345c73fc   
   XPost: misc.phone.mobile.iphone, comp.mobile.android   
   From: ithinkiam@gmail.com   
      
   "Paul M. Cook"  Wrote in message:   
   > On Thu, 10 Dec 2015 15:05:22 +0000, chris wrote:   
      
   > Most people are pretty stupid compared to the people you and I   
   > hang around with daily.   
      
   You have no idea who I around with, so please stop pretending you do.   
      
   >   
   >> No, no! That's not how science works. That's just positive selection   
   >> bias. You're finding data that fits your bias. That is what we in the   
   >> trade call 'bad' science.   
   >   
   > This is not true Chris,   
      
   You're slipping into a world of fantasy   
      
   >   
   > I really do intuit that cellphones *should* cause more accidents.   
   > Yet, *nobody* can find those higher accident rates.   
   > That's because they don't exist.   
   >   
   > How is that a "positive" bias?   
      
   It's your choice of data over any other to support your   
    'intuition' that is positive bias.   
      
   You really don't understand do you?   
      
   >   
   >> Yup. But reliable data does not prove an assumption if it is used   
   >> inappropriately.   
   >   
   > This is true.   
      
   Great!   
   >   
   > Same thing here.   
   > If there are no increased accident rates that anyone can find   
   > in any *reliable* data, that means that the accident rate is   
   > not rising at all due to cellphone use.   
      
   No. Your logic is flawed.   
      
   A theory can be disproved with data, but not with the absence of   
    data. The accident rates does not show anything, so it isn't   
    proof.   
      
      
   > Everyone will then complain that the increased accidents can   
   > be "hidden" in the overall accident data - but - they're morons   
   > because there is absolutely no way numbers as big as you would   
   > predict could possibly be *hidden* so cleverly in the reliable   
   > data.   
      
   How big so you predict the numbers to be? Why?   
      
   >> OK, now, mr hard science what can you conclude from the fact there isn't   
   >> an obvious or observable 'huge' increase in accidents? Hmm?   
   >> Does it prove anything? Can you prove anything with that data?   
   >   
   > Chris. How did you manage to *miss* entirely what the data says?   
      
   I'm not the one looking at it. You're the one missing the point   
    that just because you can't see a correlation doesn't mean that   
    cellphone use doesn't cause accidents.   
      
   > Nobody can find any increase in the accident rate.   
   > Period.   
      
   Even they did, it wouldn't prove anything. Why can't you get that?   
    It is always going to be just a correlation.   
      
   >> There is no '[sic]'. The studies, within their remits, do prove it. You   
   >> agree with them in point 1. above. You're contradicting yourself.   
   >   
   > The studies, as we all know, are manufactured experiments.   
      
   Aka a 'controlled' expt.   
      
   > They may or may not show what happens in the real world.   
   > I'm sure the peer-reviewed studies are 'good science' but what you   
   > have to do with studies is not just read the headline that the   
   > newspeople print.   
      
   And have you done that, before dismissing them?   
      
   > The studies show a pacman style experiment that clearly requires   
   > a real world test.   
      
   Problem is ethically you can't. You can't put people at real risk   
    for the purposes of an experiment. So they have to stick to   
    simulations.   
      
   > That they fail the real world test is what is so surprising here.   
      
   The census data is not a real world test.   
      
   > The fact is *nobody* on earth can show any increase in the number   
   > of accidents per vehicle mile, as reported by the US Census Bureau.   
   >   
   > That's like testing Einstein's theory, and finding out that light   
   > does *not* bend when it passes by a massive object.   
      
   It's not like it at all. The light bending experiment was looking   
    precisely at the phenomenon of interest. Not all starlight   
    aggregated over a month or whatever.   
      
      
   >> Absence of evidence is not proof of absence.   
   >> This is classic 'Argument from ignorance':   
   >   
   > You're misquoting the statistics maxim.   
   > It's not at all a "lack of evidence".   
      
   The link you snipped disagrees with you. You snip all the links I   
    provide. Why?   
      
   > There is *plenty* of good hard evidence.   
   > The US Census Bureau has been compiling those statistics for decades.   
      
   No. You have lots of data, but no evidence.   
      
   > Their very good evidence shows that there is no increase in the accident   
   > rate (which is vehicle accidents per mile driven).   
   >   
   > That is a cold hard fact.   
      
   It is a fact that there are data.   
      
      
   ----Android NewsGroup Reader----   
   http://usenet.sinaapp.com/   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)