I'm not questioning the validity of the empirical rule for its general usage, I'm simply saying that you don't need numbers when studying psychological matters like this.
Also, don't forget the personal knowledge people have prior to reading or compiling the stats, which is also various. If I tell you that Italian teenagers living in Naples are likely to commit more crimes than teenagers living in Rome, would you understand the social context behind this statement? By applying the same logic to American stats you'd get even more generic results, because the region-to-region and even town-to-town differences can change virtually everything.
SIT DOWN AND SHUT UP UNTIL YOU GET AN EDUCATION.
You can't just spout garbage about things you think you understand.
Let me take apart your asinine argument line-by-line. (It's the argument that's asinine, not you. Treading that fine line!)
"you don't need numbers when studying psychological matters like this" -- yes you do. How else do you address them? Write some poetry? Talk a bit about what you saw on the street corner? ALL SOCIAL SCIENCE IS STATISTICS. It is the only way to draw conclusions about a population.
"Also, don't forget the personal knowledge people have prior to reading or compiling the stats, which is also various." We call that 'design bias' or 'interpretation bias' and it's not a criticism of statistics. It's a separate problem that can be compensated for, just like drag on an airplane.
"I tell you that Italian teenagers living in Naples are likely to commit more crimes than teenagers living in Rome, would you understand the social context behind this statement?" Boy, if you did a correlation between local socieconomic educators, level of education, unrest, average income, AND CRIME, using some goddamn STATISTICS, you would understand the social context!
"By applying the same logic to American stats you'd get even more generic results, because the region-to-region and even town-to-town differences can change virtually everything." WHICH IS WHY IF YOU WANT TO DRAW CONCLUSIONS ABOUT A PARTICULAR REGION OR TOWN YOU USE A SAMPLE REPRESENTATIVE OF THAT REGION OR TOWN.
Elementary, my dear Mobius.
Now let's go back to other misinformed things you've said!
Those studies are specific, too specific, because they're not widespread. They take many variables in consideration, but the problem is that they don't analyze that many teenagers. As result, all major assumptions come from the classic "If 5 out in 80 teenagers do this, then 50,000 out in 800,000 do this" - it's all a game of numbers and percentages which doesn't really help... and you know why? Because there are many problems, not only one, leading to certain behaviors. Vague assumptions would end up with something like "Videogames are bad for children and teenagers", and we all know of bogus this assumption is due to the various nature of games (genre, series, etc. etc.).
By "accurate" I mean "about 100% reliable".
Far from accurate stat = several thousand people
Partially accurate stat = several hundred people
Very accurate stat = several dozen people
Do you know what a representative sample is?
Come on. Define it for me. Make me day. TELL ME WHAT A REPRESENTATIVE SAMPLE IS.
Then explain to me how to use it to draw accurate conclusions about a population.
Give me the equation that tells you the sample size (N) you need to draw conclusions with a degree of accuracy (P) about a population of size X. Go on.
Look, all you're doing is attacking bad research design. Massive, flaming strawman. If you want to attack the basis of social sciences (statistical observation), learn something about good research.
Your argument is basically '**** science/statistics. We can conclude everything we need to from common sense. In fact, math makes this less accurate and is unnecessary.'
gb2the1450s.
...
You know what? Go read "Research Design and Methods: A Process Approach", by Borden and Abbot. It addresses every one of your misperceptions in excruciating detail.