Tyler Vigen

For a humorous look at what happens if we assume correlation causes causation there's Spurious Correlations: http://www.tylervigen.com/ My favorite: "Number people who drowned by falling into a swimming-pool correlates with Number of films Nicolas Cage appeared in" With enough variables and noise you'll find things that correlate if you look hard enough. Without experiments where you control one thing and observe the other it's hard to demonstrate causation.

That is an excerpt from a book about using psychology for the purposes of persuasion in marketing. It was not a study on the memetic effects of suicide reporting, which aspects of the reporting might contribute to contagion, or a treatise on untreated depression or suicide. I read that PDF. Twice. It makes some ... great leaps of conclusion and assumption, without much in the way of investigation or data. It concludes for example that young people intentionally fatally drove their cars into trees because they read articles about suicide, but it presents no actual evidence for that case other than a statistical correlation. As is mentioned frequently on HN, correlation in statistics is tricky: http://tylervigen.com/ I don't think we have enough information at hand to suppress discussion of depression or suicide.

Please leave your anti-science anti-intellectual comments outside of hacker news. "FED was instrumental in making USD the most stable currency in the last 100 years" - baseless argument. comparing fiat money to fiat money says nothing about efficacy of fiat money. "Too bad when California deregulated their energy market..." - http://www.tylervigen.com/ Furthermore, where does this money appear from? If the government makes energy cheaper by paying for it by taxing you more, you are just lying to yourself. "anti-abortion" - please explain why third-trimester abortions should be allowed. the baby can live by being born at that point, and the mother has had plenty of time to deal with the issue. but sure, you can disagree with that, and I won't debate it here. "anti same-sex marriage" - "I think the way to fix DOMA is maybe to try to make all of our laws more neutral towards the issue [of same sex marriage], and I don't want the government promoting something I don't believe in.", Rand Paul "vaccination" - http://www.latimes.com/nation/politics/politicsnow/la-pn-rand-paul-vac… So hey, maybe you disagree with some of his positions after you found out what they actually are. Is that enough to say, you shouldn't care at all about privacy, the military industrial complex, foreign entanglements that make us less safe, more drones strikes on civilians, and a corrupt, racist criminal justice system? I recommend becoming informed and making a rational…

I think we are, collectively, greatly underestimating the value of determinism and, conversely, the cost of nondeterminism. I've been trialing a different product with the same sales pitch. It tries to RCE my incidents by correlating graphs. It ends up looking like this page[1], which is a bit hard to explain in words, but both obvious and hilarious when you see it for yourself. [1]: https://tylervigen.com/spurious-correlations

It's not exactly what you're suggesting, but Spurious Correlations is great fun: https://www.tylervigen.com/spurious-correlations

Don’t forget this awesome quote: “However, a recent paper (Samsel & Seneff, 2013), argued that glyphosate may be a key contributor to the obesity epidemic and the autism epidemic in the United States, as well as to several other diseases and conditions, such as Alzheimer's disease, Parkinson's disease, infertility, depression, and cancer.” It’s causing all the bad things? That would be amazing if true, unfortunately it sounds unlikely (chiropractic comes to mind) and the rest of the science world doesn’t believe these authors. If the graphs in the paper seem compelling, make sure to spend some time here: https://www.tylervigen.com/spurious-correlations

This is the one, and it's great fun to look at: https://tylervigen.com/spurious-correlations

> Car crashes and ATC mistakes are two different sets of data. Just because they are going up or down at the same time and you can overlay some other event on top of it means absolutely nothing on its own. A good time to dust off this classic: https://www.tylervigen.com/spurious-correlations

Here are some of my favourite graphs: https://www.tylervigen.com/spurious-correlations

This reminds me of one of my favorite websites: https://tylervigen.com/spurious-correlations

Also related: Spurious Correlations[0]. These are great and really drives home the idea that correlation neither equals or implies causation. At least one of these graphs implies that if the US wanted to reduce the number of drivers getting hit by trains, they simply just need to stop importing oil from Norway. [0]: https://www.tylervigen.com/spurious-correlations

I love “causation does not imply correlation.” We still reason so much from the implicit premise that correlations are meaningful. It’s so intuitive, but wrong. The thing that finally got me to wrap my mind around it was this: https://tylervigen.com/spurious-correlations

I don't know whether this is a truly serious paper or can be filed under "Spurious Correlations"[0] - paracetamol is widely enough used that links to the mentioned "issues" could be tenuous at best. [0]: https://www.tylervigen.com/spurious-correlations

The 'Spurious Correlations' site is a fun place to look for things like this [0]. My favorite being a r=0.666 correlation between the number of 'people who drowned by falling in to a pool' and 'Films Nicolas Cage appeared in' [0] https://www.tylervigen.com/spurious-correlations

We all know that one can make statements that while factually correct, can be misleading, distorting, or unsound in some other way. (e.g. The statement "the divorce rate is highly correlated to the per capita consumption of margarine" is factually true [1] - but left on its own might leave people with the impression one has to do with the other) [1] http://tylervigen.com/spurious-correlations The same is true here - leaving out the context that FedEx has spent ~$50b on capex in the past 10 years and recognized $8.2b of income tax expense in the same 10 year period (which is not all US fedral, and might not have been the amount remitted over that period) changes the tone of the story. Additionally, really all that is happening is that tax collections go down a bunch in one year and up for the following six years, the only gain is on time value of money)

How is it you think that is evidence somehow? You might as well conclude that having a unicorn as your school mascot is why kids do so well at that school. http://www.tylervigen.com/spurious-correlations There are so many confounding factors that could be causing students to do better in those schools. Here is my take away from your 'data': Imagine how much better those students would be doing if they also weren't getting up earlier than is natural or recommended.

A well-known refrain, I know, but correlation does not imply causation[0]. Spurious Correlations[1] has been posted here many times[2], but a few representative examples may be worth sharing just the same: Per capita cheese consumption correlates with Number of people who died by becoming tangled in their bedsheets Total revenue generated by arcades correlates with Computer science doctorates awarded in the US Japanese passenger cars sold in the US correlates with Suicides by crashing of motor vehicle [0] https://en.wikipedia.org/wiki/Correlation_does_not_imply_causation [1] https://www.tylervigen.com/spurious-correlations [2] https://hn.algolia.com/?query=spurious%20correlations&sort=byPopul…

To someone who understands statistics, why is the method conducted by the researchers: " To examine the connection between historical acts of racial violence and modern-day voting behavior among blacks in the South, Williams examined county-level voter participation data in the six Southern states where individuals are asked to identify their race when they register to vote. " Where they basically drew connections between two sets of statistics, different from what 'spurious correlations'[1] takes aim at? Although the article doesn't mention at any greater depth as to how the research was done. [1] https://www.tylervigen.com/spurious-correlations

-Which is all the reason I need to bring up Tyler Vigen’s amazing spurious correlations: https://www.tylervigen.com/spurious-correlations

This is the paper version of [1]. Worthy of further research? Sure. "Damning" and "urgent"? Probably not. [1]https://www.tylervigen.com/spurious-correlations

Time for an old favorite: http://tylervigen.com/spurious-correlations

Not only is that anecdotal, but even if it's true it is correlative in the weakest sense. There are undoubtedly countless other features for which "cultures" (how does one define this concretely btw) with low Alzheimer's rates also are outliers. Here is a great demonstration of how correlation does not imply causation: http://www.tylervigen.com/spurious-correlations.

My personal favorite: http://tylervigen.com/view_correlation?id=1597 US spending on science, space, and technology correlates with Suicides by hanging, strangulation and suffocation

http://www.tylervigen.com/spurious-correlations It is best to assume "everything is correlated with everything else", rather than jump on every correlation. On the other hand it is important to encourage people to notice correlations and coincidences, but realize not every one can be followed up on, and most of these that are noticed will be only weakly involved in any solution. Really all we know at this point is that Alzheimer's disease is associated with the presence of more stuff that stains with antibodies raised towards Amyloid-beta aggregates than usual. It isn't even clear whether these are a cause or a symptom, nor whether these stains are near 100% specific towards beta-Amyloid (ie as derived from Amyloid precursor protein) rather than other aggregations of beta-sheets (like any prion "infection").

great chart bloomberg http://tylervigen.com/spurious-correlations

No science like social science based on google search statistics and correlation ; ) And what would you make of that? https://www.google.com/trends/explore#q=i%20hate%20jews%2C%20i%20hate%… I would also recommend http://www.tylervigen.com/spurious-correlations

This was the site I was actually looking for as great examples of correlation http://www.tylervigen.com/spurious-correlations

Wow, and the graphs on http://www.tylervigen.com/ are just incredible.

Would be interesting to test this in data such as these: http://www.tylervigen.com/

Pretty big confidence intervals, unclear what the true test space is. I'm just going to leave this [1] here. [1] http://www.tylervigen.com/