Fig 1.
Dataset construction process, from parsing raw movie reviews in the movie review data, enriching metadata with the OMDb API, identifying genders of key contributors (actors, director, writer), to measuring gender representation metrics.
Fig 2.
Total reviews and unique movies by decade.
The bars indicate the number of reviews (left Y-axis), while the line graph shows the number of unique movies (right Y-axis).
Fig 3.
Distribution of movies by country.
Fig 4.
Inference pipeline of fine-tuned DeBERTa models, providing the likelihood of specific bias types in sentences and determining the presence or absence of bias for explicit sex markers.
Table 1.
Categorization of gender bias in text, detailing subtypes such as generic pronouns, sexism, and exclusionary terms, with examples illustrating their manifestation in language.
Adapted from [3].
Fig 5.
Temporal evolution of the Gender Representation Disparity Index (GRDI) from the 1990s to the 2020s, showing male-to-female representation proportions across actors, directors, and writers over time.
Fig 6.
Clustered bar chart showing gender representation proportions across movie genres.
Each cluster represents a genre, with bars for the first actor, director, and writer roles. Blue bars indicate male representation and red bars indicate female representation.
Fig 7.
Relationship between the Gender Representation Disparity Index (GRDI) and average benevolent sexism in film criticism, illustrating increased sexism as the GRDI moves toward a female-dominated cast.
Fig 8.
Relationship between the Gender Representation Disparity Index (GRDI) and average hostile sexism in film criticism, illustrating increased sexism as the GRDI moves toward a female-dominated cast.
Fig 9.
Relationship between the Gender Representation Disparity Index (GRDI) and average explicit sex markers in film criticism, illustrating a decrease in markers as the GRDI moves toward a female-dominated cast.
Fig 10.
Comparison of boxplots for benevolent and hostile sexism scores in movie review transcripts, segmented by the gender of the first actor, director, and writer, highlighting the differences in antagonism and bias when females lead movie roles.
Fig 11.
Boxplot comparison of benevolent and hostile sexism scores in movie review transcripts, categorized by movie genres.