SciTechBlog   « Back to Blog Main
September 21, 2009

Program predicts your sexuality based on your Facebook friends

Posted: 09:21 AM ET

Two Massachusetts Institute of Technology students have developed a program they claim can accurately predict sexual orientation based on a person's Facebook friends.

"Project Gaydar" scanned the profiles of more than 1,500 Facebook users who identified themselves as gay, straight, or bisexual. Analysis revealed gay men had a higher proportion of homosexual friends than straight men.

The discovery that gay men have gay friends isn't a surprise. But, using this information, the Gaydar program was applied to 947 men who chose not to identify sexual orientation in their profiles. While the students could not confirm the accuracy of all 947 predictions, they personally knew 10 of the men were homosexuals and the program identified each of these men as gay.

Facebook spokesman Simon Axten addressed the study to the Boston Globe:

In general, it’s not too surprising that someone might make inferences about someone else without knowing that person based on who the person’s friends are. This isn’t specific to Facebook and is entirely possible in the real world as well.

However, the study highlights the power of data-mining social networks. Statistical analysis that would be impossible in the "real world" can reveal more than we choose to disclose in our online profiles.

Do you worry about the loss of privacy online? Or are you happy to reveal a little personal info if that's what it takes to keep using Facebook?

Posted by:
Filed under: Facebook • Internet


Share this on:
Mark   September 21st, 2009 10:52 am ET

10 out of 947 were confirmed as homosexual because someone personally knew them? And this is being touted as legitamate research?


Patrick   September 21st, 2009 11:08 am ET

I agree, this is ridiculous.


boddhu   September 21st, 2009 11:12 am ET

CNN buys everything🙂


Charles   September 21st, 2009 11:19 am ET

10 out of 10, or 100% accuracy seems rather statisticly significant to me.

Maybe you guys are a bit nervous?


Mike D   September 21st, 2009 11:48 am ET

Cant we do something more useful like, predict who the CRIMINALS are?


alexandra lara   September 21st, 2009 11:55 am ET

ohh wow part of this article is true i guess but i don't think it maters who there friends are. but watever:]♥


Dustin   September 21st, 2009 11:55 am ET

"Project Gaydar" are you kidding me? And they say this is "research"... Sounds like a bunch of homophobes to me.


guy in KC   September 21st, 2009 11:55 am ET

You obviously didn't read correctly, they personally knew 10 were gay and the program correctly guessed their sexual orientation. They never said anything about the other guesses. It very may well have predicted correctly but the confirmation isn't possible unless the subjects confirm. I think CNN gets a bad rap because of the low intelligence of most of its viewers/readers. I agree that the project is ridiculous, but it could be a start of something important. The internet was once considered ridiculous and look at it now.


Dave   September 21st, 2009 12:36 pm ET

What is missing is a statement about predictions that someone is straight. The program could have predicted all 947 are gay and the statement that the 10 they personally knew were gay would have been equally confirmed. If they also had said that the program predicted 20 people they personally knew were straight and had been correct there as well, then I think the result would have been more meaningful.


Tanner   September 21st, 2009 12:54 pm ET

Umm.. Wow. I have at least 12 gay friends I wonder If I'm automatically gay now?

If this same thing was applied to race people would be shouting from the hills how racist this was.


Anon   September 21st, 2009 12:56 pm ET

Project Gaydar is the best project ever, and I hope they make Project Retard next. Anyone who is friends with Glenn Beck will immediately be flagged.


Andrew   September 21st, 2009 1:34 pm ET

Interesting that they can make an algorithm to guess that.

BTW some peoples reading comprehension is fail, just ignore them.


Mark W.   September 21st, 2009 1:47 pm ET

What's next, project gullible?


Scott   September 21st, 2009 1:56 pm ET

You needed a "study" and a PROGRAM for this? Let's see the statistical significance of more difficult to detect characteristics. . .I mean, I can , looking at profiles, give you, within a pretty good confidence interval, which "straight" men are likely in the closet, just by looking through their profiles–as Eve Sedgwick wrote–it's the secret that always gives itself away.


Vinu   September 21st, 2009 2:03 pm ET

Hope they will put their brains on more useful things in life.


Jeff22153   September 21st, 2009 2:16 pm ET

Umm.... wouldn't it be easier to read the "Looking For" and/or "Relationship Status" lines? I mean, c'mon, who needs a program to tell you that??


Justin   September 21st, 2009 2:18 pm ET

Clearly, this is indisputable proof of a person's sexual preference.


Josh   September 21st, 2009 2:35 pm ET

...and what are they trying to determine? That your guilty just by degrees of seperation? I smell a witch hunt.


Steven   September 21st, 2009 2:49 pm ET

I know 10 cat. What does that make me?


Rigell Leonheart   September 21st, 2009 2:56 pm ET

Go MIT! Shouldn't they be working to solve the world energy crisis rather than determining sexual orientation on facebook?


DrMatrix   September 21st, 2009 3:30 pm ET

Some of my best friends . . .


Kareem   September 21st, 2009 4:16 pm ET

So, how much money was wasted on this meaningless project in this recession? hmmm...


Melissa   September 21st, 2009 4:16 pm ET

I'd like to see this application develop even further and allow me to predict which of my friends are black.


Alberto   September 21st, 2009 4:19 pm ET

i have facebook and most of my friends i do not know them, very few have i send a request, the ones who send a request to me i accept most of the time.


Angelikk   September 21st, 2009 4:24 pm ET

Wow this isn't research, it's stupidity. Who in there right mind cares who anyone else is attracted to on facebook? Creating a program that could relieve world hunger or eliminate poverty would have been time better spent!


Whitney   September 21st, 2009 4:25 pm ET

I have gay or bi friends and family as friends and i'm straight. This kind of program isn't creating tolerance of people with different preferences, but it isn't new people do the exact same thing when they see someone walking down the street based the way a person talks, walks, and dresses. These kinds of assumptions are made if a girl is a female athlete on a collegiate team and so on. They are just that assumptions. Sometimes these assumptions are indeed correct but not always. If u wanna kno if someone is gay ask them if they say no then they aren't if they say yeah then they are. Plain and simple. I'm glad CNN reported the article; it's interesting although I wished the privacy point of view was emphasized more. The study wasn't science and it doesn't claim to be so there's no problem.


zhirow   September 21st, 2009 4:26 pm ET

Amazingly, out of all the currently posted comments 14 of them miss what the actual impact of the article is. It has nothing to do with who's gay or not. Instead of whining about the subject matter, try to actually see the articles point. It's not about who is gay and who isn't or the lame name of the project.

"However, the study highlights the power of data-mining social networks. Statistical analysis that would be impossible in the “real world” can reveal more than we choose to disclose in our online profiles.

Do you worry about the loss of privacy online? Or are you happy to reveal a little personal info if that’s what it takes to keep using Facebook?"

The article is about the power of data mining, the rest is just for the masses who troll CNNs forums, to which many of the previous comments are obviously for.

" ...using this information, the Gaydar program was applied to 947 men who chose not to identify sexual orientation..."

"... they personally knew 10 of the men were homosexuals and the program identified each of these men as gay..."

Thats the relevant part of the article, using who you know and abstract information from both you and your friends creates a profile that accurately determines things about you such as but not limited to sexual orientation. You should be concerned with your privacy and habits being mined more then trolling and sounding quite frankly, stupid.


R...   September 21st, 2009 4:28 pm ET

No Kidding,,,I expect more from CNN. Yes, if you look at a someone's friends you can tell they're gay. I'm gay and I know that,,,my friends both str8 and gay talk to one another too...

hardly a news item.


Steven   September 21st, 2009 4:35 pm ET

They could have just asked nicely instead of wasting their time and money on such socially inept project. I suppose calculated guessing is better than the facts. Just observe all the reading material in the check out lanes at your local grocery.


michael   September 21st, 2009 4:59 pm ET

i have a large amount of homosexual friends and am straight, i predict that this "program" would inaccurately predict that i am also homosexual


flirtsquirt   September 21st, 2009 5:20 pm ET

i don't need data mining. just looks at male friends pics, most the the gay guys are showing off their abs.


Chris   September 21st, 2009 5:45 pm ET

Ha... What about people like me that have hundreds of friends that they don't know in real life because they added friends to play Mafia and Vampire Wars? Out of my 600 friends, I only know 110... will the program know that?


Robert   September 21st, 2009 5:53 pm ET

Statistically insignificant. You need a sample size larger than 10.


R. Noonan   September 21st, 2009 6:22 pm ET

I think all of you missed the real point which is that data mining can determine more about you than you choose to disclose based on the e-footprint left by everyone who uses the internet.

While most of us think surfing the web is anonymous, the fact of the matter is that every connection is logged and whether you decide to lend your name to an page or not, your machine address is provided with every transaction.

This can be mined and an interconnecting profile created based on ALL of the web surfing you do not just where you think you are disclosing your identity. That's all news, information, social network or shopping sites you have simply touched.

Anyone with a Gmail account should assume that their mail is being read regularly by the NSA based on previous Echelon protocols, only far more advanced and pervasive. Ever wonder why the FBI gave up on Carnivore as a tool... its because they found a better way to get at the big picture without you knowing.


muzz   September 21st, 2009 6:34 pm ET

Some people just have to much spare time on their hands. Who cares about what other peolpe do and who their friends are, at least they have friends. I suggest thses analysts look at themselves and find something else to do.


R. Noonan   September 21st, 2009 6:37 pm ET

Sorry, all of you missed the point but the amazing Zhirow.

Peace to all the squabblers who don't get it, or are afraid of being outed by Perez Hilton using Gaydar.


fred   September 21st, 2009 6:41 pm ET

Why does someone take the time to develop such a program?

How about using that kind effort to track the course of the H1N1 virus or figure out how many more troops we need to finish the job in Iraq and Afghanistan?

Some people need to get a life! Of course I'm the last person on this planet who doesn't have a Facebook account (and never will) so who really cares!


Nola   September 21st, 2009 6:49 pm ET

I think this is actually a pretty important case. It shows how far people will go to try and find out more about people without those people knowing about it. Lets say you are an employer who is looking for specific dirt on potential hires prior to hiring them... for instance, if they party a lot, or if they use drugs or not. The same type of program could find this bogus info out and severely limit your opportunities to find a job. Facebook, myspace, etc. are great means to find out dirt on people, and I am 100% against sites like that.


Pikalos   September 21st, 2009 7:06 pm ET

I personally think it's a great finding just wish their energy was channeled towards something more contructive or entertaining like most likely to be voted 'man of the year' etc. Facebook does not need to get branded for 'gaydaring' it's members.


Chuck   September 21st, 2009 7:18 pm ET

This is ridiculous! The only thing you can tell from MY Facebook friends is that I'm a fan of Cinematic Titanic. (Frank Coniff and Mary Jo Pehl are friends. How gay is that?)


Christopher   September 21st, 2009 7:28 pm ET

This is homophobic and very insulting.


David   September 21st, 2009 7:36 pm ET

Shame on you CNN. What rubbish.


Colin   September 21st, 2009 8:20 pm ET

They're MIT students, but is this project actually related to the school in any way? If this was done for course credit (I can't imagine it's anything to do with graduate research), was it for sociology, computer science – what? Whether or not this was a waste of time and money depends on what sort of funding it received, and what skills it was meant to develop.

On a side note, this is a pretty poor example of what's actually possible with statistical analysis. The tricky thing is to get reliable predictions from data that are not so obviously linked to the quality being predicted.


Jerry of Madison, WI   September 21st, 2009 8:35 pm ET

This research was pointless. People tend to attract friends that are similar to themselves. Heterosexuals will have more heterosexual friends than gay friends. Homosexuals will still have a lot of heterosexual friends, but will have a higher percentage of homosexual friends... think they need a refund on their education tuitions. Duh! Someone who's gay will have more gay friends. They had to go to MIT to figure this out??


Jennifer   September 21st, 2009 10:31 pm ET

They needed MIT students to do this? Why didn't they just yell out their window, "Hey, do gay guys or straight guys know more gay guys?" And how do they prove their findings without documenting ongoing sexual histories of the subjects?


CNN Sucks   September 21st, 2009 11:10 pm ET

can we spend more time focusing on IMPORTANT things and not usless info like this..

this (and programs like Nancy Grace) is why CNN should be ashamed to put themselves down as a credible source of news and information


zach   September 22nd, 2009 7:01 am ET

This borders on the offensive...


cathy smithson   September 22nd, 2009 8:37 am ET

This is not science, this is not even psesudoscience. No wonder the american people do not understand science, or statistics, etc. when this is the kind of science that is reported on. I would think at least MIT students would not publish a study with such a ridiculous experimental design and interpretation, and letting their personal knowledge interfere with conclusions of the study. And I had presumed MIT were our best and brightest. Oh well looks not to be the case.


Gary   September 22nd, 2009 8:57 am ET

Hahaha excellent.

Looks like Michael and Dwight are that much closer to purchasing a gaydar for the office eh?


Betty Holland   September 22nd, 2009 11:33 am ET

You needed a survey to figure that out? Helllloooo....this idea must have been one of a group of straight men, in their 40's or older, who have had no contact with gay or bisexual individuals.


ChrisAmericaUSA   September 22nd, 2009 3:25 pm ET

This is the most stupidest thing I have ever read! To waste time wondering "who" is gay! Does knowing someone's sexual preference seriously impacts one's personal awareness. This is what I have to say "WHO CARES WHO IS GAY? I DONT!"


The American Wanker   September 22nd, 2009 4:40 pm ET

All,

While views on what this particular project from “two… students” of MIT are scattered. Last year over 6,000 students graduated from MIT. Many of you are clearly missing the bigger picture. They are working on an algorithm that is sifting through a few particular fields in the social sphere. One day soon, systems will sift through myriads of data and provide substantial information on human habits, trends and assist in future planning, insight into the day to day lives of people in regions and unfortunately targeting marketing.

Please read zhirow’s comments. I agree completely with that direction.

Why should anyone expect more or less for a news source? CNN is a world wide provide of news. Period. There is nothing deep about this snib-it of news, other then letting those not in MIT (which appears to account from more then 99.9% of Americans know something along the lines of: Hey you, info is being mined, calculated and used).

All in all, read, write your opinion, sooner or later, a databases could be cracked linking your comments on this site to your social site (if you have one) or other personal information and then the algorithm’s created, will be able to determine your education level, political view, and tell the government if you should be allowed to travel, drive a vehicle, or mate (An extreme view, but hey, might as well stir the pot a little. Discussion never hurt anyone).

Best of Luck!


Barron   September 22nd, 2009 5:13 pm ET

So wait, now straight people have to be affraid to be friends with gay people because some rediculous "program" might call them homosexuals as well? Fantastic! Another reason to alienate gays.


Steve   September 22nd, 2009 6:07 pm ET

After reading all of this dribble, I may turn gay too!


Jeff   September 22nd, 2009 6:22 pm ET

You are all missing the point.

The point of this article is not "WE KNOW IF YOU ARE GAY!"

It is about the power of data mining of social networks. Here they claim to be able to infer an undisclosed fact based upon separate freely available knowledge. Something that would, as they say be very hard if not impossible in the real world.


Jeremy   September 22nd, 2009 11:10 pm ET

Jeff and the others who have pointed it out are right. This whole thing is about the idea that a single program could in effect tell peole who they should associate themselves with, what they should do or any number of things effectively controling our lives. This is not a new idea but it is expressed in a new way using technology and will most likely have more studies based off of this one refining the idea to the point where it may be feasible for use by anyone from organizations looking for possible donors to goverment bodies looking for persons or groups.

However in regards to the MIT program's results, just using the simple assumption that gay men are gay by having gay friends and the opposite for straight men is more a fault of the researchers not using enough data. For example, were I on facebook and was included in this data mining I would be listed as a straight man as almost all of my guy friends are straight men even though I personally am not. Figure that.


AES_CIPHER   September 23rd, 2009 12:41 am ET

Its the concept that's important here not the actual application. This is pure data mining and information theory in play. I am sure they are also not interested in people's sexual orientations. What they are interested in is fact that they can use data intelligence to predict with a high certainty a particular scenario. Lots of work have been put to this including web crawling, data profiling, entropy etc to see this a success. I think what you should be worried about is what public information do facebook display to make this possible?


John-Michael   September 23rd, 2009 3:33 am ET

This is so silly. Now I'm straight because all of my FB straights. What's next?


Mike   September 23rd, 2009 7:13 am ET

The main point of this article is about the potential power of data mining social networks, not determining sexuality.


Jeremy   September 24th, 2009 5:29 pm ET

Good thing this is just a blog, not news...hope people reading this realize that.


ThisIsStupid   September 24th, 2009 5:55 pm ET

I have 87 "friends" on facebook. 70 of them are either business entrepreneur connections, those who share an interest in positive thinking and the Law of Attraction, and those who are members of my facebook game 'families" (you have to "friend" someone in order for them to be part of your army, mafia, etc.). I fail to see how this data-mining logarithm can factor this in and give accurate data.


Uhm Bob   September 25th, 2009 5:54 am ET

Maybe the two MIT students were cruising for dates? How else are computer nerds supposed to get laid?

Just a joke folks. In all reality, it is spooky that these algorithms can be developed so that YOU can be categorized. With categorization comes differentiation, and ultimately, discrimination. You could be subjected to the next round of government surveillance just because you like combat boots and read certain magazines and are therefore more likely to be a national security risk. The internet, and the data that is out there, on you and your habits, even though you are unaware that it is being mined, is scary.


richy   September 30th, 2009 11:11 am ET

I think it's a bit of a sorry app to be honest. I would imagine that it's pretty obvious statistically that on the whole, the more gay friends you have , the more likely you are to be gay.

You don't need to be a genius to guess that would be the case.

cheers
rp
http://www.wonderwebmarketing.com


joe   October 16th, 2009 2:50 pm ET

The story is about data mining.... what is wrong with you people? You see the word "gay" and your IQ drops to 60.

The purpose of the story is to tell you that even though you don't specifically state something, the data you put online can be analyzed by an algorithm that can accurately predict your preferences. This could be used to market things directly to you based on things you like or that the algorithm predicts you like. It could also be used to predict behavior among many other things.


DillyRat   October 19th, 2009 6:02 am ET

I'm not surprised so many people "missed the point." What surprises me is how they dress themselves in the morning...

Haha... I like joe -> "what is wrong with you people? You see the word "gay" and your IQ drops to 60."


Fredilly   October 19th, 2009 5:51 pm ET

It always starts with two M.I.T students. This is definitely no "GOOGLE"....wasting valuable research on Gaydar?


His_wife99   October 22nd, 2009 7:11 am ET

These are not really doctors, but more like medics or PAs, or NAs. ,


Tom   December 5th, 2010 4:54 pm ET

Totally agree with Rich, it doesn't take a rocket scientist to realise that statistically that on the whole, the more gay friends you have , the more likely you are to be gay. It like saying the more single girls you have as friends, the more likely you are to be straight.
Internet Marketing Expert


iPhone ??? DIY   October 14th, 2014 9:05 pm ET

Give a warm name for every river and every mountain.
iPhone ??? DIY http://shopping.geocities.jp/icloudjp/index.html


Leave Your Comment


 

CNN welcomes a lively and courteous discussion as long as you follow the Rules of Conduct set forth in our Terms of Service. Comments are not pre-screened before they post. You agree that anything you post may be used, along with your name and profile picture, in accordance with our Privacy Policy and the license you have granted pursuant to our Terms of Service.


subscribe RSS Icon
About this blog

Are you a gadgethead? Do you spend hours a day online? Or are you just curious about how technology impacts your life? In this digital age, it's increasingly important to be fluent, or at least familiar, with the big tech trends. From gadgets to Google, smartphones to social media, this blog will help keep you informed.

subscribe RSS Icon
twitter
Powered by WordPress.com VIP