In clinical trials involving multiple endpoints and testing the ideal situation is to have primary and secondary endpoints clearly defined and appropriate error control. This also extends to analysis of subgroups. Often in practice these goals are not achieved. This talk will discuss issues and cases that the speaker has seen and/or been involved. Criteria will be discussed for achieving appropriate error control and situations where this is not achieved will also be discussed. The presentation will also discuss what is appropriate to say and conclude from a clinical trial when the primary objectives are not achieved but some of the multiple endpoints/testing seem to indicate “something positive” about the study.