Insurance status in 1993 is the dependent variable, and the independent or predictor variables include previous experience with an earthquake, perceived risk, economic status, demographic status, ethnicity, length of residence in California, and distance from an active fault trace. The resulting equation is statistically significant, and the variables most closely related to insurance status are earthquake experience, perceived risk, and home value. The longitudinal survey population is similar to the 1993 full cross-sectional sample. The percentage of longitudinal respondents with earthquake insurance increased steadily throughout the 1989-1993 study period. One unique aspect of this series of surveys was the opportunity to monitor changes in both insurance purchase and also attitudes toward earthquake hazards in the same individuals during the four-year study period. The factor most closely associated with the adoption of earthquake insurance is perceived vulnerability to earthquake damage-the belief that one's own home will be damaged by a major earthquake within the near future.