P is for Privacy

August 20, 2010 7:48 AM

MetaFilter users are
    55% male
    68% 18-49
    83% Caucasian
    54% >$60k/yr

How do I know? Perhaps a little zombie told me.
posted by DU (92 comments total)

Well I am 83% caucasian so I have to agree.
posted by KevinSkomsvold at 7:52 AM on August 20, 2010 [6 favorites]

And I do make 54% of $60K/yr.
posted by (Arsenio) Hall and (Warren) Oates at 7:53 AM on August 20, 2010 [49 favorites]

I'm 100 percent zombie.
posted by Astro Zombie at 7:54 AM on August 20, 2010 [24 favorites]

This site attracts a less affluent audience.

There is a high index of Graduates and Post Graduates here.
posted by furiousxgeorge at 7:54 AM on August 20, 2010 [15 favorites]

I like my metafilter users like I like my women.

55% male
68% 18-49
83% Caucasian
54% >$60k/yr

Awwww yeah.
posted by slimepuppy at 7:55 AM on August 20, 2010 [30 favorites]

Nice, but can they accurately guess how many gumballs are in the MeFi jar?
posted by Slack-a-gogo at 7:56 AM on August 20, 2010

You think you know me. But you don't know me!
posted by Joe Beese at 7:56 AM on August 20, 2010

This site attracts an affluent audience.
posted by furiousxgeorge at 7:57 AM on August 20, 2010 [1 favorite]

I am not a number. I am a free man!
posted by _Lasar at 7:58 AM on August 20, 2010 [3 favorites]

NYC unemployment insurance website has some interesting stats.
posted by griphus at 7:59 AM on August 20, 2010 [2 favorites]

54% >$60k/yr

I'm really dragging down the average, here.

55% male

Being in drag will down the average here.
posted by shakespeherian at 8:00 AM on August 20, 2010 [19 favorites]

Oh dear, wrong place! Now I am dreadfully embarrassed. I only came because I thought it said effluvient audience.

*drips away, leaving only stenches and stains behind*
posted by adipocere at 8:00 AM on August 20, 2010 [10 favorites]

Well I am just out of luck almost all the way around........
posted by bjgeiger at 8:01 AM on August 20, 2010

There's probably something I'm not understanding about what the graph represents, but what the heck happened in the first part of this month?

Did we all simultaneously discover that there is an "outside"?
posted by dry white toast at 8:01 AM on August 20, 2010 [2 favorites]

I think you're conflating the 'zombie cookies' issue (which was to do with shady use of the Flash Local Shared Object) with data harvested by 'conventional' means. QuantCast hasn't been using the 'zombie cookies' technique for a year now.

But yes, install something like BetterPrivacy if you're at all concerned.
posted by le morte de bea arthur at 8:01 AM on August 20, 2010

<1% Addicts

What else do I have to do?
posted by yeti at 8:01 AM on August 20, 2010 [15 favorites]

That's weird, dry white toast, the "page views per person" graph is the exact inverse!
posted by griphus at 8:03 AM on August 20, 2010

I installed BetterPrivacy, but it doesn't seem to be much better. It just deletes my cookies periodically (timer or when I shut down or when I press a button or whatever). Wouldn't a better solution be to

1) block these tiny flash apps
1a) for sites like MeFi to not embed the app in the first place (perhaps they aren't and it's only correlated with everything else? but why does MeFi have a flash cookie?)
2) only serve cookies back up to them what baked up
posted by DU at 8:05 AM on August 20, 2010

Actually Astro Zombie, you're only 6/11 Zombie. The rest of you is out of this world.
posted by sciurus at 8:06 AM on August 20, 2010

The BetterPrivacy addon for Firefox (referenced in the article) manages these cookies and will let you delete them, automatically or by hand.

Flashblock is also handy, both for disabling the more annoying flash elements (i.e. most of them) and for preventing the hidden ones from running. (And, since it replaces flash elements with an "F" icon, you can see the hidden elements.)
posted by suetanvil at 8:08 AM on August 20, 2010 [5 favorites]

I want to know who the 1% of users in age 3-12 bracket are.

Although it might be refreshing to read askmes about crayons and hopscotch.
posted by ricochet biscuit at 8:08 AM on August 20, 2010 [11 favorites]

83% Caucasian

40 gigs online server space and a laptop, we were promised!!
posted by nomadicink at 8:10 AM on August 20, 2010 [4 favorites]

Actually Astro Zombie, you're only 6/11 Zombie. The rest of you is out of this world.

This site is really cramping my armchair statistician hobby.
posted by dry white toast at 8:13 AM on August 20, 2010

I give all my zombie friends zombie cookies so they won't eat my brain.
posted by ZenMasterThis at 8:13 AM on August 20, 2010 [2 favorites]

That's good enough for me!
posted by swift at 8:13 AM on August 20, 2010

40 gigs online server space and a laptop, we were promised!!

And some of us may even be part Yoda.
posted by divide_by_cucumber at 8:15 AM on August 20, 2010

adipocere: *drips away, leaving only stenches and stains behind*

May I interest you in a room at the Suppurate Motel?
posted by Gin and Comics at 8:15 AM on August 20, 2010 [5 favorites]

I am not a number. I am a free man!

Says you, #94866.
posted by Faint of Butt at 8:18 AM on August 20, 2010 [5 favorites]

It seems that there is an essential aspect of the way the web works that's broken with regard to privacy. Specifically, sites are much less likely to only serve pages from their own servers any more. Resources are frequently pulled from servers owned by marketing companies like Google, Yahoo, etc., and it's no surprise that these requests are being culled for demographic information. Even some of the simplest pages I've written often pull resources from Google APIs that undermine user privacy significantly.

Mozilla has done some excellent work on privacy in their browser so far, but I'm not sure I see a technical solution to the issue of third-party resource requests giving away demographic information, aside from legislation, which would be hit-or-miss.
posted by odinsdream at 8:19 AM on August 20, 2010

I've been accurately stereotyped ALL MY LIFE.
posted by nevercalm at 8:21 AM on August 20, 2010

What the hell happened to Metafilter's traffic in August? It looks like it dropped 90%, then bounced back.

I think there might be something a little wrong with their software.

---

Anyway, Quantcast and sites like it (except Alexa, as far as I know) actually get their data directly from ISPs. So when your browser makes an HTTP request to metafilter.com, that (depending on the ISP) actually gets snooped and sent to these traffic detail sites. So, no matter how high you set your privacy settings, there is actually nothing you can do -- unless the site admin turns on HTTPS.

With HTTPS turned on, you can sometimes guess which site someone is visiting using the IP address, but that doesn't work very well. Reverse IP lookups might not return the sitename. And multiple hosted sites often share IPs.

So, it doesn't really matter that much if your computer leaks data as far as Quantcast is concerned.

I wonder how much of this kind of stuff Addblock actually blocks.
posted by delmoi at 8:22 AM on August 20, 2010

Anyway, Quantcast and sites like it (except Alexa, as far as I know) actually get their data directly from ISPs.

Quantcast uses cookies.
posted by DU at 8:24 AM on August 20, 2010

Shady data collection willies aside, the data is pretty interesting. Are we really 71% child-free? Given our age distribution, that number seems significantly different than the population as a whole.

I have no kids, and no intention of spawning any. I guess on the filter, I'm the norm.
posted by overhauser at 8:29 AM on August 20, 2010

Actually Astro Zombie, you're only 6/11 Zombie. The rest of you is out of this world.

This site is really cramping my armchair statistician hobby.

Whoops...that wasn't the comment I meant to reply to at all.
That's weird, dry white toast, the "page views per person" graph is the exact inverse!
posted by dry white toast at 8:32 AM on August 20, 2010

Nice, but can they accurately guess how many gumballs are in the MeFi jar?

55% male * 2 = 1.1
posted by aaronetc at 8:33 AM on August 20, 2010 [2 favorites]

You know who else had ~1 ball?
posted by DU at 8:35 AM on August 20, 2010 [4 favorites]

slimepuppy: "I like my metafilter users like I like my women.
55% male"

How progressive of you!
posted by charred husk at 8:36 AM on August 20, 2010 [1 favorite]

And I used to think I was special and unique. Oh well...

ONE OF US!
ONE OF US!
ONE OF US!
ONE OF US!
posted by WinnipegDragon at 8:37 AM on August 20, 2010 [2 favorites]

I like to combine the swear jar with the gumball jar-- how many nickels are in the swear jar? PROTIP: Guess low, because if you're wrong, you can quickly adjust the number to match your guess.
posted by shakespeherian at 8:37 AM on August 20, 2010 [2 favorites]

It's interesting that the demographics change by country. The UK and India skew younger and more male than the US. Canada also skews younger, but is 50/50 on gender. The Netherlands is a teenage boyzone.
posted by desjardins at 8:55 AM on August 20, 2010

We were just talking about this sort of stuff - where I noted

that [the] FLASH Control Panel and allows you to change how flash behaves. You can tell it not to allow 3rd party LSO*'s, you can set the amount allowed to zero globally (second tab) or you can do it on a site by site basis (6th tab).

Adobe keeps track of the browser/LSO history in a number of locations- check wikipedia for the details/list related to LSO files on a system.

*These flash cookies are properly called LSO (Local Shared Object) and aren't cool, especially the shared part.
posted by zenon at 9:00 AM on August 20, 2010 [2 favorites]

delmoi:With HTTPS turned on, .... And multiple hosted sites often share IPs.

Those two are mutex (unless you're ok with using one certificate for all virtual servers on that IP).

When establishing an SSL connection the webserver has to use a certificate to start the crypto handshake, but it cannot know which virtual server's certificate to use. All the webserver has is an IP for the request. The server doesn't know the host name requested (and which virtual server to delegate to) until after the SSL connection is established and the "Host" parameter to the HTTP/1.1 header is sent. Apache faq.

So privacy.. yeah good luck with that.
posted by ecco at 9:01 AM on August 20, 2010 [1 favorite]

Looking at the graph, what happened in late July/early August?
posted by madcaptenor at 9:04 AM on August 20, 2010

When I refresh the MeFi page this thread shows the word "are" right after "MetaFilter users" and before:
"55% male
68% 18-49
83% Caucasian
54% >$60k/yr"

It then disappears. What is causing this, is it the zombie cookie?
posted by coolxcool=rad at 9:05 AM on August 20, 2010

Some of the cities data looks broken. More unique visits from Louth, Ireland than from Dublin? More from Leesburg, VA than from Boston?
posted by strangely stunted trees at 9:05 AM on August 20, 2010

strangely stunted trees: some ISPs show up as being in different locations from where their customers are. I seem to recall that AOL was at one point based in northern Virginia, although they're not any more; some of their network infrastructure might still be there.
posted by madcaptenor at 9:08 AM on August 20, 2010

Creepy. When quancast first appeared I asked to have my sites excluded from their guesstimations (when you don't "quantify" by adding a javascript to your site, it just guesses based on traffic calls from everywhere else in some way). Guess what, you can't have your site excluded! Chris Martelottis replied to me personally and said; "Yes, I do not have a way to do this. I'm sorry." he also added: "As an audience measurement service we operate much like Nielsen for TV - if a program airs, we rate it."

Don't get me started on Nielsen's guesses.
posted by dabitch at 9:10 AM on August 20, 2010 [1 favorite]

"This site attracts an affluent audience"

Come for the articles stay for the "foobies"
posted by pianomover at 9:16 AM on August 20, 2010

Metafilter users are probably also a little more likely than the general net populace to use Flashblock and BetterPrivacy. (Also NoScript, Cookie Monster, RefControl, and TinFoil Hat. That last isn't a Firefox extension -- I just wear it.)
posted by Zed at 9:26 AM on August 20, 2010 [2 favorites]

These stats are pretty suspect. Every single website I looked at "atrracts an affluent audience."

But maybe that stat is true because only affluent people have the luxury of time needed to surf the web.
posted by jnrussell at 9:34 AM on August 20, 2010

coolxcool=rad: “When I refresh the MeFi page this thread shows the word "are" right after "MetaFilter users" ... It then disappears. What is causing this, is it the zombie cookie?”

No. It doesn't do that for me, nor (presumably) for anybody else.

The word "are" in that sentence is the link of the post. If you aren't seeing it after the page loads, that probably means your browser is coloring links wrong. Try highlighting the page where the word "are" should be (to see if it's actually there) or visiting this page with another browser.
posted by koeselitz at 9:42 AM on August 20, 2010

I'm not sure what this says about QuantCast's accuracy.
posted by public at 9:42 AM on August 20, 2010

Quantcast uses cookies.

They may use cookies, but they also get data on sites that don't participate.
posted by delmoi at 9:55 AM on August 20, 2010

Guess what, you can't have your site excluded! Chris Martelottis replied to me personally and said; "Yes, I do not have a way to do this. I'm sorry." he also added: "As an audience measurement service we operate much like Nielsen for TV - if a program airs, we rate it."

Actually, that's not true look. My guess is they actually charge to remove your site.
posted by delmoi at 9:58 AM on August 20, 2010

68% 18-49

Well, that's just shocking.
posted by ChurchHatesTucker at 9:59 AM on August 20, 2010

Does anyone know where Quantcast get any sort of data to actually back up the claims they make?

Doesn't everyone fill out demographic information by picking the highest number for every category?[1] Is everyone but me out there filling out survey forms with their true information?

[1] Why yes, I am a 80+ year old Female making > $500,000 a year who spends 12 or more hours online daily.
posted by madajb at 10:09 AM on August 20, 2010 [1 favorite]

I tend to visit this site from work. So, neener neener.
posted by chavenet at 10:10 AM on August 20, 2010

These stats are pretty suspect. Every single website I looked at "atrracts an affluent audience."

Nonsense. Take a look at the stats for icantaffordacomputer.com.
posted by cortex at 10:22 AM on August 20, 2010 [1 favorite]

Can some explain like I'm an 11 year old how quantcast determines the ethnicity, age, income etc of visitors to these particular websites? I am confused.
posted by modernnomad at 10:23 AM on August 20, 2010 [1 favorite]

I seem to recall that AOL was at one point based in northern Virginia, although they're not any more

The very first time I tried using a log analyzer at my then-work, c. 1999, I was baffled by how many people were coming from Virginia. Like WTF? Until I realized (read?) that most AOL traffic showed up that way. Those were the days.

Don't get me started on Nielsen's guesses.

My mom (and kid sis) were a Nielsen family for a while, at first TV and then internet. I think they had some sort of software installed on mom's computer. I tried to talk my sister into making lots and lots of visits to friends' blogs. :)
posted by epersonae at 10:28 AM on August 20, 2010

As a disembodied green head floating inside a box, it appears I am still statistically insignificant.
posted by Kirk Grim at 10:32 AM on August 20, 2010

MetaFilter users are
55% male
68% 18-49
83% Caucasian
54% >$60k/yr

When the sample is derived from
a) people willing to provide personal information
b) people who are not using privacy filters
c) people who have forgotten the study that says most people don't like to participate in studies.

How do I downgrade from 55% male, and upgrade to $60k/yr? These are useless statistics unless I am given the magic mathematical formula to change my status!
posted by tzikeh at 10:33 AM on August 20, 2010

These stats are pretty suspect. Every single website I looked at "atrracts an affluent audience."

The Goatse.fr mirror doesn't seem to. I can't imagine why.
posted by benzenedream at 10:35 AM on August 20, 2010

<1% Addicts

Ooh, I can see my house from here!
posted by quin at 10:37 AM on August 20, 2010

why does MeFi have a flash cookie?

I don't think MeFi has a flash cookie. They are installed at other sites, but then track ALL your internet usage regardless of what site you visit.

Kind of creepy, I have to admit.
posted by hippybear at 10:56 AM on August 20, 2010

I am not a number.
posted by TwelveTwo at 10:56 AM on August 20, 2010 [2 favorites]

delmoi > Actually, that's not true look. My guess is they actually charge to remove your site.

Wow, new feature! You need to join (and install their javascript) to hide your sites data, so I have to collect my visitors data via Quantcast (do not want), in order to not show people the Quantcast data. That's nice.
posted by dabitch at 11:01 AM on August 20, 2010 [1 favorite]

I know for a fact that, first thing in the morning, I'm 68% male.
posted by KevinSkomsvold at 11:08 AM on August 20, 2010

Suddenly, I feel poor!
posted by sonic meat machine at 11:12 AM on August 20, 2010

I'm not sure what this says about QuantCast's accuracy.

What seems inaccurate? I can totally buy that a bunch of teenage males go to that site to "window shop." It's not a profile of their actual customers.
posted by desjardins at 11:19 AM on August 20, 2010

So this is what Chrome is talking about when they suggest you kill Flash cookies, I guess. I'm glad I do that on a semi-regular basis now.
posted by immlass at 11:20 AM on August 20, 2010

You know who else had ~1 ball?
posted by DU at 10:35 AM on August 20 [+] [!]

Lance Armstrong?
posted by MultiFaceted at 11:23 AM on August 20, 2010

How reliable/unreliable is this data? What doesn't it account for? And most importantly, How does it work!? Is it a joke? A guess?

Seriously, despite the privacy and other issues, the stats have some pretty useful applications for research. Even if one uses the data to demonstrate how inaccurate the data actually are.
posted by iamkimiam at 11:32 AM on August 20, 2010

Was anyone actually questioned about income or number of children? No? Then this is a chart of guesswork.
posted by Cranberry at 11:56 AM on August 20, 2010

In all seriousness, this sort of thing really bothers me. Cookies are a very useful technology, but "tracking cookies" like the Quantcast crap make people think that cookies in general are bad, and end up getting them blacklisted in spyware programs and disabled by IT...
posted by sonic meat machine at 12:02 PM on August 20, 2010

Another privacy crisis easily negated by 15 seconds of browser changes (or zero seconds, if you already browse safely).

This all goes back to my perfect-world "If you can't be bothered to learn how to put air in your tires, you're too f-ing lazy to be allowed to drive on the same roads as me" theory.
posted by coolguymichael at 12:17 PM on August 20, 2010

Cookies are a very useful technology
And so are scripts.

But due to tracking and malware I put betterprivacy and noscript on all relatives' computers. They don't like it because it disables normal website operations like banking. And I don't like it because of the training involved to allow access to their required sites.

But the piece of mind is worth it to keep down on the malware removal and to preempt the questions they ask about web privacy. "Yup, your protected" - somewhat.
posted by Exad at 12:21 PM on August 20, 2010

If they just called them cupcakes, everyone would accept them.
posted by Kabanos at 1:11 PM on August 20, 2010

If they just called them cupcakes, everyone would accept them.

Or cakeballs.
posted by spilon at 1:51 PM on August 20, 2010

MetaFilter may not have a professional white background, but MeFites sure do.
posted by decagon at 2:06 PM on August 20, 2010

All you guys are WHITE!? I am cancelling plans for a dim sum MeFi meetup.
posted by helmutdog at 2:38 PM on August 20, 2010 [1 favorite]

What to I lose if I block 3rd party LSOs? Persistent game data (maybe?). Anything else?
posted by willnot at 3:35 PM on August 20, 2010

What to I lose if I block 3rd party LSOs? Persistent game data (maybe?). Anything else?

I blocked 3rd party LSOs and made all storage permission only after the first privacy thread a few weeks ago. So far, the only thing I've lost is... um... nothing? The only real change that I've noticed is that certain flash video websites don't store what the volume setting is from the last time I used the site. And that isn't a 3rd part LSO kind of setting, anyway, as it belongs to the website I'm visiting, not a 3rd party.
posted by hippybear at 4:08 PM on August 20, 2010

What to I lose if I block 3rd party LSOs? Persistent game data (maybe?). Anything else?

Adobe Connect, a somewhat common Flash-based web conferencing solution, will no longer work. I just found that out today!
posted by me & my monkey at 4:13 PM on August 20, 2010

The data is derived through guesswork, if some very refined guesswork. All the "tracking cookie" is is a unique identifier(generally speaking, the Quantcast cookie is going to be dropped from a third party domain that does not have access to the website you are actually visiting). As you visit more sites, it starts to develop a better picture of who you are based on your browsing habits.

Tracking cookies aren't bad inherently, as long as users understand what they are. Specifically, what Quantcast was doing was bad because they were duplicating values in an LSO and a browser cookie. Most people think "okay, clear cookies in browser" stops their future activity from being tracked, but since Quantcast was referencing the LSO value to repopulate the next browser cookie with the same unique value and tying previous browsing activity to future activity...yeah, not cool at all. I feel that we're getting to the point where the average internet user gets the concept of "clear recent history"(at the very least, it's really accessible in almost all common browsers), but education on "flash cookies" just isn't there yet.

On the bright side, Adobe did integrate Flash into the "private browsing" function in browsers that have it, which means that if you use private browsing, your Flash stored data will be wiped when your browser data is at the end of the session. Nothing is tied to the standard "clear cookies" browser function, though.
posted by sawdustbear at 5:32 PM on August 20, 2010

Very good fresh air interview from yesterday describing the tracking companies and the exchanges which sell the information about "you.".
posted by shothotbot at 7:33 PM on August 20, 2010

MetaFilter users are

100% Special Snowflakes
posted by crossoverman at 12:44 AM on August 21, 2010

If you click on "contact" down mwah at the lower right it has a link to Federated Media where you can buy ads. It says:

Audience:

* 67% male
* 78% 18-39
* 32% HHI above $75k
* 29% managers or above
* 35% IT professionals, developers or engineers
* 47% publish their own blog
posted by vapidave at 2:20 PM on August 21, 2010

If you don't like Adobe's little secret flash cookies, you could install an extension like 'Better Privacy' and roust them out.

Or you could just replace the 'Macromedia' folder/directory with a file titled 'Macromedia'. (You might even put a little text in it, like 'screw you Adobe' or 'Om mani padme hum'.) Voila, no more LSO's in your life.
posted by Twang at 12:50 AM on August 22, 2010

Or you can use the Flash Control Panel linked in a comment above to set your global settings to 0 space, no 3rd party installs, and "ask me" for every download... That seems to have done the trick for me. I've checked, and there's nothing there that I'm not okay with.
posted by hippybear at 9:09 AM on August 22, 2010

If they just called them cupcakes, everyone would accept them.

What, because people hate cookies so much?
posted by maryr at 11:00 AM on August 22, 2010

Firefox Private Browsing Data Leaves Little to the Imagination
posted by homunculus at 1:23 PM on August 24, 2010