We do not know whether the WIR sample represents the population of individuals in recovery. The WIR sample is mostly white and relatively well educated. However, comparisons to other samples of individuals in recovery (e.g., National Alcohol Survey) show no differences in demographics across samples (Subbaraman et al. in press). Furthermore, the correlates of non-abstinence we found were similar to those found in NESARC, suggesting similarities between our sample and large, nationally representative samples of individuals in recovery (Dawson et al. 2007). The WIR data do not include current dependence diagnoses, which would be useful for further understanding of those in non-abstinent recovery. In addition, the WIR quality of life measure is based on a single question; future studies could use instruments that detail various aspects of mental and physical functioning. WIR is also cross-sectional by design, though it did include questions about lifetime drug and alcohol use. Finally, the WIR survey did not ask about preferential beverage (e.g., beer, wine, spirits), usual quantities of ethanol and other drugs consumed per day, or specifics regarding AA involvement; because these factors could impact the recovery process, we will include these measures in future studies.
