Based on attachment theory, this study builds a theoretical framework. This theory explains the psychological bonds people have with people, places, and activities (Moore, 2002; Scannell and Gifford, 2010; Shen et al., 2022; Tsaur et al., 2014). This theory has been successfully extended to human interactions with the environment (Barrable, 2025). Barrable (2025) explains how an emotional connection with nature helps to strengthen feelings of security, belonging, and identity. There is an essential distinction between gardening as a pastime and as an obligation. Research on psychological stress has confirmed that personal responsibility is a significant stressor (Rink et al., 2023). High personal responsibility for task completion results directly in psychological tension and physiological stress reactions (Iwanaga et al., 2000). It shows that high-stakes situations with great responsibility can be discouraging. In our research model, we assumed that perceiving gardening as a voluntary leisure activity rather than a mandatory task determines its psychological impact. Leisure orientation fosters positive attachment through repeated and meaningful engagement, enhancing personal satisfaction and wellbeing (Shen et al., 2022; Valencia et al., 2022). According to attachment theory, connections with safe and comforting objects provide a sense of security and basic comfort, which reduces stress (Shen et al., 2022; Valencia et al., 2022). When people develop a strong attachment to gardening through a leisure-oriented approach, it becomes an integral part of their identity (Kiesling and Manning, 2010) and coping repertoire (Fjaestad et al., 2023). The immersive and mindful nature of gardening requires focused attention on the present moment (Macaulay et al., 2022). This helps break the cycles of thought that are central to stress and anxiety (Shen et al., 2023). Thus, we constructed a conceptual model that includes both behavioral engagement (frequency) and psychological orientation (attitude toward leisure), which predicts attachment to the activity. As a result, attachment to the activity mediates the relationship between these antecedents and improvements in mental wellbeing, as well as reductions in stress. These findings have implications for the future development of horticultural ecotourism. The basic idea of horticultural ecotourism programs is that they serve as low-commitment and easily accessible entry points. It promotes a leisure mindset, which gives them special significance (Fox, 2017). Visitors participating in ecotourism can enjoy gardening without bearing full responsibility for the long-term care of the garden or the harvest. Being a “guest” is a psychological trait that relieves the pressure associated with performance (Jordan et al., 2019). It turns the activity into pure leisure and is perfect for fostering attachment to it, which has a positive impact on mental health (AlAli et al., 2024). Thus, studying gardening is aimed at optimizing psychological outcomes through the promotion of a leisure-oriented approach to gardening, which is relevant to developing future tourism products.

A high frequency of visits to green areas and private gardens positively impacts physical health and the prevention of chronic diseases (Chalmin-Pui et al., 2021; Hong et al., 2019). Chalmin-Pui et al. (2021) reported that gardening at least 2–3 times a weekly is associated with the most significant health benefits. This includes enhanced physical activity levels and improved overall wellbeing. Garden settings provide fresh air exposure, suppress pathogens, enhance immune responses, and positively influence various physiological systems (Andersen et al., 2021). The importance of garden spaces has become particularly apparent during the COVID-19 pandemic. For example, Corley et al. (2021) found that home gardens served as core health resources for older adults during the lockdown. We propose that repetitive behavioral engagement directly contributes to attachment to an activity.

In line with attachment theory, recent studies suggest that attachment may develop not only toward places but also toward recurrent, meaningful activities embedded in everyday life, such as gardening. Prior research has shown that the core dimensions of attachment theory positively and significantly influence individuals’ pro-environmental behavior and overall wellbeing (Shen et al., 2022). Building on this theoretical foundation, the present research focuses on one attachment dimension, activity attachment, within the context of gardening. Gardening is a long-term, place-based, and practice-oriented activity. It involves repeated interactions with natural elements, making it particularly conducive to the development of emotional bonds through sustained engagement. In this context, activity attachment refers to the extent to which individuals feel emotionally involved and connected to gardening-related practices and experiences. These activities may occur in private or community gardens and include planting, cultivation, and harvesting. These include other routine practices that collectively shape a broader gardening experience (Hume et al., 2022). By emphasizing activity attachment, this study extends existing attachment research and highlights gardening as a distinctive form of activity through which psychological connections and wellbeing can emerge. For example, research in the physical activity domain suggests that individuals’ attachment orientations significantly shape how and why they engage in sustained activities, including their preferences for activity type, intensity, and social context (Hill et al., 2024). This study offers relevant insights into the engagement patterns in gardening. Regular engagement enables individuals to develop familiarity with the activity, skills, and emotional connections associated with it (Shen et al., 2022; Tsaur et al., 2014). Our assumptions also align with the recreation specialization theory, which suggests that greater engagement in leisure activities strengthens place identity and place dependence (Huo et al., 2025).

In high-income countries, gardening is often a leisure activity, whereas in low-income countries, it is a means of providing a livelihood for the family. For instance, many people in Australia and England enjoy gardening or flower cultivation as peaceful and satisfying hobbies (Cheng and Pegg, 2016; Fox, 2017). Alaimo et al. (2024) found that engagement in gardening among new gardeners in the USA resulted in nearly universal benefits, including food production, physical activity, and a core “gardening triad” of psychological rewards, such as the nurturance of plants, feelings of responsibility, and a restorative connection to nature. However, a realist review by Flores et al. (2025) clarifies that motivation is highly context-dependent and that the path to these advantages is not foreseeable. Food security, curiosity, and self-efficacy are essential motivators (Flores et al., 2025). In addition to individual motivations, gardening as a leisure-oriented activity also carries important social and experiential meanings. Empirical evidence shows that time spent in gardens and a stronger connection to nature are associated with higher levels of social cohesion, trust, and interaction, particularly in urban contexts (Oh et al., 2022). These findings suggest that leisure gardening facilitates personal wellbeing and socially embedded experiences. Similarly, research indicates that individuals’ motivations and preferences are strongly shaped by the characteristics of the activities they engage in, and that voluntary and enjoyable participation is essential for generating positive experiences (Valencia et al., 2022). In gardening, engaging in these practices by choice can enhance enjoyment, meaning, and emotional involvement. Together, these studies support the view that when gardening is approached as a leisure activity rather than an obligation, it is more likely to foster meaningful engagement and attachment. In contrast, a behavioral study by Iwanaga et al. (2000) suggests that a sense of high responsibility for one’s task enhances psychological responses. Therefore, the benefits of gardening are most effectively realized when this activity is perceived as an intrinsically motivated leisure pursuit rather than obligatory work (Rubel et al., 2024). Attachment theory suggests that the voluntary and enjoyable nature of leisure-oriented approaches promotes a strong attachment to activities (Shen et al., 2022). Thus, we hypothesized that a psychological orientation toward gardening is essential.

The study was conducted in Shymkent, the third-largest city in Kazakhstan. It has a population of approximately 1.2 million. There are three reasons for selecting Shymkent. First, the city became a third metropolis in 2018. Second, despite rapid urbanization, Shymkent maintains close cultural ties to traditional dacha gardening methods entrenched in Soviet traditions (Yussupova et al., 2019). This creates a unique context for examining the psychological mechanisms of gardening. Third, the study coincides with Kazakhstan’s “Roadmap for Development of Dacha and Gardening Farms for 2024–2026” (Zakon.kz, 2023), which aims to transform private garden plots into accessible green spaces and enhance community wellbeing. However, this policy lacks localized empirical evidence. Our research addresses this gap by examining the psychological impacts of gardening among urban residents in various forms, such as with indoor plants, in private yards, on community plots, or at suburban dachas.

This study used a quantitative cross-sectional design to test a complex mediation model examining the relationships between gardening frequency, leisure orientation, activity attachment, and psychological wellbeing outcomes. The cross-sectional design of the study allowed us to capture current gardening practices and psychological states among a diverse urban population at a single point in time. Subsequently, we demonstrated the relationships between the variables. The design also allowed for a comparison of the results between gardeners and non-gardeners, as well as an examination of the variability within the gardener population. Structural equation modeling (SEM) was selected as the primary analytical approach. It helps in the simultaneous testing of multiple direct and indirect paths, provides an overall assessment of model fit, and accounts for measurement error in latent constructs.

We obtained ethical approval from the Shandong Xiehe University Ethics Committee (approval number: LLSC-KY03-2025028). Data were collected between March and April 2024. All procedures conformed to the Declaration of Helsinki (1964), and the participants provided informed consent before participation.

The non-probability sampling snowball method was employed. This method is essential because Shymkent lacks a comprehensive sample of urban gardeners. The author’s personal experience with the region’s dacha culture provided access to local connections and an understanding of a culturally sensitive approach to research. To ensure diversity among representatives, we collected data by phone and in person from owners of gardening dachas and local agricultural institutions, all of whom were aged above 18 years. The first contacted respondents were asked to forward the survey to those they thought might be suitable. The research assistants used brief personal storytelling techniques to explain the purpose of the study and encourage participants to participate. As noted by Mamirkulova et al. (2020) and Abbas et al. (2025), this is effective in the Central Asian cultural context. Indeed, the cultural characteristics contributed to a high response rate of approximately 70.3%. Of the 300 questionnaires sent, 210 valid responses were received for analysis.

For the current research, data were collected using a structured self-administered questionnaire. Following Brislin (1970), the questionnaire was first translated from English into Kazakh and Russian languages. An independent back-translation was then conducted to ensure conceptual equivalence and cultural appropriateness for the population of Shymkent, Kazakhstan. Some items were adapted to reflect the local cultural context, as direct translations could not fully capture specific meanings. For example, minor adjustments were made to the direct use of the word “dacha” when describing gardening activities. This term has exceptional cultural significance in Kazakhstan, and the words “garden” or “plot” do not fully capture it. The final questionnaire consisted of five sections: demographic information, gardening status and profile, leisure orientation, activity attachment, and outcome scales.

The study used a sequential screening and measurement procedure to accurately classify participants and ensure that the measurement of all psychological constructs was appropriate (see Figure 2).

Step 1: Demographic Information. All participants provided standard demographic information (gender, age, education, profession, and living status).

Step 2: Primary Gardening Status Screening. All potential participants first answered a screening question to determine their eligibility for the study’s core focus and to establish the primary comparison groups: “In the past 12 months, have you engaged in any gardening activities? This includes caring for indoor plants, balcony or container gardens, private yard gardens, community garden plots, or dacha plots in suburban areas.” Respondents answering “Yes” were classified as gardeners and proceeded to Step 3A. Those answering “No” were classified as non-gardeners and proceeded to Step 3B.

Step 3A: Gardening Profile. Gardeners were asked to specify their engagement frequency: “On average, how often have you engaged in gardening in the past 3 months?” Response scale: 1 = Once per week, 2 = Twice per week, 3 = Three times per week or more.

Step 3B: Non-gardeners completed a two-part question: (a) selecting the main reason they do not garden from a list of barriers (e.g., lack of space, time, and skills), and (b) naming their primary alternative leisure activity in an open-ended text field (e.g., reading, watching TV).

Step 4: Measurement of Psychological Constructs. All participants then completed the core psychological scales, with instructions tailored to their group to provide concrete references.

For Gardeners, the items in the Leisure Orientation and Activity Attachment scales referred specifically to their gardening activity.

For Non-Gardeners, the same scales referred to the specific alternative activity they named in Step 3 B.

The following scales were administered:

Leisure Orientation: A 4-item scale adapted from Fox (2017). Participants rated their agreement with statements such as “I approach this activity primarily as a form of leisure and relaxation,” on a 7-point Likert scale (1 = Strongly Disagree to 7 = Strongly Agree).

Activity Attachment: A 5-item scale adapted from Shen et al. (2022). Participants rated statements such as “This activity brings me good memories” on the same 7-point Likert scale.

Mental wellbeing: Measured using seven items adapted from the WHO-5 wellbeing index (de Wit et al., 2007) on a 7-point Likert scale.

Perceived Stress: measured using an adapted 10-item perceived stress scale (PSS) from Cohen et al. (1983) on a 7-point frequency scale (1 = never to 7 = very often). Negatively worded items were reverse-coded, such that higher final scores indicated lower perceived stress. All the scale items are listed in Table 1.

Five academic experts were invited to review the initial version of the questionnaire. Their comments helped us refine the question wording and scale structure and adjust for cultural specifics in Kazakhstan. Next, a pilot test was conducted with a small group (n = 20) that was representative of the target population. After official data collection (n = 210) was completed, the internal consistency of all multi-item scales was assessed using Cronbach’s alpha, and all scales were above 0.75. The results confirmed the reliability of the measurements in the study sample.

The collected data were analyzed using a two-stage analytical approach with IBM SPSS Statistics Version and IBM SPSS AMOS.

As shown in Table 2, the demographic profile of respondents (n = 210) shows a predominance of women (60.5%) and middle-aged and older individuals, with 61.4% of the participants aged 41 years and older. Educational attainment was distributed almost evenly, with a slight majority (55.7%) holding less than a bachelor’s degree. The occupational profile indicates that government employees and homemakers are the major participants (29.0% each). Living status revealed a strong family orientation, since 49.0% of the respondents lived with their families.

Table 3 presents the characteristics of the gardener and non-gardener participants in this study. Nearly two-thirds of the participants (64.3%, n = 135) reported engaging in gardening within the past 12 months. In particular, the most common frequency was gardening three times a week (30.5%). Among those who did not garden (35.7%, n = 75), the primary barriers were lack of space (24%), lack of time (20%), and lack of knowledge/skills (16%). Sixteen percent of non-gardeners did not specify reasons, and some reported motivational barriers, such as not wanting to take responsibility for the process (10.7%), a lack of interest (6.7%), or high costs (6.7%). When examining alternative leisure activities among non-gardeners, nearly half (49.3%) primarily engaged in passive activities such as watching television and movies (29.3%) or using social media and the Internet (13.3%). Active leisure activities, such as sports/physical education (16.0%) and reading (20.0%), accounted for approximately one-third of the alternative options. The distribution indicates that the leisure profiles of non-gardeners differed significantly from those of gardeners, for whom gardening was invariably the primary leisure activity.

Principal component analysis with Promax rotation was conducted to examine the factor structure of all the scale items. The Kaiser-Meyer-Olkin measure (0.870) and a significant Bartlett’s Test of Sphericity (p < 0.001) confirmed the excellent suitability of the data for factor analysis. The analysis revealed a factor that accounted for 66.68% of the total variance. Moreover, the pattern matrix revealed that all items loaded above 0.70 on their intended factors, without significant cross-loadings. However, we removed one item of mental wellbeing (MWB5) and one item of activity attachment (AA5), as well as four items of perceived stress (PS4, PS5, PS7, PS8) due to a loading factor below 0.70 in the factor analysis. High communalities across the remaining items indicated that the retained items were well explained by the extracted factors. In addition, it provides confirmation of scale validity and reliability.

The author applied Harman’s one-factor test to assess the common method bias (CMB). This test revealed eight factors that cumulatively accounted for 36.12% of the variance, which was below the 50% threshold. This confirmed that the CMB was not an issue in our data. Furthermore, it was supported by a simple marker variable method, which also confirmed the minimal standard method variance (Simmering et al., 2015). No significant demographic differences were observed (p > 0.05).

All scales demonstrated excellent reliability (Cronbach’s α ranging from 0.843 to 0.948), well exceeding the 0.70 threshold. Respondents reported a high average level across all constructs (means 5.45–5.69 on 7-point scales), with a significant negative skew (skewness −1.09 to −1.56). This indicates that the distributions are concentrated toward the positive ends of the scales (see Table 4).

The composite reliability (CR) values ranged from 0.844 to 0.940, and all average variance extracted (AVE) values exceeded 0.50, supporting convergent validity. Items within each scale measure the same underlying construct (Nunnally and Bernstein, 1994). Discriminant validity was confirmed using the Fornell–Larker standard. Relatively, the square root of each construct’s AVE (diagonal values) was greater than its correlations with all other constructs (off-diagonal values), and the MSV was lower than the AVE (see Table 5).

The measurement model also demonstrated a good fit: χ² = 185.187; degree of freedom = 129, TLI = 0.972; CFI = 0.977; RMSEA = 0.046. These results collectively demonstrate that the measurement model possesses adequate reliability, convergent validity and discriminant validity. The final structural equation model demonstrated an excellent fit with the data: χ²(146) = 222.14, p < 0.001, CFI = 0.969, RMSEA = 0.050 (90% CI: 0.036–0.063), GFI = 0.901, and NFI = 0.915. All fit indices exceeded the commonly recommended thresholds, with CFI and TLI values of 0.90 or above and RMSEA values below 0.08, indicating strong support for the theoretical model (Hair et al., 2019; Hu and Bentler, 1999) (see Table 6).

Structural equation modeling (SEM) analysis revealed a comprehensive network of significant relationships supporting the proposed model. All proposed paths were statistically significant (p < 0.05) (see Table 7).

Direct Effects of Activity Attachment: Both gardening frequency (β = 0.225, p = 0.004) and leisure orientation (β = 0.209, p = 0.009) were significant positive predictors of activity attachment. The analysis supports H1a and H1b and indicates that more frequent engagement and a leisure-oriented mindset each contribute independently to forming a stronger emotional bond with the activity.

Direct Outcomes of Activity Attachment: Activity attachment was a significant mediator and positively predicted both mental wellbeing (β = 0.256, p < 0.001) and lower perceived stress (β = 0.241, p < 0.001). Hence, H2a and H2b support the suggestion that a stronger bond with gardening is directly associated with higher wellbeing and lower stress levels.

Direct effects of past circumstances on mental health outcomes: The frequency of gardening had a significant positive direct effect on mental wellbeing (β = 0.226, p = 0.002) and perceived stress (β = 0.378, p < 0.001), independent of activity attachment. Hence, H3a and H3b are supported. Similarly, leisure orientation had a significant positive direct effect on mental wellbeing (β = 0.193, p = 0.011) and perceived stress (β = 0.217, p < 0.001), independent of activity attachment. Therefore, the results support H4a and H4b.

Mediating Role of Activity Attachment: The analysis confirmed activity attachment as a significant partial mediator for all paths (see Table 8). Significant indirect effects were found for gardening frequency on wellbeing (β = 0.057, 20.1% of total effect) and stress (β = 0.054, 12.5%), supporting H5a. In addition, for leisure orientation on wellbeing (β = 0.053, 21.5%) and stress (β = 0.050, 18.7%), supporting H5b.

To assess the impact of gardening frequency on psychological outcomes, a series of one-way ANOVAs was performed. It compared four groups that involved non-gardeners and those who gardened once, twice, or three times per week. The analysis found that the most significant factor determining psychological benefits was whether a person gardened at all (See Table 8).

As shown in Table 8, statistically significant differences were observed between the gardening frequency groups for all three outcome variables. Post-hoc Tukey HSD tests revealed three key outcomes, which are shown below. A large and significant effect was found for perceived stress (η² = 0.33). The non-gardener group reported significantly higher stress levels than all the gardening groups (all p < 0.001). In particular, there were no significant differences in stress among the groups that gardened once, twice, or thrice a week. These results indicate that any regular gardening engagement, regardless of frequency, offers substantial stress-reduction benefits. This accounted for approximately 33% of the variance in perceived stress. A different correlation was found for mental wellbeing (η² = 0.13). The Non-Gardener group had significantly lower wellbeing than the twice-a-week (p < 0.001) and thrice-a-week (p < 0.001) groups. However, it was not substantially different from the once-weekly group (p = 0.263). The results suggest a frequency effect, which suggests that to improve mental wellbeing significantly, gardening should be done twice a week or more often compared to people who do not garden. For activity attachment (η² = 0.10), there was a tendency towards lower stress. The non-gardener group reported significantly lower attachment than all the gardening groups (all p < 0.05), with no differences among the gardening frequency groups. The results suggest that regardless of how often one gardens, a psychological bond with gardening develops through engagement.

The current research presents three interrelated theoretical contributions, all of which argue that activity attachment is a central concept in nature-based wellbeing studies. First, we extended attachment theory by introducing activity attachment as a novel construct in the gardening study. Previous work has focused on place attachment (Barrable, 2025; Scannell and Gifford, 2010); however, we validated that psychological bonds with an activity itself serve as a potential mediator for mental health benefits. It establishes activities as legitimate attachment objects for the elderly. Second, this study integrates leisure studies with environmental psychology. We found that intrinsic motivation from recreation research (Fox, 2017; Murtagh and Frost, 2023), which is measured as leisure orientation, significantly predicted activity attachment and wellbeing outcomes in developing countries. Third, the results found a dual-pathway model. SEM shows that frequent gardening for leisure purposes has a substantial direct effect on gaining immediate benefits and (b) an indirect effect through attachment, which gains additional benefits.

A key public health strategy is to reimagine gardening in public communication. This suggests that it can be transformed from a chore into a desired form of relaxation and rejuvenation. This reimagining aims to overcome the key barriers to participation related to time and motivation. Support the national dacha roadmap and infrastructure. It recognizes that it functions not simply as an agricultural program but as an essential public health measure. Its success in expanding access directly influences the improvement of the population’s mental health. Policymakers could develop a new green gardening sector, such as U-pick farms, as they are comparably low-cost and leisure-oriented activities. It has scientifically proven stress-reducing benefits.

The present study has limitations that guide future research. A sample from a single city, such as Shymkent, can limit the ability to generalize the findings. Future studies should employ longitudinal designs and conduct studies in other cities across diverse settings. Our study did not test controlled confounding variables, which may have influenced the relationship between activity engagement and wellbeing. Future research should consider including relevant demographic, environmental, and psychological covariates to account for these effects. Moreover, some survey items were removed during the factor analysis because of low factor loadings. Although the instruments were adapted for local relevance, these modifications may limit comparability with prior studies and affect the measurement of constructs. A key methodological opportunity is a multi-group SEM analysis comparing our model between gardeners and non-gardeners. Testing for measurement and structural invariance will allow us to determine whether the strength of psychological relationships (from activity attachment to wellbeing) is similar across activities. This will allow us to determine whether these benefits are a universal function of a favorite leisure activity or whether they are specifically enhanced by gardening. Third, our findings show that almost half of the non-gardeners engaged in passive leisure activities. It stresses that future research should test gardeners and non-gardeners as two distinct categories. They can compare the wellbeing outcomes and underlying mechanisms of gardening with other activity categories, such as passive screen time, social leisure, and sports. Such a comparative approach will allow us to explore whether the benefits of gardening are related to a restorative connection with nature, an active and creative process, or a combination of both.