This paper presents a statistical analysis of structural fragility curves, considering both empirical and analytical approaches. Empirical fragility curves are developed using bridge damage data from the 1995 Kobe earthquake, while analytical curves are constructed through nonlinear dynamic analysis. The fragility curves are represented by two-parameter lognormal distribution functions, with parameters estimated using the maximum likelihood method. The paper also discusses methods for testing the goodness of fit of these curves and estimating confidence intervals for the parameters (median and log-standard deviation). An analytical interpretation of randomness and uncertainty associated with the median is provided. Empirical fragility curves for bridges in the Hanshin Expressway Public Corporation (HEPC) are developed based on damage data from the Kobe earthquake. The curves are expressed using two-parameter lognormal functions, with the median and log-standard deviation estimated via maximum likelihood. The peak ground acceleration (PGA) is used to represent seismic intensity. The fragility curves are constructed using 770 single-support reinforced concrete columns from two viaducts. The curves are validated using statistical methods, and the results show that the hypotheses cannot be rejected at the 10% significance level. Analytical fragility curves are developed for two representative bridges in the Memphis area using nonlinear dynamic analysis. The bridges are modeled with specific structural characteristics, and the fragility curves are derived based on the probability distributions of material properties. The curves are validated using ground motion time histories and spectral analysis. The results show that the fragility curves are consistent with the lognormal assumption. The paper also discusses statistical procedures for testing the goodness of fit of fragility curves and estimating confidence intervals. The goodness of fit is tested using the chi-square distribution, and the results indicate that the hypotheses cannot be rejected at the 10% significance level. The confidence intervals for the parameters are estimated using Monte Carlo simulation techniques. The paper concludes that the fragility curves developed for the HEPC bridges and Memphis bridges are statistically valid and can be used for seismic risk assessment. The combined and composite fragility curves are also discussed, showing that they can be derived from individual fragility curves of different bridge types. The lognormal assumption is used for these curves, and the results indicate that the assumption is valid for the given data. The study highlights the importance of statistical analysis in the development of fragility curves for seismic risk assessment.This paper presents a statistical analysis of structural fragility curves, considering both empirical and analytical approaches. Empirical fragility curves are developed using bridge damage data from the 1995 Kobe earthquake, while analytical curves are constructed through nonlinear dynamic analysis. The fragility curves are represented by two-parameter lognormal distribution functions, with parameters estimated using the maximum likelihood method. The paper also discusses methods for testing the goodness of fit of these curves and estimating confidence intervals for the parameters (median and log-standard deviation). An analytical interpretation of randomness and uncertainty associated with the median is provided. Empirical fragility curves for bridges in the Hanshin Expressway Public Corporation (HEPC) are developed based on damage data from the Kobe earthquake. The curves are expressed using two-parameter lognormal functions, with the median and log-standard deviation estimated via maximum likelihood. The peak ground acceleration (PGA) is used to represent seismic intensity. The fragility curves are constructed using 770 single-support reinforced concrete columns from two viaducts. The curves are validated using statistical methods, and the results show that the hypotheses cannot be rejected at the 10% significance level. Analytical fragility curves are developed for two representative bridges in the Memphis area using nonlinear dynamic analysis. The bridges are modeled with specific structural characteristics, and the fragility curves are derived based on the probability distributions of material properties. The curves are validated using ground motion time histories and spectral analysis. The results show that the fragility curves are consistent with the lognormal assumption. The paper also discusses statistical procedures for testing the goodness of fit of fragility curves and estimating confidence intervals. The goodness of fit is tested using the chi-square distribution, and the results indicate that the hypotheses cannot be rejected at the 10% significance level. The confidence intervals for the parameters are estimated using Monte Carlo simulation techniques. The paper concludes that the fragility curves developed for the HEPC bridges and Memphis bridges are statistically valid and can be used for seismic risk assessment. The combined and composite fragility curves are also discussed, showing that they can be derived from individual fragility curves of different bridge types. The lognormal assumption is used for these curves, and the results indicate that the assumption is valid for the given data. The study highlights the importance of statistical analysis in the development of fragility curves for seismic risk assessment.