The gender genie thinks the current version of my novel (all 128,000 words of it) was written by a...female, but only by a narrow margin (my net female score was 175,322 and my net male score was 171,174--not sure that's even statistically significant).
In case anyone was wondering, the gender genie is an algorithm that is available at http://bookblog.net/gender/genie.php. You can paste a work of fiction or nonfiction (they recommend something that is at least 500 words long) into a window and click a button. It spits out a couple of scores and tells you whether the program thinks you are male or female. It assigns a certain number of points (male or female) to certain words. For instance, the word 'with' has a feminine score of 42, while the word 'it' has a masculine score of 52.
The program has a list of words (16 for feminine and 17 for masculine) that its creators state are used more often by one gender versus the other. I assume that this is based on some kind of statistical analysis that they've performed on a large number of writing samples. Some of the word assignments seem in line with stereotypes. "With" is feminine, because women are supposed to be all about cooperation...right? 'Above' and 'below' have male points assigned, because men are all about hierarchies, right? Yes, I'm being a bit tongue in cheek here. But what about 'it,' 'at' and 'around' (also words with masculine scores)? Those words don't seem stereotypically masculine. And while 'she' and 'her' refer to female people or animals, of course, don't men ever write novels with lots of female characters? Can a man 'cheat' the gender genie by writing a story from a feminine point of view? And why aren't male pronouns assigned 'masculine' scores? They're not on the word list at all. Does that mean that women refer to male characters or animals about as often as men do?
So I have some questions about this thing. My novel comes out having a net feminine score, but barely. Since one of my protagonists is male, I'm not thinking that this is a 'bad' thing. I don't doubt that word usage probably varies a bit by gender, but it seems like there might be other things (such as pattern of word use or use of certain word combinations) that would vary more between the sexes. And wouldn't culture play a role too, even among native English speakers?
I also wonder whether some of these gender differences in word use stem from the subjects that male and female authors most often choose to write about. The only way to test that would be to have large groups of males and females asked to write about a narrowly proscribed topic (fiction or non fiction) and then to analyze their word choices.
I am guessing that topic may have a significant effect, because it seems to with my own writing. I find that when I break my novel down and run individual chapters through the gender genie, some come out strongly masculine and a larger number come out modestly feminine. The 'masculine' chapters tend to be the ones where there is some kind of action or my male character is dealing with some kind of personal issue, while the 'feminine' chapters are ones that are written from my female character's pov or that involve a lot of dialog from anyone's pov.
My non fiction is almost unilaterally 'masculine.' This blog entry, for instance, comes out as having a higher 'masculine' score (1160 to 925), as do the handouts I write for my students (I teach biology, so the handouts are written in very straightforward, 'technical' language). My short story 'Forever Home' is written in first person from a female pov, but gender genie also thinks it was written by a male (1713 to 1589).
I suspect that my writing style is not highly 'gendered' whatever that is, so my word usage depends on what I am writing about. The authors of the program say that it accurately guesses an author's gender most of the time, but I wonder what percentage of people come out near the middle of the distribution or cross over into the other gender. Are they people who tend to write about subjects that are usually chosen by the opposite sex? Is there a 'feminine' way of writing a physics textbook or a 'hard' science fiction story and is there a 'masculine' way of writing a romance novel or an article on prom dresses? I'm not trying to be shamelessly sexist here, but I am curious about whether the creators of this program controlled for non fiction topics and for fiction genres or not.
I also wonder if it would be possible to create a similar algorithm that can, with reasonable accuracy, guess a person's culture, age or nationality (if you're checking for British English versus American, spelling of certain words and use of certain slang terms is a give away, but if you corrected for that, is there still a difference in word usage?
In any case, I'm a bit skeptical about what it is the gender genie is really telling us about our writing, but it is fun to play with.