A computerized adaptive test for placement of students in postsecondary French second language courses is evaluated for unidimensionality of its three component tests: reading comprehension of a short paragraph; selection of the appropriate statement in a given situation; and a "fill-in-the-blank" section. A variety of statistical procedures were used to assess the components' unidimensionality, including a structural equation approach, factor analysis, nonparametric approach, and item response theory approach. Results of each of these analyses are explained and synthesized. It is concluded that the varying and sometimes conflicting results raise questions about the concept of unidimensionality, and that unidimensionality is a matter of degree rather than a yes/no issue, depending heavily on expert judgment. A 49-item bibliography is included, and procedures and results of the different analyses are appended. (MSE)