Evaluating gender prejudices in readers
Graduate Researcher • September 2017 - August 2019 •
Research design, literature search, study development, data analysis, report writing, knowledge translation
Historically, and into today, many women authors feel the need to write under male or gender-neutral pseudonyms in order to be published. This can be seen in famous cases such as the Brontё sisters who, in the mid 1800's, published under the names Currer, Acton, and Ellis Bell and J.K. Rowling who was explicitly told to not use her actual name so people would read the Harry Potter series. But why is this?
Methods: A/B testing, in person and online testing, survey, statistical analysis
Tools: Qualtrics, R, OneNote
1. Do people judge books written by women more harshly than those written by men?
2. Are negative evaluations of romance novels related to their associations with women?
To test this, we used an A/B design wherein 166
participants were given a series of passages from books and told they were either written by men or women (denoted by the name of the author). From here, they were asked to evaluate each passage on a series of evaluation criteria.
We then used Linear Mixed Effects Modeling and Bayesian statistics to examine the effects of author gender and see if these results differed based on participant gender.
Results from this study showed no difference in how people evaluate passages based on the gender of the author. When including participant gender we again saw no difference.
Next, we wanted to build on this to look at genre. Specifically, we were interested in the romance genre which is overwhelmingly read, and written, by women and, despite being the second most popular book genre, is viewed almost universally as being of poor quality.
We repeated our design from Study 1, this time adding information about the genre of the books as being from either romance or literary fiction. Other than the inclusion of genre, Study 2 mirrored Study 1.
Again, we used Linear Mixed Effects Modeling and Bayesian statistics this time looking at the effects of author gender and book genre and again testing whether the results differed based on participant gender.
As with Study 1, we found no difference in how people evaluate passages based on the gender of the author, replicating our findings from Study 1. Further, there was no effect of book genre or participant gender.
Many women feel the need to publish under male names in order to be published, with one woman reporting that she was 8 times as likely to get a response when she used a male pseudonym. However, these results suggest that this bias is not based on the reality of readers.
One possibility is that a bias does exist, just not on the part of readers. It could be that publishers think that people don't want to read books written by women. If this is the case, these results suggest that publishers are making incorrect assumptions by thinking boys will want to read books written by women or selling books written by women for less than those of men.
Overall, more research needs to be done to understand why books written by women aren't as valued as those written by men and why romance is viewed so negatively, but our research suggests the bias against women authors is not reflected in how readers actually feel about their works and, if publishers are making their decisions based on this belief it would be in the best interest of these companies to update these beliefs.
This project resulted in a peer-reviewed publication on which I am the senior author and was shared in an invited post on Character and Context, the blog for the largest association of Social and Personality researchers in the world. I have also been able to share this work at multiple international conferences with diverse audiences.