Compared to being in recovery for less than one year, individuals in the 5–10 years, 10–20 years, or 20 years or more recovery groups all had significantly lower odds of non-abstinence (Table 2, Model 4). Effects for shorter lengths of recovery (i.e., 1–2 years, 3–5 years) were in the same direction yet not significant at the traditional P<0.05 level. Furthermore, the odds of non-abstinent recovery appear to decrease in a non-linear fashion such that the effect size for those with 20+ years of recovery is notably smaller than the effect size for those with 1–2 years (OR = .28 vs. OR = .76). These results imply that the longer an individual is in recovery, the more likely he/she is to abstain. Figure 1 illustrates the relationship between length of time in recovery and probability of abstinence.
