Re: Steve Gough sent to jail


"Rik" <rt100@xxxxxxxxxxxxxxxxxxxxxxx> wrote in message
news:rk36o118qmbr5t8ulk5ghu6fk2i19ilaf9@xxxxxxxxxx
> On Tue, 22 Nov 2005 12:18:05 GMT, Steve Doerr
> <REVERSEdoerr.steve@xxxxxxxxxxxxxxxx> wrote:
>
>>Isn't it the case that the accuracy (margin of error) of a survey
>>depends on the *absolute* size of the sample, not on its size relative
>>to the population sampled? I'm sure I've read that somewhere.
>>
>>That's probably a gross over-simplification of the actual statistical
>>theory involved, but I'm no statistician.
>
> There's quite a sample calculator site at
> http://www.surveysystem.com/sscalc.htm which demonstrates that a
> sample of 4 million in a population of 60 million will return a 99%
> confidence that the sample is representative of the population as a
> whole to within a .01 (nought point zero one) percentage point.

I think you have left out the condition that the sample is drawn at random.
(In the case of SG I am happy that the people he passed and did not complain
do represent a fairly random sample).


.

Relevant Pages

  • Re: Steve Gough sent to jail
    ... >>Isn't it the case that the accuracy (margin of error) of a survey ... >>theory involved, but I'm no statistician. ... > whole to within a .01 (nought point zero one) percentage point. ...
    (uk.rec.naturist)
  • Re: Steve Gough sent to jail
    ... >Isn't it the case that the accuracy (margin of error) of a survey ... >depends on the *absolute* size of the sample, ... >theory involved, but I'm no statistician. ...
    (uk.rec.naturist)
  • Re: Linking cognitive science to juggling
    ... evidence for your hypothesis. ... on the other hand am a statistician. ... of course assumes a random sample from the population. ... Reading the study design (I'm always on the lookout for bad study design ...
    (rec.juggling)
  • Re: Convience sampling to random sampling
    ... for a survey colleagues made a convience sample. ... observation from the random sample with one or more observation from ... you want to assume that the refusers are like ...
    (sci.stat.consult)
  • Re: Survey Sample Size
    ... > The survey is a series of questions. ... > determine upfront how aggressive we need to be at soliciting responses? ... For example, if key question 1 requires 25 samples, key ... up with a statistician. ...
    (sci.stat.consult)