Process capability indexes are widely used in industries. Nevertheless, important statistical issues that deal with the appropriate use of process capability indexes have only appeared in the technical literature and are thus unknown to many users. In this paper we will first review various kinds of process capability indexes and then try to understand how much uncertainty is associated with them in relation to the assumption of normal distribution, the estimate of ${sigma}$, the sampling error, the size of number of observations available and the presence of data correlation. Finally, we will propose a proper use of process capability indexes in an effort to assess a complex process.