Saturday, 11 July 2009

Correlation and causality

I didn't realise it from the beginning. I am almost certain that I only became aware of this since last year. But since then, I feel that my thought-process has become much sharper.
What am I talking about?
I'm talking about the fact that the causality and correlation are totally different.
It is one thing to say that A is somehow related to B, and completely another to say that A is the cause of B.

Let's think of an example. Meet Lucy.
Lucy is 6 year old girl, who was the flower girl at her aunty, Julie's wedding, just about a year ago. Since then, Lucy saw Julie only a few times, but each time, Julie's belly was getting bigger and bigger, until just yesterday, when Lucy visited Julie with mum. Lucy was told that Julie gave a birth to a baby boy. Lucy wasn't sure what was happening, but when she saw the little red baby sleeping in the cot, Lucy somehow knew what her mum said about "giving birth".
Having seen a peaceful and cuddly looking baby cousin for the first time in her life, Lucy wished that she would have her own baby one day. And seeing that it all started with Julie getting married about a year ago, Lucy wished that one day she'll get married so she will have a baby, just like this little baby cousin.

Now, pause and think about that last sentence.

What is happening in Lucy's mind here?
Lucy have seen the correlation between Julie's marriage and the baby cousin. There is no doubt about that, and there is no mistake in that.
But Lucy went further and concluded that "getting married" is the cause of "having a baby".
We "grown-ups" know that this conclusion is a mistake. While there is no need to deny the correlation between the marriage and having a baby, it is obvious to us that the marriage does not necessarily result in having a baby. Also, we see many people having a baby without getting married. (Please reserve your moral, ethical, and religious position induced comments here. I'm merely pointing out what I, nay, we all observe today.) So, making a connection, (ie. correlation) between marriage and giving a birth is fine, but to conclude that marriage makes (ie. causality) one to give birth is wrong.

Unfortunately, I see this kind of foggy thinking way too prevalent.
People get confused about correlation and causality all the time, making incorrect judgement of situation and comments. Media people, whom I expect (or, after seeing so much junk, should I say that I just hope? wish?) to have a clear(er) view on this thing don't fare better at all.
Here's just one example of that in the SMH: Dump the toy boy for a lasting union. (Ok, I admit, it's in the Life & Style section, not a fair place for the SMH some may say.)
It tries to convince the reader that their advice is reliable by saying, "The advice comes not from an agony aunt in a women's magazine but from some of the country's top demographers." But all the conclusion they seem to make is purely from statistical data, and statistical data alone, not logical thinking on top of the statistical data. You saw that there was a correlation between a demographical condition to a lasting marriage, but you cannot simply say that you've got to keep that that condition in order to have a lasting marriage. You must first prove that the condition is the cause of the lasting marriage before you can give such advice. And I hope that the readers will look for the logical proof before trusting their advice.

Anyway, I don't know why I wrote this much about that silly "tabloid-like" article. Maybe it was my passion for this understanding about marriage that did it. Or was it my passion for understanding the difference between the correlation and causality? See, figuring out the cause is quite difficult.

No comments: