Press "Enter" to skip to content

Psychological analysis of 1200 redditors (MBTI)

Before you begin, here is an info-graphic on how to read the correlated data.
(The data are taken from 1200 responses to the reddit survey).
The items are more closely described below (you can click on every image to zoom it).

In this image, you can see that people who think they are intellectually superior to others correlate with having no friends. (or vice versa – it is a correlation, not causation – so, people who have more friends maybe do not feel intellectually superior. Either way is possible, take what you want from it.

Red is negative correlation, blue is positive. Size of the square determines the strength of the correlation (every correlation I have used in the graphs is statistically significant with p < 0.05). You can see the numeric value itself in the lower part of the matrix.

Another negative correlation you can see in this example is that people who feel superior do not feel emotions very much. Or a positive correlation between people who said they feel their breathing fully and who meditate regularly. Another interesting one may be that people who reported breathing deeply also feel their emotions more. And people who feel their emotions correlate with having more friends. Social empathy.
Anyway, you can see more of these in the article, I have posted an analysis like this for every MBTI type – if you are interested only in the MBTI analysis and not in the process of analysis or analytics of other variables, jump straight to it!

Work Pipeline

First I exported the Google Forms data to Google Spreadsheets, and from there to a .csv file (classic Excel format).

I decided I will learn the R language for this.
Searched for some youtube tutorials, as you do.. and found two interesting videos.
https://www.youtube.com/watch?v=6hjNwMkNlpA – absolute basics, file loading
https://www.youtube.com/watch?v=igPQ-pI8Bjo – correlations

Of course I installed R language binaries and RStudio first.

Okay, so now we are ready.
Next step was to sanitize and edit the data, so the correlations would yield some non-gibberish data. So, for example – for question „How hard is for you to feel your emotions?“ I had to change the „Easy“, „Neutral“, „Hard“ to integers 0,1,2 – so the correlation would work. I did this manually for every possible variable. Later I learned that R language can, of course – a statistical language, do this with a simple function.

One question about emotions looks like this in the survey

I coded the values like this: Very hard = 0, Hard = 1 ….. Very Easy = 5.
Every question is coded logically. So, the more emotions you feel, the higher value etc. You should have no problem reading the graphs, because I tried to do it so it can be intuitively understood and the words self-explanatory. If you think you will be fine then feel free to skip the table below to the analysis itself.
The table with coded values

Question in surveyCoded wordTransformation or details
How often do you do sports/fitness?sportsnot active = 0
3x week = 5
What kind of activity do you do?activity-
How old are you?age18-20 = 18
>75 = 75
Is it hard or easy to feel your emotions fully and understand them?emotionsvery hard = 0
very easy = 5
Do you have sedentary job / hobby?sedentaryyes = 1
neither = -1
Do you have many close friends?friends>3 3
<3 2
0 0
Are you often invited to parties?partyguyshy 0
heart of party 2
Do you create parties and invite people?partyleadall the time 4
never 1
Do you usually start a conversation or do you expect people will text you (fb)?convostartstart 1
wait 0
Do you feel like you are doing what you want in life?happinesslike stones throwed 0
prison 1
enjoy fully 5
Do you meditate often?meditationtrue 1
false 0
Do you consider yourself to be a spiritual person?spiritualtrue 1
false 0
Do you find conversation about deep issues stimulating or draining?deepconvolove it 3
hate it 1
Are you aware of moments where you stop breathing or breathe shallow?breathfull breath 2
not at all 0
Do you consider yourself intellectually superior to other people?superioryes 2
ok 1
no 0
Do you eat/drink lots of sugar?sugaravoid 0
try to 1
cola 3
What is your body build?body
Do you have lots of hobbies?hobbiesmore 3
less 2
no 0
Were you bullied or did you do the bullying?bulliedbullied 1
okay 0
victim -1
Do you have a problem with authority?conformistconform 1
defy 0
Do you have a condition?conditionsmapped conditions like conditions.Depression and conditions.ADHD
Do you drink a lot of water?watermore 2 = 3
les 2 = 1
Have you ever had a relationship longer than 4 months?longrelationshipyes 3
complicated 2
no 1
Has anyone ever tell you "I love you"?lovedyoutrue 1
false 0
Do you think you get enough hugs in your life?hugsno 0
okay 1
yes 2
Do you often feel alone?aloneyes 2
sometime 1
no 0
Do you have a job?job
Do you know which MBTI type are you?mbti
1 of 4 MBTI : Are you Extroverted or Introverted (E vs I)m_ei
2 of 4 MBTI : Intuitive N or Sensing Sm_is
3 of 4 MBTI: Thinking T or Feeling Fm_tf
4 of 4 MBTI : Judging J or Perceiving Pm_pj

Implementation Intention

I have used multiple libraries for this, namely

caret package – https://github.com/topepo/caret/
corrplot package  – https://github.com/taiyun/corrplot

I will provide the final script and the csv table with the answers at the end of the article, so you can also play with the data and do you own correlations if you want 🙂

The script language is pretty easy, but I had to learn a bit about the workings of the statistical analysis. The videos above helped a ton!
Also, the layout in RStudio IDE is awesome for fast prototyping

Correlation Analysis

Let us begin with friends and happiness, then we go into individual MBTI characteristics.

This one is focused on „Happiness„.
Happiness is negatively correlated i.e. with having no job, being skinny and being introverted. Positively then with having more hobbies, more friends throwing parties, enjoying deep conversations, starting conversations and being extraverted. Also, a weird negative correlation is between being skinny and getting hugged. Maybe it’s the uncomfortable bones 😀

When you’re looking, you never find.

Here I focused on the question „friends“.
We can see a correlation between having friends and throwing more parties and being a „party guy“. People with friends also report to be more happy / or is it that happier people have more friends? Also Extraverted people throw more parties and often create parties, and do more sports.
Now for other interesting variables. People who have condition.ADHD and condition.SeasonalAllergies reported to be focused at their breath less and also drink less water. Or for an obvious one, people who feel alone report lower levels of happiness. And they are not happy with their job (job.happy.with.it), did not have very long relationships and they are not as extraverted.
Now for sugar! People who reported drinking more water eat less sugar, or vice versa.
And people who do sports a lot do not eat sugar that much, or it can be that people who eat sugar a lot do not do sports very much. And sugar is also correlated to people not loving you (???), seasonal allergies and ADHD, being a party guy and throwing parties. Hmmm. At least one positive correlation is between having more friends and eating sugar (???).

Another one, focused on „age“.
So, first thing I noticed was that old age and skinny body are negatively correlated 😀
Older age is also correlated with having been in long term relationship, someone telling you they love you, throwing more parties, having more friends, more hugs, being happier in life and feeling less alone.

Now for the question about feeling alone.
It is positively correlated with being skinny, eating sugar, hating your job or not having one. And it is in negative correlation with being a Sensing type, having no medical conditions, having long relationships, feeling happy, getting more hugs, being happy with your job, drinking more water, having fit body, having ADHD, having hobbies and meditating often.
So if I introduce some bias into this – it could be that sensing people tend to live their life rather than sit home thinking and they have no medical conditions (maybe because they do more sports (below you can see that Sensors love running and yoga)). Feeling alone and shorter relationships is an obvious one, as is not feeling happy. With that comes getting less hugs. No job or not being happy with one – job provides colleagues and eating lunch together, thus people with no job may report feeling more alone. For having ADHD and feeling less alone.. weird. Maybe ADHD people tend to generally have more random impulsive energy and thus are interesting for others but on the other side, it is very hard to keep friends when you get overwhelmed with managing your everyday life, not even going into holding long term social connections.
And the last two – having hobbies and meditating. Well, when you take a photography seminar or do yoga, you come into a social circle of 30 people on average, 2 times a week. When you meditate you can look for local meditation seminars with other people looking to calm their mind, and maybe you can even meet some spiritual friends!

MBTI analytics

Extraversion

Extraversion correlates with enjoying going to parties, starting conversations rather than waiting for being talked to, enjoying deep conversations, having more friends, feeling your emotions more deeply, having ADHD, being more spiritual, having had a long relationship, doing sports, being fit but also with being sedentary (you can notice a downward trend in sports->fit->sedentary in the correlations values).
Negatively correlated it is with having no job, being a conformist and not doing any activities. Interesting artifact in the pattern in the first quadrant is positive correlation between being a conformist and feeling your emotions. Which reminds me I cannot forget to add a gender question next time.

Introversion

Introversion is in positive correlation with depression, social anxiety, skinny body and obedience to authority. Woah. Now for something positive please. It correlates with having less ADHD!!
As you can see there is more nagatively correlated items, like being a party guy, having fit body, feeling your emotions, having friends, starting conversation, meditating and enjoying deep conversations.
And now for random artifact – people who meditate more are less hyperactive.. or can it be that people who are hyperactive do not meditate that much?

Sensing

Sensing people seriously enjoy Yoga (3 items!).. or is it that people who do yoga become more sensing?
We can also see that sensing type is correlated with being more happy and being obedient to authorities.
Negatively then to feeling alone, not doing any activities, feeling superior and being extroverted.
But the one observation that just confirms my real life experience with sensors is that they absolutely hate deep conversation. Why is that?
You can notice a strong negative correlation in superior-conformist.

Intuitive

Intuitives are negatively correlated to obedience to authority, doing cardio or yoga and feeling happy. They are positively correlated  with feeling alone, feeling superior, enjoying deep conversation and not doing any sports.

Thinking

Thinkers. They are positively correlated with feeling intellectually superior to others, being intuitive, being fit and having no illnesses.
They are negatively correlated to feeling alone, being bullied, being fat, being spiritual, doing yoga, being a conformist and feeling their emotions.

Feeling

In the Feeling type we can notice positive correlations in being a conformist, feeling your emotions (…), doing yoga and walking, being more spiritual, being bullied and feeling more alone.
Negative correlation for Feeling types is in feeling superior to others and in having no medical conditions.

Perceiving

Perceiving types are positively correlated to being intuitive, being extroverted, being a party guy, feeling superior to others, and being of thinking type.
They are negatively correlated to being spiritual, feeling emotions and very strongly to obedience to authorities (as are thinking, intuitive and extroverted types).

Judging

Judging types are in positive correlation with feeling type, being a conformist, feeling their emotions, being spiritual and doing yoga and with older age.
They are negatively correlated to being a party guy, extraversion, being intuitive and to feeling superior.

And now for something completely different.

Meditation

People who meditate are in positive correlation with being more spiritual, feeling their breath more deeply, having no illnesses, enjoying deep conversations, being extraverted, having more hobbies and they are negatively correlated to being fat, being a conformist and eating sugar.

And that is all from the MBTI analysis! I tried to analyse the medical conditions, but it yielded too few results, example:

Reddit post analysis

Most people answered the survey between 17-21 GMT+2. Good to know!

Here is how many percent of subreddit subscribers viewed the post/survey. ENTP, ESFJ and ESTP win!

Here is how many people who viewed the post also upvoted it. Why are they way you are, ESFP and ISFJ? I have posted 2 times to each of their subs over the span of three days, but no, just got downvoted to zero, both times.

And now the final metric – comments. ENTP and INTP win!!

Source code and data

As promised, here is the mbti source code (rename to .r extension and run it in RStudio) and the table data.

I hope you enjoyed the article and learned something new.
More articles like this to come.

– John Blomqvist.

Be First to Comment

    Napsat komentář

    Vaše emailová adresa nebude zveřejněna. Vyžadované informace jsou označeny *