To enlist the cooperation of the laboratories, it was necessary to ensure the confidentiality of results and the anonymity of the laboratories and that the resultant data would not be used for other than the stated research purposes. When the study found that a number of laboratories were performing poorly, the funding agency argued for the desirability of identifying the participating laboratories and reporting their performance on individual test specimens, despite the previous promises of confidentiality. After considering arguments for and against destroying the code that enabled the matching of laboratories with results, the research staff decided to dismantle the codebook and return this identifying information to each participating laboratory. This compromise relieved the central office of the ability to link laboratory names with results and gave only the individual laboratories the capability of revealing their own test results. This action raised a number of ethical questions relating to the public interest, the relationship between the private citizen and the State, and the conditions under which promises of confidentiality should or should not be broken. Researchers are urged to consider such issues before entering into agreements for confidentiality. Some recommendations for reducing such confidentiality dilemmas are provided. 36 references.