A Masculine-Sounding Woman
I want to loop back to something that happened about a week ago. The twins and I did Note to Self’s Privacy Paradox together. It was pretty good. I wish they had spent more time covering the pros and cons of some of the suggestions — it’s fine to point out an add-on, but the listener should also know how that add-on may affect their Internet experience — but, overall, it was a good reminder of how we are tracked every single second of the day.
One of the activities included a site where you could plug in a sample of your writing and it would analyze you in the same way trackers analyze and categorize you as you surf the web. Not every “fact” that data gathererers like Facebook have about you are correct. Some of them have been inferred by what you like or how long you spend on a site or how you word things.
This site, Apply Magic Sauce, states that it “accurately predicts psychological traits from digital footprints of human behaviour.” I used the cut-and-paste feature to feed it a bunch of random blog posts. The magic sauce is fairly certain that I sound like I’m in my mid-twenties and my psychological gender is masculine.
I’m more liberal (yes), more impulsive than organized (no), more contemplative (yes), more competitive (no), and more laid back (no).
Though it gave a deeper description that felt a little more accurate later on the page.
Lastly, it told me my leadership potential.
I’ve never taken that personality test, so I’m not sure whether I actually am an INTP, though it sounded pretty accurate reading the description.
What made me so amused was that each post I fed into the site came back with a high probability of being male. Other facets of my personality changed. I wasn’t always an INTP. I wasn’t always in my twenties. But I always male. The lowest rating I got was 75% (“your probability of being male is 75%”), but most of the time it was in the high 90s. Even my Valentine’s Day post yielded a 94% chance of identifying as male.
It made me wonder how I am categorized by various sites. Do they know I’m a woman? Do they know I’m introverted? That I prefer solids to prints, paint to wallpaper, indoors to outdoors? Do they know I always eat yogurt for breakfast? That I care about environmental issues?
By the way, this post stating above that I am a woman still yielded an 82% chance of being male. And it tells me I sound like I’m 22 years old. But I’m still an INTP, so there’s that.
Go feed one of your posts into the site and tell me how it categorized you.
10 comments
I think that most women who read a lot, especially certain genres, will come out being ‘male’ when their writing is sampled. It is because the majority of the books we read are written by men, and it effects your ‘voice’. I also come out as male, and I am often confused as a male on the internet. I wonder why you are considered in your 20’s? You do sound like a hip chicke 🙂
OMG, everything else about it is so uncannily like me that I am almost convinced I am 97% male! It got everything– age, Myers-Briggs, etc.
I assume part of why I’m coding as male is because I make an effort to get rid of qualifiers in my writing. Strunk and White will always code male. In my speaking I still use a lot of those softeners (a lot compared to a male economist, anyway) because I have to.
Didn’t go try it yet, but I will say that Twitter promo tweets keep trying to fix me up with a man.
Hmm. Interesting.
I tried it with my latest post and it said “79% male”. Weird weird weird. I think its BS. The rest of it was pretty spot on: neuroticism, hard-working & organized, liberal/
OK I just put in the next to the last post and: 84% female. So…yeah. I guess you need to put more data in to get a more accurate prediction.
I wrote a short story from the viewpoint of a 15-year-old girl, so I decided to subject it to the test. Turns out she talks like she’s 31 (bugger) but at least her gender came out right (although only slightly).
Interesting!! I fed a number of posts in and it flip-flops between male and female, but I’m always in my 30s (true). My last post (where I announced pregnancy) still coded as male, so clearly content is not relevant…
I came out totally female on one post, and “female but not afraid to show masculine side” on another. I was 31 for the mostly female post and 26 for the middle of the road one, which was interesting. Both came out INTP, but I wonder if because they are blog posts I am self-reflecting all over the place and so that’s one reason? I’ve never done the Meyer-Briggs in its entirety, but those real reliable facebook ones put me in similar boats, although usually I’m Judging. I am apparently easily stressed and emotional. Also contemplative, liberal & artistic, competitive, and impulsive & spontaneous (which is hilarious since I need a plan and last minute surprises stress me out). What I find interesting is how they determine gender. What makes writing more “male” or more “female?” It troubles me a bit to think on that one. Also that in the wake of others being proclaimed male, there’s a tiny part of me that feels like being told I’m female means my writing is overly emotional, with too many extended sentences, interjections, and qualifiers. I say I’m proud of my femaleness! 🙂 Definitely an interesting thought experiment.
I’m glad you did this because I’ve been meaning to test out Apply Magic Sauce (I first saw this in conjunction with how the election was won using it) but first I’d need to investigate its privacy requirements.
I tested with 2 blog posts. I am a 37 year old male or a 27 year old female. I am INTP and ISTJ, neither of which are what I think I actually am.
And after I took the test, I cleared my cookies and chanced my browser’s cookie policy.