Download Sampling distributions • sample proportion ( ˆ p ) fraction (or

Survey
yes no Was this document useful for you?
   Thank you for your participation!

* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project

Document related concepts

Sufficient statistic wikipedia , lookup

History of statistics wikipedia , lookup

Bootstrapping (statistics) wikipedia , lookup

Taylor's law wikipedia , lookup

Statistical inference wikipedia , lookup

Resampling (statistics) wikipedia , lookup

Sampling (statistics) wikipedia , lookup

Student's t-test wikipedia , lookup

Gibbs sampling wikipedia , lookup

Transcript
Chapter 18
1
Sampling distributions
• sample proportion ( pˆ )
fraction (or percentage) of measurements in a
sample that are “successes”:
€
pˆ = (# successes in sample) / n
• population proportion ( p )
€
fraction (or percentage) of the measurements in the
population that are “successes”; we also write
q = 1 − p to represent the fraction of “failures” in the
population
€
• sampling error/variability
variability in values of a statistic when it is
measured in different samples of the same size
randomly selected from the same population
• sampling distribution of a statistic
distribution of all possible values of the statistic
computed for every possible choice of (randomly
selected) sample of a fixed size n chosen from the
population
Chapter 18
2
• sampling distribution model for pˆ
A SRS of size n with sample proportion pˆ is
selected from a population, chosen
so that it is
€
€
(1) small enough to be no more than 10% of the size
of the population (the 10% Condition), but
(2) large enough that it includes at least 10
successes and 10 failures (the Success/Failure
Condition),
then the normal model governs the sampling
distribution of values of pˆ . The random variable
represented by our measurements of pˆ has an
expected (mean) value
of E( pˆ ) = p, equal to the
€
population proportion, and a standard
deviation of
€
pq
€
SD( pˆ ) =
.
n
€
€
€
That is, the sampling distribution model for pˆ is

pq 
N  p,
.
n 

€
Moreover, this normal model is a better description
of the sampling distribution when the sample size n
€ is true because of…
is larger. This
Chapter 18
3
the most fundamental theorem in all of statistical
theory…
• the Central Limit Theorem
Regardless of the population from which we are
sampling or the statistic we may be measuring, the
distribution of the sampling statistic comes closer
to being normally distributed the larger the sample
size n becomes, and while the mean of the
distribution does not depend on n, the standard
deviation will decrease as n gets larger; therefore,
the statistic we are sampling will become a better
and better approximation of its true mean value for
larger and larger n.
Chapter 18
4
• sampling distribution model for y
A SRS of size n with sample mean y is selected
from a population with population mean µ and
standard deviation σ. The sample is chosen so that
€
its measurements are independent of each other, a
€
condition that is difficult to check, but reasonable
to assume, unless the sample is drawn without
replacement, in which case it should be small
enough to be no more than 10% of the size of the
population (the 10% Condition).
Then the normal model governs the sampling
distribution of values of y . The random variable
represented by our measurements of y has an
expected (mean) value of E( y ) = µ , equal to the
population mean, and a standard deviation of
€
σ
€
SD( y ) =
.
n
€ €
That is, the sampling distribution model for y is
€
€

σ 
N µ ,
.

n
€
Moreover, this normal model is a better description
of the sampling distribution the larger the sample
€
size n is.
Chapter 18
5
• standard error
In practice, we won’t generally know the value of
the standard deviation of the sampling distribution
of our statistic – after all, it depends on population
parameters: for instance,
SD( pˆ ) =
€
SD( y ) =
pq
n
σ
n
depends on p and q;
depends on σ.
€
But we can estimate them. When we do, our
€
estimates
are called standard errors:
€
ˆˆ
pq
;
n
s
SE( y ) =
.
n
SE( pˆ ) =
€
€
€
€