| Abstract: |
Measuring impacts of extreme events on population well-being is complicated if
data are not representative of the pre-event at-risk population or a
representative sample of the population is not followed post-event. The
sources and nature of some important biases are documented using data from the
Study of the Tsunami Aftermath and Recovery (STAR) which documents the
evolution of population well-being before and after a major natural disaster,
the 2004 Indian Ocean tsunami. The baseline, collected 10 months before the
tsunami, is representative of the at-risk population. Respondents have been
followed and re-interviewed multiple times in the fifteen years since the
tsunami achieving high follow-up rates. We empirically document the importance
of construct samples that represent the pre-event at-risk population, rather
than an opportunistic sample of those in the vicinity of the event. Pre-event
characteristics condition where and in what circumstances people live
post-event. Post-event well-being is associated with post-event living
conditions in the short-term, and that over time, the link weakens between
short-term living arrangements and post-event well-being. Failure to follow-up
all respondents, especially those who move away from the location of the
event, yields biased estimates of impacts of the event. |