A Lexical Universe for Verbal Reasoning Tests.

Autor/inn/en	Breland, Hunter M.; Jones, Robert J.
Titel	A Lexical Universe for Verbal Reasoning Tests.
Quelle	(1992), (16 Seiten)Verfügbarkeit
Sprache	englisch
Dokumenttyp	gedruckt; Monographie
Schlagwörter	College Freshmen; Comparative Analysis; Computational Linguistics; High School Students; High Schools; Language Tests; Mathematical Linguistics; Statistical Analysis; Test Construction; Thinking Skills; Verbal Tests; Vocabulary; Word Frequency; Word Recognition + Suchen Sie Ihr Suchwort? Studienanfänger; Linguistics; Computerlinguistik; High school; High schools; Student; Students; Oberschule; Schüler; Schülerin; Studentin; Language test; Sprachtest; Mathematische Linguistik; Mathematische Sprache; Statistische Analyse; Testaufbau; Denkfähigkeit; Mündliche Prüfung; Wortschatz; Word analysis; Frequency; Wortanalyse; Häufigkeit; Worterkennung
Abstract	The usefulness of a "lexical universe" in the development of verbal reasoning tests was studied. The lexical universe, which is a comprehensive listing of words and their frequencies based on a collection of machine-readable texts exceeding 14 million words of running text, is an outcome of a 1987 College Board study of the language in materials encountered by high school students and first-year college students. This report describes an analysis of the materials collected and presents statistics on the following: total frequencies; an index of dispersion across text categories; an estimate of frequency per one million tokens; and a standard frequency index based on a logarithmic transformation. Additional tables show some sample words and statistics for the College Board corpus as compared to the American Heritage corpus, and a comparison of frequency statistics for a selection of 20 words based on 4 different corpora. The difficulty of acquiring and maintaining a large vocabulary is discussed, along with the importance of word frequency and word frequency counts. It is concluded that the major advantage of word frequency counts is that they yield difficulty estimates for all words that are encountered in the corpus. Contains 24 references. (LB)
Erfasst von	ERIC (Education Resources Information Center), Washington, DC