Structural equation modeling reveals decoupling of ecological and self-perceived outcomes in a garden box social-ecological system

Abstract

It is well known that green urban commons enhance mental and physical well-being and improve local biodiversity. We aim to investigate how these outcomes are related in an urban system and which variables are associated with better outcomes. We model the outcomes of an urban common-box gardening-by applying the Social-Ecological Systems (SES) framework. We expand the SES framework by analyzing it from the perspective of social evolution theory. The system was studied empirically through field inventories and questionnaires and modeled quantitatively by Structural Equation Modeling (SEM). This method offers powerful statistical models of complex social-ecological systems. Our results show that objectively evaluated ecological outcomes and self-perceived outcomes are decoupled: gardening groups that successfully govern the natural resource ecologically do not necessarily report many social, ecological, or individual benefits, and vice versa. Social capital, box location, gardener concerns, and starting year influenced the changes in the outcomes. In addition, the positive association of frequent interactions with higher self-perceived outcomes, and lack of such association with relatedness of group members suggests that reciprocity rather than kin selection explains cooperation. Our findings exemplify the importance of understanding natural resource systems at a very low "grassroot" level.

Figure 1

A simplified schematic diagram of the social-ecological system (SES) framework applied for urban gardening system in this study. Sub-systems Actors, Governance, Resource and Evolutionary theory may influence the Outcomes, which itself is composed of Self-perceived and Ecological outcomes. The Self-perceived and Ecological outcomes may or may not be associated with each other in this system. To facilitate dialogue between evolutionary biology and SES literature and combine variables from somewhere between actors and governance, we treat Evolutionary theory as a subsystem.

Figure 2

A research flowchart of the study.

Figure 3

The model for ecological and self-perceived outcome –latent variables and their reflective indicators in an urban gardening system. The covariance between the latent variables, factor loadings for each reflective indicator (all significant) and error terms for the reflective indicators are presented (unstandardized estimates / standardized estimates bolded) (see Table S5 online for p-values and standard errors for each estimate). The error terms for the reflective indicators are shown after the arrows pointing to the indicators and they signify the variance not explained by the respective latent variable they are loading onto.

Figure 4

The SEM results for the variables belonging to the sub-systems Actors, Governance, Resource and Evolutionary theory, predicting the two latent variables Self-perceived and Ecological outcomes described by their reflective indicators. The regression coefficients presented by arrows pointing to the latent variables show the significant (p value < 0.05) associations between the predictors and the outcomes (note that the non-significant predictors were not omitted from the model, see Table 2). The estimates presented are unstandardized estimates and their standard errors are presented in brackets. Error terms signifying the variance not explained by the model are included for latent variables and the reflective indicators.

Figure 5

Map of Turku city area presenting the locations of the urban box gardening groups. The boxes were distributed around Turku quite evenly (on the left all the locations), however concentrating on the city center (augmented map on the right). The black circles represent groups, who answered the questionnaire and the green triangles the groups, who didn`t answer the questionnaire.