In article <908hn6$bbm$[EMAIL PROTECTED]>, <[EMAIL PROTECTED]> wrote:
>I think I may have to be clearer about what I have and what I want. I
>am sampling names (for a mail survey) from a database of registered
>voters. I want the names to be distributed geographically in
>proportion to the population of the 21 counties of New Jersey.
>I have 21 pairs of numbers: for each county, the percentage that county
>has of the total NJ population, and the percentage that county has of
>the total of all NJ registered voters. For 7 counties, the percentages
>of population and voters are equal; for the remainder the absolute
>value of the disparity ranges from .3 to 1.7.
>So, by eye, it seems reasonable to assert that the counties register
>voters at roughly the same rate. I guess the null is that there is no
>difference in the proportions of county:state for voters and
>population.
>Sorry if this is silly or trivial; I'm just finishing up Intro. Stat.
The first question to ask is whether you want to ask the
question; what are the consequences of a "bad" sample?
You can get a sample with the right geographic distribution
from the registered voters, but it will not be a random
sample of the population, and would not be even if your
assumptions are correct.
However, there is no good reason to expect the agreement
you seem to want. For someone to be eligible to register
to vote, the requirements include being 18 years old,
citizenship, and residence for a certain length of time.
In addition, the person has to want to register. Your
null hypothesis, as you have stated it, is almost
certainly false.
Furthermore, if you have a county of population of 40,000,
the standard deviation of the number in a random sample with
a given proportion is bounded by 100, which is only .25%.
Overall, I suggest that you learn decision theory, and
do not use inappropriate recipes.
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Dept. of Statistics, Purdue Univ., West Lafayette IN47907-1399
[EMAIL PROTECTED] Phone: (765)494-6054 FAX: (765)494-0558
=================================================================
Instructions for joining and leaving this list and remarks about
the problem of INAPPROPRIATE MESSAGES are available at
http://jse.stat.ncsu.edu/
=================================================================
