Statistics, as a discipline, is probably the most powerful tool in the scientific arsenal, when used properly and professionally, as well as
UNDERSTOOD properly. But it's not for the faint of heart. It requires intelligence and training to use properly. Mark Twain lived in an era where
modern statistics did not exist. He was using the popular sense of the term, where people used numbers in a light-hearted fashion and proclaimed they
were statistics. OP is in the same ballpark here. It's really a superficial criticism. Indeed, when you are talking about modern physics issues like
quantum mechanics, it IS statistics.
The fact is, when a poll, for example, is done properly in a truly random format, polling about 1,000 "likely voters" is accurate to a "95% confidence
level." This means that the poll will be accurate, completely accurate, 19 out of 20 times. There is a 5% possibility that the conclusion is
inaccurate due to chance. If you want a "99% confidence level," you have to have more people in the poll, so it costs more. All these presidential
polls you are seeing are at a 95% confidence level. They are reasonably accurate. Ron Paul is not going to win.
The problem comes in when statistics is used improperly, often with the dubious "help" of inaccurate computer programs and inaccurate data. A good
example is the statistics used to support the theory of global warming. Note: I'm not going to cite sources here because this is a theoretical
discussion and it would slow me down considerably.)
The famous "Hockey Stick" used by Al Gore, et al to prove CO2 is killing us is an illusion provided by the computer program used to construct it. It
turns out that if you feed ANY "red noise" into the program, you get a hockey stick. To understand that, you need to know what red noise is. You
already know that "white noise" is random data. Red noise is similar, except that it is random FROM THE LAST POINT registered. A good example of that
is the stock market. A given stock "closes at $20.50." The next day it starts there and ends either up or down. It could be either, but "red noise"
means it started at the previous close. The Hockey Stick looks very precise, but if you plug in red noise as the data, you always get a hockey stick.
It's a bogus hockey stick. It is "statistics" at its worst.
Another example is temperature data used by NASA
to prove that
temperatures are rising. Look at the chart. It turns out if you take out the "adjustments" NASA put into the temperature data, global warming goes
away. This is another example of the manipulation of data in using statistics to skew results.
So, yeah! Statistics can be misused to manipulate public perception, and has been. But don't mistake that fact with the idea that statistics itself is
bogus. It's not at all. And a superficial condemnation of the science of statistics is as bogus as the misuse of statistics by those who do not know.
For the record, I have taken three college-level statistics courses and several seminars. I am a complete beginner. With a statistics textbook in
front of me and using a cookbook approach and a lot of time I can probably figure out a standard deviation issue, but to perform statistical
measurements in an accurate manner is beyond pretty much everyone, including college professors, except professional statisticians themselves.
edit on 4/15/2012 by schuyler because: (no reason given)