The dynamics and influential factors of intentions and actual behaviours in flood adaptation

Floods pose significant risks to societies worldwide. Private flood adaptation is considered important to reduce flood risk. Investigating the influential factors on individual adaptation behaviour is thus essential. Many behavioural theories hypothesise a vital role of the adaptation intention toward adaptation behaviour. However, the literature shows a substantial gap between intention and behaviour, referred to as intention behaviour gap. This could be because most existing research is based on cross-sectional data, which does not reveal the changes in attitudes, intentions, and behaviour over time. For example, implemented measures might reduce the intention and behaviour, but these changes cannot be captured by only one survey time point. Our research thus deploys a two-wave panel survey with 401 respondents from Central Vietnam to (1) examine the dynamics of behaviour and intention over time, (2) examine the role of intention on actual behaviour and vice versa, (3) find influential predictors explaining intention and behaviour, and statistically compare the predictors.

Linear mixed models (LMMs) show that adaptive behaviour and intention of three groups of measures, namely, preparing devices, retrofitting houses, and adapting livelihoods, have significantly increased after half a year, except for the intention of preparing devices. The most influential factors in explaining behaviour and behavioural change are housing situations, personality traits, social norms, coping appraisals, and intention. For intention, socio-demographic characteristics, risk perceptions, social norms, and personalities are more important. It is noteworthy that the influential factors are highly measure-specific. Specific models show a clear difference in predictors between intention and behaviour. Bivariate LMM and statistical comparisons further confirm that only a handful of predictors could be used as interchangeable proxies between behaviour and intention. For example, out of 18 examined factors, only wishful thinking, knowledge, and moving permanently show similar influence on both the intention and behaviour of retrofitting houses. By contrast, house type, respondent’s age, building a new home, and house located in an urban area show significantly different influences; the remaining factors are uncertain to use as interchangeable proxies. These results suggest carefully reconsidering the use of research on intention to draw policy recommendations for behaviour in the flood risk domain.