The Gender of Contributors

The materials on this page are still very much a work in progress. At the MJP, we don’t catalogue the gender of an author or editor when we assemble our MODS records describing the contents of individual journals. But we do, of course, include the author’s first name (or names), as well as authors’ terms of address (like Mrs.), and from that information we can infer (albeit with a sizable risk of still getting it wrong) what the author’s gender probably was.

Elli Mylonas, from Brown’s Center of Digital Scholarship, along with David Trejo, a Brown undergraduate, have created a program that helps us do this. Drawing from a dataset made up of a list of names (last name, first name), the program identifies the first name, checks it against two lists of gendered names from the census bureau, and then decides (if a name gets hits from both lists) which of the two genders is more likely, based on the popularity of the name for men or women. When the program has trouble deciphering the name, or can’t assign it a simple gender, it labels it as “undetermined”. Here’s the link to the gender-identifying website that Elli and David have created; you can paste into it a series of names you want to process. And here’s what a spreadsheet of the processed names looks like:

To calculate the overall number of authors who belong to each gender category (female, male, or undetermined), we can create a pivot table from the spreadsheet, with the authors’ names in the vertical axis, the three genders in the horizontal axis, and the authors’ last names (which correspond to the overall number of contributors) as the table’s value. You can see such a pivot table by pressing on the “gender numbers” tab just above the spreadsheet.

Then, taking the totals of each gender listed in the bottom row of the pivot table, we can create pie charts like the ones that appear below on this page. The information in these charts is derived from datasets in which each author is listed just once; thus, they ideally give us a picture of the gender of all of the contributors to a magazine. If we worked with a list of authors that weighed the number of contributions each made, we could visualize instead the percentage of a magazine created by either gender.

In preparing the data represented in these charts, we’ve manually checked and corrected the genders that the program automatically assigned to contributors, and we’ve also eliminated any obvious duplicate references to the same contributor. Still, the gender of a large percentage of the contributors remains undetermined—a testament, in part, to the commonplace practice in these journals of identifying contributors by aliases or initials.

The data in the above spreadsheet, by the way, happens to represent all the contributors to The Egoist, The Little Review, and Others combined from July 1915 through July 1919. Here’s what our data says about their gender:

And here are three more charts depicting the gender-makeup of the contributors to each of these magazines:

How You Can Make Charts Like These

To chart the gender of contributors to an MJP journal, download from the MJP file repository on Sourceforge the third .txt dataset (detailing “every contributor”) for that journal. Since the info within each dataset is ordered chronologically, you can tailor the dataset to a specific date range by deleting any strings that fall outside your desired range—just be sure to hang onto the first string which contains the field headers.
Next, create a spreadsheet from the .txt file, and then create a pivot table from the spreadsheet that gives you an alphabetized list of every contributor to the journal.
Now copy that list of names into the online name disambiguation device, and press “Disambiguate!”
Copy the output from the “Results” box into a blank .txt file, and add pipe-divided field headers (which you can copy from gender tool website, if you like) into the top line.
If you want to increase the accuracy of the results, you should now review the list of names and change any gender assignments that you know are wrong (some research here into the authors would be helpful).
Now create a new spreadsheet out of the corrected .txt file, and create a new pivot table from the spreadsheet that, like the pivot table on the page above, counts the number of gendered contributors to the journal.
Then on Google documents, create a new spreadsheet of just two columns: In column A, make “Gender” the header in the first row and enter “female,” “male,” and “undetermined” in the next three rows below it; and in column B, make “Number” the header in the first row and enter, in the rows that follow, the total number of female, male and undetermined contributors, which appears at the bottom of your second pivot table.
Finally, use the “insert chart” option on Google Documents to visualize the spreadsheet data as a pie chart, which you can then insert into the spreadsheet and also export.