dc.description.abstracteng | Urbanization has been recognized as one of the major forces in shaping the 21st century on
planet earth. The by far largest increase in the share of humans living in cities is expected to
take place in the Global South. The consequences of this development for the environment are
not fully understood: In the aggregate, cities consume and pollute natural resources far
beyond their immediate environments, and they might do so more than income differentials
suggest. However, there are also examples of demand-pattern changes conducive to
environmental conservation.
Learning from such examples to mitigate overall environmental degradation appears decisive,
yet ambiguity of real-world urban resource depletion is matched by a lack of theoretical
models facilitating understanding of the underlying processes: Social-Ecological Systems
literature may be considered one of the most profound theoretical approaches to predict the
complexity of urban human-environment interactions. But the social subsystem within which
urbanization originates remains simplistic in most urban SES models to date. This may reflect
three problems with integrating known psychological mechanisms into such models: Southern
urban contexts are prone to alter mechanisms that often originate in Northern lab settings (a).
Causal quantification is moreover often incomplete in general (b). a and b could mutually
reinforce each other in Southern urban contexts, where interaction and feedback with other
such mechanisms will likely bias simple aggregation of extant literature (c).
This thesis hence hypothesizes that predictive abilities of urban SES models, and eventually
policy leverage, can be increased by integrating culturally and environmentally
contextualized, as well as less statistically biased psychological mechanisms into these
models. This is operationalized as revisiting a simple framework that so far predicts a
transition from ‗green-loop‘ to ‗red-loop‘ dynamics with urbanization more or less
deterministically; that is, decreasing salience of natural environments may not automatically
result in reduced pro-environmental attitudes (path i), that translate even less into stewardship
behavior (path ii). Chapter I introduces this overarching research question and analytical
framework, against both an empirical and a theoretical background.
Empirical evaluation proceeds with latent variable structural equation models of moderated
mediation, where feedback is controlled through spatial lags of exogenous variables utilized
as instruments. They are calibrated on a random sample of 1200 decisionmakers stratified for
urbanization along two rural-urban gradients in Bangalore / India. The water problems of this
rapidly growing Southern megacity exemplify global trends, much like its globalization-
ii
induced social change could point to more general developments. Measurement is achieved
by shortened psychological scales, combined to decomposed games and manifest indicators.
As a first operatinalization of the research question, chapter II hypothesizes that loss of salient
nature experience may be substituted by alternative, urban drivers of pro-environmental
attitude (path i above). That is, policymakers may not have to take preference-driven
transitions toward a red-loop as given, once culturally and environmentally contextualized
quantification of societal feedback allows for more realistic urban SES (problems a and c
above). This is achieved via jointly modeling psychological mechanisms mediating the effect
of urbanization onto environmental concern – beyond potential nature experience loss.
Separately documented, Northern urban features like anonymity suggest reinforcement of
such loss. But Southern urban dynamics like modernization-induced value change could
acquire a positive meaning neglected by extant, simplistic urban SES models; they could even
interact with and change the meaning of more traditional constructs.
Positive aggregate effects of urbanization on environmental concern replicate extant literature
here. However, they emerge only after instrumenting and due to unexpected, uniquely Indian
feedback between mediating urban features. These findings are interpreted as contributing not
only to more differentiated predictions of urban SES, but also to generalizability of known
aggregate effects within urban psychology.
The second essay (chapter III) questions the undifferentiated loss of all types of nature
experience for everyone under urbanization-related income growth (path i above). It
hypothesizes that urban SES models could benefit from integration of literature relating
positive emotional connections with nature to environmental concern. Extant models may be
drawing on decreasing negative degradation experiences only, as this effect already features
prominently in income-concern literature. But especially memories of positive experiences
could exhibit more differentiated loss patterns under income growth.
Chapter III therefore investigates persistent influence of remembered, positive experiences on
environmental concern across an urbanizing, Southern income distribution (problems a and c
above). Differential influence of this type is found to bridge the minimum of an otherwise U-
shaped income-concern relationship. I.e., particularly pronounced effects of nature experience
at intermediate income allow for monotonic increases of concern in income, because
beneficial psychological traits are present particularly among those urbanizing individuals that
abandon agriculture. Like the second essay, these results imply positive equilibrium dynamics
for Bangalore‘s SES, and can thus not explain the current empirically obvious red loop. Such
iii
differential bridging may however be exploited particularly by policy-makers at the rural-
urban interface, and it also explains previous sample-sensitivity of the income-concern link.
A third essay (chapter IV) finally tests for statistical confounding in cultural relativity of one
of the urbanization-induced moderators of the society-ecology link (problems a and b for path
ii above). Generalized social trust is known to increase in urbanization, and it is conducive to
acting upon preferences originating in environmental concern. This relationship is assumed to
be mediated by collectivism. Yet the link from collectivism to generalized trust still lacks
quantification beyond mere correlation. Only this would facilitate integration into urban SES
models besides extant, mostly negative moderators of path ii, and justify promoting certain
cultural traits for sustainability.
Previous results are found to be explained by a complex interplay of confounders.
Collectivism only affects generalized trust negatively under absence of voluntary cross-group
interaction, and given either reverse causality through bad experiences with generalized
morality, or emphasis of the hierarchical vertical dimension of collectivism over horizontal,
distance based sociability. As not only cross-group experiences but also collectivism increase
with Indian urbanization, these results once again imply dynamics toward positive social-
ecological equilibria in Bangalore.
The thesis concludes with a comparative discussion in chapter V. All three essays exemplify
benefits from inclusion of culturally and environmentally contextualized psychological
mechanisms into SES models of urbanization. More complex models can then point to
specific policy leverages against red-loop transitions, beyond reinforcing Southern roots of
nature experience. But additional research is still needed for truly realistic calibration of such
models. Future studies should explore further moderators of path ii, but also alternative
psychological constructs that truly predict pro-environmental behavior best in a given context.
Alternative representations of natural environments besides water could strengthen internal
validity further, whereas external validity would benefit from replication in less globalized
Southern cases. | de |