Abstract:
From the late 1960s through the 1970s, researchers worldwide have shown interest in the
exploration of gender representation in children’s literature, including books, stories, and
educational materials. A significant representational discrepancy was witnessed and proved
between both genders in central characters, illustrations, titles, and text in different children’s
stories and books through several studies conducted over the years. Several methods have been
used for the detection of gender bias, yet most of these methods followed a manual frequency-based qualitative and quantitative content analysis approach that focuses on the word-level
detection of gender bias in language. This study, however, presents an advanced automated
computer-driven approach that can detect different gender bias categories at a phrase-level and
sentence-level. This study applies its automated methodology and finds countless instances of
gender bias patterns investigated in more than 200 children’s books and stories, most of which
are still read to and by children today. It also tries to explore any relationship existing between
the gender bias categories detected and some attributes collected, such as “author’s gender”,
“book genre”, and “year of publication”. This study finds significant effects of the” author’s
gender” and “book genre” on the use of the different types of gender bias categories where male
authors tend to display a greater bias in language towards males as compared to female authors.
This research also presents the previous work that has been done in the field of gender research
in children’s literature and discusses the negative impact that a gendered language has at a
micro-level and macro-level. Finally, this work aims to enhance the existing detection
approaches, especially for the identification of gender bias existing at the level of the language,
and it presents an automated machine-led content analysis approach for this purpose.