The Bollen-Stine bootstrap is the only option offered in the Amos program as a means of addressing the issue of non-normality. However, the only information that appears to be provided when this option is requested is the Bollen-Stine p-value, which represents the proportion of times that the model chi sq for the bootstrapped samples exceeded, was about equal to, or was less than the chi sq for the observed data.

Given that there appears to be no correction to the SEs or to the overall chi square value for the hypothesized model, can anyone explain exactly how the Bollen-Stine bootstrap truly addresses the issue of non-normality?