Beyond social experiments: Simulation of descriptive littering norms through agent-based modelling
Year: 2019
Type: dissertation
Abstract: Agent-based modelling presented itself as a suitable method for capturing littering behaviour because of its capacity for simulating autonomous agent behaviour and its spatial explicitness. Complex models, which have the potential for generating predictive output, require extensive validation and availability of detailed input data. Because neither is available for this research topic, the model was kept relatively abstract and of limited validity. A remaining merit of agent-based modelling is that it enables the exploration of theoretical assumptions and that it could facilitate the exposal of gaps in theory. During the modelling process and the accompanying literature study, vagueness in both qualitative and quantitative assumptions was found indeed.\nThe most influential and well-rounded explanation of littering behaviour, as described by the focus theory of normative conduct, was operationalised in this model. The theory details how the state of the local environment affects behaviour, as already observed in the well-known broken windows theory. However, it was found in the newer theory that increased littering occurs in dirty environments and decreased littering in clean environments only when the descriptive norm is activated in an individual, which occurs when they observe somebody else disposing of litter. The extent of the influence of descriptive norm has been detailed in multiple studies and was approximated in the output of the model developed in this project. An important finding of this research is that the configuration of personal norms as stochastic variables and activated norms as corresponding multipliers is a valid interpretation of the theory through which it seems empirical data can be reproduced.\nThroughout the modelling process and particularly in the final experiments, the importance of the activated descriptive norm was further emphasised. It was found to be more influential to overall littering rates than the traversing crowd’s age composition or the accessibility of bins. Moreover, because people also activate each other’s descriptive norm by disposing of their litter properly (in a bin), increased visibility of bins (due to placement or street layout) can counter-intuitively even lead to more littering in the long run. Although this finding is unforeseen, it is not inconsistent with the theory and is encouraged to be studied in practice.\nSeveral uncertainties also arose from the modelling process. The values and behaviours of numerous parameters of varying importance are not described in theory and were therefore given estimated definitions. Most were tested in sensitivity analysis or experiments and seemed to be represented reasonably, but the sheer number of uncertain variables suggests further research into specific aspects of littering behaviour is in order. Under-researched variables include:\n-\tthe distribution of personal norms,\n-\tthe conditions in which inter-personal norm activation occurs,\n-\tthe definitions of clean and littered areas,\n-\tthe distance people are willing to walk to bins and its correlation to personal norms.\nThe main objective of this research was to gain insight into littering behaviour through agent-based modelling by simulating empirical findings and theoretical assumptions in settings that are otherwise difficult to examine in practice. The agent-based model was successfully developed: its output is realistic, in line with theory and partially validated. By extension, this also means that the theory on which it was based has a logical foundation; even when its assumptions were simulated in unstudied scenarios, the results seemed reasonable. Moreover, several uncertainties with regard to underexposed assumptions were found in the modelling process. Agent-based modelling has therefore proved valuable in producing insights for littering research. A modest contribution to solving the societal and environmental problem of littering has thusly been attempted to be made.
Author: J.B. Bakker
Language: English
Subfield: Management, Monitoring, Policy and Law
Field: Environmental Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W3042306797