Don’t jump to conclusions…

Cause and effect is not always this obvious…

In sci­ence and sta­tis­tics it is a com­mon and dire warn­ing that “cor­re­la­tion does not imply cau­sa­tion”.  Every­one under­stands the con­cept of cause and effect: hit the bil­liard balls and watch them roll away in the oppo­site direc­tion you hit them.  How­ever, some­times the cor­re­la­tion is not so obvi­ous… some­times a cor­re­la­tion doesn’t really exist but is only per­ceived to.  See­ing should not always be believ­ing, study­ing is impor­tant.  Before jump­ing to a con­clu­sion, make sure you have accu­mu­lated and con­sid­ered as much data as pos­si­ble.  When it seems that a link exists, it is cer­tainly worth look­ing into.  Unfor­tu­nately, many are eager to make a dis­cov­ery so they might dis­cover a link that is only loosely con­nected, or not even con­nected at all (some other force is caus­ing the effect that is not being con­sid­ered).  Coin­ci­dences exist, and just because it would be con­ve­nient for two events to be con­nected, that does not neces­si­tate it to be so.

Some­times, there are other vari­ables at play that may be the real cause of a cer­tain phenomenon.

How did I lose weight when I switched back to reg­u­lar Coca-cola?

Felic­ity switched off diet soft drinks such as Coke Zero back to reg­u­lar ones like Sprite and two weeks later she noticed that she lost 5lbs!  She started to tell her friends about this and began to assert that reg­u­lar soft drinks are actu­ally health­ier than the diet ver­sions.  Vera, a more skep­ti­cal friend of hers objects, and notes that she has been work­ing out a lot more lately, and that she has also been eat­ing lighter meals at work. Vera argues that these are the likely causes of the weight loss.  She even chal­lenges her to switch back to diet pop and see if the weight loss increases.  Unfor­tu­nately, Felic­ity already told every­one about her amaz­ing ‘dis­cov­ery’ and now her rep­u­ta­tion is on the line.  Heck, she even con­vinced a few oth­ers to join her in her anti-diet belief, and they even con­firmed with her more or less sim­i­lar results (except for Rick, who gained weight, but for­get him).  She even half wrote out the out­line for a book that she is prob­a­bly maybe going to write about the sub­ject someday.

Felic­ity is fal­la­ciously declar­ing a cor­re­la­tion with­out proper ver­i­fi­ca­tion, such as what Vera offered.  Trag­i­cally, she may be so firm in her belief that she never dares to test if the cor­re­la­tion she assumed was real was actu­ally caused by other fac­tors after all.  Maybe she will find out the hard way when she stops work­ing out and then the weight loss trend reverses itself.  She prob­a­bly would have lost even more weight if she kept going with the diet soft drinks all along.  Worse, she prop­a­gated her fal­la­cious the­ory with­out proper iden­ti­fi­ca­tion and ver­i­fi­ca­tion and risked the health of every­one who heard about her the­ory.  Maybe after she notices that her friends (who are not work­ing out very much) end up gain­ing weight, she will be forced to agree with the data that her “dis­cov­ery” is not the true cause of the effect.

Some­times, coin­ci­dences really do happen!

Do Vac­cines directly cause all sorts of other diseases?

Pigs are fly­ing all around the world right now, and some peo­ple refuse to believe it.  The H1N1 Swine Flu Vac­cine is cur­rently a hot table of con­tro­versy amongst some along the fringe.  They scare you with zil­lions of cor­re­la­tions of peo­ple get­ting a vac­cine of some sort and then soon after becom­ing ill or dis­abled in some way.  In an ABC News article, David Spiegel­hal­ter won­ders “why don’t we ever see a head­line ‘Man wins lot­tery after flu jab’?”  Dr. Steven Black and his col­legues at the Cin­ci­nati Children’s Hos­pi­tal in Ohio “cal­cu­lated what might be expected even if there were no vac­ci­na­tion cam­paign.”  For exam­ple, 397 out of 1 mil­lion preg­nant women will have a mis­car­riage (known as spon­ta­neous abor­tion) within a day of hav­ing a vac­cine.  Inter­est­ingly enough 397 out 1 mil­lion preg­nant women will suf­fer the same fate within a day of drink­ing a cup of Bub­ble Tea.  This is because this is the nor­mal expectancy out of such a large sam­ple size.  So if the reports are higher that 400, then there may be some­thing abnor­mal going on, oth­er­wise it can be assumed that there is no cor­re­la­tion between the mis­car­riages and the vaccine.

Is there is a cor­re­la­tion between a child get­ting a vac­cine and then devel­op­ing autism?  On his blog Neu­ro­log­ica, Steven Novella dis­cusses an inter­est­ing study pub­lished in Envi­ron­men­tal Health Per­spec­tives called CHARGE (The Child­hood Autism Risk from Genet­ics and the Envi­ron­ment) which set out to “com­pare blood total Hg con­cen­tra­tions in [2–5 year olds] with autism or autism spec­trum dis­or­der (AU/ASD) and typ­i­cally devel­op­ing (TD) con­trols in population-based sam­ples” and con­cluded that “total Hg in blood was nei­ther ele­vated nor reduced in CHARGE Study preschool­ers with AU/ASD as com­pared with unaf­fected con­trols, and resem­bled those of nation­ally rep­re­sen­ta­tive sam­ples.”  So, basi­cally, they could not draw a link between expo­sure to mer­cury and devel­op­ing autism, via vac­cines.  Fish con­sump­tion was also con­sid­ered in the results.  Novella also reminds us that “when thimerosal was removed from the vac­cine sched­ule in the US there was no cor­re­spond­ing change in the rate of autism diagnoses.”

It is never a bad idea to dou­ble check and see if there is a real con­nec­tion or not, but when the evi­dence is lack­ing for a cor­re­la­tion it risks becom­ing a fal­lacy to believe in such a link.

Some­times, it is easy to con­fuse an effect to being the cause itself.

If you wear it, they will come!

Why is it that when a lot of peo­ple wear­ing Pearl Jam T-Shirts gather together, all the sud­den Pearl Jam shows up on stage and sings for a cou­ple of hours?  Did the peo­ple gath­er­ing cause Eddie Ved­der to make an appear­ance?  Or per­haps the cor­re­la­tion is back­wards!  Per­haps Pearl Jam was play­ing that night, and a zil­lion fans decided to show up as a result.

This is per­haps also a rein­forced cor­re­la­tion, as Pearl Jam may not con­sider doing a show if they did not think many fans would show up.  Then again, they might just do it anyway.

Some­times, we can only be sure of some­thing by a process of elimination.

Are twins con­nected by some invis­i­ble, quan­tum, or psy­chic force?

When deter­min­ing if a causal cor­re­la­tion is valid or not, beware that it seems to be impos­si­ble to observe causal rela­tion­ships directly.  Sci­en­tific exper­i­ments attempt to approx­i­mate what is known as the coun­ter­fac­tual state of the world.  Accord­ing to Wikipedia, “a coun­ter­fac­tual is a con­di­tional (or “if-then”) state­ment indi­cat­ing what would be the case if its antecedent were true”.  For exam­ple, if one wanted to test if twins have some spooky or genetic dis­po­si­tion to be just as intel­li­gent, as long as one of them stud­ies, a sci­en­tist could set up a test that pits one twin against the other.  One twin will study all night long, while the other goes to the foot­ball game and par­ties all night.   Sure enough, if it turns out that the stu­dious twin scores a bet­ter grade than her hun­gover sis­ter, we can rule out a cor­re­la­tion of some sort of psy­chic con­nec­tion between them.

A good sci­en­tist will hypoth­e­size a cor­re­la­tion, study it for casual rela­tion­ships, cross out false cor­re­la­tions, and finally arrive at a the­ory com­prised of the hypoth­e­sizes that still fit the data.