Basic concepts of statistically distributed flaws and moisture-enhanced growth of cracks under stress are used in a nonparametric bootstrap analysis to assess the reliability of dual-pane glass aircraft windows. Statistical distributions of window strengths, evaluated by rapid loading in an inert environment, are analyzed as three-parameter Weibull distributions using a maximum likelihood procedure. Strength distributions for a variety of pane surface conditions are evaluated, for example, as-ground and polished surfaces, which are typical of a `protected' inner window pane, and outer pane surfaces with various types of simulated in-service `damage,' e.g., airborne dust impact, windblown sand, and cleaning/handling scratches. The crack growth parameters needed to assess the time-dependent crack growth behavior are determined at room temperature in a water environment via dynamic fatigue, or constant stressing rate tests on indented specimens. Predicted lifetimes at a 95% confidence level are ascertained for various window scenarios at a 99% survival probability via a Monte Carlo simulation using a nonparametric bootstrap procedure.