As personal computers (PC’s) have become more powerful, sophisticated data analysis programs have evolved for this platform. SAS/STAT for Personal Computers, a derivative of the mainframe application, was leased by the Customs Research Laboratory to do canonical and discriminant analysis in November 1988. In January 1995, we acquired the Unistat 3.0 package. There are obvious differences in the ease of use of the two packages. This paper will discuss the analytical differences that were observed by processing a single data set in both packages. Implications of using this type of data treatment on measurements from thermal analysis experiments will be discussed, along with the underlying analytical questions posed by the sample set. A basic question will be addressed : does the choice of the software affect the ultimate results of the analysis?