Statements on conflicts of interest provide important information for readers of scientific papers. There is now compelling evidence from several fields that papers reporting funding from organizations that have an interest in the results often generate different findings from those that do not report such funding. We describe the findings of an analysis of correspondence between representatives of a major soft drinks company and scientists researching childhood obesity. Although the studies report no influence by the funder, the correspondence describes detailed exchanges on the study design, presentation of results and acknowledgement of funding. This raises important questions about the meaning of standard statements on conflicts of interest.