Methodological observations concerning word rankings and z-score refinements

Abstract

Abstract This article evaluates word rankings suggested by Ary L. Goldberger, Albert C. Yang, and C. Peng as a means of establishing the authorship of texts in the light of Delta, developed by John Burrows at about the same time. The tests carried out with high ranking function words and results established with the more modern approaches of Rolling Delta, Rolling Classify, and the General Imposters method give clear evidence that word rankings only return crude and unreliable results that cannot keep up with non-traditional modern methods. Even though the stylistic difference between Marlowe and Shakespeare plays could be stated, word rankings failed to recognize Shakespearean stylistics in The Jew of Malta, Edward II, and Doctor Faustus. It was only through the use of z-scores that a wider vocabulary provided a larger degree of differentiation.