The NLS dataset does not contain author gender information, so genders were inferred from first names and may not be 100% accurate. There are five gender categories authors are sorted into: male, mostly male, androgynous, mostly female, and female. An author's gender category is determined by how often people with their first name identify as male or female.

For example, 'Orlagh' is frequently a female first name and rarely a male first name so is classified as 'female'. On the other hand, 'Cameron' is usually a male first name, but is sometimes a female first name so is classified as 'mostly male'. First names with a relatively even split are classified as androgynous.

Authors in the dataset are sometimes an organisation or body rather than a person or group of people. In this case, the gender is classified as 'unknown'. Cases where the gender is unknown are not shown in the graph and make up about 1 in 4 classifications. Very rare names, particularly non-english names, are sometimes classified as 'unknown' too.

The gender inference process was done by a tool developed by an external source. More details can be founded in this project's associated paper.