AIMS: Considers whether participant non-response at household, individual and item-level, may have had an effect on the results of the British Gambling Prevalence Survey 2007. METHODS: Cross-referencing responses with results collected from both respondents and non-respondents and from other data sets. FINDINGS: Overall, the findings reported in the BGPS 2007 were robust against potential non-response biases. Infrequent gamblers were less likely to answer the problem gambling screen than frequent gamblers. This meant that the prevalence of problem gambling in some population sub-groups was marginally lower than previously reported. However, this had no impact at the overall population level. CONCLUSIONS: The level of problem gamblers remained at 0.6%, the same as stated in the BGPS 2007 report.