naisho,
statistical math? how does that enable someone to form a non-biased opinion? just curious.
Statistics to sociological methodology, is like money is to working. Let me make a few similar synonyms contrats - a lot of GRE/GMAT prep I've had the past few months is still lingering in my thoughts, and I thought this would be able to explain the way I see it better:
Statistics:Sociology is to...
Religion:Martyr
Retardedness:Bigotry (I made this one up)
Equality:Manifesto
Scam:Quack
Did you get that? Basically to me, if statistics are done properly, especially thorough depending on the circumstances of the debate/subject, you have very good information regarding the outcome of people's opinions on something, or in a rhetorical sense, you could understand what is right or wrong in the sense of numbers.
For example, if we were to poll the entire US on how they viewed GWB's behavior during his stay at 8 years of the White House - this is nearly impossible, it would have to be precise with little room for mistakes. Now if you conduct the statistical data properly, centralizing it as in:
Poll Asian minorities living in California of their standpoint on GWB's term:
2,000,000 minorities polled - (exact ethincity, US citizen status, city, social status, age of respondent are extra variables that help pinpoint the answer to the cause - the more, the better). The problem is actually conducting a mass survey in statistical data - you need hope that you get your numbers as straight as possible, relying on forms of polling like phone, mail/paper, 'net polls, all have a negative effect of not producing correct answers.
80% of respondents felt that GWB did not do a good job
15% of respondents felt that GWB did OK
5% of respondents declined to choose an answer or were indifferent
of the 80% negative respondents, 50% were aged 40-60
of the 80% negative respondents, 20% were aged 18-39
of the respondents aged 40-60, 90% had an education from a 4-year college or higher
it goes into detail and then you can draw a logical conclusion like:
50%(show exact number) Asian citizens in California aged 40-60 with a 4 year education or higher dissapproved of GWB's term held in office.
In a non-domestic or politics issue type of statistical response, you can produce information easily based on what is presented to you and make sure that there are no loopholes that prove it incorrect. Such as:
Idaho produces 100,000 tons of Potatoes a year.
This year stores have reported receiving up to 40,000 tons of potatoes/ farmers have stated a natural crop disaster/drought is destroying their potatoes. (Let's make this a simple-simple statement instead of going into extra factors like other cases not accounted for)
Therefore - In the previous year, Idaho's potato production dropped 40% due to a natural disaster, with 1% accountability for other means.
The problem is, media/people tends to misuse statistics sometimes, illegally I call it in a way. Then the effect is that people don't understand it properly.
In the two examples I gave above, you'd end up seeing stuff like:
- 50% of californians polled disapproved of GWB's behavior (false, misuse of statistics due to minorities do not represent the complete CA)
- The USA lost 40% of its potato production recently due to a natural disaster in Idaho (false, IDAHO lost 40% of potatoes)
The way I see it, the correct numbers never lie, it's the words that make it do should the person report them erroneously.