Citizen science (CS), the involvement of citizens in scientific research and/or knowledge production, is being widely discussed in the scientific community and in policy as an instrument to address major global problems including insect population decline [1
]. Insect pollinators have consequently become a widespread research topic for CS projects. CS engagement may include volunteers observing pollinators [5
], collecting presence/absence data on nests [8
], and projects investigating pollination [9
]. Aside from a valuable source of data for researchers [11
], CS is also proposed as an approach to increase interest in and appreciation of insect pollinators [12
Changes in peoples’ (pro-)environmental behaviour (PEB) are also among the anticipated outcomes of CS participation [14
]. PEB is gaining in importance, as people are increasingly important actors in pollinator conservation [16
]. This is especially the case in the management of home or public gardens. Here, CS is promoted to support a change in the role of gardeners as stewards through pro-pollinator (or ‘pollinator-friendly’) habitat management [19
]. In addition, participation in a pollinator CS project can be considered as PEB [21
]. Despite this important link between CS participation and PEB, there are few insights about the factors that motivate gardeners’ pro-pollinator engagement.
Research on the motivations for participation and retention of participants in pollinator CS suggests that, while project dependent, educational motivations are the most important driver for volunteer participation [22
]. Research on volunteering in an insect CS project showed the surveyed participants were driven by both altruistic (pro-social) and egoistic (self-serving) motivations, but rated the pro-social functions as more important for their engagement [24
]. The methods also seem to be important, with complex projects having relatively low rates of retention and completion [9
]. In a recent qualitative study on self-reported reasons for participating in pro-environmental CS activities, Tsybulsky [21
] found that the influence of internal reasons (desire of social interaction, contribution, education, leisure as well as personal satisfaction and pleasure) were more important than external or demographic reasons to explain participation. While these assumptions are mostly based on self-reported motivations, we lack information on the underlying drivers of people’s behaviours.
Theoretical concepts such as Cognitive Hierarchy [25
] propose that underlying drivers of people’s behaviours are attitudes and behavioural intentions. Following Cognitive Hierarchy, attitudes are defined as “an association, in memory, of an evaluation of an object” [26
]. Behavioural intentions are close antecedents of behaviour [27
] and often used as proxy of actual behaviour [28
]. However, behavioural intentions do not necessarily predict actual behaviour [29
], and other factors may influence PEB such as emotions [30
], nature-relatedness [31
] and identity [32
Emotions influence mental processing including perceptions [33
] and are often referred to as the root cause of human response to wildlife [35
]. Furthermore, emotions related to behavioural intention are indicated to predict behaviour [36
]. Environmental identity (to what extent one sees oneself as someone whose actions are environmentally friendly) plays an important role in PEB [37
]. Van der Werff et al. [37
] found that it is less likely that PEB is driven by intrinsic motivation (i.e., behaviour is satisfactory and enjoyable by itself [38
]), but by a strong environmental identity, which makes people feel morally obliged to perform PEB. In addition, PEB is linked to human–nature connection [39
], which can be described by peoples’ nature-relatedness, i.e., how people perceive their relationship with nature [31
In relation to pollinators, Knapp and colleagues [18
] showed peoples’ nature-connectedness, more specifically the diversity of interactions with nature, and their sense of being able to help pollinators are important predictors of people’s pro-pollinator actions. To a lesser extent, socio-psychological factors may be predictive of behaviour, including: attitude towards pollinators; knowledge of pollinators and conservation actions; perceived social pressure to help pollinators; and the environmentally friendly self-identity. To our knowledge, however, no previous work investigates multiple socio-psychological factors in relation to behaviour towards pollinators, including participation in pro-pollinator CS. Our work relates socio-psychological factors hypothesized to explain PEB-intentions and -behaviours towards pollinators (Figure 1
), in our study, of urban gardeners.
We surveyed participants and non-participants of a wild pollinator CS project in community gardens in Berlin and Munich, Germany to investigate how identity, nature-relatedness, emotions and attitudes may predict participation in the CS project as a proxy for pro-pollinator behaviour and their intentions to get involved in pro-environmental behaviour. With this, we aim to better inform public engagement in pollinator conservation.
2. Materials and Methods
We surveyed urban gardeners in 18 community gardens in Berlin, Germany and 15 community gardens in Munich, Germany. These gardens were part of an ongoing CS research project around wild pollinators and pollination function. In the project gardeners were asked to observe the pollination function over the course of the growing season of selected tomato, pumpkin and/or pepper plants by documenting when their plants (1) flowered, (2) were pollinated (closed flowers), and (3) had fruits. Gardeners measured the size of the fruit upon harvest. At the same time, researchers documented the management of the gardens (plant diversity, ground cover) and the wild pollinators in the gardens three times over the summer. Scientists and gardeners used different methods to study together the effect of garden features on pollinator diversity and pollination efficiency in the community gardens (Figure 2
). The project initially started in Berlin in 2020 and was expanded to Munich in 2021 using the same CS methods to gather more comprehensive results. In 2020, 44 participants submitted data at the end of the gardening season. Exchange and contact between the gardeners and scientists took place mainly online due to the Corona pandemic in three organized Q&A sessions and during the field research on plant and pollinator diversity.
The survey questionnaire was distributed in online and in paper format after multiple project introduction and training meetings (April/May 2020 in Berlin, April/May 2021 in Munich) with the gardeners of the community gardens. In addition, non-participating gardeners in particular were surveyed in the course of the respective first project season until November 2020 in Berlin and until July 2021 in Munich. The survey instrument was delivered in German and took between 15 and 20 min to complete. Participation was anonymous, voluntary and questions could be skipped.
We measured the following variables in the survey instrument: nature-relatedness, emotions, attitudes, gardeners’ identity and behavioural intentions (Table 1
). We gave the surveyed a short introduction with an illustrated overview showing that honeybees, wild bees, bumblebees, hoverflies, butterflies, wasps, flies and beetles are pollinators. To measure discrete emotions towards pollinators, we selected six single items of discrete emotions that relate to the concept of basic emotions [40
] and are often studied in relation to wildlife: joy, fear, disgust, interest, fascination, anger and compassion [41
]. To assess attitudes towards pollinators, we developed three items based on the definition of Fazio et al. [43
] as in an ‘association, in memory, of an evaluation of an object’ attitudes to: (i) I like pollinating insects; (ii) I like pollinating insects in my garden; (iii) I think pollinating insects are interesting animals; and (iv) I think that pollinating insects are worth protecting. Specific identities are likely to be linked to behaviours related to identity [32
]. Therefore, we focused on gardener identity based on their reported main reasons for urban gardening [44
]: (i) gardening for food production; (ii) nature conservation/engagement, (iii) and social capital. Here we asked gardeners to report their level of agreement on a 5-point Likert scale to the following question “I consider myself as a gardener who appreciates …” with three items: (1) large harvest, (2) nature/species conservation, (3) social interactions. To assess PEB specifically concerning pollinator conservation, we assessed four items, which were adapted from Jacobs and Harms [46
]. We measured the following behavioural intentions to get involved in pollinator conservation on a 5-point Likert scale: learning, implementing measures for pollinating insects, motivating others, and donating money. Following the same premise as Tsybulsky [21
] we considered participation in the wild pollinator CS project as PEB.
Respondents indicated whether they were participating in the CS project before answering the questionnaire. Depending on participation or non-participation, we asked an open question about the motivation for participation, respectively the reasons for non-participation. For socio-demographics, respondents indicated, among other things, their age, gender and the size of the place they grew up.
The developed survey instrument was tested in a pilot with N = 12 people from different socio-demographic backgrounds and adapted when necessary (e.g., items were not clearly understood).
The analysis was performed in the R Statistical Environment (R version 4.0.2). We calculated the internal consistencies of latent constructs (i.e., attitudes, nature-relatedness and behavioural intentions) using Cronbach’s α. Emotions were single items and consequently, no internal consistency needed to be calculated. The latent constructs attitudes (Cronbach’s α = 0.78), nature-relatedness (Cronbach‘s α = 0.82), and behavioural intentions (Cronbach’s α = 0.84) showed acceptable internal reliability and consequently, all items were used for further analyses. For each latent construct, we calculated average scores of the associated items as composite indices. The composite index of attitudes were the averaged four items measuring attitudes towards pollinators, of nature-relatedness the averaged six items of the NR-6 and of behavioural intentions the respective averaged four items about intentions to get involved in pollinator conservation. We used participation in the CS as a proxy for behaviour. Thus, we included a binary response variable indicating participation (1) and non-participation (0) (hereafter as behaviour).
We created four different sets of generalized linear models to analyse how the (latent construct of) gardeners’ pro-pollinator behaviour intentions and behaviour were explained by the gardener’s identity, emotions towards pollinators, attitudes towards pollinators and nature-relatedness. We separated cognitive concepts (e.g., attitudes) and emotions in our models given that the Cognitive Hierarchy focuses on cognitive processes and does not include emotions. Consequently, and given that the sample size did not allow us to include all variables in one model, we ran each model separately. First, with behaviour as binary response variable (binomial distribution) and behavioural intentions, identity, attitudes and nature-relatedness as explanatory variables. Second, again with behaviour as response variable and the six emotions as explanatory variables. Third and fourth, we used behavioural intentions (quasipoisson distribution to take into account the underdispersed count data) as response and again, identity, attitudes and nature-relatedness as well as the six emotions as explanatory variables, respectively.
Additionally, we analysed the motivation for participation and the reasons for not participating using qualitative content analysis [48
]. The coding categories were based on recent literature on participation in pollinator CS [21
]. Two researchers discussed and adopted the coding categories to achieve a mutual understanding and sufficient coding consistency.