Comparing Two Measures for Formality

Formality is an important dimension of language style. Texts of different genres tend to have different degrees of formality. F-score (formality-score) is a most popular measure for formality to differ genres. It uses a method of combining proportions of words of different types, with nouns, adjectives, articles and prepositions as positive elements, and adverbs, verbs and interjections as negative elements. This measure is limited to the word level; however, indicators of genre are not restricted to the word level. This paper proposes a new composite component score of formality, “CF-score”, which consists of a combination of five CohMetrix component scores: narrativity, referential cohesion and deep cohesion, syntactic simplicity, word concreteness. F-score and CF-score are compared with TASA corpus in diverse text genres. Results of General Linear Model showed CF-score formality measure was more powerful and more rational with text genres than F-score.