Maximizing the potential of worksites for smoking intervention remains elusive. We hypothesized that long-term effectiveness of group intervention would be enhanced when offered within an "enriched milieu" (full program) compared with relative isolation (group only). The data failed to support the hypothesis. Although sustained abstinence rates were higher at full-program sites (50%) than at group-only sites (44%), the difference did not achieve statistical significance. Initial (35% vs. 47%) and 12-month (18% vs. 22%) quit rates at full-program and group-only sites also failed to demonstrate the "benefit" of the "enriched milieu."