Gender constitutes a central variable in studies of language variation, as it both reflects and challenges prevailing conceptions of how men and women are socialized into distinct conversational styles. The present study examines gender-related differences in spoken American English through an analysis of the Switchboard Dialogue Act Corpus, a large dataset comprising 1155 spontaneous telephone conversations involving 1058 male and 1252 female speakers. Grounded in variationist sociolinguistics, a corpus-based quantitative approach was employed to identify gender-related patterns across lexical, grammatical, and pragmatic dimensions. The analysis revealed that women produced significantly more turns and utterances, whereas men produced longer utterances and lexical items. Lexically, both genders shared most vocabulary, though notable contrasts emerged: men more frequently used uh, while women preferred um; women also employed we and family-related terms more often, whereas men used wife with higher frequency. Grammatically, men showed a preference for nouns, numerals, and determiners, while women favored pronouns, verbs, and the to-infinitive form. Pragmatically, men tended to use statement-opinion, whereas women were more inclined toward statement-non-opinion. Contrary to earlier claims, no significant gender difference was observed in the use of tag questions. These findings demonstrate that gender differences in spoken language are real yet subtle, highly context-dependent, and best captured through large-scale corpus evidence rather than through stereotypical assumptions or broad generalizations.