Environmental quality or behavioural cognitions? Identifying variables that can be used to classify intended civic engagement against transportation noise exposure Natalie Riedel, Emily Mena, Heike Köckler, Birgit Reineke, Annette Peters, Lars Schwettmann, Kathrin Wolf, Gabriele Bolte, and Ute Kraus Natalie Riedel, 1 Emily Mena, 2 Birgit Reineke, 3 Gabriele Bolte 4 University of Bremen, Institute of Public Health and Nursing Research, Department of Social Epidemiology Grazer Str. 4 D-28359 Bremen Germany Heike Köckler 5 Hochschule für Gesundheit, Department of Community Health, Professorship on Place and Health Gesundheitscampus 6 – 8 D-44801 Bochum Germany Annette Peters, 6 Kathrin Wolf, 7 Ute Kraus 8 Helmholtz Zentrum München – German Research Center for Environmental Health (GmbH) Institute of Epidemiology Ingolstädter Landstr. 1 D-85764 Neuherberg Germany Lars Schwettmann 9 Helmholtz Zentrum München – German Research Center for Environmental Health (GmbH) Institute of Health Economics and Health Care Management Ingolstädter Landstraße 1 D-85764 Neuherberg, Germany Martin Luther University Halle-Wittenberg

1 natalie.riedel@uni-muenster.de (email address of current affiliation) 2 e.mena@uni-bremen.de 3 birgit.reineke@uni-bremen.de 4 gabriele.bolte@uni-bremen.de 5 heike.koeckler@hs-gesundheit.de 6 peters@helmholtz-muenchen.de 7 kathrin.wolf@helmholtz-muenc h en . de 8 ute.kraus@helmholtz-muenchen . de 9 lars.schwettmann@helmholtz-mu e n c h e n . de

a Shea mar ce 21-24 AUGUST SCOTTISH BENT caso

Department of Economics D-06099 Halle (Saale) Germany

ABSTRACT The European Environmental Noise Directive opens the floor for the public to participate in the development and review of noise action plans. Regarding health inequities related to environmental noise exposure, we aimed to identify determinants that influence residents’ civic engagement against transportation noise exposure. We defined four variable groups representing either environmental quality with indicators of (1) noise exposure and (2) environmental resources or behavioural cognitions with indicators related to (3) civic engagement and (4) generalised emotions and cognitions. We used data collected from 3,743 participants in the population-based KORA study in the Augsburg region, Germany, and employed Conditional Inference Trees (CIT) to identify variable sets within these four groups that are most likely to classify civic engagement intentions. A final CIT model encompassed all variables. Participants’ social characteristics were included in each of our CIT models. Results suggest interactions between environmental quality (related to sound/atmosphere in the sleeping room) and income and between income and emotions (positive affectivity) with respect to engagement intentions. When considered simultaneously, engagement- specific cognitions appeared as more relevant for engagement intentions than environmental quality. This finding indicates that noise action planning should take behavioural cognitions into account when designing participation processes. 1. INTRODUCTION

The European Environmental Noise Directive (END; Directive 2002/49/EC) is the legal basis for an environmental policy dealing with ambient noise on the local level through noise action planning. While opening the floor for the public to participate in the development and review of noise action plans, noise action planning is not provided with legally binding standards for environmental quality. Thus, a socially inclusive and effective public participation is all the more important for local planning to reach a fair and sound weighing of various interests and stakes. Against the background of unequal exposures and health impacts, participation can deliver additional evidence relevant for sustainable and equitable local developments. Moreover, participation can confer citizens a sense of control over their environment if they feel recognised by the planning authorities and perceive fairness throughout the planning process. However, materially and socially deprived population groups often participate less effective in environmental decision processes. [1] For this reason, it is crucial to make the link between distributional and procedural dimensions of environmental justice explicit. This is particularly true for environmental noise, given its health burden [2] and its impact on health inequities. [3, 4] In this study, we therefore focus on determinants of intention for civic engagement against transportation noise exposure (called hereafter intended civic engagement) as a highly relevant predictor of active coping in participatory processes. [1, 5] Building on a conceptual model of cognitive-motivational determinants of civic engagement, [6] we defined four components representing specific lines of argumentation that address determinants of intended civic engagement. The first two lines of argumentation referred to environmental quality in residents’ environment and included (1) a component describing exposure to transportation noise as uncontrollable stress stimuli and (2) a component dealing with environmental resources like access to a quiet side at home that provide controllability through cognitive restoration. The second two lines of argumentation were about controllability perceptions resulting from behavioural cognitions, with (1) one component introducing engagement-specific cognitions and (2) one component elaborating on generalised emotions and cognitions. By discussing aspects of both environmental quality and behavioural cognitions, we stress the inter-relatedness of perceived environmental control and procedural control.

2. METHODS

2.1. Statistical analysis We made use of the Conditional Inference Tree (CIT, also referred to as C-tree [7]) to investigate intentions for civic engagement as an outcome of the four sets of variables as defined by the environmental quality and the behavioural cognition components. Our CIT procedure recursively partitions the data into mutually exclusive subgroups in order to identify variable sets that are most likely to classify civic engagement intentions. This means that a binary split is induced on a predictor variable (root node) in order to maximise the homogeneity between individuals concerning intended civic engagement on the one hand and to maximise heterogeneity between the two resulting child nodes on the other hand. We allowed the CIT to have up to four nodes from the root (the first splitting variable) down to the furthest child. The tree structure as represented by variable nodes shows interactions between the splitting variables that follow a hierarchical order. We fitted one CIT-model for each environmental quality component and each behavioural cognition component, all encompassing sociodemographic characteristics and related socio-economic circumstances, because these variables reflect social power relations. It follows from the hierarchical tree structure that the first binary split affects the selection and sequence of the next splitting variables. This is why we performed sensitivity analyses in models where a socio-demographic characteristic related to potential discrimination generated the first split. Discrimination can be related to sociodemographic characteristics like ethnic origin, gender, religion, disability, age or sexual orientation, as proposed by Article 3 of the Basic Law for the Federal Republic of Germany and the General Act on Equal Treatment. In this case, the CIT model was re-run without the respective socio- demographic characteristic in order to reveal what mechanisms might lie behind it to induce intentions for civic engagement. To account for the potential inter-relatedness between the components, a final CIT-model included the variables from all four components. We employed the partykit package [7] and R 3.6.1 (R Core Team 2013) in order to perform the CIT analyses.

2.2 Data Our analysis was based on 3,743 participants from the population-based KORA study in the Augsburg region, Germany (54 % women, aged 43 – 92 years). KORA is a research platform that was initiated to study associations between living and health conditions among the population with German citizenship in the City of Augsburg and two adjacent rural counties. [8] In 2019, a thematic questionnaire was sent to a large subset of KORA participants. The KORA study adhered to ethical standards (KORA-Fit EC No 17040), which includes participants’ written informed consent.

2.3 Outcome variable In our questionnaire, participants indicated their degree of agreement to the statement “I have plans to take action against traffic noise exposure” on a bipolar six-point Likert scale. These ‘actions against traffic noise exposure’ were specified by coping behaviour directed at planning authorities or political institutions (e.g. addressing complaints to the administration, participating in a survey, or joining a local initiative). To keep our analysis manageable, we dichotomized the outcome variable (agreement vs. disagreement).

2.4 Variables of the components of environmental quality We used both objective and subjective information to capture noise exposure and environmental resources. The noise component was operationalized by noise indicators according to the END as well as participants’ perception and appraisal of the noise stimulus as gathered from questionnaire

items, including residents’ expectation of the future development of road traffic noise in the residential environment and noise annoyance at home. Variables of the environmental resource component referred to the dwelling (like pleasantness of sound or atmosphere in the sleeping / living room as reported by participants in the questionnaire) as well as to the presence and quality of green and blue areas in the residential environment as measured by the Normalized Difference Vegetation Index (NDVI) and participants’ ratings on questionnaire items.

2.5 Variables of the components of behavioural cognition The questionnaire variables for the engagement-specific cognitions component were adapted from the Theory of Planned Behaviour [9] and the Conservation of Resource Theory [10] as suggested by Köckler. [1] They were complemented by variables on traffic-related cognitive distortions and perceived effectiveness of participants’ engagement. Variables of the generalized emotions and cognitions component were selected based on the Cognitive Activation Theory of Stress (CATS) [11] and its application in the noise context [6], including behavioural outcome expectancies, affectivity, and noise sensitivity.

2.6 Other variables Available sociodemographic variables relevant for potential discrimination were gender (binary male vs. female), age, background of migration, and formal disability status. Furthermore, socio-economic circumstances included, among others, household net equivalent income, and perceived financial situation. Self-rated health, hearing impairment, and frequency of living at the primary residential address were considered as confounders. 3. RESULTS

More than one in ten participants was positive about intended civic engagement.

3.1. Results for the CIT models on environmental quality components (noise exposure and environmental resources)

The CIT model on the noise component showed a potential interaction between participants’ expectations of future road traffic noise levels in their residential environment (first splitting variable, root node) and traffic-related vibrations (child node) as well as noise annoyance (child node). The CIT model on the environmental resources component suggested an interaction between gender (root node) and the environmental quality related to the sleeping room (child node). The sensitivity CIT model without the gender variable produced an interaction between the sleeping room variable (new root node), household net equivalent income (child node) and perceived financial situation (child node).

3.2. Results for the CIT models on behavioural cognitions components (engagement-specific cognitions component and generalized emotions and cognitions component) The CIT model on the engagement-specific cognitions component estimated several interactions between variables within the frame of the Theory of Planned Behaviour and perceived effectiveness of civic engagement already performed. The CIT model on the generalized emotions and cognitions component showed gender as first and only splitting variable. In the sensitivity model without the gender variable, we observed an interaction between household net equivalent income (new root node) and positive affectivity (child node) and perceived financial situation (child node).

3.3 Results for the final model including variables from all components When the variables from all components were considered simultaneously, the CIT procedure produced the same results as in the model for the engagement-specific cognitions component.

4. CONCLUSIONS

Public participation can help local planning arrive at more equitable and sustainable decision-making in the context of noise action planning and beyond (i.e. local planning). Our findings point to the need to recognize residents’ characterizations of the noise stimulus, dwelling-related environmental resources, engagement-specific cognitions and generalized emotions over and above END noise indicators, while being aware of discrimination potentially related to socio-demographic characteristics. Engagement-specific cognitions appeared as more relevant for showing intentions for civic engagement than environmental quality in the final CIT model across components. We may assume that health inequities can originate not only from unequal exposure distributions, but also from cognitive-behavioural mechanisms and their interplay with sociodemographic characteristics (e.g. gender) and related socio-economic circumstances (e.g. income, measured both objectively and subjectively). Though highly exploratory, our approach emphasizes the necessity to integrate perceptions of environmental and procedural control. Noise action planning should take behavioural cognitions into account when designing participation processes. 5. ACKNOWLEDGEMENTS

The KORA study was initiated and financed by the Helmholtz Zentrum München – German Research Center for Environmental Health, which is funded by the German Federal Ministry of Education and Research (BMBF) and by the State of Bavaria. The questionnaire and noise data from the Bavarian Environmental Agency presented in this study was financed within the research project ‘Exploring cognitive-motivational determinants of health (inequities) in the context of the European Environmental Noise Directive’ by Natalie Riedel. This project was funded by the German Research Foundation (Deutsche Forschungsgemeinschaft DFG RI 2781/1-1, project number 387821120). Selected variables were financed by the BMBF project INGER Integrating gender into environmental health research (01GL1713A). 6. REFERENCES

1. Köckler, H. Umweltbezogene Gerechtigkeit. Anforderungen an eine zukunftsweisende Stadtplanung . Peter Lang GmbH (2017). 2. World Health Organisation (WHO) Regional Office for Europe, Joint Research Centre. Burden of disease from environmental noise: Quantification of healthy life years lost in Europe (2011). 3. Dreger, S., Schule, S. A., Hilz, L. K. & Bolte, G. Social Inequalities in Environmental Noise Exposure: A Review of Evidence in the WHO European Region. International Journal of Environmental Research and Public Health , 16 (6), 1011 (2019). 4. Dreger, S. & Bolte, G. Inequalities in self-reported noise annoyance . In World Health Organisation (WHO) Regional Office for Europe (Ed.) Environmental health inequalities in Europe: second assessment report. 66-71 (2019). 5. Riedel, N., Scheiner, J., Jöckel, K-H., Schüz. B. & Bolte, G. Is older residents' exposure to road traffic noise associated with civic engagement for noise protection? A cross-sectional path analysis. Journal of Transport and Health , 20 , 101007 (2021). 6. Riedel, N . Kamp, van, I., Köckler, H., Scheiner, J., Loerbroks, A., Claßen, T., Bolte, G. Cognitive-Motivational Determinants of Residents' Civic Engagement and Health (Inequities) in the Context of Noise Action Planning: A Conceptual Model. Environmental Research and Public Health, 14 , 578 (2017). 7. Hothorn, T., Hornik, K. & Zeileis, A. Unbiased Recursive Partitioning: A Conditional Inference Framework. Journal of Computational and Graphical Statistics , 15 , 651-674 (2006). 8. Holle, R., Happich, M., Löwel, H. & Wichmann, H. E. KORA--a research platform for population based health research. Gesundheitswesen 67 Suppl 1 , S19-25 (2005). 9. Ajzen, I. The Theory of Planned Behaviour. Organizational Behavior and Human Decision Processes , 50 , 179–211 (1991).

10. Hobfoll, S. E. Conservation of Resources. A New Attempt at Conceptualizing Stress. American Psychologist 44 , 513–524 (1989). 11. Ursin, H. & Eriksen, H. R. The cognitive activation theory of stress. Psychoneuroendocrinology 29 , 567–592, (2004).