Often in statistics it is not possible to determine the exact value of some treatment effect or other outcome of interest. However, it may be possible to determine the set of possible values for the outcome of interest. We call this the identified set. The approach is called set identification as opposed to point identification. It is common in problems that the set of interest lies in an interval of the real line. Such a set has a highest possible value, the upper bound, and a lowest possible value, the lower bound. Thus it is common to refer to set identification as "bounds."