If research on personality is to be productive and progressive, it is necessary to make objectives explicit and to link abstractions with common experiences. Six levels of classification outline these objectives. Level 1 involves the characterization and classification of persons. Level 2 studies attributes. Level 3 is concerned with inferred processes. Level 4 studies the processes in construing behavior and in making attributions. Level 5 looks for behaviors common to most people. Level 6 examines behaviors in sequence. Generalizability (external validity) is low for the first three levels and increases for the remaining ones. Research at the earlier levels shows little likelihood of developing into a basic science; the possible degree of agreement between observers is limited. Whether this classification seems adequate or not, each person researching in personality must make explicit his phenomena, methods and goals. (Author)