Numerical Taxonomy for Educational Researchers

The grouping of pupils into homogeneous subgroups is an educational practice of long standing. Elementary pupils are commonly assigned to reading groups; high school students are enrolled in differentiated curricula on the basis of their vocational goals, and considerable emphasis is placed upon grouping of pupils in modern instructional programs. These groupings are usually based upon procedures such as teacher judgment, a test score distribution, or some judgmental combination of test scores and pupil characteristics. Given the importance of pupil grouping to present educational practice, it is surprising that the use of informal grouping procedures is widespread when well-developed algorithms for grouping exist. These algorithms have been developed in the biological sciences to yield taxonomies of plants and animals, in information retrieval for the classification of documents, and in the social sciences for the grouping of persons. Although many of these grouping algorithms can be performed by hand, they are widely available in the form of computer programs. The potential for application of these techniques in educational practice and research is considerable, yet the actual use of such algorithms in educational settings has been rare. The goal of the present paper is to review this methodology in order to acquaint educational researchers with these simple, but useful, grouping algorithms.