I Spent A Month Of My Life Studying Social Media, Here’s What I Found

An In-Depth Study Into Medium’s Demographics

A Thinker
Crazy Ones
6 min readJun 4, 2016

--

Apparently 9 out of 10 Medium users don’t like Donald Trump’s hair (more on that later)…

Let me give you a short back-story. “Medium is bascially a startup founder’s paradise — you should check it out.” That was my introudction to the platform by a good friend of mine. I believed that for a while too! Until… Last month, I read an article that drew out the idea that Medium’s community was much wider than the self-anticipated “startup” niche. In fact, it displayed some shocking numbers at just how many topics were discussed every day more so than their “startup” counterpart. While Medium may have started as a topic-focused platform, it became clear to me that this was no longer the case — But the story doesn’t end there. After reading this, I began to wonder what type of people make up Medium. Obviously as an entrepreneur, I was destined to create my own way to find this out. And so, on May 3, 2016, I began “The Medium Study: The demographics behind the platform”

After 1 month, many messages, lot’s of Google searches, and 329 Medium users (including myself), here’s what I found out…

Who are these people, and how did I get their info?

Good question. The 750 Medium users that I started with were broken up in the following way.

I started by finding the top 25 tags/categories on Medium (these are just examples)
  • First, I identified the top 25 categories (or tags) commonly used on Medium. Those included (in no particular order): Startup, Investing, Leadership, Humor, Advice, Privacy, Politics, Poetry, Reading, Education, Energy, Journalism, Future, Finance, Culture, Startup, 2016 Election, Health, Design, Books, Internet of Things, Life Lessons, Music, Parenting, and Technology.
I collected different types of users. Note: I didn’t use Jon as one of them
  • Second, I collected 5 users that were the top writers from the respective tag/category (sometimes it double crossed and so I went down to #6, etc). I collected 10 writers that had most recently published a ‘quality’ article under the given tag (this was a little hard to find, but I did my best). Then, I collected 8 users, randomly selected, but ones that had written at least one article under the given tag (again, if users double crossed across tags, I only included them in one). Finally, I collected 2 users, randomly selected, that had recently commented on an article pertaining to the given tag. All the users I collected I made sure to identify as “active” within the past two weeks (this would make them easier to get in contact with).
  • At the end, I had 30 diverse users from 25 different categories across Medium. This part actually didn’t take as much time as you’d think, but the next part would be the most challenging…
An introduction/conversation on LinkedIn that I had with a Medium user I interview, Justin. I am blurring his name and details because I am using his personal data in this study.
  • My next task was getting in contact with as many of these users as possible. You may think think that doing this would be extremely hard, but many were easily contacted through LinkedIn (an easy Google search would commonly pull up the matching profile). Through LinkedIn or other emails I found listed by their names online, I was able to contact 589 users out of the 750 I originally identified. Out of those 589, 328 returned my stalker-like messages (just kidding). Not too bad ratios!
  • For most of the 328, I asked questions over email or LinkedIn, but talked for about 5 minutes on the phone with somewhere between 35–50 of them. However, even with just messaging between LinkedIn and email, I was able to garner around the same amount of data than that of talking over the phone.
  • In total, that leaves me with detailed information on 329 diverse Medium users (including myself). It was time to collect the results.

In total, that leaves me with detailed information on 329 diverse Medium users (including myself).

It’s time to display the results…

Note: These percentages are rounded up or down (.5 scale) from a decimal

Education

  • 88% of users are college graduates at a 4-year university
  • 94% of users are high school are high school graduates (9% of that 94% are graduates by a GED)
  • 16% of users have a masters degree (28% of those degrees were in computer science)
  • 4% of users have a doctorate degree
  • Fun Fact: 35% have “political science” as one of their undergrad majors

Race and Ethnicity

  • 68% of users are white, 14% are black, 13% are Asian, and 5% other (based on how they identify).
  • 94% have American citizenship (6% of this percent have dual citizenship).
  • 92% were born in The United States.

Age

  • 77% said they are under 50 years old, 55% said they were under 30, and 6% said they were under 18.

Political Standings (for those in the US)

  • 65% identify as Liberal/Democrat, 27% as Republican/Conservative, 7% as Independant/Unsure, and 1% as Libertarian or other.
  • 58% said they voted in the last Presidential election in 2012 (note that a few were under 18 at the time).
  • 90% said they plan to vote in the upcoming election.
  • Of those who plan to vote or have voted, 56% say they support Hillary Clinton, 24% said Bernie Sanders, and 23% said Donald Trump.

Yes Or No Questions

  • Are you currently self-employed or work at a startup?

33% Yes, 77% No

  • Are you or do you consider yourself religious?

87% Yes, 12% No, 1% No Answer

  • Do you support a ban on assault rifles?

55% No, 43% Yes, 2% Unsure

  • In general, do you believe The United States still has serious racial issues?

89% Yes, 2% No, 9% Unsure

  • Do you think Facebook purposely altered their news feeds?

58% Yes, 17% No, 25% Unsure or Don’t know well about the story

  • Do you approve of Donald Trump’s hair?

81% No, 19% Yes (highest comment was that it was too orange)

Finally… Medium Approval Ratings

  • In general, do you approve of and enjoy using Medium?

88% Yes, 10% No, 2% Unsure

77% No, 13% Yes, 10% “I’ve heard the name but don’t know who he is” (Sorry Ev)

That’s a lot of data

So, I kinda lied when I said I would only take 5 minutes to ask questions — it was more like 10 to 15. However, almost all the people I contacted were excited to see what types of data I would draw from this project. Whether we can draw accurate conclusions from this or not, no data on Medium’s demographics at the scale have ever been recorded and displayed to the Medium community, and I think this is the first social media platform study ever to conclude what percentage of users approve of Donald Trump’s hair. Because of that, it was worth it. Unfortunately for The Donald, the answer wasn’t in his favor this time around.

Conclusions

I leave the most precious of questions up to you… for I am just the data man — for now.

Sincerely,

A Thinker

--

--

A Thinker
Crazy Ones

Enjoying life — ignoring the bullshit. Mysterious lead editor at Crazy Ones. https://medium.com/crazy-ones/.