The Matilda Effect in Science Communication: An Experiment on Gender Bias in Publication Quality Perceptions and Collaboration Interest

• The perceived scientific quality of an abstract was found to be significantly dependent on both the gender of the author and the gender-type of the topic, regardless of the gender of the participant.
  • Abstracts from male authors were rated as having significantly greater scientific quality than abstracts from female authors (estimated mean 5.33 vs. 5.26 on a 10-point scale).
  • Abstracts from male authors pertaining to male-typed topics were rated as having significantly greater scientific quality than abstracts from female authors on male-typed topics (5.46 vs. 5.20).
  • Abstracts from male authors pertaining to male-typed topics were rated as having significantly greater scientific quality than abstracts from male authors pertaining to female-typed topics (5.46 vs. 5.22).
• The interest in scientific exchange and collaboration was found to be significantly dependent on both the gender of the author and the gender-type of the topic, regardless of the gender of the participant.
  • Abstracts from male authors on male-typed topics had significantly greater interest for collaboration than if they worked on female-typed or gender-neutral topics.
  • Abstracts from female authors on female-typed topics had significantly greater interest for collaboration than if they worked on gender-neutral topics.
  • Abstracts on male-typed topics with male authors had significantly greater interest for collaboration than if they had female authors.
• Stronger support for gender equality (i.e. lower endorsement of traditional gender roles) was associated with significantly higher ratings for the perceived scientific quality for abstracts from female authors; however, there was no difference for abstracts by male authors or for interest in scientific exchange and collaboration.

Overall, this study showed that identical work attributed to male authors is considered more scientifically robust than if it is attributed to female authors, which is posited to be due to gender stereotypes and expectations about performance in science. This was particularly true for research topics considered to be more stereotypically male. This significant difference could be predicted by the respondents’ general attitudes toward gender: participants who held more traditional attitudes towards gender were less likely to rate female authors’ work as scientifically robust.

Email requests sent to 39 different graduate programs across the country garnered online participation from 243 graduate students (70% female) enrolled in 20 communication graduate programs. These students were asked to evaluate abstracts presented at the 2010 International Communication Association (ICA) conference, where each paper was assigned a fake male or female name. The topic of each paper was categorized as “female-typed,” “male-typed” or “gender neutral.” Each participant reviewed 15 abstracts, where 6 related to the research question and the remaining 9 were included as controls. For each participant, the 6 abstracts were drawn from a pool of 12 available abstracts—each participant received two of each gender-typed category, and each pair had one paper authored by two female names and the other paper authored by two male names.  Participants evaluated each abstract on a scale from 1 to 10 (not at all to very) in terms of scientific quality (8 items: rigorous, by competent authors, influential, important, innovative, publishable in a prestigious journal, high-quality contribution reflective of expertise) and collaborative interest (2 items: interest in discussing research with the authors, and in collaborating with the authors). Participants also completed a 14-item questionnaire on attitudes towards gender roles, which was interspersed with 11 general questions to distract from the questionnaire’s purpose.