Our study has limitations common to most large epidemiological studies. First, the NESARC sample only included civilian households and group living populations aged 18 and older. Therefore, information was unavailable in the military at the time of the survey, which may have different rates of PTSD or AD. Second, because of questions about the validity of the diagnosis of alcohol abuse (Hasin et al., 1999; Keyes and Hasin, 2008), we limited our analyses to alcohol dependence. Thus, our results may not generalize to individuals with PTSD and alcohol abuse. Third, impulsivity was assessed only with one item, with may have led to limited reliability, increased error variance, and decreased statistical power. Even with this broad assessment, our results suggest differences in impulsivity across the three groups, consistent with previous analyses of the NESARC (Chamorro et al., 2012). Fourth, we assessed suicide only among individuals who had been sad or depressed for at least two weeks, which may have excluded impulsive suicide attempts. However, the results of previous studies suggest that the number of individuals with history of suicide attempts who
