The objective of this research is to identify factors that influence both the decision (yes or no) and level of flood insurance among coastal homeowners in the southeast U.S. Recently flood damage has dramatically increased (Flood), and Crossett et al. (2004) report that coastal populations are growing. And in spite of rising costs of living in coastal areas, people are willing to pay more for access to ocean views and other natural amenities associated with coastal living (Bin and Kruse, 2006). Although the federal government provides flood insurance programs and encourages at-risk residents to insure their property from flood, rates of uptake remain low (Burby, 2001; Kunreuther, 2006; Landry and Jahan-Parvar, 2009). The National Flood Insurance Program (NFIP) was created to provide often subsidized premiums to cover losses which private insurance markets failed to offer. However, as Kunreuther et al.(1978) argue, many people do not bother to prepare, and have a low willingness to pay for coverage, even if subsidized (Kunreuther 1996). However, of those who have previously experienced flooding, they tend to insure their properties more (McClelland, Schulze, and Coursey 1993). Based on previous literature, we identified key factors to establish testable hypotheses regarding effect on flood insurance demand. These include: income, previous flood experience, the presence of a mortgage, home location (both flood zone status and distance from the shore), participation in CRS, the distance from the coast, the house construction year as well as measures of respondent risk preferences and perceptions. Data on flood coverage level and the above explanatory variables were obtained via revealed-preference online survey method, contracted through Knowledge Networks (KN) during August-September 2010. We chose to contract with KN for several reasons. First, they are, to our knowledge, the only survey firm that can legitimately say they have a true probability based sample for an online survey because they recruit by phone and/or mail (randomly selected using random-digit dialing (RDD) or by using address-based sampling); additionally they provide internet access to households that do not have it. KN was also contracted to overcome the typical of low response rate when surveying the general public. KN uses an online panel (called the “Knowledge Panel”). KN Panel members that were homeowners were sampled from 95 counties in Gulf Coast and Florida Atlantic Coast counties in AL, FL, LA, MS, and TX, with an 47% response rate (720 observations), with 67% from FL, 24% from TX, 5% from LA, and 4% collectively from AL and MS. As expected, insurance purchase is positively affected by the individual’s risk perception, their risk preference, whether or not they have a mortgage, flood zone residence, their income, CRS, previous flood experience, and the year of construction of house. Coefficients of mortgage and risk perception, income, flood zone are significant at 0.05 the level. Additionally, the coefficient of distance from the coast is only significant at the 0.1 level.