A Longitudinal Analysis of Racial and Gender Bias in New York Times and Fox News Images and Articles

TL;DR

This work examines the frequency and prominence of appearance of racial and gender groups in images embedded in news articles, revealing that racial and gender minorities are largely under-represented, and when they do appear, they are featured less prominently compared to majority groups.

Abstract

The manner in which different racial and gender groups are portrayed in news coverage plays a large role in shaping public opinion. As such, understanding how such groups are portrayed in news media is of notable societal value, and has thus been a significant endeavour in both the computer and social sciences. Yet, the literature still lacks a longitudinal study examining both the frequency of appearance of different racial and gender groups in online news articles, as well as the context in which such groups are discussed. To fill this gap, we propose two machine learning classifiers to detect the race and age of a given subject. Next, we compile a dataset of 123,337 images and 441,321 online news articles from New York Times (NYT) and Fox News (Fox), and examine representation through two computational approaches. Firstly, we examine the frequency and prominence of appearance of racial and gender groups in images embedded in news articles, revealing that racial and gender minorities are largely under-represented, and when they do appear, they are featured less prominently compared to majority groups. Furthermore, we find that NYT largely features more images of racial minority groups compared to Fox. Secondly, we examine both the frequency and context with which racial minority groups are presented in article text. This reveals the narrow scope in which certain racial groups are covered and the frequency with which different groups are presented as victims and/or perpetrators in a given conflict. Taken together, our analysis contributes to the literature by providing two novel open-source classifiers to detect race and age from images, and shedding light on the racial and gender biases in news articles from venues on opposite ends of the American political spectrum.

Figures (12)

• Figure 1: Summary of racial classifier results. (A) Confusion matrix of predictions made by the model for different racial groups. (B) Examples of images that were correctly or incorrectly classified by the model when predicting racial group. (C) Results of t-SNE dimensionality reduction of the images with regards to racial group.
• Figure 2: Summary of age group classification. (A) Confusion matrix of predictions made by the model for the different age groups in the dataset. (B) Examples of images that were correctly or incorrectly classified by the model when predicting age group. (C) Results of t-SNE dimensionality reduction of the images with regards to age group.
• Figure 3: Race and gender in news images. (A) Racial and gender representation in the photos included in articles under the Art, Sport, Food, and Travel categories. Chi-squared tests are used to measure differences between NYT and Fox. (B) Normalized area occupied by face of different racial and gender groups across all the aforementioned categories. Independent t-tests are conducted on normalized area values of NYT and Fox. $ns: p > 0.05$, $*: p < 0.05$, $**: p < 0.01$, $***: p < 0.001$, $****: p < 0.0001$
• Figure 4: Proportion of non-neutral articles which mention a given race and convey a particular emotion. Chi-squared tests are used to compare the proportion of articles with a given emotional tilt in NYT and Fox News. $ns: p > 0.05$, $*: p < 0.05$, $**: p < 0.01$, $***: p < 0.001$, $****: p < 0.0001$
• Figure 5: Normalized sentiment difference of articles which mention a particular racial group in New York Times and Fox News. Darker shades of green represent a more positive average sentiment, while darker shades of pink represent a more negative average sentiment.