We investigated the magnitude of self-medication with antibiotics in a peri-urban area of Southern Kerala State, India and factors influencing this practice. First, a random sample of 400 households was surveyed in one primary health centre area near Trivandrum. We found 69.3% (95% CI = 64.8-73.8) of households had at least one person using a pharmaceutical product during the two-week recall period; antibiotics formed almost 11% of the medicines consumed. Next, pharmacy based interview and observation data were collected from 405 antibiotic purchasers sampled from 11 out of the 12 private pharmacies in the area. Seventy-three of these 405 customers purchased antibiotics without a prescription (18%; 95% CI = 14.3-21.7). By combining the household survey and pharmacy observations, we estimate that almost half of 1% (0.41%; 95% CI = 0.24-1.16) of the population, or four people per 1000, is engaged in self-medication using antibiotics in Kerala in any two-week period. Our data show that people least likely to follow this practice are from higher income families, having more education and higher status occupations and receiving the benefits of medical insurance. Conversely, logistic regression analysis indicated that risk of buying antibiotics without a script was associated with education at secondary level or below, the perception that it is expensive to consult a doctor and low satisfaction with medical practitioners. Keralites' self-medication patterns are interpreted broadly using social, cultural, historical and economic perspectives. Solutions to the problem of antibiotic misuse are suggested, proceeding on several fronts: among practitioners, suppliers and marketeers of medicines, and among the population of pharmaceutical consumers themselves.