Five labs – See the personality behind your posts | Hacker News Reader

Five labs – See the personality behind your posts

labs.five.com

48 pointsuptown25 comments

This is the sort of thing that makes me glad I stopped using Facebook a long time ago. Of course, you could equally easily apply the technique to people's writings on HN, but there's a much lower probability of that being leveraged into a standard social media app.

This is the sort of thing that makes me glad I use Facebook. I find it interesting and don't know how else something similar could be achieved.

I actually just did that with your HN account, fairly easy:

100% Openness. 58% Extraversion. 63% Agreeableness. 30% Neuroticism. 31% Conscientiousness.

Sorry about the late reply, I was gone for a few days. I know this would work on any batch of content such as my HN postings, but it's an extra step you take as an individual. In the context of Facebook there seems to be a much greater possibility of it becoming infrastructure - 'rate your friends!'

damn, apparently I'm fucking negative and I keep telling everyone I'm fucking negative

The correlation with other people is pretty interesting. I was impressed that my family members showed up there, I suppose it makes sense.

I went through the celebrity list, and the person I was most similar to... "Paul Graham and you are 89% similar"

Personality is great... but could it detect mental illness, instead, so that (a) people don't waste time in hopeless clashes with, and/or taunt, the suffering; and (b) the ill might get help?

This doesn't measure your personality, but the content that you choose to share on Facebook. I work out regularly and am very hard working, but I choose not to post about those things. Does this mean that I am less conscientious?

The list of words used for the Extraversion trait is all slang and poor grammar. Is it not possible for someone to be extroverted while typing status updates in complete sentences?

I don't understand how using "fuck" a lot can be directly correlated to neuroticism. There are so many situations where it can be used that assigning the whole word to a signal for neuroticism seems like a misstep to me.

From the original article (as linked in a sibling comment), you can see the following result which is really interesting [1]. I do use the word 'fuck' when speaking and I've never been bothered by it but I can say that I don't think I've ever used in a facebook post. Which aligns with the study as I've previously done the Five Factor Inventory and have a low neuroticism score.

[1] http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3783449/figure/p...

From the same study, Fuck is also highly correlated with age and gender. They also found a 25% less strong correlation between swear words and neuroticism, compared to a previous, given study. I'd imagine correlating swear words is especially hard because of their varied use. (For extreme happiness vs anger, for example)

source and methodology: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3783449/

Ah, yeah, admittedly I didn't do much searching around for the source. I'd imagine the research is better than my intuition though. Thanks for that.

Also, being religious seems to be highly correlated with "agreeableness", as the example words include: Christ, church, praise, blessed, Christmas, psalm, and prayers.

You might be more prone towards agreeableness around others sharing your faith, but the more ensconced you are, the less likely you'll be open towards people with differences of opinion.

EDIT: Searching through the linked research paper, I couldn't find a single mention of "faith", "religion", or "religious", so I guess that was perhaps not considered.

EDIXx2: I think an interesting followup (to the research and possibly the tool) would be an attempt to discern how siloed away each person is and assigning a score to weight their other values. If you're in an echo chamber of everyone being #blessed, it's a bit different from seeing a wide range of opinions and activity. Even something simple like the mean distance of friends from the user's location, maybe with a foreign multiplier, I dunno.

EDITx3: Not even just distance, facebook has that about section as well. Obviously, many people have non-standard entries there, so some by-hand correction would need to be employed, but you could at least get a sense of someone's network variety.

Maybe Five Labs correlated those words with Family or Inclusive, based on the chart in the research paper, but it definitely isn't clear. None of the charts show a strong correlation between "faith" words and a certain personality, at least from my view.

This is what the app says Paul Graham's predicted personality is: http://labs.five.com/pg

INVENTIVE • SOLITARY • SENSITIVE • EFFICIENT • ANALYTICAL

This actually worked pretty well. I havent been posting to FB in recent years so I wonder if this has changed.

Someone should do this for my HN comments. I am usually less sarcastic/funny on HN than FB.

Deleted comment

Well, there you go. Thanks for the enlightenment.

Yawn! How useless and contrived is the statement, "You're x% Inventive"?

This trivializes introspection. People are really deriving insight from this?

I must admit, I'm curious as to which facebook posts they got Mahatma Gandhi's 2,284 words from.

it would be great if you could apply this to tweets or something more generic like a personal blog. (i feel like tweets would be easy... blog would be harder to parse reliably).

As someone who uses twitter a fair amount, and barely uses facebook at all - this won't really work for me.

My Facebook persona is stern, reserved, and curious. I'm not that easy to describe, though ;-)

Aha! This is for English speaking people.

Reply on news.ycombinator.com