mind. The buffer zone is a conceptual area in which information is temporarily stored before undergoing evaluation by the multi-filtering system (Vuong et al. 2022).

Due to the interconnectedness of biological systems, the mindset is not a steady information collection. It is continually updated due to the dynamic nature of cellular information exchange and the adaptability of neurological systems. The updating processes in human minds use a process known as “live-wiring” rather than the more prevalent “hard-wiring” observed in simpler systems, which rely more heavily on predetermined genetic information (Eagleman, 2015). This phenomenon can be attributed to neuroplasticity (Galván, 2010; Mateos-Aparicio and Rodríguez-Moreno, 2019). As a result, the mindset’s content is likely influenced by the information availability and accessibility in the surrounding environment as well as the multi-filtering process.

For absorbed information from the environment to become highly-trusted information (core values or beliefs), it needs to pass through the multi-filtering system of the mind. It should be noted that although core values and beliefs are considered highly trusted information, they are not identical but are deemed to have a higher level of vitality toward the system than other types of information in the system (those that are newly absorbed or not yet evaluated), as they are synthetic information formed through interactions between information within the mind and that newly absorbed from the environment (Vuong and Nguyen, 2024a, 2024d). When new information enters the mind, it is treated in two different ways: Information integration and differentiation (Vuong and Napier, 2015). If the newly absorbed information is aligned with the existing highly trusted information in the mindset, it will be quickly synthesized and integrated into the mindset (Nguyen et al. 2022). If the newly absorbed information exhibits substantial discrepancies with the existing core values, it will undergo evaluation through a differentiation process in which the cost-benefit assessment will determine whether the emerging information is rejected or accepted (Vuong et al. 2022). In some cases, when the perceived benefits of accepting new information surpass those of keeping existing core values, the new information will replace the core values. Generally, if new information is perceived as potentially beneficial, it will likely be assimilated into one’s mindset and subsequently affect thinking processes, filtering systems, and behaviors. Conversely, if the information is perceived as costly, it will likely be dismissed. When perceived costs and benefits are ambiguous, the information will be temporarily stored in a buffer zone until sufficient data is available for evaluation (Vuong et al. 2022).

Based on the cognitive processes presented above, it is reasonable to assume that Vietnamese urban people’s belief in biodiversity loss depends on the availability and accessibility of biodiversity-related information and the multi-filtering process. The visual elaborations of the information-based cognitive processes behind the emergence of biodiversity loss belief among Vietnamese urban residents are displayed in Fig. 1. In urban settings, particularly in Vietnam, the availability of green spaces, such as public parks, is diminishing as a result of urban development, hence significantly limiting the opportunities for urban dwellers to obtain biodiversity-related information in public settings (Nguyen, 2017; VietNamNet Bridge, 2018). There exist two major pathways via which urban residents in Vietnam might obtain information pertaining to biodiversity. The first way involves the utilization of information transmitters, such as the internet, books, Netflix documentaries, movies, and other comparable media (Boissat et al. 2021; Jones et al. 2019; Vuong and Nguyen, 2023). Another way entails pet ownership and engaging in different kinds of planting and home gardening, among other possibilities. The first approach cannot provide urban residents with firsthand experience of the benefits of biodiversity and the negative consequences of biodiversity loss. In contrast, the latter approach enables people to interact directly with animals or plants, allowing them to perceive these species’ services. When urban residents can perceive the benefits provided by biodiversity, their mindsets might favor biodiversity-related information during the information-seeking, -selecting, and -filtering processes, making them more likely to believe in biodiversity loss. Thus, we assume that urban residents who own more types of pets and in-house plants are more likely to believe in biodiversity loss (see Scenarios B and D in Fig. 1).

Fig. 1
figure 1

Information-based psychological process behind the emergence of biodiversity loss belief among Vietnamese urban residents.

Full size image

Besides the access to the diversity of pets and in-house plants, we also expected the aesthetic experience of the house induced by pets/plants to contribute to the mind’s filtering process. Aesthetics has several different and complementary definitions that are contingent on the specific discipline within which it is delineated, such as art, philosophy, social science, and cultural history (Tribot et al. 2018). In the current study, we examined the urban residents’ aesthetic experience of their houses due to plants/animals, so the definition of landscape aesthetics is preferable. Landscape aesthetics is defined as aesthetic benefits people receive through their senses and interactions with the landscapes (Swaffield and McWilliam, 2013). The aesthetics can be divided into two complementary approaches (Tribot et al. 2018): (1) the transmitter approach, which is associated with the inherent value of a landscape, as evaluated by its biophysical attributes that stimulate an aesthetic response (Fry et al. 2009), and (2) the receiver approach, which describes the landscape through the lens of human perception (Chatterjee, 2014; Müderrisoğlu and Gültekin, 2015).

From the mindsponge information-processing perspective, people’s aesthetic experience can be redefined as a positive perception of aesthetics—an outcome of the information process using information absorbed through the observation of the surrounding environment and highly trusted information stored within the mindset. The absorbed information through the sensory system reflects the biophysical characteristics of the environment. However, when the information enters the mind, it will become a subjective value that is more or less influenced by the previously existing information, especially the core values (Vuong et al. 2022). For that reason, the aesthetic experience of a person is dependent on not only the biophysical attributes of the surrounding environment but also the information priorly existing within the mind.

According to the Mindsponge Theory, the outcome of an information process will be used as input for the subsequent information process (Vuong, 2023). Therefore, the aesthetic experience with the house induced by pets/plants will also contribute to assessing information related to biodiversity. If the experience is positive, biodiversity-related information will be perceived as more beneficial and more likely to be integrated into the mindset (see Scenarios C and D in Fig. 1). On the contrary, if the experience is negative, the information will be perceived as costly and more likely to be ejected from the mind (see Scenarios A and B in Fig. 1). Based on this reasoning, we assumed that urban residents’ aesthetic experience of their house due to plants/animals would moderate the relationship between access to the diversity of plants and pets and belief in biodiversity loss. To validate the assumptions in this subsection, we select the variables and construct the model in the following subsection.

Model construction

Variable selection and rationale

This study used primary data from a dataset of 535 urban residents from 35 cities across Vietnam (Nguyen, 2021). The dataset was peer-reviewed and published in Data Intelligence with the aims of increasing transparency, having the dataset validated by other experts, facilitating the reproduction of analysis results, and offering resources for other researchers to create additional knowledge in the study field (Vuong, 2017, 2018, 2020b). The dataset is about urban residents’ multifaceted perceptions toward biodiversity–human interactions in Vietnam, which is constructed with six major categories, namely: (1) wildlife product consumption, (2) general biodiversity perceptions, (3) biodiversity at home and neighborhood, (4) public park visitation and motivations, (5) national park visitation and motivations, and (6) socio-demographic profiles (Nguyen, 2021). Although the dataset has many variables, only a proportion of variables in the second and third categories of the dataset was employed in the current dataset.

Before the survey was distributed, it was designed based on the interview results of urban residents in the two largest cities in Vietnam. A total of 38 individuals residing in Hanoi and Ho Chi Minh City were engaged in comprehensive interview sessions conducted between November 15 and December 26, 2020. The interviewees’ age, gender, occupation, and prior experiences with nature were considered during the interviewee selection process to diversify their viewpoints. The interview process was completed when it reached “theoretical saturation,” when no new information or perspectives emerged from the data collected (Creswell and Poth, 2018).

After the questionnaire design was finalized, the survey was collected and targeted at urban residents across Vietnam using the non-discriminative snowball sampling technique through a web-based platform, Google Forms. To elaborate, the first batch of participants was recruited from the researchers’ network, and then they were asked to distribute the online questionnaire to their friends or relatives; new referrals were again encouraged to pass the online questionnaire to their friends or relatives who resided in urban areas. The sampling was stopped when there were no new referrals. Five hundred eighty-one responses were acquired between June 18 and August 8, 2021.

However, only 535 responses were eligible for analysis after the four-step quality check. The first step was checking whether the respondents were residing in the urban areas using their responses regarding residency; those living in non-urban areas were excluded (n = 27). Second, respondents who were under 18 years old were excluded (n = 13). Third, repetitive responses were excluded by checking the email address (n = 3). Fourth, those who gave identical answers to a set of questions using the same response in the Likert scale and selecting all answers to checkbox questions were simultaneously excluded (n = 3).

Despite using snowball sampling, we tried to keep the sample’s diversity of origins, gender, age, and educational levels to reduce possible biases (see Table S1 for the description of respondents’ socio-demographic details). For that reason, responses from 35 cities across Vietnam were collected, with most of the participants residing in Ho Chi Minh City (n = 347, accounting for 64.86%) and Hanoi Capital City (n = 107, accounting for 20%). The proportion of females and males was relatively equal, with 57.08% and 42.92%, respectively; people from various age ranges were included, with respondents between 23 and 50 having the highest proportion (64.48%); most respondents’ educational levels were undergraduate (61.68%) and post-graduate (22.55%); a majority of the respondents (46.11%) had the monthly income around 5–15 million VND (approximately $196–$589 at the current exchange rate). According to The General Statistics Office of Vietnam (2024), the average monthly income in urban areas in 2020 was 5.590 million VND (approximately $220), so the income distribution of the study’s sample was relatively aligned with the population’s actual distribution. Finally, all the responses were encoded and saved under a comma-separated value format. A more thorough description of the interview’s and survey collection’s design and results are available in the following study (Nguyen, 2022).

To uphold the ethical standard of the study’s design, we instructed the participants to read and agree with the consent form, which stipulates the research purposes, questionnaire contents, and participants’ confidentiality prior to undertaking the survey. Since this study was not funded, we are not constrained by contractual obligations and can prioritize safeguarding the participants without compromise. Our institutes do not mandate ethical approval for social survey research. Additionally, to the best of our knowledge, formal ethical review boards for conservation research practices were unavailable in Vietnam during survey collection due to the lack of expertise and resources (Brittain et al. 2020).

We relied on established sample size standards to determine the minimum sample size benchmark for the data collection. In particular, VanVoorhis and Morgan (2007) suggest that the reasonable sample size for regression analysis is around 50. Meanwhile, Green (1991) recommends acquiring the sample size calculated based on the inequality N > 50 + 8 m (where m is the number of independent/predictor variables) for testing the multiple regression. A study with five predictor variables, like the current study, requires a minimum sample size of 90. If we follow the sample size determination method based on a confidence level of 95%, a 0.5 standard deviation, and a 5% margin of error, the recommended sample size should be at least 385. Since the dataset’s sample size is well above these mentioned sample size standards, it is plausible to say that it is adequate for data analysis.

For the research objective of the current study, four variables were employed for the statistical analysis (see Table 1). To measure the urban resident’s belief in biodiversity loss, we employed BioLossBelief variable, which reflects how the respondent viewed the biodiversity loss phenomenon. The respondents’ self-reported numbers of pet and plant types at home were utilized to estimate the respondent’s access to biodiversity-related information. Those self-reported numbers are represented by AnimalDiversity and PlantDiversity variables, respectively. Finally, the aesthetic experience of the respondent was represented by the variable HomeAesthetic, which was measured by the question: “How much does the presence of plants/pets affect the aesthetic of your house?” If the respondent answered ‘very negative effect’ or ‘ negative effect,’ they would be encoded as ‘negative effect.’ In contrast, if the respondent answered ‘very positive effect’ or ‘positive effect,’ they would be encoded as ‘positive effect.’

Table 1 Variable Description.
Full size table

Statistical model

To validate the assumptions presented in Subsection “Theoretical foundation”, we constructed model 1:

$${BioLossBelief} \sim {normal}\left(\mu ,\sigma \right)$$
(1.1)
$$\begin{array}{ll}{\mu }_{i}={\beta }_{0}+{\beta }_{1}* {{AnimalDiversity}}_{i}+{\beta }_{2}* {{PlantDiversity}}_{i}\\\qquad+\,{\beta }_{3}* {{HomeAesthetic}}_{i}+{\beta }_{4}* {{AnimalDiversity}}_{i}* {{HomeAesthetic}}_{i}\\\qquad+\,{\beta }_{5}* {{PlantDiversity}}_{i}* {{HomeAesthetic}}_{i}\end{array}$$
(1.2)
$$\beta \sim {normal}\left(M,S\right)$$
(1.3)

The probability around \(\mu\) is determined by the form of the normal distribution, whose width is specified by the standard deviation \(\sigma\). \({\mu }_{i}\) indicates the level of urban resident \(i\)’ belief in the occurrence and significance of biodiversity loss; \({{AnimalDiversity}}_{i}\) indicates the number of pet types that the urban resident \(i\) owned; \({{PlantDiversity}}_{i}\) indicates the number of plant types that the urban resident \(i\) owned; \({{HomeAesthetic}}_{i}\) indicates whether the urban resident \(i\) felt their home aesthetic due to plant/animal; \({\beta }_{4}\) and \({\beta }_{5}\) indicate the coefficients of the non-additive effects of \({{AnimalDiversity}}_{i}\), \({{PlantDiversity}}_{i}\), and \({{HomeAesthetic}}_{i}\) on \({BioLossBelief}\). If the coefficients \({\beta }_{4}\)’s and \({\beta }_{5}\)’s distributions are significant, and the associations between the species diversity and biodiversity loss belief are considered conditional on aesthetic feeling. Model 1 has seven parameters: the coefficients, \({\beta }_{1}\)– \({\beta }_{5}\), the intercept, \({\beta }_{0}\), and the standard deviation of the “noise”, \(\sigma\). The coefficients are distributed as a normal distribution around the mean denoted \(M\) and the standard deviation denoted \(S\). The logical network for Model 1 is presented in Fig. 2.

Fig. 2
figure 2

Model 1’s logical network.

Full size image

Analysis and validation

Bayesian Mindsponge Framework (BMF) analytics was employed in the current study for several reasons (Nguyen, La, et al. 2022; Vuong et al. 2022). First, the method integrates the logical reasoning capabilities of Mindsponge Theory with the inferential advantages associated with Bayesian analysis, as these two approaches exhibit a high degree of compatibility (Nguyen et al. 2022). Second, Bayesian inference is a statistical approach that treats all the properties (including the known and unknown ones) probabilistically (Csilléry et al. 2010; Gill, 2014), enabling reliable prediction of parsimonious models. Nevertheless, utilizing the Markov chain Monte Carlo (MCMC) technique still allows Bayesian analysis to deal effectively with various intricate models, such as multilevel and nonlinear regression frameworks (Dunson, 2001). Third, Bayesian inference has various advantages in comparison to the frequentist approach. One notable advantage is the ability to utilize credible intervals for result interpretation instead of relying solely on the dichotomous decision based on p-values (Halsey et al. 2015; Wagenmakers et al. 2018).

In Bayesian analysis, selecting the appropriate prior is required during the model construction process. Due to the exploratory nature of this study, uninformative priors or a flat prior distribution were used to provide as little prior information as possible for model estimation (Diaconis and Ylvisaker, 1985). A prior-tweaking technique was also employed to check the robustness of the posterior distributions. Specifically, we reran the analysis using two priors reflecting different levels of beliefs in the studied parameters: the prior reflecting our disbelief in the associations (i.e., normal distribution with mean being 0 and standard deviation being 0.5) and the prior reflecting our belief in the associations (i.e., normal distribution with mean being 0.5 and standard deviation being 0.5). If the estimated posteriors remain unchanged, the results can be deemed insensitive to different priors.

Moreover, incorporating priors into the model also helps address the multicollinearity issue. Leamer (1973) suggests that in Bayesian inference, the multicollinearity problem is “the weak data problem associated with large standard errors of estimated coefficients and, in a Bayesian analysis, the coincidence of prior and posterior distributions on certain subspaces.” Hence, effective resolution of multicollinearity can be achieved when prior information is substantial. Specifically, research indicates that the Bayesian method utilizing informative priors outperforms Ridge regression in addressing multicollinearity (Adepoju and Ojo, 2018; Jaya et al. 2019).

After the model was fitted, we employed the Pareto-smoothed importance sampling leave-one-out (PSIS-LOO) diagnostics to check the models’ goodness-of-fit (Vehtari and Gabry, 2019; Vehtari et al. 2017). LOO is computed as follows:

$${LOO}=-2{{LPPD}}_{{loo}}=-2\mathop{\sum}\limits_{i=1}^{n}\log \int p\left({y}_{i}\left|\theta \right.\right){p}_{{post}\left(-i\right)}(\theta )d\theta$$

\({p}_{{post}\left(-i\right)}(\theta )\) is the posterior distribution based on the data minus data point \(i\). The k-Pareto values are used in the PSIS method for computing leave-one-out cross-validation, which helps identify observations with a high degree of influence on the PSIS estimate. Observations with k-Pareto values greater than 0.7 are often considered influential and may be problematic for accurately estimating leave-one-out cross-validation. Commonly, a model is considered fit when the k values are below 0.5.

If the model had a good fit with the data, we would proceed with the convergence diagnoses and result interpretation. In the current study, we validated the convergence of Markov chains using statistical values and visual illustrations. Statistically, the effective sample size (n_eff) and the Gelman–Rubin shrink factor (Rhat) can be used to assess the convergence. The n_eff value represents the number of iterative samples that are not autocorrelated during stochastic simulation, while the Rhat value is referred to as the potential scale reduction factor or the Gelman–Rubin shrink factor (Brooks and Gelman, 1998). If n_eff is larger than 1000, it is generally considered that the Markov chains are convergent, and the effective samples are sufficient for reliable inference (McElreath, 2018). As for the Rhat value, if the value exceeds 1.1, the model does not converge. Typically, the model is considered convergent if Rhat = 1. Visually, the Markov chains’ convergence was validated using trace plots, Gelman–Rubin–Brooks plots, and autocorrelation plots.

The Bayesian analysis was performed on R using the bayesvl open-access package, which provides good visualization capabilities (La and Vuong, 2019). Considering the issues of data transparency and the cost of reproduction, all data and code snippets of this study were deposited onto Zenodo (Vuong, 2018): https://zenodo.org/records/13897554.

Results

PSIS-LOO diagnostics shows that all k values are below the threshold of 0.5, indicating that the constructed model fits well with the dataset (see Fig. 3). Thus, the estimated results can be interpreted.

Fig. 3
figure 3

PSIS-LOO diagnostic plot estimated using uninformative priors.

Full size image

Table 2 shows the estimated posteriors of the analytical model’s parameters and convergence diagnostic values (i.e., n_eff and Rhat). All the n_eff values are above 1000, and all Rhat values equal 1, so the model’s Markov chains can be deemed convergent. The convergence of the Markov chains is also validated by the trace plots (see Figure S1), Gelman-Rubin-Brooks plots (see Figure S2), and autocorrelation plots (see Figure S3) in the Supplementary.

Table 2 Estimated posteriors.
Full size table

Table 2 presents the estimated posterior distributions of the constructed model, and Fig. 4 illustrates them. Since the constructed model is complex, it is necessary to visualize the findings before interpreting them. However, before interpreting the results, the reliability and robustness of the findings need to be evaluated.

Fig. 4
figure 4

Coefficients’ posterior distributions estimated using uninformative priors.

Full size image

Figure 4 illustrates the posterior distributions of the constructed model on an interval plot. The thin blue lines represent the probability mass that is outside the highest credible zone, whereas the thick blue lines indicate the probability mass contained within the 89% Highest Posterior Density Intervals. (HPDI). As seen in Fig. 4, the 89% HPDI of Homeaesthetic, PlantDiversity, and PlantDiversity*Homeaesthetic are entirely distributed on either the positive or negative side of the axis. This suggests that the estimated results of these coefficients are highly reliable. However, a proportion of the distributions of AnimalDiversity and AnimalDiversity*Homeaesthetic is still located on the opposite side of the axis, implying that the association between AnimalDiversity and BioLossBelief and the moderation effect of Homeaesthetic on the relationship between AnimalDiversity and BioLossBelief are weakly reliable.

Even when rerunning the computation using different types of priors reflecting our belief and disbelief in the associations, the estimated posteriors’ tendencies remain unchanged (see Table 2). This implies that the estimation is robust toward changing prior beliefs and multicollinearity issues.

Employing Eq. (1.2) and the estimated mean values of parameters in Table 2, we calculated degrees of biodiversity loss belief. For clarity, the estimated degrees of biodiversity loss belief based on plant and pet diversity are plotted in Fig. 5A, B, respectively. Both Figures suggest that the effects of plant and pet diversity on urban residents’ biodiversity loss share a similar pattern. To elaborate, for urban residents who felt their houses’ aesthetic due to plants/animals, the diversity of plants or pets is not associated with the belief in biodiversity loss. For those who did not feel their houses’ aesthetic due to plants/animals, the diversity of plants or pets is positively associated with the belief. Nevertheless, the association of plant diversity is more significant and clearer than that of pet diversity.

Fig. 5: Estimated degrees of biodiversity loss belief.
figure 5

a Among urban residents who own a plant (plants) but not a pet. b Among urban residents who own a pet (pets) but not a plant.

Full size image

Discussion

Using the Bayesian Mindsponge Framework (BMF) analytics on a dataset of 535 urban residents, the current study found that the aesthetic feeling induced by plants/animals is positively associated with the belief in the occurrence and significance of biodiversity loss. This positive association supports our assumption that urban residents’ aesthetic experience contributes to the information multi-filtering process. When the aesthetic feeling is positive, it will add more perceived value to pet- or plant-related information (including biodiversity-related information) in the subsequent information-seeking, -selecting, and -filtering processes, making urban residents more likely to recognize the occurrence and significance of biodiversity loss. The finding is also aligned with previous research, which underscores the positive influence of aesthetic experience on environmental values and the intention of pro-environmental behavior (Li et al. 2022; Mtutu and Thondhlana, 2016; Wang and Yu, 2018).

We also found a positive association between the diversity of plants and pets and the biodiversity loss belief. Still, the association is conditional on the urban residents’ aesthetic feeling of the house induced by plants/animals. Specifically, the diversity of plants and pets is only positively associated with biodiversity loss belief when urban residents feel that their houses’ aesthetics are negatively affected by plants/animals. However, species diversity has no significant association with the belief in biodiversity loss among those with positive aesthetic feelings. The positive association between species diversity and biodiversity loss belief when urban residents have a negative aesthetic feeling induced by plants/animals can be explained through the information-processing lens of Mindsponge Theory (Vuong, 2023).

The theory suggests that individuals are more inclined to absorb information that resonates with their existing core values (or highly trusted information). Aesthetics is not the only ecosystem service provided by the diversity of plants and pets so that people might perceive benefits other than aesthetics. Although the negative aesthetic feeling induced by plants/animals can make people less likely to absorb biodiversity-related information, other perceived benefits of plant and pet diversity might offset this effect. Studies have suggested that pet ownership and home gardening can result in multiple advantages, such as favorable physical outcomes, improved mental health, connection to nature, place attachment, attention restoration, etc. (Kaplan and Kaplan, 1989; Kruger et al. 2014; Raymond et al. 2019; Samus et al. 2022; Wood et al. 2015; Zhang et al. 2021). The result also aligns with the concept of personalized ecology, which advocates that personal experiences with nature can shape individual perspectives through different stimuli (e.g., sight, sound, smell, taste, or touch) (De Garine-Wichatitsky et al. 2021; Gaston, Soga, et al. 2018). For instance, engagement with birds has been identified as a contributing factor to humans’ health and well-being due to their aesthetics and vocal prominence. Some types of birds can be attracted to feeders, facilitating positive nature interactions in close proximity (Belaire et al. 2015; Gaston, Cox, et al. 2018).

Besides the personalized ecology, the hypothetical Biophilia, which assumes that humans have an innate tendency to seek connections with nature and other life forms, might also be applicable in this situation (Kellert and Wilson, 1995; Wilson, 1986). Perhaps Biophilia exists in people’s mindsets or has been reinforced through genetic information over thousands of years surviving and evolving within the natural environment, making them more likely to recognize the benefits provided by plant and pet diversity (Barbiero and Berto, 2021; Nguyen, Le, et al. 2023). Due to such offsets, people with negative aesthetic feelings about their houses are still likely to absorb more biodiversity-related information when they access a higher level of plant and pet diversity. Despite the explanation above, it remains unclear why species diversity does not affect the belief when urban residents possess a positive feeling about their houses’ aesthetics. Thus, future studies should be conducted to elaborate on this difference.

Notably, the association between pet diversity and biodiversity loss belief is less significant and reliable than that of plant diversity. This could be because plants are often perceived as more stationary elements in the environment, contributing to the overall ambiance without dynamic changes (Huey et al. 2002). Pets, conversely, are more energetic and active members of the living environment (Kateryna et al. 2023). Therefore, as the diversity of pet species increases, the task of pet care concurrently becomes more intricate and challenging, especially when those pets are exotic animals other than dogs and cats. The house’s cleanliness and tidiness will be negatively affected without adequate management, adding the perceived costs of pet diversity into the urban residents’ mindset. These perceived costs might subsequently make the effect of pet diversity on the absorption of biodiversity-related information less consistent and more subject to context. For example, Nguyen, Nguyen, et al. (2023) found that when urban residents feel comfortable at home, they are more likely to believe in the occurrence and significance of biodiversity loss as the pet diversity increases. Nevertheless, when urban residents feel uncomfortable, pet diversity is negatively associated with their belief in biodiversity loss.

Based on the current study findings, we suggest that enhancing the house’s aesthetics through in-house planting or pet ownership can be a crucial way to improve urban residents’ belief in the occurrence and significance of biodiversity loss. When urban residents recognize that biodiversity loss is real and can cause severe consequences to their lives, they will be more likely to seek and absorb information associated with biodiversity conservation, which gradually helps build up the eco-surplus culture among urban residents, “a set of pro-environmental attitudes, values, beliefs, and behaviors that are shared by a group of people to reduce negative anthropogenic impacts on environments as well as conserve and restore nature” (Nguyen and Jones, 2022a; Vuong, 2021). As urban residents are often perceptually disconnected from nature, the eco-surplus culture will help safeguard them from the environmental crisis and extinction denialism and improve their willingness to support pro-environmental actions and policies. Even when the aesthetics cannot be achieved, increasing the urban residents’ connections to biodiversity by planting more types of plants in the house can also help raise their belief in biodiversity loss and subsequently build up the eco-surplus culture. When such eco-surplus cultural values are widely shared within society, they can help stimulate the shift in wildlife values and change attitudes toward wildlife management. For example, it was found that anthropomorphism is strongly related to mutualist values, which regard wildlife as companions in the individual’s social community and worthy of respect (Manfredo et al. 2020). Such values also help drive public support for policies and management practices prioritizing environmental protection and biodiversity conservation (Manfredo et al. 2021). Despite the potential of the eco-surplus culture, the global lack of green space, disproportionately affecting low-income and racial minority communities in cities (Schell et al. 2020), can significantly hinder the socio-cultural transitions. Thus, apart from aesthetic factors and biodiversity, the expansion of green space also needs prioritizing.

Although the influence of education and prior experience on aesthetic appreciation of natural landscapes and properties has not been tested, it is evident that aesthetic appreciation in arts is significantly influenced by complex webs of meaning derived from their self-hood, personal experience, and socio-cultural surroundings (Carlson, 2005; Harland et al. 2000; Jorgensen, 2011; Vuong, 2024). Thus, increasing opportunities for nature interactions (e.g., pets, plants, public parks, nature-based recreation) and environmental education can be a good strategy to shape urban residents’ aesthetic perception. Effectively implementing the strategy can lead to a loop of aesthetic experience, eco-surplus culture, natural conservation, and biodiversity (Nguyen and Jones, 2022a; Nguyen, Nguyen, et al. 2024; Tribot et al. 2018; Vuong, 2021; Vuong and Nguyen, 2024b). This strategy is especially applicable and essential for young generations that are born and nurtured in urban areas, as they are the populations that have limited interactions with the natural environment but will be responsible for environmental decision-making in the future (Vuong, 2020a). The strategy also shares some similarities with the conservation behavior feedback model, which suggests that participation in citizen science can help reinforce the attitudes and intentions to engage in conservation behaviors through a “greener sense of self” process (Toomey and Domroese, 2013).

Vietnam’s rapid urbanization presents a dual challenge as cities expand and green spaces dwindle, creating a growing disconnection between urban dwellers and nature. Our study emphasizes the need to enrich urban residents’ interactions with plants and animals in their immediate surroundings. Drawing upon Vietnam’s cultural heritage, which values harmony with nature, traditional Vietnamese gardens stand as exemplary models for integrating natural elements into urban areas (Tran et al. 2023; Vuong and Nguyen, 2024c). These gardens often incorporate eco-friendly features adaptable to contemporary high-rise housing, promoting sustainability and nurturing a connection with nature (Hong Na and Park, 2011). Our study also advocates for enhancing aesthetic experiences involving plants and animals to deepen urban residents’ connection to nature, aligning with cultural norms and fostering a stronger bond with the environment.

Moreover, urban planning plays a crucial role in shaping environments and residents’ relationships with nature (Colding et al. 2020). By integrating nature-based solutions into urban planning initiatives, cities can grow sustainably while enhancing residents’ quality of life (Mouratidis, 2021). In Vietnam, there is a growing recognition of the importance of incorporating green infrastructure and nature-based solutions into urban development to mitigate biodiversity loss (Mabon and Shih, 2021). Our findings underscore the potential benefits of integrating green spaces, community gardens, and pet-friendly amenities into urban planning initiatives, thereby enhancing residents’ aesthetic experiences and promoting a harmonious coexistence with nature.

Addressing potential challenges posed by stray pets or invasive plant species in urban ecosystems is also essential (Abdulkarim et al. 2021). Stray pets can disrupt the balance of urban ecosystems by competing with native wildlife for food and resources, preying on native animals, and spreading diseases (Neuenkamp et al. 2021). Invasive plant species can outcompete native plants, leading to declining biodiversity and ecosystem health. They can also alter soil conditions and disrupt natural water cycles (Perry and Cox, 2024). Control and management measures, such as removal programs for stray animals and eradication efforts for invasive plant species, are essential for minimizing their impact. Public awareness and education are vital for encouraging responsible behavior among residents, including proper pet ownership and gardening practices. Enacting and enforcing legislation and regulations can effectively control the spread of stray pets and invasive plants.

The study has limitations, so we present them here for transparency (Vuong, 2020b). Firstly, the study’s emphasis on urban areas in Vietnam may restrict the findings’ applicability to other urban areas in different countries and rural settings. The establishment of causal relationships between variables like plant diversity, aesthetics, and beliefs about biodiversity loss is also hindered by a cross-sectional dataset collected at a particular time, so the result interpretation should be correlational. There are also concerns about the long-term impacts of exposure to varied plant species and aesthetic improvements on pro-environmental actions, besides belief in biodiversity loss. Future research is suggested to provide a more thorough investigation of these relationships and their implications for urban conservation efforts by incorporating various populations, using longitudinal designs, and considering confounding factors.

In general, the study suggests that urban residents’ perceived values of nature can be enhanced through their interactions with in-house plants and animals, especially when such interactions are associated with aesthetic experiences. Appropriate integration of green spaces, community gardens, and pet-friendly amenities into urban planning initiatives might facilitate the residents’ aesthetic experiences with nature, which can prevent the dwindling connection to nature and foster stronger bonds and harmonious coexistence with nature. Simultaneously, management measures are also needed to control the spread of stray pets and invasive plants effectively. Although the study is conducted with samples in Vietnam, its implications might serve as helpful references for other urban areas around the globe, especially those with rapid urbanization like Vietnam.