Tag Archives: statistics

My name is danah and I’m a stats addict.

I love data and I hate stats. Not stats in abstract — statistics are great — but the kind of stats that seem to accompany any web activity. Number of followers, number of readers, number of viewers, etc. I hate them in the way that an addict hates that which she loves the most. My pulse quickens as I refresh the page to see if one more person clicked the link. As my eyes water and hours pass, I have to tear myself away from the numbers, the obsessive calculating that I do, creating averages and other statistical equations for no good reason. I gift my math-craving brain with a different addiction, turning to various games — these days, Yushino — to just get a quick hit of addition. And then I grumble, grumble at the increasing presence of stats to quantify and measure everything that I do.

My hatred is not a new hatred. Oh, no. I’ve had an unhealthy relationship with measurement since I was a precocious teenager. I went to a school whose grades came in five letters — A, B, C, D, F (poor E, whatever happened to E?). Grades were rounded to the nearest whole number so if you got an 89.5, you got an A. I was that horrible bratty know-it-all student who used to find sick and twisted joy in performing at exactly bare minimum levels. Not in that slacking way, but in that middle finger in the air way. For example, if you gave me a problem set with 10 questions on it, where each question was harder than the last, I would’ve done the last 9 problems and left the first one blank. Oh was I arrogant.

The reason that I went to Brown for college was because it was one of two colleges I found out about that didn’t require grades; the other college didn’t have a strong science program. I took every class that I could Pass/Fail and loved it. I’m pretty sure that I actually got an A in those classes, but the whole point was that I didn’t have to obsess over it. I could perform strongly without playing the game, without trying to prove to some abstract entity that I could manipulate the system just for the fun of it.

When I started my blog back in the day, I originally put on a tracker. (Remember those older trackers that were basically data collectors in the old skool days??) But then I stopped. I turned it off. I set it up to erase the logs. I purposefully don’t login to my server because I don’t want to know. I love not knowing, not wondering why people didn’t come back, not realizing that a post happened to hit it big in some country. I don’t want the data, I don’t want to know.

Data isn’t always helpful. When my advisor was dying of brain cancer, he struggled to explain to people how he didn’t want to know more than was necessary. His son wanted the data to get his head around what was happening, to help his father. But my advisor and I made a deal — I didn’t look up anything about his illness and would take his interpretation at face-value, never conflicting whatever he said. He was a political scientist and an ethnographer, a man who lived by data of all forms and yet, he wanted to be free from the narrative of stats as he was dying.

As we move further and further into the era of “big data,” I find that stats are everywhere. I can’t turn off the number of followers I have on Twitter. And I can’t help but peek at the stats on my posts on Medium, even though I know it’s not healthy for me to look. (Why aren’t people reading that post!?!? It was sooo good!!) The one nice thing about the fact that 3rd party stats on book sales are dreadfully meaningless is that I have zero clue how many people have bought my book. But I can’t help but query my Amazon book sale rank more often than I’d like to admit. For some, it’s about nervously assessing their potential income. I get this. But for me, it’s just the obsessive desire to see a number, to assess my worth on a different level. If it goes up, it means “YAY I’M AWESOME!” but if it goes down, no one loves me, I’m a terrible person. At least in the domain that I have total control over — my website — I don’t track how many people have downloaded my book. I simply don’t know and I like it that way. Because if I don’t know, I can’t beat myself up over a number.

The number doesn’t have to be there. I love that Benjamin Grosser created a demetricator to remove numbers from Facebook (tx Clive Thompson!). It’s like an AdBlocker for stats junkies, but it only works on Facebook. It doesn’t chastise you for sneaking a look at numbers elsewhere. But why is it that it takes so much effort to remove the numbers? Why are those numbers so beneficial to society that everyone has them?

Stats have this terrible way of turning you — or, at least, me — into a zombie. I know that they don’t say anything. I know that huge chunks of my Twitter followers are bots, that I could’ve bought my way to a higher Amazon ranking, that my Medium stats say nothing about the quality of my work, and that I should not treat any number out there as a mechanism for self-evaluation of my worth as a human being. And yet, when there are numbers beckoning, I am no better than a moth who sees a fire.

I want to resist. I want serenity and courage and wisdom. And yet, and yet… How many people will read this post? Are you going to tweet it? Are you going to leave comments? Are you going to tell me that I’m awesome? Gaaaaaaaah.

(This post was originally published on September 24, 2014 in The Message on Medium.)

Teens Don’t Tweet… Or Do They?

Yesterday, Mashable reported Nielsen’s latest Twitter numbers with the headline Stats Confirm It: Teens Don’t Tweet. This gained traction on Twitter turning into the trending topic “teens don’t tweet” which was primarily kept in play all day yesterday with teens responding to the TT by saying “I’m a teen” or the equivalent of “you’re all idiots… what am I, mashed potatoes?”

I want to unpack some of what played out because I’m astonished by the misinterpretations in every which direction.

We have a methodology and interpretation problem. As Fred Stutzman has pointed out, there are reasons to question Nielsen’s methodology and, thus, their findings. Furthermore, the way that they present the data is misleading. If we were to assume an even distribution of Twitter use over the entire U.S. population, it would be completely normal to expect that 16% of Twitter users are young adults. So, really, what Nielsen is saying is, “Everyone expects social media to be used primarily by the young but OMG OMG OMG old farts are just as likely to be using Twitter as young folks! Like OMG.”

We have a presentation problem. Mashable presented this report completely inaccurately. First off, Nielsen is measuring 2-24. My guess is that there are a lot more 24-year-olds on Twitter than 2-year-olds. Unless Sockington counts. (And she’s probably older than 2 anyhow.) Regardless, the Nielsen data tells us nothing about teens. We don’t know if young adults (20-24) are all of those numbers or not. If all 16% of those under 24 on Twitter were teens, teens would be WAY over-represented in proportion to their demographic size.

We have a representation problem. The majority of people are not on Twitter, regardless of how old they are. Those who use Twitter are not a representative percentage of the population. Geeks are WAY over-represented on Twitter. Celebs and celeb-lovers are WAY over-represented on Twitter. Newshounds are WAY over-represented on Twitter. And while Joe the Plumber has an account on Twitter, I doubt it’s him. Age is not the right marker here.

We have an interpretation problem. Saying that 16% of Twitter users are 24 and under is NOT the same as saying that 16% of teens are on Twitter. We don’t know what percentage of youth (or adults) are on Twitter. If you want to compare across the ages, you need to know what percentage of a particular demographic is using the technology.

We have an impression management problem. There are teens on Twitter. Thousands of them. Saying “Teens Don’t Tweet” gives the wrong impression because there are plenty of teens who do tweet (as they so kindly vocalized on Mashable and on Twitter). Still, just because they suddenly became vocal doesn’t mean that those who are there are representative of teens as a whole. Furthermore, the presence of teens on Twitter doesn’t mean that Twitter is a mainstream tool amongst teens. It’s not.

Given all of these problems, I immediately dismissed the Nielsen report and the Mashable post as irrelevant and meaningless. Then it became a Trending Topic. So while I had a million things to do yesterday, I spent 6+ hours reading the messages of the people who added content to the trending topic, reading their posts about other things, going to their profiles on other sites, and simply trying to get a visceral understanding of what youth were engaged enough on Twitter to respond to the trending topic. What I found fascinated me. I’m still coding the data so you won’t get any quantitative data just yet, but I want to give you a sense of my impression.

Teens On Twitter

The majority of teens who responded to the Trending Topic simply responded to the statement “Teens Don’t Tweet” by noting that they were a teen and they tweeted. Others just noted that the trending topic was dumb. Many didn’t know why the term had become a trending topic, were unaware of the Mashable article or Nielsen study, and thought that Twitter chose the trending topics. (I was in awe of how many teens commented that Twitter was stupid for making such a lie a trending topic. Some thought it was Twitter’s attempts to tell them they didn’t belong. One did ask if it was a trap to get teens to come out of the closet about their real age.)

Many of the teens who responded to the TT were not American or Canadian. I saw bunches of Brazilian teens, some Indonesian teens, and a smattering of teens from Europe, China, and Mexico. Many of their Twitter streams mixed English and the local language of their country. English dominated the responses but I did see non-English responses to the English trending topic.

About half of the teens included a link to a non-Twitter page in their bio. The pages were really mixed. Among the SNSes, MySpace dominated, but there were some Facebook links and links to Piczo and Multiply. There were also links to YouTube, Blogspot, LiveJournal, Deviant Art, and personal homepages.

Very few of the teens put their age in their bio, although quite a few made their age available in the content or through links. Teens posted messages like “I’m 16 and I’m on Twitter.” And birthdays are a big enough deal that I was seeing things like, “I can’t wait until I’m 16 and can get a car. Only 3 months to go!” And of course there’s MySpace.

Most of the teens on Twitter followed on the order of 40-70 other people (with fewer followers). Who they followed included a smattering of other teens and a collection of big names – celebs, bloggers, geeks. There wasn’t much discussion on their feeds about the number of people following them but they frequently highlighted how many tweets they had. I was surprised by how many of them would write a tweet saying nothing more than “this is my 1207th tweet!” Their content is primarily phatic in nature with an eye for updating as often as possible.

The most salient visceral reaction that I got when looking at the teens’ Twitter streams was that teens on Twitter seemed to fit into three categories: 1) geeky teens, tech teens, fandom teens, machinema teens; 2) teens who are in love with the Jonas Brothers/Miley Cyrus, musicians, or another category of celebs; 3) multi-lingual foreign teens with friends/followers around the world who seemed to participate in lots of online communities.

While I can’t make any meaningful conclusions until I spend more time with the data, it seems to me that the teens on Twitter – or at least the teens responding to the trending topic – are not representative of teens as a whole. That’s not a bad thing. They’re geeks and passionate creators and trendsetters and pop culture addicts. I don’t get the sense that they’re dragging their friends into Twitter, but rather, focusing on using Twitter to engage with other people who share their interests or people that they admire.

Anyhow, I’m continuing to track this but I thought I should just report out what I’m seeing in case it’s of use to anyone but me.

Be warned: This blog post was written in brain-dump style to get some general impressions out there while I analyze the data. My goal is to give you a sense of what I’m seeing, assuming that you aren’t staring at thousands and thousands of tweets by teens. Please don’t interpret it as a “report” or a “study” or anything other than what it is: a blog post.