Saturday, March 24, 2012

Statistic in daily life

If you are curious enough, or criticize enough, may be, you will find that statistic is all around us. Especially, nowaday, doing a poll and get some statistic is very easy. how how to understand the number and the meaning behind, that may be, or i would say, no many people really know what does it mean.

 for example, in recently newspaper, geologist said that, there are 70% chance that Tokyo will have magnitude 8 or higher in next 4 years. it seems very easy to understand. but if i ask, what is the chance in coming year? i am afraid only few people can answer. and how about if there is none in next 4 years, what is the chance? will it increase or decrease? ok, that may sound only geologist will know.

there is many polls on public opinion, but how large is the voters is enough to represent the public? if a city has 1 million people, is a poll on 1000 people enough? moreover, some poll will show numbers, what is the error of those numbers? say, in the US president election, John has 45% and Mary has 41% support. if the error is 10%, than it is hardly to say John is winning. however, if the error is 0.1 %, John is definitely winning. OK, although in a real vote, there is no error, since all error vote will not be counted. but in any survey, there must be some error, just like any scientific measurement. and it is the survey, make social science becomes a science.

OK, the above example may be not very related to yourselves. how about medical researches or reports? say, in 1 test, your AIDS result is positive, does it mean you really have AIDS? because every tester, there is a blind point, which is called false positive or false negative. what about the test is a false positive? that totally chance your life. moreover, occasionally, some medical researches will say something like, an apple a day can reduce the chance for getting cancer by 50%. what is this 50%? if the chance of getting cancer is just 1%, from 1% to 0.5%, is that really a difference? ( a side track, that's why, in science research, we like to say how many order the result was improved, 1 order means 10 times, 2 orders means 100 times, that is more significant.)

anyway, i wish i have time to share these in future and talk about each by each.

No comments: