Previous research has shown a male bias in the media. This study tests this statement by examining how the pronouns She and He are used in a news media context. More specifically, the study tests whether He occurs more often and in more positive semantic contexts than She, as well as whether She is associated with more stereotypically and essential labels than He is. Latent semantic analysis (LSA) was applied to 400 000 Reuters' news messages, written in English, published in 1996-1997. LSA is a completely data-driven method, extracting statistics of words from how they are used throughout a corpus. As such, no human coders are involved in the assessment of how pronouns occur in their contexts. The results showed that He pronouns were about 9 times more frequent than She pronouns. In addition, the semantic contexts of He were more positive than the contexts of She. Moreover, words associated with She-contexts included more words denoting gender, and were more homogeneous than the words associated with He-contexts. Altogether, these results indicate that men are represented as the norm in these media. Since these news messages are distributed on a daily basis all over the world, in printed newspapers, and on the internet, it seems likely that this presentation maintains, and reinforces prevalent gender stereotypes, hence contributing to gender inequities.