(Almost) all Quiet Over One and a Half Years: A Longitudinal Study on Causality Between Key Determinants of Private Flood Mitigation.

Previous studies do not agree on the strengths and directions of the effects between risk appraisal, nonprotective, and protective responses in private flood mitigation. This may be due to the widespread use of cross-sectional survey designs, which infer causality from theoretical considerations alone. The present longitudinal study, in contrast, builds on the logic that cause precedes effect to confirm the direction of effects. Drawing on two-wave survey data from 554 flood-prone households in Austria, cross-lagged autoregressive models analyze pairwise combinations between risk perception, fear, five nonprotective responses (fatalism, denial, wishful thinking, reliance on social support, reliance on public protection), and seven specific protective responses (ranging from coordination with neighbors to structural modifications of the building). These factors show substantial temporal stability, in particular for nonprotective responses and fear. Only in very few instances can effects over time be confirmed statistically. Nonprotective responses emerge as the major drivers; foremost, denial, and reliance on public protection limit private flood mitigation. This overall null finding on causality may trace back to the 1.5 years' time span and the absence of any policy intervention or flood disaster between survey waves, and the high stability of protection motivation theory components. This finding puts into question the theoretically assumed causal relationships and the effects found in cross-sectional studies. The high trait-like stability requires perseverance in risk management efforts to change attitudes and capabilities. Finding nonprotective responses as key determinants in an overall picture of stability suggests that this factor merits a stronger role in future risk research.