If a study involves inordinate physical risk or breach of privacy, it must be reformulated. In addition, a poorly conceived study that does not lead to a satisfactory answer (e.g. when the sample size is too small, etc.) is also non-ethical.