To determine if smoking cessation counseling is affected by physician or patient demographics, patient exit interviews and chart audits were administered between January 1991, and February 1992, in an inner-city teaching clinic. Patients who received medical care on an ongoing basis in the clinic were eligible for the study. Patients with dementia or inability to verbally communicate in English were excluded. According to patient self- report, 26% (167) of the 642 patients who participated were never asked about smoking. Of the 168 current smokers, 23% (39) were never advised to quit. Male patients were more likely to report being asked about smoking on the day of the clinic visit (odds ratio = 2.37, 95% confidence interval [CI] = 1.62 to 3.46). Male smokers were more likely to report being advised to quit (odds ratio 2.39, 95% CI = 1.25 to 4.58), as were white smokers (odds ratio 3.66, 95% CI = 1.37 to 9.82). According to patient report (smokers and nonsmokers), white physicians were more apt to ask about smoking on the day of the clinic visit (odds ratio = 2.68, 95% CI = 1.31 to 5.48). Physician and patient demographic characteristics may significantly influence the rate of smoking cessation counseling. Additional studies ate needed to confirm the presence of such biases.