The Gender Genie

[Mark Johnson]

A while back in Most attractive setting for female players, there was a little consideration of how to make RPG text more palatable to a larger range of people.  I do not want a repeat of that thread, but I have continued to ponder the problem.

Recently, I came across the Gender Genie, an algorithm which separates "male" from "female" writing.  The notion behind this is that there are a set of "male" and "female" words and constructions that can be analyzed to reveal the sex of the author of any particular work.  There is more about the algorithm here.

Personally, I think the terms "objective" and "subjective" or "categorical" and "personal" would be preferable here.  I find the idea of "male" and "female" writing (even if empirically valid across a statistical sample) anti-individualistic.  And it sounds rather sexist as well.  For the sake of clarity, I will try to reference their terms as well mine.

Not suprisingly, most RPG game text is overwhelmingly "male."   This could be a reflection of the gender of the writers, the intended audience, tradition and function (by its nature instructive text is probably more "categorical" than "personal").  

Is this anything that needs to be considered when crafting game text?

Has "male" game text been a barrier to female entry into the hobby?

If your target audience is largely female would you consider writing differently?  

If you are a woman, do you find it necessary to adjust your writing to fit RPG text traditions?  

What techniques can be used or have been used to broaden the appeal and readability of the game?  Example: Restate major points in both "objective" and "subjective" styles with out being redundant.

If your game is about simulating "objective" reality (some Sim-types) is it better to use highly "objective" language and if it is fundamentally about relationships (some Narrativist games) do you adjust your language accordingly?

How have you creatively dealt with other gender based issues("he"/"she") in your writing as it pertains specifically to RPGs?

[Mark Johnson]

Ironically, the Gender Genie said the above post was

"Female Score: 621
Male Score: 547

The Gender Genie thinks the author of this passage is: female!"

Definitely to be taken with a grain of salt.

[Ian Charvill]

The important thing to remember is that Male/Female is not a gender split - it's a sexual one.

Therefore you might in some sense be able to split masculine from feminine prose but that may not tell you much more than any form of stylistic jusdgement.

That is: male/female is an objective classification of things; masculine/feminine is subjective: what one culture will judge to be a feminine trait, another might consider masculine (additionally different contexts within the same culture will render different judgments: e.g. 30-40 years ago, cooking is womens work, but most chefs were men).

[Rich Forest]

Hi Mark,

As Ian has pointed out, the news article and the "Gender Genie" are conflating sex with gender (male/female vs. masculine/feminine).  This is common, and perhaps to be expected, from popular versions of academic texts.  Still, it's important to remember that the researchers are most likely making a more careful assertion.  I don't have access to the original academic articles, but the titles of both articles refer to "gender," "genre," and "text" and really aren't making the strong assertions made by the program.  So your concerns about the possible sexist implications of the program may be well-founded based on the "Gender Genie" webpage and the online news article, but remember that the original articles are probably a lot less troubling.  

Now, on to your main questions:

Is this anything that needs to be considered when crafting game text?

Has "male" game text been a barrier to female entry into the hobby?

If your target audience is largely female would you consider writing differently?

If you are a woman, do you find it necessary to adjust your writing to fit RPG text traditions?

I'm going to skip the first question because I'm not sure how to answer it—I'm reading it more as a statement of your overall concern rather than a specific question you want answered.

Now, I think the second question, "Has 'male' text game text been a barrier to female entry into the hobby" might be based on a fallacious assumption.  It implies that women will tend to read text that has the traits associated with a feminine style of writing.  The problem is, we don't know that.  It may be true.  It may not.  The program doesn't claim to tell us anything about how women read.  It claims to analyze gender-typical traits in writing.  

The third question, "If your target audience is largely female would you consider writing differently?" is interesting.  I think, frankly, that it would be hard for me to do this with any degree of confidence.  If I'm writing for a female target audience, again, I'm not guaranteed that this audience necessarily wants to read something that has a more feminine writing style.  But even if I assume this to be true, I'd have to know my stuff, otherwise there would be the danger that my intentional attempt at a feminine style would produce something that missed the mark.  As an analogy, think of an American doing a "British" accent.  Sure it's convincing... to other Americans.  And with trying to write in a more feminine style, we're talking about a number of really subtle differences.  Can you consistently even read a text and tell if it was written by a man or a woman?  I don't know that I can.  That article is talking about using a computer to do this—it can analyze the information on a really large scale, very quickly, and make predictions based on that.

Now, I'm not a woman, so I can't properly answer your fourth question directly.  What I can say is that I suspect both men and women have to adjust their writing to fit RPG text traditions.  Different types of text require different types of writing.  RPGs are no different in this regard.  Academic writing for publication is famously in line with many typically masculine uses of language.  This does not stop many, many women from successfully becoming professors and publishing articles in international journals.  The fact is, nobody comes out naturally and writes like that.  It takes years and years of practice to learn to so effectively obfuscate what you're really saying :)  So the question, I suppose, is "to what degree do you have to change?"  Do women writers have to change their style more than males to get a typical RPG style?  

And how important is it, anyway?  Is it the main factor that is influencing sales?  Is it even an important factor in sales?

Also, while I really don't know the strengths and weaknesses of the "Gender Genie," it's worth pointing out that the version on that website is based on a simplified version of the algorithm offered by the researchers, and that text type is very, very important to its function, as is the size of the sample.  The more words, the more accurate it should be, if it is based on sound data.  Too few words, and you can't expect confident generalizations.  And the type of text is key, too.  In checking your own text, you identified it as "non-fiction," and the program identified it as exhibiting slightly more feminine than masculine language traits.  But I would argue that the language of an internet post, even on a site like the forge, is a lot closer to "blog entry" than it is to "non-fiction."  If you choose blog entry, you get very, very different results for your first post:

Female Score: 238
Male Score: 970

I guess the reason this is important to note is that it supports my point about the importance of text type on shaping the language used.  Adapting to RPG writing is not just a masculine style vs. feminine style issue.  Many of the typical traits of RPG writing may not be particularly masculine, necessarily.  Many of them may be there because of the need to clearly explain game rules.  I would be interested to see if any female game writers have any anecdotal evidece to report about how they've changed their style, but I suspect it'll be hard to parse out what changes were made to make the language more "masculine type" in fitting with masculine language traits in RPG writing, as opposed to what changes had to be made for other reasons.  This whole thing is even more difficult to talk about because we haven't established very clearly just what traits we're talking about when we say "masculine/feminine language."

Rich

[Mark Johnson]

I certainly agree with Ian's point about "gender" usage.  I wanted to maintain some continuity with the article and program hence the quotes around "male" and avoiding the term masculine since it was not one that they used.  When using terms masculine/feminine I definitely would recommend using Ian's breakdown.

Rich, what a thoughtful and a tad overwhelming post.   I just found the papers in question.  I will get back to your points after I have absorbed these, since they do seem to have a lot of direct bearing on your points.

Gender, Genre, and Writing Style in Formal Written Texts

Automatically categorizing written texts by author gender

And a different iteration of the algorithm.

[Sylus Thane]

What techniques can be used or have been used to broaden the appeal and readability of the game?

I have people read what I'm working on as I go along. Not just the people I game with but anyone I can. This way I can get the best possible range of responses, male or female.

Usually what happens goes something like this:

1. I write out a section of the book.

2. I hand it over to my wife to read. Watch for her reaction. Does she just kind of skim through or is she glued to the page?

3. I hand it over to one of my male friends and watch for the same reaction. If I can't watch their reaction as they read, such is the age of internet friends, I wait to see how much they talk about it. What portions keep their attention the most etc.

4. I hand it to someone who does not game and see what their reaction is. Do they find it engaging? Does it perk their interest to the point they may wish to try it out. Until he passed away my dad was really great for this as his wealth of life experience let him throw all sorts of out of the park comments my way. Now I use my sister for this as she is not a gamer and has increasingly good writing skills.

5. I look back over everything I wrote with everyones reactions in mind and compare my own reactions with them. Did I like as much portions they liked? Were there any portions that I truly liked that no one else seemed to notice?

6. I go back through and rewrite the portions that people seemed to dislike and hopefully make them engaging for all the people who read for me.

All of this helps in trying to gauge that feminine masculine line and not be too far to either side. I want anyone to read my book not worrying about whether or not it is aimed at their gender, but instead simply reading to see if it perks their preferences.

Now unless I decide to write a game or supplement for gun toting maniacs I aim for both genders. Of course a gun toting maniac book should sell very well, the girls almost always have the bigger guns. :)

And Gender Genie says:
Female Score: 477
Male Score: 963

Not bad, now if they could just reverse it to see what kind of gender the audience you were writing it for was instead that would be really cool.

Jason

[Valamir]

Heh, what an interesting little algorithm.  I just ran a few samples of my writing through the thing, and they ranged from 2:1 to 7:1 Male to Female.

Given the recent discussions about the "maleness" of the Universalis mechanics I'm forced to pause and say...hmmmmm

[Emily_Dresner]

Just out of curiosity, I pumped in one of my blog entires and it spat out this:

Female Score: 347
Male Score: 758

So either I'm more guy than most guys (possible), or these theories on gender and language is a bunch of bunk.   I tend to lean toward the latter.  

Either your text is smooth, easy to read, full of active tense goodness, and decent verb usage, or your text is garbage.  I wrote for a long long time about gender and gaming, and I know the gaming industry obsesses about the problem, but the fact is: most women don't like gaming, just like they don't like horror, or gore, and they won't like Kill Bill.  _I_ like all those things, so it's a matter of taste, not a matter of "is my writing more male or more female."

[Rich Forest]

Mark, you definitely do your homework.  Now I'll have to read those articles to make sure I don't fall behind when you return to the topic.  I'm still curious about what (or perhaps whether) we can find here that may be useful to RPG writing.  I'll leave that job to you (see how I deftly avoid all the heavy lifting?).

Jason has a point, I think, as well.  Get lots of people to read what you write and give you feedback.  Listen to their feedback, and make further decisions about what changes you want to make (and what changes you don't) based on that feedback.  This is one of the wonderful effects of the indie game design board here at the Forge, of course, and I think doing it here and in real life is one practical approach to widening the appeal of your work.

And Emily, your point about women in gaming is well taken.  Certainly, gaming is associated with a number of other fandoms that are also famously unpopular with women, and I suspect members of those fandoms obsess equally about widening the appeal of their hobby/interest to women.  Frankly, many subcultures and fandoms in general seem to occasionally obsess about widening their appeal, period, and gaming is not unique in this regard.  I'm also a fan of mixed martial arts, for example, and I lurk around MMA boards from time to time, and sure enough, there is regular talk about "widening the appeal of the sport."  So yeah, that's the way it is.  And that's not only the way it is for us--that's the way it is for a lot of little fan groups.  Still, I'm not sure that this means we should just sit back and accept it, necessarily.  I'm not opposed to seeing folks try to figure out ways to widen the appeal of roleplaying, or mixed martial arts, or anything I'm interested in.  Roleplaying is fun.  Of course, I like the idea of more people sharing my interest in it.  Even if it seems unlikely.

I have to disagree with you a bit on your final take on studies of gender and language, though, Emily.  You've presented it as an either/or thing, "So either I'm more guy than most guys (possible), or these theories on gender and language is a bunch of bunk."  I don't think it's necessarily that extreme.  There are other possibilities here as well.  Firstly, the websites are using simplified versions of the algorithm.  Secondly, the study doesn't claim to have nailed down the differences between masculine and feminine usages of English—it only claims an 80% success rate, and it is to be expected that it doesn't work in every case.  One exception does not really prove anything, in and of itself.  The goal of the study is primarily descriptive, regardless, and it is talking about tendencies of language use.  We cannot make 100% predictions of many things in language, but we can describe tendencies.  Keep in mind that there are more and less careful claims being made here, so the source is important.  One of the big names in talking about gender and language is Deborah Tannen, and her work is often criticized by her colleagues as a bit "popular," in the sense that she writes popular books and makes lots of money (so perhaps they're a bit jealous), but also that she tends to draw her conclusions in rather broad strokes in said popular books.  However, it's important to note that this stuff is not just the result of people sitting around thinking about language and writing.  It's the result of studies of text, sometimes large corpora of text, and it's descriptive.  You may be uncomfortable with the idea that men and women in English speaking cultures tend to use language in some different ways, but that doesn't make the work bunk.  That said, I'm always leery of the website/news article versions of these kinds of work because they often simplify and generalize in ways that are unsupportable, so your skepticism is not entirely unfounded.

Hm, I suppose I'm ranging a bit widely from stuff that's applicable to gaming, at this point, so I'll let up and wait for a bit of clarification from Mark on where he wants to head with this.

Rich

[Mark Johnson]

The summaries of the studies that were published online did not horribly misrepresent them.  Basically it said that women tend to write in a style more resembling fiction writing (use of dialogue, personal pronouns, emphasis on relationships) whether they are writing fiction or non-fiction.  Conversely that men's writing tends to more resemble non-fiction (emphasis on classification, description, its).  The authors assert that their readers felt that the women's writing tends to feel more "involved."  

I still think that perhaps we call these two styles "objective" and "subjective" or somesuch and simply observe that (according to this study anyway) that men and women tend to favor one or the other.  On the other hand, you brought up Deborah Tannen.  In researching this topic I found an interesting story which goes somewhat against the survey but might be fruitful for our purposes.

Tannen once had a group of students analyze articles from men's and women's magazines, trying to see if they could guess which articles had appeared in which class of publication. It wasn't hard. In men's magazines, the sentences were always shorter, and the sentences in women's magazines had more ''feeling verbs,'' which would seem to bolster Koppel's findings. But here's the catch: The actual identity of the author didn't matter. When women wrote for men's magazines, they wrote in the ''male'' style. ''It clearly was performance,'' Tannen notes. ''It didn't matter whether the author was male or female. What mattered was whether the intended audience was male or female.''

This still doesn't address the question in your first post of course of what effect this exactly has on the ability of a text to communicate to a reader.  If we are (subconsciously) adjusting our prose to assume a male reader might it not (subconsciously) repel the female reader?  I am not aware of any research directly adressing that.

At this point, the question becomes, what can an RPG creator do if this is an issue:

-deliberately write "male", even though the text has the risk of being uninvolving and impersonal.

-deliberately write "female" or hire a female writer

-attempt to write in an "androgynous" style (with the risk of turning off people who like a more informative style and those who prefer a more personal style)

-tailor each section according to its needs.  Perhaps the tutorial and advice sections have a more personal approach, while the rules sections are written in a dryer "objective" style.  I think that many games actually do take this approach somewhat.  Even in D&D, the advice oriented DMG has tended to have a more conversational tone than the rules oriented Player's Guide.  Fantasy Wargaming by Bruce Galloway perhaps represents the epitome (or nadir) of this approach with over a hundred pages of personal anecodtes, book recommendations, funny game happenings, reflections on medieval life, etc.  before word one about the rules.

-Sylus' suggestion of having a variety of people "test drive" different sections of the text.  Undoubtedly, this is a sound suggestion regardless of your motivations.  It is astounding that everyone doesn't work this way.

My question is do the above really work or is there a better way to address this problem?  Or is it time to put the genie back in the bottle?

[jrs]

Mark,

I am having difficulty with the presumption that a feminine writing style is necessarily more appealing to women.  My understanding of the Gender Genie exercise is that it attempts to describe a style of writing and subsequently the gender of the writer, but says nothing about the preferences of the reader (Rich made this point above).  Stating that a masculine writing style might "(subconsciously) repel the female reader" is a stretch.  

To illustrate my point about preferences,  I conducted my own little experiment.  I selected a game that on my first reading elicited the reaction, "Must play now!"  Since I have a play-test version on my computer, I dumped it into the Gender Genie algorithm.  It's described as male writing.  (No surprise, the author is male.)  Honestly, I can't say that the writing style had an effect on my interest in the game--it could have been the mechanics or the setting or the fact that I knew someone else interested in playing.  I can say that the writing style was not detrimental to my interest.  

I also happen to think that an objective, non-fiction writing style, which the researchers describe as indicative of male writing, is more condusive to describing how to play a game.

Julie    

Words: 204
(NOTE: The genie works best on texts of more than 500 words.)

Female Score: 192
Male Score: 538

The Gender Genie thinks the author of this passage is: male!

And the Gender Genie is wrong.

[AnyaTheBlue]

I'll just second Julie's comments.

I ran some of my posts here on the Forge through the Gender Genie.  If I said it was nonfiction, the GG identified me as Female.  If I said it was a blog post, it identified me as Male.  Um.

I think all it's doing is showing up a trend.  It doesn't explain where that trend comes from.  Cultural differences?  Socialization differences?  Biological differences?  Socio-economic differences?  You could probably find similar trends in people (of both sexes) taken from New York vs. people taken from L.A.

Trends are interesting, but they say nothing about individuals.  For every trend, or 'majority statistic' there will be actual individuals who by definition do not fall into the majority.  Or, to put it another way, yeah, you can maybe find something out about the aggregate behavior of women from something like this (women in US society, anyway).  But it doesn't necessarily tell you anything particular about the behavior of, oh, Meg Ryan or Laura Bush.  As individuals, they may or may not be in your 'majority'.  So, really, how useful is this information?

[Meredith]

My sense of the frustration with Gender Genie is its inherent fallibility as a quantitatively-based survey measure to predict individual responses.  It feels a little like the frustration qualitative researchers have with quantitative research.

So what sorts of qualitative research have been done on game writing?  I have to admit I did not go read those two gender articles - are they qualitative?  And can the Forge serve as a qualitative research tool?  Perhaps a section devoted to in-depth exploration of individuals' experiences - not looking for theoretical predicability (as often happens in the Actual Play forums), but for general understanding?

[Valamir]

Interestingly I find the genie's precision to be pretty laudable.  Assuming the self reported numbers can be trusted, a 75% success rate of an abbreviated version of an algorithm that's only expected to be 80% right is pretty darn good.  If you could guess the stock market right 75% of the time you'd be a very rich person.

But the power of the algorithm is not so much in its predictive qualities (who really cares if you can guess the authors gender).  But rather in highlighting differences in word useage patterns.  

There are only really two coefficients being tested, gender and fiction vs non vs blog.  One would have to really know how the algorithm was derived to identify common fallacies caused by data mining and other such things.  

But it seems to me that a certain group of people use a certain collection of words more often than another group, and this can be demonstrated to be fairly consistant over time, that there's something interesting going on there.  After all, if people write differently, they likely communicate differently.  And if they communicate differently then the thought processes behind that communcation are likely different, and if those differences are consistantly identifiable then it would seem to me to offer some interesting insight into the way people think.

How is that relevant to RPGs...probably not very given the whole thing is still too crude to draw any sort of "if I write like this my game will be more accessible to women" conclusions from.  But it does highlight a basic fact that word choices are significant.

[AnyaTheBlue]

Ralph,

How is that relevant to RPGs...probably not very given the whole thing is still too crude to draw any sort of "if I write like this my game will be more accessible to women" conclusions from.  But it does highlight a basic fact that word choices are significant.

These are excellent points.  However, the problem I have is that it doesn't tell you why those word choices were made.  The difference in word choice between the groups is important, but we don't know what difference between the groups might be the actual cause of the difference.  Is it because one group is biologically male and one is biologically female?  Is it because one group has more engineers than the other?  Is it because one group has more parents than the other?  Is it because one group has more poor people than the other?  We don't know.  Yes, biological sex is one of the observable differences between the populations, and whatever differences are present is about 80% correlated with biological sex.  But that doesn't mean that the biological sex is the cause of the word choice differences.  It just means both might be caused by some related other thing that differs between the groups.

I'm not sure I'm explaining this well.  Am I making sense?