statistics

Do you think the following foods should be banned?

1.  A food that 98% of cancer patients have had daily
2.  A food that 96% of rapists ate the day they commited their crime
3.  A food that 94% of the people who died of heart stroke had had daily

If your answer is yes, it might be a good idea to study a bit of statistics.  The food described in 1~3 is all bread.

In this world full of data and numbers, it's better to be able to interprete what exactly you're looking at.

I'm only writing this to review what I've studied lately, but from what I've learned, the important part is comparison.  You can't tell anything from 1~3 unless you compare them with the data of healthy people.  If there is no difference, you can't necessarily say the food has anything to do with the conditions above.

The basic hypothesis is "there is no difference between the data of two (or more) groups （2つの集団のデータに差はない）", and the only situation in which you can say there *is* a difference is when you can prove the hypothesis wrong; or in other words, when the probability of the hypothesis being correct is 5% or less （差がある（危険率5％）） which means there *is* a difference with the probability of 95% or higher.  If the probablity of the hypothesis being correct is 6%, you can't say there is a difference （差があるとは言えない）.  But the 5% line is changed depending on what you're looking at; it's just usually set at a low level because you want to be careful when you say there is a difference.

Like when you develop a cancer drug.  You probably want to know that the drug really does make a (good) difference before you let your patient go through all the painful side-effects.

Usually, we tend to analyze data assuming that there *is* a difference between two different groups -- like in the example above (normal vs. abnormal), when in fact, there can be many common traits between two groups.

Statistics can probably prevent us from being decieved by our own preconcieved ideas or even hopes.

And apparently, data scientists are said to be the "sexiest professionals of the century".  Well, at least in Japan!