A Study of the Relationship of Certain Variables to Sex Characteristic Identification From the Speech of Heterosexual and Homosexual Individuals.

Abstract

This study explored relationships of speaker sex and masculinity-femininity judgments and 12 measures of rate, fundamental frequency, and intensity from taped reading and spontaneous speech samples of female and male heterosexual and homosexual individuals. Phase One was designed to determine judgment reliability and any procedural variables that might influence judgments. Sex judgments were more accurate on the speakers' second performances. Both sex and masculine-feminine judgments were more accurate on reading than on spontaneous speech. Analysis of judgments from type-scripts of spontaneous speech yielded a significant judge sex-by-training interaction. In Phase Two, 20 listeners judged sex and masculinity-femininity from taped reading and spontaneous speech samples of female and male heterosexual and homosexual speakers. The mean fundamental frequency low of females judged male/undecided was higher than those judged correctly. The mean intensity standard deviation of females judged female intensity standard deviation of female judged female was greater than that of females judged male/undecided. All male speakers were correctly identified. Females judged masculine displayed a greater mean number of syllables per second, a lower mean fundamental frequency high, and a more restricted mean fundamental frequency range than those correctly described. Males judged feminine displayed a higher mean fundamental frequency mode and a greater mean intensity standard deviation than those correctly described. Female speakers were judged male more often than males were judged female. Homosexuals were incorrectly described as masculine or feminine more often than heterosexuals. Males averaged more syllables per second than females. Mean fundamental frequency low, high, and mode were higher for females and males. Mean fundamental frequency range was greater for females. Mean intensity low and mean intensity range were significantly different for heterosexuals and homosexuals. Significant sex-by-type interaction occurred on mean intensity range and on mean syllables per second. Mean syllables per second, words per minute, percent pause time, fundamental frequency range, and intensity standard deviation were significantly different for reading and spontaneous speech. Suggestions for clinical application and future research were made.