Correlated: Discover surprising correlations

'Each day, a new survey is posted — something like "Do you like math?" or "Have you ever gone on a blind date?"

'At the end of the day, the results of the survey are compared with the results of all previous surveys, and the two outcomes with the strongest link are highlighted.'

Insert face in palm.

No. You cannot do this. Nearly all of the correlations found on this blog are pure noise. This is the data-science equivalent of staring at a cloud, noticing it looks like a balloon dog, and thinking "Jesus, why would anyone go to the trouble of making a cloud that looks like a balloon dog? Think of the engineering effort involved! Why isn't this in the news?"

Google for "multiple comparison issues" and "Bonferroni correction" if you want more technical detail on how stupid this is.

Hi, I'm the OP. I'd just like to encourage the critics in this thread to read "A response to Correlated's critics," which is linked to from the homepage:

http://coding.pressbin.com/108/A-response-to-Correlateds-cri...

I really hope it becomes standard for websites like this one to no longer require a full signup to use it. Maybe use a combination of sessions, IP addresses, browser fingerprints, etc. to differentiate anonymous users. I understand that the creator(s) of the site want to gather as much contact information as possible, but those who care to be a part of the community will do just that, while the rest (like myself) just want to try it out immediately... without filling out another sign up form.

I submitted my answer to the corn on the cob question but didn't make it any further than that because it immediately asked me to fill in a handful of fields to sign up. If I was able to continue without having to do that, I would have probably stuck around longer to provide the site with answers to other questions (providing value) and possibly even gotten sucked in and if I saw that the site was well-made, become a long-lasting member of the site's community.

It's important to change the expression "X% of people" to "X% of people browsing the Internet with nothing better to do than answer silly questions". For example:

In general, 61 percent of people who search for a blog on correlations and submit answers to poll questions there claim to pee in the shower. But among those people who have nothing better to do than click buttons on blogs and who claim they would use a bidet if available, 79 percent claim they pee in the shower.

With some level of rigor you can correct for self-selection and lying, but I doubt they are doing that here.

Everyone hating on this: lighten up. I am pretty sure they know that these correlations are nonsense. It is still a funny site.

Is it safe to take a survey of 50 people seriously?

28% of HN people looking at the OP site will regret wasting 2 minutes of their life.

Holy cow. I hope that shower statistic isn't really true!