BACKGROUND:Telephone surveys have been used widely in public health research internationally and are being increasingly used in Ireland and the U.K. METHODS:This study compared three telephone surveys conducted on the island of Ireland from 2000 to 2004, examining study methodology, outcome measures and the per unit cost of each completed survey. We critically examined these population-based surveys which all explored health related attitudes and behaviours. RESULTS:Over the period from 2000 to 2005 the percentage of calls which succeeded in contacting an eligible member of the public fell, from 52.9% to 31.8%. There was a drop in response rates to the surveys (once contact was established) from 58.6% to 17.7%. Costs per completed interview rose from 4.48 euro to 15.65 euro. Respondents were prepared to spend 10-15 minutes being surveyed, but longer surveys yielded poorer completion rates. Respondents were willing to discuss issues of a sensitive nature. Interviews after 9 pm were less successful, with complaints about the lateness of the call. Randomisation from electronic residential telephone directory databases excluded all ex-directory numbers and thus was not as representative of the general population as number generation by the hundred-bank method. However the directory database was more efficient in excluding business and fax numbers. CONCLUSION:Researchers should take cognisance of under-representativeness of land-line telephone surveys, of the increasing difficulties in contacting the public and of mounting personnel costs. We conclude that telephone surveying now requires additional strategies such as a multimode approach, or incentivisation, to be a useful, cost-effective means of acquiring data on public health matters in Ireland and the U.K.