Correlation vs. causality

The BBQ Pit
#1

Why can’t so many people understand the difference between these two things?
Just because two things happen together with some level of predictability,
it does not mean that one CAUSES the other!

On Friday mornings I put out the trash can, and whaddaya know, later that day, the garbage truck comes. Is it BECAUSE I put out the trash? NO, it’s because both events are caused by the fact that Friday is trash day in my neighborhood.

Last night on 20/20 the lead story was on cell phones possibly causing ill health effects.
Now without getting into a big debate on that subject (That’s why I posted to the Pit and not GD. I happen to think there may be a problem and we need more independent research), one of the studies cited to support their proposal that there is a health risk from cell phones claimed that their research showed that people with brain tumors on the right side of their brain typically held their cell phones to their right ear, and those with brain tumors on the left side of their brain typically held their cell phones to their left ear, and that this meant that the cell phones caused the brain tumors.

Well ya think it might be because, for example, mostly right-handed people hold their phones with their right hand and ditto for lefties? And that right or left handedness, or some other “third party” cause might be the reason for the tumor? Perhaps right-handed people have a propensity to slap themselves more often on their right side (the DOH! effect), or talk on their regular (wired? cordless?) phones on that side.

I’m not saying it’s not true, heck it could be the cell phones,
but this was purported science leaping to these causal conclusions without looking beyond what they were looking to find.
There are a million other examples of this around.
I’m just frustrated by the pseudo-science.
And all the people who take it at face value.

Thanks for listening. ::takes deep breath::

#2

You damn technology freaks are all the same. Friggin’ cell phone apologists.

-andros-
on his $8,000 workstation

#3

OK, andros, I’ll bite….

::calming music playing softly in the background::

I’ll state again what bugged me is not the attack on cell phones.
(Disclosure: Yes, I’m in the cellular biz – as if my username didn’t give it away – but not a shill for it. I go against most in my industry and say there actually might be a problem.)

I just used the cell phone story because it’s my latest example of the phenomenon/ignorance that bugs me. (I try to stay on top of developments in my industry and so when someone called me to say the show was on, I tuned in.)

And as for calling me a technology freak, well I’m kind of flattered. Always thought of myself as a bit of a Luddite. I have no idea how much my computer cost. It was a gift.

Appreciate your efforts to get a real flame going. :slight_smile:

#4

Another example of the problem Wireless cites (so we can debate the actual issue rather than the evils of cell phones) is the link between breast implants & autoimmune diseases. Implants haven’t been proven to CAUSE autoimmune diseases, though women with implants have higher rates of those diseases. Sadly for Dow Corning, a jury couldn’t make the distinction & slapped them with some hefty damages…

#5

Well, I tried. But I can’t disagree with you. My wife’s PhD is in cognitive psych–she’d been a Causaion/Correllation prophet for years.

There’s no debate here. People who think that breathing must cause cancer since cancer patients all breathe is running 100% clue-free.

Freakin’ Luddities. All the same. :wink:

#6

The Journal of Irreproducible Results once ran a wonderful list of high statistical correlations between clearly unrelated events – like the rainfall on the plains of Spain each month correlating precisely with the number of automobile accidents in Perth, Australia.

It was a wonderful list, I no longer subscribe to that remarkable Journal, but I’d bet they have more recent such.

Of course, that’s not confusing correlation with cause, that’s just noting that statistical correlations are nothing more than statistical correlations.

#7

www.junkscience.com

Best site I’ve found that attacks the “correlation = causation” myth. (Although some of its articles can be vituperatively biased.)

#8

Tracer, junkscience is one of those sites I hit a couple times each week.

“The problem with the world is that everyone is a few drinks behind.” - Humphrey Bogart

#9

Here is another example of correlation without causation:

There are greater liquor sales in areas where teachers are given higher salaries.

Although there is a significant correlation, there is no evidence of causation. Both teacher salaries and liquor sales might be caused by a third factor, such as affluence.

This really should be moved into Great Debates. I don’t see it progressing into real flaming.

#10

andros,

I always thought breathing caused death because everyone who died was breathing just a monent ago. I’m trying to cut down myself.

#11

Oh, almost forgot – the book Science Without Sense, whose cover appears on that website, can be read in its entirety starting at http://junkscience.com/news/sws/sws-preface.html .

#12

Dex, the Journal of Irreproducible Results is no more, but its editorial board is now responsible (or irresponsible, if you prefer) for the Annals of Improbable Research. They keep some articles on their website at www.improbable.com . I haven’t seen any correlation/causation lately, but they do a lot of important science - recently, a paper about developing a double-strength placebo, a most important and oft-neglected part of double-blind studies.

Never attribute to malice anything that can be attributed to stupidity.
– Unknown